Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Unique functions for Notch4 in murine embryonic lymphangiogenesis

Unique functions for Notch4 in murine embryonic lymphangiogenesis In mice, embryonic dermal lymphatic development is well understood and used to study gene functions in lymphangiogenesis. Notch signaling is an evolutionarily conserved pathway that modulates cell fate decisions, which has been shown to both inhibit and promote dermal lymphangiogenesis. Here, we demonstrate distinct roles for Notch4 signaling versus canoni- cal Notch signaling in embryonic dermal lymphangiogenesis. Actively growing embryonic dermal lymphatics expressed NOTCH1, NOTCH4, and DLL4 which correlated with Notch activity. In lymphatic endothelial cells (LECs), DLL4 activation of Notch induced a subset of Notch effectors and lymphatic genes, which were distinctly regulated by Notch1 and Notch4 activation. Treatment of LECs with VEGF-A or VEGF-C upregulated Dll4 transcripts and differentially and temporally regulated the expression of Notch1 and Hes/Hey genes. Mice nullizygous for Notch4 had an increase in the closure of the lymphangiogenic fronts which correlated with reduced vessel caliber in the maturing lymphatic plexus at E14.5 and reduced branching at E16.5. Activation of Notch4 suppressed LEC migration in a wounding assay significantly more than Notch1, suggesting a dominant role for Notch4 in regulating LEC migration. Unlike Notch4 nulls, inhibition of canonical Notch signaling by expressing a dominant negative form of MAML1 (DNMAML) in Prox1+ LECs led to increased lymphatic density consistent with an increase in LEC proliferation, described for the loss of LEC Notch1. Moreover, loss of Notch4 did not affect LEC canonical Notch signaling. Thus, we propose that Notch4 signaling and canonical Notch signaling have distinct functions in the coordination of embryonic dermal lymphangiogenesis. Keywords Lymphangiogenesis · Notch · VEGF-C · Dermis Introduction Lymphangiogenesis is the process by which new lymphatic vessels sprout off pre-existing vessels. Sprouting of new lymphatic vessels requires coordinated lymphatic endothelial Ajit Muley and Minji Kim Uh have contributed equally to this work. * Carrie J. Shawber Wihuri Research Institute, Biomedicum Helsinki, cjs2002@cumc.columbia.edu Haartmaninkatu, 8, 00290 Helsinki, Finland Translational Cancer Medicine Program, Faculty Department of Obstetrics and Gynecology, Columbia of Medicine, Helsinki Institute of Life Science, University University Medical Center, New York, NY 10032, USA of Helsinki, FI‑00014 Helsinki, Finland Department of Pharmacology, Columbia University Medical Departments of Molecular Medicine and Experimental Center, New York, NY 10032, USA Medicine, Sapienza University, 00185 Rome, Italy Department of Physiology and Biophysics, University Department of Surgery, Columbia University Medical of Illinois Chicago, Chicago, IL 60612, USA Center, New York, NY 10032, USA Laboratory of Stem Cell and Neuro‑Vascular Biology, Cell and Developmental Biology Center, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD 20892, USA Vol.:(0123456789) 1 3 206 Angiogenesis (2022) 25:205–224 cell (LEC) proliferation, directional migration, and cell–cell lymphangiogenesis in the postnatal mouse ear [11]. In adhesion to form a properly patterned and functional net- embryonic dermal lymphangiogenesis, Notch1 deletion work. In murine dorsal skin, lymphangiogenesis begins at in LECs did not affect lymphatic branching, but increased embryonic day 12.5 (E12.5) at the side of the trunk and fol- lymphatic vessel caliber which was proposed to be second- lows dermal blood vessel development to meet at the midline ary to an increase in LEC proliferation and decreased LEC around E15.5 (Fig. S1a) [1]. Dermal lymphangiogenesis in apoptosis [14]. More recently, it was shown that loss of one mouse embryos is well characterized allowing for analysis copy of Dll4 was associated with reduced embryonic dermal of lymphatic endothelial signaling pathways, such as Notch. lymphangiogenesis in mice [15], a phenotype opposite to The Notch family of signaling proteins consists of four that seen in retinal angiogenesis [7, 8]. Additional studies cell surface receptors (NOTCH1-4) that are bound and acti- are needed to clarify the differences in the lymphangiogenic vated by membrane-bound ligands of the Delta-like (Dll1, phenotypes observed upon disruption of lymphatic endothe- 4) and Jagged (Jag1, 2) families expressed on neighboring lial Notch signaling. cells. Upon ligand activation, the extracellular domain of Here, we examined the roles for Notch4 and canonical NOTCH is released, which induces conformational changes Notch signaling in embryonic dermal lymphangiogenesis. that expose two proteolytic cleavage sites (TACE and We demonstrated that NOTCH1, NOTCH4, and DLL4 are γ-secretase/presenilin) that in turn releases the intracellular expressed, and Notch signaling active in embryonic der- cytoplasmic domain (NICD) from the cell surface [2]. In mal lymphatic endothelium. VEGF-A and VEGF-C sign- the canonical Notch signaling pathway, NICD transits to the aling differentially regulated Notch pathway gene expres- nucleus, binds the transcriptional repressor RBPjκ, where sion and activity in cultured LECs. Mice nullizygous for it recruits an activation complex including Mastermind- Notch4 displayed an embryonic dermal lymphangiogenic like (MAML) and HDACs, and activates RBPjκ-dependent phenotype characterized by increased LEC migration and transcription of Notch effectors, such as those in the HES/ reduced branching. In contrast, inhibition of canonical Notch Hey families. Notch also signals via a less well-understood signaling increased lymphatic vascular density consistent non-canonical RBPjκ-independent pathway that has been with an increase in LEC proliferation. Together, these data suggested to not require nuclear localization of NICD [2]. demonstrate that dermal lymphangiogenesis is dynami- During development of the blood vascular system, Notch cally regulated by Notch and requires both NOTCH1 and signaling is essential for arterial endothelial specification, NOTCH4 functions, as well as canonical and non-canonical vascular smooth muscle cell differentiation and viability, Notch signaling. and sprouting angiogenesis [3–5]. Studies of murine retinal angiogenesis have shown that VEGF-A, via activation of VEGFR2, upregulates DLL4 expression in the filopodia- Materials and methods extending tip cell located at the vascular front [4, 6–8]. Dll4 signals to the neighboring Notch-expressing stalk cell, where Cell culture/constructs Notch activation downregulates VEGFR2 and VEGFR3 expression and inhibits the tip cell phenotype. During reti- HeLa cells were maintained in 10% FBS DMEM. Human nal angiogenesis, inhibition of DLL4 or NOTCH1 leads to umbilical vein endothelial cells (HUVEC) were isolated a hypersprouting phenotype characterized by an increase in as previously described and maintained in EGM2 (Lonza) tip cells at the expense of the stalk cells, increased VEGFR2 [16, 17]. Neonatal human dermal lymphatic endothelial cells and VEGFR3 expression, and decreased vascular outgrowth (HdLECs) were either purchased (Promocell) or isolated as [6, 7, 9]. Although it has been shown that VEGF-C induces previously described [18] and maintained on fibronectin- DLL4 in cultured LECs [10, 11], the mechanisms by which coated plates in EGM2-MV2 (Lonza; complete medium) Notch regulates dermal lymphangiogenesis remain to be supplemented with 10 ng/mL VEGF-C (R&D). To activate elucidated. Notch signaling, HdLECs were lentivirally infected [19] We previously demonstrated that NOTCH1 and NOTCH4 using pCCL.pkg.wpre vector encoding N1IC, N4/Int-3, are expressed in the postnatal maturing dermal lymphatics or GFP. N1IC encodes the constitutively active cytoplas- [12]. Studies of postnatal lymphangiogenesis have shown mic domain of NOTCH1. N4/Int-3 encodes an activated that pharmacological inhibition or genetic manipulation Notch4 allele generated by MMTV insertion [20]. Tran- of Dll4/Notch1 signaling can result in both increased and scripts and protein expression was confirmed by quantita- decreased lymphangiogenesis [11, 13]. Neutralizing anti- tive (q)RT-PCR and Western analyses of samples collected bodies against NOTCH1 or DLL4 suppressed lymphangi- post-infection. ogenesis in the postnatal mouse ear, tail dermis, and a wounding model [13]. In contrast, an inhibitory soluble DLL4 extracellular domain fused to FC (Dll4FC) stimulated 1 3 Angiogenesis (2022) 25:205–224 207 HdLEC assays DLL4‑ligand activation assay and mRNA sequencing VEGF treatment of HdLECs: confluent HdLEC monolay - Tethered Ligand Assay: The recombinant extracellular ers were starved overnight in 1% FBS in EBM2 (Lonza) domain of the Notch ligand hDLL4FC (Sino Biologicals or in human endothelial SFM (Fisher Scientific) followed Inc.) or IgG-FC (Sino Biologicals Inc.) were coated onto by either 1 or 5  h in EBM2/SFM containing 100ng/mL a 24-well plate (Corning) on a fibronectin matrix (Sigma). VEGF-A (R&D), 100ng/mL VEGF-C (R&D), or 500ng/ Following an overnight incubation at 4 °C, primary ECs C156S mL VEGF-C (R&D) prior to RNA isolation. Assays (at 80% confluency) were trypsinized and seeded onto the were performed at least 3 times. For detection of AKT and coated plates and incubated at 37 °C with 5% CO for 6 h. ERK activity, HdLECs were serum starved overnight in Experiment was performed in triplicate. SFM containing 1% FBS and 0.1% BSA, followed by 5 h RNA was isolated using the RNEasy Mini Kit (Qiagen), in SFM alone. Cells were then switched to SFM contain- quantity and integrity measured using a Bio-analyzer (Agi- ing 0.1% BSA and either 100ng/mL VEGF-A or 100ng/mL lent TapeStation 4200, UIC Genome Research core) prior VEGF-C for 20 min prior to fixation with cold 4% PFA. to RNA sequencing. TLA HdLEC samples were sequenced Assays were performed in duplicate for two different lenti- at a ~30 million paired-end (PE) read depth with 150-base viral transductions. fragments by Novogene (https:// en. novog ene. com/). Raw Migration Assay: HdLECs were seeded in triplicate on a reads from in vitro screens were mapped to the Human data- fibronectin-coated (Thermofisher) 12-well plate in complete base (ENSEMBL/GRCh38) using STAR (version 2.5.0a) medium. The following day (0-hour time-point), a scratch and processed with Samtools (version 1.4.1). The counts through the confluent monolayer was made across each well obtained by FeatureCounts (version 1.5.2) were analyzed by using a 200 µl pipet tip, and medium was changed to EBM2 DESeq2 (version 1.18.1) to identify differentially expressed containing 100ng/ml VEGF-C. For migration assays with genes. The RNAseq datasets generated in the current study mitomycin C, confluent monolayers were pretreated with are available in the NCBI Gene Expression Omnibus reposi- 10 µg/mL mitomycin-C (Sigma) for 45 min prior to scratch- tory at https://ww w.nc bi.nlm .nih .go v/ge o (Accession Num- ing. Cells were maintained in EBM2 containing 100ng/ml ber GSE183631). VEGF-C and 0.1 µg/mL mitomycin-C while migration was assessed. Growth into the scratch was documented at 0, 4, 8, Gene expression analyses 12, and 25 h with a Zeiss Axiovert 40 CSL inverted micro- scope or at 0, 4, 8, and 24 h using an Olympus IX83 micro- RNA was isolated using the RNEasy Mini Kit (Qiagen) and scope. Cell migration rate was determined using imageJ soft- reverse transcribed using the VersoTM cDNA Synthesis ware [21] and calculated as the percentage of cell-free area Kit (Thermo Fisher) or First Strand Synthesis Kit (Invitro- at different time-points relative to the initial wound area. gen). qRT-PCR was performed in triplicate for each gene Assays were performed at least 2–3 times for two independ- (Table S1), using ABsoluteTM Blue QPCR SYBR Green ent lentivirally generated HdLEC populations. Master Mix (Thermo Fisher) or Sybr Green Master Mix (Applied Biosystems) and 7300 Real-Time PCR System Co‑culture Notch reporter assay (Applied Biosystems) or CFX96 PCR Cycler (Biorad). Gene-specific qRT-PCR standards were used to determine HdLECs (90% confluent) were lipofected (Lipofectamine transcript levels and normalized to β-actin expression [12]. 2000; Invitrogen) with the Notch reporter plasmid PCRs were set up in triplicate and performed at least 3 times. pGL3.11CSL [12] containing 11 repeats of Notch/CSL For the validation of mRNA sequencing data, qPCR was (RBPjκ) cis-elements, and phRL-SV40 renilla (Promega) to done using SYBR Green master mix (Applied Biosystems) normalize lipofection efficiency. HeLa cells were lipofected and primers specific to genes of interest (Table S2). The with pCR3 plasmids encoding either DLL4-FLAG or JAG1- mean cycle threshold (Ct) values from the triplicate run for FLAG with empty vector serving as a control. 24 h after each sample were analyzed using β-actin as the reference lipofection, HeLa and HdLECs were co-cultured together at gene. ΔΔCt method [22] was used to calculate the relative a 1:1 ratio on fibronectin-coated plates in EGM2. 24 h after expression using the following steps: (1) Normalization to co-culture, a luciferase reporter assay was performed using reference gene: ΔCt = Ct – Ct . (2) Relative expres- GOI GOI BA the Dual-Luciferase Reporter Assay System (Promega) and sion between conditions: ΔΔCt = ΔCt – ΔCt . The GOI EXP CNT a TD20/20 luminometer (Turner Designs). Luciferase val- analysis was done using Microsoft Excel and Prism. ues were normalized to Renilla values. Each condition was performed in triplicate, 4 times. 1 3 208 Angiogenesis (2022) 25:205–224 a Zeiss Axioskop2 Plus and Zeiss AxioCam MRc camera Western blotting with Zeiss Zen software, or an Olympus IX83 Inverted Sys- tem Microscope and Olympus cellSens software. Confocal NOTCH4 expression in Notch4 null mice was determined by Western blot. Fresh E14.5 hindlimbs were lysed in RIPA microscopy was performed with a Zeiss LSM 510 META Confocal Microscope and the LSM software. buffer (Invitrogen) containing protease and phosphatase inhibitors (Thermofisher) on ice and protein concentra- For cell immunochemistry, cells were fixed in 4 % PFA on ice for 15 min, permeabilized and blocked with 0.1% tion determined by BCA protein assay kit (Pierce). Equal amounts of protein were separated on an 8% SDS-PAGE Triton X-100, 2% BSA, 3% donkey serum in 1 × PBS for 1 h at room temperature. Cells were then incubated over- gel and transferred to a nitrocellulose membrane. The mem- branes were blocked with 3% nonfat milk and 3% bovine night with primary antibody at 4 °C followed by an incu- bation with Alexa Fluor-conjugated anti-donkey secondary serum albumin (Jackson ImmunoResearch) in Tris-Buff- ered Saline Tween-20 and probed with antibodies against antibodies (Invitrogen) for 1 h at room temperature. Slides were washed in 1 × PBS and mounted with Vectashield the cytoplasmic domain of NOTCH4 [23] and β-ACTIN (Abclonal). Horse radish peroxidase-conjugated secondary with DAPI mounting media. Images were captured with an Olympus IX83 Inverted System Microscope and Olympus antibodies (Life Technologies) were used, detected with SuperSignal West Pico Chemiluminescent Substrate (Pierce) CellSens software. Images were analyzed with ImageJ or Adobe Photoshop. and images captured with Biorad Chemdoc MP. Tiled 10x images were used to quantify lymphatic and blood vascular density, distance between migration fronts, fronts Mouse studies per unit length, and branch-point per unit length. 20x images −/− T2 were used to determine lymphatic vessel caliber, distance Notch4 nullizygous (N4 ) [24], Prox1CreER [25], and f l/f l DNMAML [26], CBF:H2B-Venus (NVR reporter pur- to first branch-point, number of Prox1+ cells per field, and sprout morphology. Vascular density was determined chased from Jax Labs) [27] and Prox1-tdTomato lymphatic reporter (ProxTom) [28] mice were used for these studies. as positive signal area normalized by total area. Distance between migrating fronts was determined as the mean dis- Studies were performed in mice with a mixed background, as well as a pure C57BL6j background. For studies using tance between the 2 lymphatic fronts measured at multiple T2 points (≥ 3) [1]. For analysis of migration of Notch4 mutants, Prox1CreER , tamoxifen in corn oil was administered via +/− oral gavage (10 mg/40 g) at E12.5. 3 or more independent Notch4 embryos were used to normalize between litters, as they were present in all litters analyzed and the distances litters were assessed for each analysis. Number of embryos analyzed is presented in the figure legends. between migration fronts were not statistically different than WT embryos. Sprouting fronts, defined as the sprouts that Immunohistochemistry & Imaging reside at the leading edge of the migrating front per unit length was determined as the number of sprouting fronts at E14.5 and E16.5 dorsal skin was dissected, fixed for 2 h the leading edge of the migration front normalized to the vertical length (posterior-anterior; Fig. S1a). Length of the in 4% PFA, and then immunostaining initiated. Alterna- tively, embryos were incubated overnight in 4% PFA and sprout was determined as length from tip of sprout at lym- phangiogenic front to the first branch-point. Lymphatic ves - then stored in 1 × PBS at 4 °C. E14.5 tissues were incu- bated for 2 h at room temperature in blocking buffer (10% sel caliber was determined by measuring the width of lym- phatic vessels in the maturing lymphatic plexus and adjacent donkey serum, 0.3% Triton X-100, 1 × PBS), incubated in primary antibody (Table S3) diluted in blocking buffer to the first branch-point from the migrating front. Sprout morphology at the lymphangiogenic front was determined overnight at 4 °C, and then incubated with the appropri- ate Alexa Fluor secondary antibody (Invitrogen) diluted in by counting total number of blunt-ended sprouts (rounded, lacking multiple filopodia) and spiky-ended sprouts (elon- blocking buffer overnight at 4 °C. E16.5 dermal tissues were washed in 1 × PBS containing 0.2% Triton X-100 and 20% gated with multiple filopodia) normalized to the total num- ber of sprouts assessed. Branch-points per unit length in DMSO for 4 h at room temperature and immunohistochem- istry performed as described in Cha et al. 2016 [29]. For maturing lymphatic plexus was determined as the number of branch-points per field normalized to the total length of immunostaining of sections, 5-micron sections were stained as previously described [30]. Tissue was mounted using lymphatic vessels per field. To measure Prox1+ LEC num- ber, Prox1+/LYVE1+ LECs were scored and mean number Vectashield with and without DAPI (Vector Laboratories). Images were captured using a Nikon SMZ-U Zoom 1:10 per field determined. To determine the significance between control and one experimental group, a two-tailed student’s microscope and Nikon 4500 digital camera, Nikon ECLIPSE E800 microscope and NIS Elements software, Nikon DXM t test was used. For analyses of more than two groups, one- way analysis of variance (ANOVA) was used to determine 1200 digital camera, and Image ProPlus v.4.01 software, 1 3 Angiogenesis (2022) 25:205–224 209 significance by unpaired t test. For analyses of multiple con- expressed through-out the endothelium of the sprouts at the ditions and cell populations, two-way ANOVA was used and lymphatic front which overlapped with DLL4 (Fig. 1b, c). Dunnett’s multiple comparison test performed to determine Analysis of E14.5 dermal cross-section confirmed that the significance between groups. A p < 0.05 was considered LYVE1+ dermal lymphatic endothelium expressed both significant. NOTCH1 and NOTCH4 (Fig. S1b, c). Outside of the lym- phatics, NOTCH1 expression was observed in the epider- Lymphangiography mis and blood endothelium, while NOTCH4 was expressed in the epidermis and a subset of LYVE1+ macrophages Lymphangiographies were performed as described on E17.5 (Fig. 1b, c, Fig. S1b, c) . embryos [15]. Briefly, 2 µL of 0.4% Trypan blue solution To determine where Notch is actively signaling dur- (Sigma) was injected into the dermis in periorbital region ing lymphangiogenesis, the dermal lymphatics in E14.5 with a 36G beveled needle attached to a Nano l syringe embryos carrying alleles for the Prox1-tdTomato (ProxTom) (WPI). The embryos were imaged using AMSCOPE ster- LEC reporter [28] and the Notch Venous Reporter (NVR) eomicroscope (AMSCOPE) with camera attachment 1 min [27] were assessed. Notch activity was observed throughout after injection. the lymphatic vascular plexus at both the lymphangiogenic front, defined as the LECs that make up the sprout from tip to the first branch-point, and the mature plexus where the Results vessels have begun to remodel (Fig. 2a). At the lymphangi- ogenic front, Notch activity was often observed in several Embryonic dermal lymphatics expressed NOTCH1, LECs located at the tip cell positions in spiky-ended sprouts NOTCH4, and the Notch ligand, DLL4 with filopodia (Fig.  2b), consistent with the broad expres- sion of NOTCH1, NOTCH4, and DLL4 at the front (Fig. 1). Cultured human LECs, HdLECs, express NOTCH1-4 and Notch activity was also observed in blunt-ended sprouts. In the Notch ligands, DLL4 and JAGGED1 (JAG1) [18], while the mature lymphatic plexus, Notch activity was observed NOTCH1 and NOTCH4 are expressed in the postnatal day 4 at branch-points (Fig.  2a, c). Taken together, the expres- (P4) murine dermal lymphatic vessels [12]. Notch signaling sion data suggest that DLL4 signaling via NOTCH1 and/or has been shown to be active in the E15.5 dermal lymphatics NOTCH4 has a role in regulating dermal lymphangiogenic [14], but it is not known which Notch proteins and ligands growth and maturation. are expressed in dermal LECs at this time. To study the role of Notch signaling in embryonic dermal lymphangiogenesis, Profiling of DLL4/Notch signaling in human dermal we determined the expression of NOTCH1 and NOTCH4, lymphatic endothelial cells and the angiogenic Notch ligands, DLL4 and JAG1, as well as Notch activity in E14.5 dorsal skin. This time-point is Expression studies suggested that DLL4 is the major ligand characterized by the presence of two LYVE1+ lymphatic for Notch signaling in the lymphatic endothelium during fronts migrating toward the midline which precedes a matur- dermal lymphangiogenesis. To determine if DLL4 or JAG1 ing lymphatic plexus (Fig. S1a). At this time-point, two could activate Notch in LECs, co-culture assays were per- CD31+ angiogenic fronts have fused at the midline to form formed in which endogenous Notch activation was deter- a connected blood capillary network. mined using a Notch-response CSL luciferase reporter [12]. At E14.5, DLL4 was expressed in both in the developing HeLa cells were engineered to express DLL4, JAG1, or lymphatics and blood vessels of the dermis (Fig. 1a). Unlike both (Fig. S2a) and then seeded at a 1:1 ratio with HdLECs the retina, where DLL4 expression is restricted to 1–2 tip containing a CSL-luciferase reporter. DLL4-expressing cells at the angiogenic front [6–8], high DLL4 expression HeLa cells upregulated Notch signaling nearly five-fold was observed in multiple LECs within the sprouts at the over co-cultures using parental HeLa cells, while JAG1 lymphangiogenic front. DLL4 was also expressed in the only modestly increased Notch signaling in HdLECs (Fig. arterial vessels and blood capillary network, with strongest S2b). Co-culture with HeLa cells co-expressing DLL4 and expression observed in the large arteries, consistent with JAG1 induced Notch signaling similar to the co-cultures its expression in the vasculature of the intestinal villi [10]. with DLL4 alone, suggesting that JAG1 did not interfere Unlike DLL4, JAG1 was not expressed in dermal lymphatics with DLL4 signaling. Together with the expression studies, at E14.5 (Fig. 1a). JAG1 expression was limited to the blood these data suggest that DLL4 functions as a ligand for LEC vasculature, where the highest expression was observed in NOTCH. the larger caliber arteries in a pattern consistent with vas- To further assess DLL4/Notch signaling, HdLECs were cular smooth muscle cell and endothelial cells. Staining for seeded on DLL4FC-coated or FC-coated (control) plates. NOTCH4 and NOTCH1 demonstrated that they were both After 6  h, RNA was collected and mRNA sequencing 1 3 210 Angiogenesis (2022) 25:205–224 Fig. 1 Embryonic dermal lymphatics expressed NOTCH1, NOTCH4, wild-type skin wholemount stained for LYVE1 and NOTCH4. White and DLL4. a E14.5 wild-type skin wholemounts stained for LYVE1, arrowheads mark lymphatics at the front. Yellow arrowhead marks CD31 and DLL4 or JAG1. Higher magnification of boxed areas a NOTCH4+ macrophage. Scale bars, 50  μm. c  E14.5 ProxTom presented to the right. White arrowheads mark sprouts at the lym- skin wholemount stained for DLL4 and NOTCH1. White arrow- phangiogenic front. White asterisk marks the blood vascular plexus. heads mark lymphatic sprout at the front. Yellow arrowhead marks Yellow arrowhead marks an artery. Scale bars, 100 μm. b  E14.5 NOTCH1+/DLL4+ blood vessel. Scale bars, 20 μm performed. Relative to HdLECs seeded on FC-coated plates, signaling induces genes responsible for pattern specification, HdLECs seeded on DLL4FC significantly altered the expres- neurogenesis, and chemotaxis (Fig. 3d). sion of 675 genes with a padj < 0.05 (Fig. 3a, Table S4, S5). To assess if DLL4-induced genes were downstream of 69 genes were induced 1.2-fold, while 68 genes were sup- Notch1 or Notch4 activation, HdLECs were generated to pressed 1.2-fold. Analysis of the top 30 upregulated genes express activated forms of NOTCH1 (N1IC) or NOTCH4 revealed that DLL4/Notch signaling upregulated the expres- (N4/Int-3). Quantitative RT-PCR was performed for Notch sion of known direct effectors of Notch signaling, Hey1, effectors of the Hes and Hey gene families, as well as lym- Hes4, Dll4, and Hes1, as well as key lymphangiogenic phangiogenic genes. Of the Notch effectors assessed, N1IC genes, Ackr3, Cxcr4, Ccl2, EphrinB2, Gja4 (Cx37), Gja1 and N4/Int-3 both induced the expression of Hes1, Hes4, (Cx43), and Sema3g (Fig. 3b) [31–38]. DLL4/Notch signal- Hey2, and HeyL, with the strongest induction observed for ing also downregulated lymphangiogenic genes, including Hes4 and HeyL (Fig. 3e). In contrast, only N1IC induced Apln and Adm (Fig. 3c) [35, 39]. Further analysis of the 675 the expression of Hes5. Notch1 and Notch4 activation sig- altered genes demonstrated that DLL4/Notch signaling both nificantly induced the expression of the majority of lym- upregulated and downregulated genes of the Notch pathway phangiogenic genes assessed, except for Cxcr4, Bmp2, and (Fig. S3a) and lymphangiogenesis (Fig. S3b). GO: Biologi- Tgfrb2 (Fig. 3f). Similarly, both suppressed the expression cal Pathway (BP) analysis indicates that LEC DLL4/Notch of Prox1, Podoplanin, and Lyve1 (Fig. S3b, S4b). Cxcr4 was significantly induced by N1IC, while N4/Int-3 suppressed 1 3 Angiogenesis (2022) 25:205–224 211 Fig. 2 Notch activation observed throughout the embry- onic dermal lymphatic vascular plexus. E14.5 ProxTom;NVR skin wholemounts stained for LYVE1. a Low magnifica- tion image demonstrating Notch activity throughout the developing lymphatic plexus. Blue arrowheads mark sprouts at the lymphangiogenic front with Notch activity. White arrowheads mark regions of high Notch signaling in the maturing plexus. Scale bars, 500 μm. b High magnifica- tion of spiky-ended lymphatic sprout. White arrowheads mark tip cells with Notch activity. Yellow arrowheads mark stalk cells with Notch activity. c High magnification of the maturing plexus. White arrowheads mark LECs with Notch activity. b, c Scale bars, 100 μm its expression. Neither Tgfrb2 and Bmp2 were induced. determined the effect of VEGF-A and VEGF-C on Notch While Ackr3 and Ccl2 were significantly induced by N1IC genes, ligands, and effectors in HdLECs. Serum-starved relative to the GFP controls, N4/Int-3 was a much stronger HdLECs were treated with either VEGF-A, VEGF-C or C156S inducer of both these genes (p < 0.0003, p < 0.0001 N4/Int-3 VEGF-C . In HdLECs, VEGF-A binds and activates C156S vs. N1IC, respectively). Together these data support over- VEGFR2, VEGF-C binds and activates VEGFR3, and lapping and distinct downstream signaling for Notch1 and VEGF-C activates both VEGFR2 and VEGFR3 [40]. After Notch4 in LECs. 1 h, VEGF-A and VEGF-C significantly induced Dll4 tran- scripts, which correlated with an increase in Hey1, Hey2, C156S VEGF‑C induced Dll4 expression and Notch and Hes1 transcripts (Fig. 4a, b). VEGF-C or VEGF-C activation in HdLECs both induced Dll4 expression, as well as the Notch effector, Hes1 (Fig. 4a, b). Notch1 was modestly induced by VEGF- During sprouting angiogenesis, VEGF-A/VEGFR-2 sign- A only, while Notch4 expression was unaffected (Fig.  4a). aling upregulates DLL4 in blood endothelial tip cells to To determine if VEGF-A and VEGF-C differentially activate Notch signaling in the adjacent stalk cell [6–8]. As regulate Notch signaling in BECs and LECs, HdLECs and we observed DLL4 expression and Notch activity in LECs HUVEC were serum starved and treated with either VEGF- of the sprouts located at the lymphangiogenic front, we A or VEGF-C. After 5  h, VEGF-A only induced Dll4, 1 3 212 Angiogenesis (2022) 25:205–224 Fig. 3 DLL4/Notch signaling regulated Notch and lymphangio- direct targets of Notch signaling and f  lymphangiogenic genes sig- genic genes via Notch1 and Notch4. HdLECs were seeded on either nificantly induced in the DLL4-HdLEC assay in HdLECs expressing DLL4FC- or FC-coated plates, and RNA isolated after 6 h, followed GFP, N1IC, or N4/Int-3. Data presented for two independent trans- by mRNA sequencing. Experiment was performed in triplicate. ductions done in duplicate and gene expression determined by delta a Volcano plot of genes downregulated and upregulated by DLL4FC CT method and presented relative GFP controls ± s.e.m. One-way relative to FC controls. b  Top 30 genes upregulated and c  down- ANOVA performed and significance determined by unpaired t test. e, regulated by DLL4FC in HdLECs. b, c  *mark Notch pathway and f *p < 0.05 or blue asterisks indicate p < 0.003 N4/Int-3 or N1IC rela- blue asterisks indicate lymphangiogenic genes. d  Top GO pathways tive to GFP controls. Red stars indicate  p < 0.0003 N4/Int-3 relative for biological processes for DLL4-induced genes. e  qRT-PCR for to N1IC 1 3 Angiogenesis (2022) 25:205–224 213 Fig. 4 VEGF and NOTCH signaling modulated each other in LECs. N1IC-, and N4/Int-3-HdLECS were treated with VEGF-A or VEGF- a Notch1, Notch4, and Dll4 expression, and b Hey1, Hey2, and Hes1 C for 20  min and then stained for either  e  phospho-AKT (pAKT) expression determined by qRT-PCR of HdLECs treated for 1  h or g  phospho-ERK (pERK). Scale bars, 25 μm. Quantification of C156S with no growth factor (nGF), VEGF-A, VEGF-C, or VEGF-C . mean,  f  pAKT and h  pERK expression normalized by area. Mean Experiment done in triplicate. Data presented as mean fold induc- data presented for two independent transductions and experiment tion relative to HdLECs treated with no growth factor ± s.e.m. *p < performed in duplicate. Data presented as fold expression relative to 0.05 HdLEC treated with VEGFs relative to HdLEC treated with no GFP-expressing HdLECs treated with  nGF ± s.e.m. f  pAKT: two- growth factor. c  Notch1, Notch4, and Dll4 expression, and d  Hey1, way ANOVA: p < 0.0001, Dunnett’s multiple comparison test *p < Hey2, and Hes1 expression determined by qRT-PCR of HdLECs 0.002 VEGF-A-treated N4/Int-3 vs. GFP HdLEC. h pERK: two-way treated for 5  h with nGF, VEGF-A, or VEGF-C. Data presented ANOVA: p < 0.001, Dunnett’s multiple comparison test *p < 0.03, as fold induction relative to nGF ± s.d. *p < 0.05 VEGF-treated **p < 0.0001 VEGF-C-treated N1IC or  N4/Int-3 vs. GFP HdLEC. HdLEC relative to HdLEC treated with no growth factor. e–h  GFP-, *p < 0.003 VEGF-C-treated N4/Int-3 vs. N1IC HdLEC 1 3 214 Angiogenesis (2022) 25:205–224 −/− which was associated with an increase in Hey1 and Hes1 Analysis of E14.5 Notch4 dermal wholemounts in HdLECs (Fig. 4c, d). VEGF-C induced Dll4 in HdLECs revealed that the distance between the two migrating (Fig.  4c), which correlated with an increase in Notch1 lymphatic fronts was decreased relative to wild-type and +/− transcripts (Fig. 4c), as well as Hey2 and Hes1 induction Notch4 littermates (Fig. 5a, b). This correlated with an (Fig. 4d). In HUVEC, VEGF-A induced Dll4 and Notch1 increase in the number of lymphatic fronts migrating toward −/− expression, whereas VEGF-C modestly decreased Dll4, the midline in the Notch4 dermis (Fig. 5c). Further analy- Notch1, and Notch4 transcripts (Fig. S4a). Thus, VEGF-A sis of the lymphangiogenic sprouts at the migration front and VEGF-C dynamically and temporally induced Notch revealed that the length from the front to the first branch- activity and specific Notch effectors via induction of Dll4 point did not differ between mutants and controls (Fig.  5d). in cultured LECs. We next evaluated the lymphatic vessel caliber at the lym- To determine if Notch1 or Notch4 activation altered sign- phangiogenic front and in the maturing plexus. The caliber aling downstream of the VEGFs, HdLECs expressing either of the vessel adjacent to the first branch-point at the front did N1IC, N4/Int-3 or GFP were serum starved overnight and not differ between mutant and control mice (Fig.  5e, f). How- then stimulated with VEGF-A or VEGF-C. After 20 min, ever, a significant reduction of vessel caliber was observed −/− HdLECs were assessed for AKT and ERK activation by in the maturing lymphatic plexus of Notch4 . Although immunofluorescent staining for phospho-AKT and phospho- the lymphatic vessel diameter was reduced in the maturing −/− ERK (Fig. 4e–h.). As compared to the GFP controls, activa- Notch4 plexus, branching was similar between mutants tion of AKT was reduced in N4/Int-3 HdLECs treated with and controls (Fig. S6a). To determine if the reduced dermal VEGF-A. Activation of AKT by VEGF-C was unaffected lymphatic vessel caliber was due to a change in LEC pro- −/− in HdLECs expressing either N1IC or N4/Int-3. Levels of liferation, wild-type and Notch4 E14.5 dermal skin were ERK activity were unaffected by constitutive activation of stained for the proliferation marker, KI67, and LYVE1. Der- −/− Notch1 or Notch4 in all VEGF-A-treated HdLECs (Fig. 4e, mal LEC proliferation was similar between Notch4 and f), whereas ERK activation was reduced in N1IC and N4/ control mice (Fig. S6b, c). As NOTCH4 is also expressed Int-3 HdLECs treated with VEGF-C relative to controls by the blood vasculature [20, 24], we evaluated the underly- −/− (Fig. 4 g, h). Previous studies have shown that Notch signal- ing dermal blood vascular network in E14.5 Notch4 and +/− ing alters the expression of the VEGF-A and VEGF-C recep- Notch4 embryos. Consistent with prior studies, the den- tors, VEGFR2 and VEGFR3 [9, 12, 18, 41]. The reduction sity and branching of the blood vasculature were unaffected in AKT and ERK activation downstream of VEGFs may be in the Notch4 nulls (Fig. S7) [24, 42]. secondary to Notch signaling effects on VEGFR expression. We next evaluated the dorsal dermal lymphatic phenotype −/− Therefore, the expression of Vegfr2 and Vegfr3 was deter- at E16.5 in wild-type and Notch4 embryos. The lymphatic mined in HdLECs with N1IC and N4/Int-3 by qRT-PCR. fronts had reached the midline and merged in both wild-type −/− Both N1IC and N4/Int-3 downregulated Vegfr2, while they and Notch4 embryos. Analysis of the lymphatic plexus induced Vegfr3 (Fig. S4b). Taken together, the data suggests revealed that it was disorganized with an increase in the that the decrease in AKT activation by VEGF-A and ERK distance between branch-points, while there was no differ - −/− activation by VEGF-C may be due to reduced VEGFR2 lev- ence in the mean vessel caliber in Notch4 dermis relative els, and not VEGFR3. to controls (Fig. 5g–i). Lymphatic valves were observed in both mutants and controls. A recent study has shown that Embryonic dermal lymphangiogenic defects reduced branching in the dermal lymphatic plexus was asso- in Notch4 mutant mice ciated with an increase in blunt-end sprouts due to reduced VEGFR3 signaling, which in turn led to a less branched A  prior report has shown that loss of LEC Notch1 in network [43]. In contrast, increased VEGFR3 signaling embryos leads to increased LEC proliferation and tip cells was associated with reduced blunt-ended sprouts and a [14]. To determine the role of Notch4 in embryonic lym- more densely branched network [44]. Since we observed −/− phangiogenesis, we evaluated Notch4 mice and compared a decrease in the dermal lymphatic branching in the E16.5 +/− −/− their lymphatic phenotype to that of Notch4 and wild-type Notch4 embryos, we assessed the sprout phenotype at littermates. To confirm that NOTCH4 protein is absent in E14.5 during active lymphangiogenesis. The sprouts of con- −/− the Notch4 embryos, western blots using lysates collected trol embryos uniformly expressed LYVE1 and were elon- from E14.5 embryos and staining of P4 dermal tissue with gated with numerous filopodia consistent with a lymphangi- an antibody against the intracellular domain of NOTCH4 ogenic phenotype (Fig. S8). In contrast, LECs in sprouts −/− were performed. As compared to wild-type littermates, in the Notch4 lymphatic vasculature were often rounded −/− NOTCH4 expression was absent in Notch4 tissues (Fig. with reduced and blunted filopodia. S5a–c). NOTCH1 expression determined by immunostain- Thus, Notch4 mutant mice had a distinct dermal lym- ing of P4 dermis was unaffected (Fig. S5d). phatic phenotype from that observed in mice with LEC 1 3 Angiogenesis (2022) 25:205–224 215 Fig. 5 Loss of Notch4 altered embryonic dermal lymphangiogenesis. lymphangiogenic front and in the maturing plexus. Data presented −/− a–f Dermal lymphatic phenotype was determined for E14.5 Notch4 ± s.e.m. One-way ANOVA: p = 0.019, t test: *p < 0.03, front analy- −/− +/− +/− +/− −/− (N4 ), Notch4 (N4 ) and wild-type (WT) wholemounts stained sis- WT (n = 7), N4 (n=13), N4 (n = 7), plexus analysis - WT −/− +/− +/− −/− for LYVE1. a Representative image of N4 and N4 dermal who- (n = 6), N4 (n = 9), N4 (n = 8).  f  Representative image of the −/− +/− lemounts. Red dotted line marks leading edge of lymphangiogenic E14.5 maturing lymphatic plexus in N4 and N4 littermates. front. Scale bars, 100 μm.  b  Quantification of the distance between Scale bars, 100 μm. g–i Dermal lymphatic phenotype determined for +/− −/− −/− migration fronts, normalized to the N4 controls. Data presented E16.5 Notch4 (N4 ) and wild-type (WT) wholemounts stained −4 ± s.e.m. One-way ANOVA: p = 0.014, t test: *p < 3 × 10 , **p < for VEGFR3. g Representative image of the E16.5 dermal lymphatic +/− −/− −/− 0.001. wt (n = 5), N4 (n = 13), N4 (n = 4).  c  Quantification plexus in N4 and WT littermates. Red arrowheads mark lym- of the number of lymphangiogenic sprouts normalized by length phatic valves. Scale bars, 100 μm; h quantification of mean distance of the front. Data presented ± s.e.m. One-way ANOVA: p = 0.023, between branch-points. Data presented ± s.e.m. t test: *p < 0.03 WT +/− −/− −/− t test: *p < 0.02. WT (n = 6), N4 (n = 11), N4 (n = 6).  d  Quan- (n = 4), N4 (n = 6).  i  Quantification of the average vessel caliber tification of the sprout length from the migration front to first of the dermal lymphatic plexus at E16.5. Data presented ± s.e.m. +/− −/− branch-point. Data presented ± s.e.m. WT (n = 7), N4 (n = 8), WT (n = 6), N4 (n = 8) −/− N4 (n = 6).  e  Quantification of the average vessel caliber at the 1 3 216 Angiogenesis (2022) 25:205–224 Fig. 6 Notch signaling inhibited LEC migration. a Conflu- ent N1IC-, N4/Int-3-, Hey1-, Hey2-, or GFP-expressing HdLECs were scratched and representative images for 0 and 25 h presented. Scale bars, 2.5 μm. b Quantification of percent open wound area at 0, 8, and 25 h. Data presented ± s.e.m. two-way-ANOVA: p < 0.0012, t test: *p < 0.002 N1IC or N4/Int-3 vs. GFP at 8 and 25 h, *p < 0.002 N4/ Int-3 vs. N1IC at 25 h. c Con- fluent N1IC-, N4/Int-3-, or GFP-expressing HdLECs were treated with mitomycin C and scratched and representative images for 0 and 24 h presented. Scale bars, 100 μm. d Quanti- fication of percent open wound area at 0, 4, 8, and 24 h. Data presented ± s.e.m. two-way ANOVA: p < 0.0001, Dunnett’s multiple comparison test *p < 0.0001 N1IC or N4/Int-3 vs. GFP. t test: *p < 0.0001 N4/ Int-3 vs. N1IC Notch1 deletion [14]. Rather than increased vessel diameter N1IC and N4/Int-3 expression inhibited LEC migration due to increased proliferation and branching due to increased (Fig. 6a, b). We next evaluated the effect of overexpress- sprouting lymphangiogenesis, the embryonic dermal lym- ing the downstream Notch effectors HEY1 and HEY2 phatics in Notch4 nulls had increased front closure, an early on HdLEC migration (Fig. 6a, b). Ectopic expression of decrease in the lymphatic vessel caliber, and reduced branch- HEY1 and HEY2 suppressed migration relative to con- ing without a change in LEC proliferation. trol HdLECs. Further analysis revealed that N4/Int-3 was a significantly stronger inhibitor of LEC migration than NOTCH4 activation preferentially inhibited LEC N1IC at 25 h. To insure the difference in migration was migration not due to changes in LEC proliferation, N1IC-, N4/Int- 3- and GFP-expressing HdLECs were treated with mito- To determine the effects of Notch1 and Notch4 signal acti- mycin C and LEC migration determined. Similar to the vation on LEC migration, a monolayer-wounding assay initial migration assay, both N1IC and N4/Int-3 suppressed was performed using HdLECs expressing either N1IC or HdLEC migration with N4/Int-3 suppressing migration N4/Int-3. Relative to control GFP-expressing HdLEC, both significantly more than N1IC at 24 h (Fig.  6c, d). 1 3 Angiogenesis (2022) 25:205–224 217 −/− LEC In contrast to the Notch4 phenotype, DNMAML der- Inhibition of lymphatic endothelial canonical Notch signaling increased dermal lymphatic vessel density mal lymphatics were dilated relative to control littermates LEC (Fig. 8c). One out of 5 DNMAML embryo lymphatics Notch4 has been shown to signal via RBPjκ-dependent was leaky (Fig. 8d), which was not observed in controls or −/− Notch4 lymphangiographies. These data demonstrate (canonical) and RBPjκ-independent (non-canonical) downstream pathways [45–47]. To determine the effects of that the Notch4 null and mice with a loss of LEC RBPjκ- dependent Notch signaling have distinct phenotypes at LEC specific loss of canonical Notch signaling on dermal T2 lymphangiogenesis, we used the inducible Prox1CreER E17.5, as well as at E14.5 (Figs. 5 and 7). driver to induce expression of a DNMAML transgene [26]. −/− DNMAML encodes a dominant negative form of Mammalian Canonical Notch signaling is unaffected in Notch4 embryonic dermal lymphangiogenesis Mastermind-like 1 (MAML1) that binds the NOTCH/RBPjκ complex to form an inactive complex and blocks the recruit- T2 As we observed a difference between the embryonic der - ment of transcriptional co-activators. Prox1CreER mice fl/fl −/− LEC were crossed with DNMAML mice to generate Prox1CreE mal lymphatic phenotypes of Notch4 and DNMAML T2 fl/+ LEC fl/+ mutants, we evaluated canonical Notch signaling by intro- R ;DNMAML embryos (DNMAML ) and DNMAML control littermates. To circumvent effects on early lymphatic ducing the NVR and ProxTom alleles into the Notch4 null background. Loss of Notch4 did not change canonical Notch specification caused by loss of Notch signaling in LECs [18], tamoxifen was administered to pregnant females at E12.5, signaling at the lymphangiogenic vascular front, nor the maturing lymphatic plexus (Fig. 9, S6d). This data suggested just as sprouting lymphangiogenesis begins, and the dermal lymphatic phenotype analyzed at E14.5. Unlike the Notch4 that Notch4 is not necessary for canonical Notch signaling in the embryonic dermal lymphatics. nulls, the closure of the migration fronts was the same LEC between DNMAML and control (Fig. 7a, b). The number of sprouts along the migrating front and the length of the sprout to the first branch-point were similar between mutants Discussion and controls (Fig. 7c, d). In contrast, the lymphatic density LEC was nearly 25% greater in the DNMAML compared to NOTCH1 and NOTCH4 are expressed and Notch signal- LEC ing active in the embryonic and early postnatal dermal lym- controls (Fig. 7e). The increase in the DNMAML dermal lymphatic density correlated with an enlargement of the phatic vasculature [12, 14], suggesting a role for both Notch proteins in embryonic lymphangiogenesis. Loss of Notch4 lymphatic vessel caliber at the lymphangiogenic front and in the maturing plexus (Fig. 7f, g). As compared to controls, was shown to exacerbate the Notch1 null embryonic blood LEC vascular phenotype, suggesting that Notch1 and Notch4 DNMAML dermal lymphatics had an increase in the num- ber Prox1+/LYVE1+ LECs (Fig. 7h), while branching in the have overlapping functions in the blood endothelium [24]. In contrast to the blood endothelium, we found that loss of mature plexus was unaffected (Fig.  7i). The increase in vas- cular density was specific to the lymphatics as blood vessel Notch4 led to a distinct embryonic dermal lymphangiogenic LEC phenotype, than that observed in mice with LEC deletion density was unchanged in DNMAML mutants (Fig. S9). Thus, we found that inhibition of RBPjκ-dependent Notch of Notch1 [14], or inhibition of canonical Notch signaling, presented here. At E14.5, Notch4 null embryos displayed an signaling resulted in increased lymphatic vessel density and caliber associated with an increase in LECs, suggesting that increase in the closure of the lymphangiogenic fronts to the midline, reduced vessel caliber in the maturing plexus, and canonical Notch signaling suppressed LEC proliferation in the embryonic dermal lymphatics. an increase in blunt-ended sprouts, while LEC proliferation was unaffected. By E16.5, the dermal lymphatic plexus in −/− LEC −/− Notch4 and DNMAML embryos display distinct Notch4 has reduced branching and tortuous lymphatic vessels, which may be secondary to the increase in blunt- lymphatic phenotypes at E17.5 ended sprouts at E14.5. In cultured LECs, constitutive acti- vation of Notch4 was a stronger inhibitor of migration than To assess the functionality and patterning of the dermal lymphatics, lymphangiographies were performed on E17.5 Notch1 activation and induced a subset of lymphangiogenic −/− LEC genes. In contrast, loss of LEC Notch1 at E10.5 increased Notch4 , DNMAML , and control littermates. Dye was injected within the dermis in the periorbital region embryonic dermal lymphatic density, due to increased LEC proliferation and decreased LEC apoptosis [14]. Similar and uptake by the lymphatics assessed after 1  min. The −/− Notch4 lymphatic plexus had reduced branching with to the Notch1 LEC knockout, we demonstrate that LEC expression of DNMAML, which inhibits canonical Notch/ tortuous vessels relative to the more uniform lymphatics of −/− controls (Fig.  8a). One of the 8 Notch4 embryos ana- RbpJκ signaling, increased the dermal lymphatic vascular density consistent with an increase in LEC proliferation and lyzed had blood-filled dermal lymphatics at E17.5 (Fig.  8b). 1 3 218 Angiogenesis (2022) 25:205–224 viability. Distinct functions for Notch1 and Notch4 have promote vessel maturation while having no effect on vascu- been described for endothelial progenitor cells, where Dll4 lar density [49]. Taken together, we propose that Notch1 and signaling via Notch4 specifically induced EphrinB2 and Notch4 signal dynamically to regulate lymphangiogenesis increased proliferation and migration of cultured cells [48]. and control migration and branching, versus proliferation More recently, it was proposed that endothelial Dll4/Notch1 and cell viability by distinct mechanisms. signaling induces Hey2 to suppress proliferation and tip cell Our studies suggest that Dll4 signaling via Notch1 and formation, while Jag1 activates Notch4 to induce Hey1 and Notch4 have overlapping and unique transcriptional targets 1 3 Angiogenesis (2022) 25:205–224 219 ◂Fig. 7 Loss of canonical Notch signaling in LECs increased der- further studies are necessary to understand the complexity T2 fl/fl mal lymphatic density. Prox1CreER and DNMAML mice were of Notch1 and Notch4 signaling in LECs. crossed, tamoxifen administered at E12.5 and dorsal dermis analyzed Prior studies have shown that VEGF-C induces DLL4 in at E14.5. a  Representative images of LYVE1 staining of Prox1Cre T2 fl/+ LEC fl/+ LECs leading to Notch activation [10]. We expanded these ER ;DNMAML (DNMAML ) and DNMAML (control) der- mis. Red dotted line marks leading edge of lymphatic fronts. Scale studies to understand the role of time and specific VEGFRs C156S bars, 1000 μm.  b  Quantification of the distance between migration in this process. Using VEGF-C which specifically binds fl/  + fronts, normalized to the  DNMAML controls. Data presented VEGFR3 and VEGF-A which binds VEGFR2, we found LEC ± s.e.m. Control (n = 7), DNMAML (n = 9).  c  Quantification that VEGFR2 signaling was a stronger and faster inducer of the number of lymphangiogenic sprouts normalized by length LEC of the front. Data presented ± s.e.m. Control (n = 8), DNMAML of Dll4/Notch signaling than VEGFR3 signaling. We also (n = 6).  d  Quantification of the distance between migration fronts, discovered that the induction of the Hes and Hey genes was fl/  + normalized to the DNMAML controls. Data presented ± s.e.m. time- and VEGF-dependent in HdLECs. VEGF-A induced LEC Control (n = 8), DNMAML (n = 6).  e  Quantification of average Hes1 at 1 h which persisted until 5 h, while significant Hey1 LYVE1+ vessel density normalized by area. Data presented rela- tive to control ± s.e.m. t test *p < 0.002, control (n = 7), DNMAML- upregulation was not observed until 5 h. In LECs, VEGF- LEC LEC (n = 9).  f  LYVE1, CD31, and PROX1 staining of DNMAML A was a stronger inducer of Hey1, while VEGF-C induced mutant and control dermal wholemounts. Images represent low Hey2. This differential response of Hes and Hey genes to (left) and high (middle) magnification of the lymphangiogenic front VEGF-A and VEGF-C was specific to LECs, as VEGF-C and the maturing plexus (right). Scale bars, 500 μm (left), 100 μm (middle, right).  g  Quantification of the average vessel caliber at the had no effect on Dll4, or Notch gene expression in HUVEC. lymphangiogenic front and in the maturing plexus. Data presented Our studies also suggest that additional Hes and Hey gene LEC ± s.e.m. t test: *p < 0.04, **p < 0.01, Control (n = 8), DNMAML family members than those studied in the blood vasculature, (n = 6).  h  Quantification of the number of PROX1+/LYVE1+ LECs Hes4 and HeyL, have a role in transmitting Notch signaling per field (pf). Data presented as ± s.d. ***p < 0.001. Control (n = 8), LEC DNMAML (n = 6). i) Quantification of the average number of in LECs. branch-points normalized to unit of vessel length. Data presented ± Our studies also revealed that Notch1 and Notch4 differ - LEC s.e.m. control (n = 3), DNMAML (n = 5) entially altered signaling downstream of VEGF-A/VEGFR and VEGF-C/VEGFR. Constitutive activation of Notch4 in HdLECs. Constitutive Notch1 and Notch4 activation blocked AKT activation by VEGF-A, and ERK activation by in LECs both induced the expression of Notch effectors VEGF-C, whereas Notch1 activity only modestly suppressed (Hes1, Hes4, Hey1, Hey2, HeyL) and lymphangiogenic VEGF-C activation of ERK signaling. Together, these data genes, such as EphrinB2, and Cx37 [50], as well as down- suggest that Notch4 has a role in modulating VEGFR2 and regulated essential genes in lymphangiogenesis, Podopla- VEGFR3 signaling down the PI3K/AKT and RAS/MAPK nin, and Prox1. In contrast, Hes5 expression was induced pathway. by N1IC, and not N4/Int-3. DLL4/Notch-induced genes The dermal lymphatic phenotypes were distinct between −/− LEC involved in chemokine signaling were also differentially Notch4 and DNMAML , suggesting Notch4 signals at regulated by Notch1 and Notch4. Expression of Cxcr4 was least in part via a non-canonical pathway. Notch4 has been induced by Notch1 activation, but suppressed by Notch4 shown to signal via canonical (RBPjκ-dependent) and non- signaling. In LECs, CXCR4 signaling promoted wound- canonical (RBPjκ-independent) Notch pathways in multiple induced and VEGF-C driven lymphangiogenesis in vivo, cells types [45–47]. In endothelial cells, Notch4 activation while in vitro it induced chemokine-driven LEC migration blocked LPS-induced apoptosis via RBPjκ-independent [31, 38]. Thus, it is possible that loss of Notch4 led to an upregulation of Bcl2 [45]. In mice, NOTCH4 activation increase in CXCR4 expression, which in turn increased the in the ductal epithelium required RBPjκ for physiological LEC migration toward the midline. N4/Int-3 was also a sig- alveolar development, but not for breast cancer development, nificantly stronger inducer of Ccl2 and Ackr3, than N1IC. suggesting Notch4 functions via both canonical and non- LEC-derived CCL2 has been shown to promote the recruit- canonical pathway in the breast endothelium [46, 47]. We ment of monocytes and macrophages to sites of lymphangi- observed that canonical Notch signaling was unchanged in ogenesis, where they deliver VEGF-A and VEGF-C [33, the embryonic dermal LECs in Notch4 nulls suggesting that 36]. In murine lymphatic development, ACKR3 functions to the Notch4 dermal lymphatic phenotype did not occur via a suppress LEC growth by scavenging adrenomedullin (ADM) RBPjκ-dependent mechanism. However, it is possible that [35]. Interestingly, we found that Dll4/Notch signaling also the variable phenotypes are due to differences in the pen- suppressed Adm expression, suggesting that Dll4/Notch4 etrance of global Notch4 loss versus a tamoxifen-induced signaling suppresses ADM signaling to regulate lymphatic cell mosaic expression of DNMAML in LECs. development. While our gene expression studies begin to An increase in the closure of the two lymphangiogenic elucidate some of the mechanisms by which Dll4/Notch1 fronts was observed in Notch4 mutants that correlated with and Dll4/Notch4 signaling regulates lymphangiogenesis, reduced vessel caliber in the absence of a change in LEC proliferation. This phenotype is consistent with an increase 1 3 220 Angiogenesis (2022) 25:205–224 Fig. 8 Loss of Notch4 and dele- tion of LEC canonical Notch signaling resulted in distinct lymphatic phenotypes at E17.5. a, b Lymphangiography of E17.5 wild-type (n = 6) and −/− Notch4 (n = 8) embryos. Representative images of wild- −/− type and Notch4 embryos. Boxed area enlarged to the −/− right. b Notch4 embryos with blood-filled dermal lymphatics (white arrowheads). Boxed area enlarged to the T2 right. c, d Prox1CreER and fl/fl DNMAML mice were crossed and tamoxifen administered at E12.5 and lymphangiog- raphy performed at E17.5. Control (n = 12), DNMAML- LEC (n = 5). c Representative fl/− images of DNMAML control LEC and DNMAML embryos. Boxed area enlarged to the LEC right. d DNMAML embryo with leaking dermal lymphatic vessels (red arrowheads). Boxed area enlarged below in LEC migration toward the midline. In HdLECs, ectopic NOTCH4 peptide that suppresses Notch1 signaling by Notch4 activation inhibited LEC migration significantly functioning as a ligand trap [52]. However, a loss of canoni- more than Notch1 activation. This inhibition of LEC migra- cal Notch signaling was not observed in the lymphatics of tion by Notch4 may occur via non-canonical Notch signal- Notch4 mutant mice, which would be predicted if Notch1 ing, as expression of DNMAML, an inhibitor of canonical signaling was inhibited in the model. Moreover, the Notch4 Notch signaling, did not affect the closure of the lymphangi- mutant dermal lymphatic phenotype is distinct from that ogenic fronts. Notch4 may suppress LEC migration via its observed in mice with Notch1 deleted in the LECs [14], as LEC interactions with Wnt/β-catenin signaling. Non-canonical well as the DNMAML mice. The dermal lymphatic phe- Notch4 signaling has been shown to antagonize Wnt/β- notype however may be due to loss of Notch4 in non-LECs, catenin signaling in stem and progenitor cells [29, 51]. In such as macrophages, and a conditional Notch4 allele needs LECs, loss of β-catenin signaling reduced LEC migration to be developed to better understand the cell type specific toward the midline and increased dermal lymphatic vessel requirement for NOTCH4 in lymphatic development. caliber [29, 51], phenotypes opposite to that observed in Together with published data, our studies suggest that −/− Notch4 embryos, suggesting that Notch4 via a non-canon- Notch1 and Notch4 function distinctly in embryonic dermal ical signaling suppresses LEC migration. lymphangiogenesis via a RBPjκ-dependent and -independ- Western analysis of embryo lysates and immunostaining ent pathways. We propose that Dll4/Notch1 signaling via of tissue sections using an antibody against the cytoplas- a canonical pathway suppresses LEC proliferation, while mic domain of NOTCH4 demonstrated a loss of NOTCH4 Notch4 signaling suppresses LEC migration and branching, expression in the Notch4 nulls. It has been suggested that possibly via a RBPjκ-dependent mechanism. Further studies this Notch4 null line expresses a truncated extracellular into the mechanistic interaction between Notch1 and Notch4 1 3 Angiogenesis (2022) 25:205–224 221 Fig. 9 Canonical Notch signaling in LECs is unaltered in the Notch4 mutant dermal lymphatics. +/− +/ +/− Notch4 ;ProxTOM ;NVR males were bred with +/− +/ +/− Notch4 ;ProxTom ;NVR females, and E14.5 ProxTom; NVR and −/− Notch4 ;ProxTom;NVR dermal wholemounts stained for LYVE1. a Representative images of blunted-ended and spiky-ended sprouts at the lym- phangiogenic front of wild-type −/− (WT) and Notch4 dermis. b Representative images of branch-points with Notch sign- −/− aling in WT and Notch4 in the maturing lymphatic vascular plexus. Scale bars, 50 μm Acknowledgements The authors thank Valeriya Borisenko and Marina in LECs and lymphatic development and homeostasis are Vorontchikhina for technical assistance, June Wu for critical reading necessary, as a number of therapeutics that are pan-Notch fl/fl of the manuscript, and Warren Pear (DNMAML ), Tom Gridley inhibitors or target specific receptors or ligands are currently −/− T2 (Notch4 ), Guillermo Oliver (Prox1CreER ), and Hong Young in clinical trials or the research pipeline for use in the clinic. Kwon (Prox1-tdTomato) for providing mice. Supplementary Information The online version contains supplemen- Author contributions AM and MKU share first authorship. AM, MKU, tary material available at https://doi. or g/10. 1007/ s10456- 021- 09822-5 . YM, JKK, CJS contributed to the study conception and design. Mate- rial preparation, data collection, and analysis were performed by AM, 1 3 222 Angiogenesis (2022) 25:205–224 MKU, GSD, BS, JMJ, AM, SWY, JDM, CK, MG, GR, CJS. The first and reviews in molecular. Cell Dev Biol 26(3):225–234. https:// draft of the manuscript was written by AM, MKU, and CJS and revised doi. org/ 10. 1002/ bies. 20004 by CJS. 6. Hellstrom M, Phng LK, Hofmann JJ, Wallgard E, Coultas L, Lindblom P, Alva J, Nilsson AK, Karlsson L, Gaiano N, Yoon K, Rossant J, Iruela-Arispe ML, Kalen M, Gerhardt H, Betsholtz C Funding This study was funded by the NIH/NCI (R01CA136673; (2007) Dll4 signalling through Notch1 regulates formation of tip CJS, JKK), NIH/NIDDK (R01 R01DK107633; CJS), NIH/ cells during angiogenesis. Nature 445(7129):776–780. doi:https:// NHLBI (RO1HL112626; JKK), the DOD pre-doctoral fellowship doi. org/ 10. 1038/ natur e05571 (W81XWH-10-1-0304; MKU), and the Lipedema Foundation (CJS). 7. Lobov IB, Renard RA, Papadopoulos N, Gale NW, Thurston G, These studies used the resources of the Herbert Irving Comprehen- Yancopoulos GD, Wiegand SJ (2007) Delta-like ligand 4 (Dll4) is sive Cancer Center Flow Cytometry Shared Resources funded in part induced by VEGF as a negative regulator of angiogenic sprouting. through Center Grant P30CA013696. Proc Natl Acad Sci USA 104(9):3219–3224. doi:https:// doi. org/ 10. 1073/ pnas. 06112 06104 Declarations 8. Suchting S, Freitas C, le Noble F, Benedito R, Breant C, Duarte A, Eichmann A (2007) The Notch ligand Delta-like 4 negatively Conflict of interest Jan Kitajewski has received research funding from regulates endothelial tip cell formation and vessel branching. Eisai Pharmaceuticals (CU12-3625 and UICID#084028 Eisai Ltd. Re- Proc Natl Acad Sci USA 104(9):3225–3230. doi:https:// doi. org/ search Collaborative Agreements). All other authors declare that they 10. 1073/ pnas. 06111 77104 have no conflict of interests. 9. Tammela T, Zarkada G, Wallgard E, Murtomaki A, Suchting S, Wirzenius M, Waltari M, Hellstrom M, Schomber T, Peltonen Ethical approval Isolation of HUVEC and HdLEC from anonymous R, Freitas C, Duarte A, Isoniemi H, Laakkonen P, Christofori G, discarded specimens and received IRB exempt status by Columbia Yla-Herttuala S, Shibuya M, Pytowski B, Eichmann A, Betsholtz University IRB (AAAA7338). All procedures performed in studies C, Alitalo K (2008) Blocking VEGFR-3 suppresses angiogenic involving human participants were in accordance with the ethical stand- sprouting and vascular network formation. Nature 454(7204):656– ards of the institutional and/or national research committee and with 660. doi:https://doi. org/ 10. 1038/ natur e07083 the 1964 Helsinki declaration and its later amendments or comparable 10. Bernier-Latmani J, Cisarovsky C, Demir CS, Bruand M, Jaquet ethical standards. Mouse studies were approved by Columbia Univer- M, Davanture S, Ragusa S, Siegert S, Dormond O, Benedito sity IACUC (AC-AAAE2653, AC-AAAD0577, AC-AAAP9603, AC- R, Radtke F, Luther SA, Petrova TV (2015) DLL4 promotes AAAP0452, AC-AABB9551). All procedures performed in studies continuous adult intestinal lacteal regeneration and dietary fat involving animals were in accordance with the ethical standards of the transport. J Clin Investig 125(12):4572–4586. doi:https:// doi. institution or practice at which the studies were conducted. org/10. 1172/ JCI82 045 11. Zheng W, Tammela T, Yamamoto M, Anisimov A, Holopainen T, Kaijalainen S, Karpanen T, Lehti K, Yla-Herttuala S, Alitalo Open Access This article is licensed under a Creative Commons Attri- K (2011) Notch restricts lymphatic vessel sprouting induced by bution 4.0 International License, which permits use, sharing, adapta- vascular endothelial growth factor. Blood 118(4):1154–1162. tion, distribution and reproduction in any medium or format, as long doi:https://doi. org/10. 1182/ blood- 2010- 11- 317800 as you give appropriate credit to the original author(s) and the source, 12. Shawber CJ, Funahashi Y, Francisco E, Vorontchikhina M, provide a link to the Creative Commons licence, and indicate if changes Kitamura Y, Stowell SA, Borisenko V, Feirt N, Podgrabinska were made. The images or other third party material in this article are S, Shiraishi K, Chawengsaksophak K, Rossant J, Accili D, included in the article's Creative Commons licence, unless indicated Skobe M, Kitajewski J (2007) Notch alters VEGF responsive- otherwise in a credit line to the material. If material is not included in ness in human and murine endothelial cells by direct regulation the article's Creative Commons licence and your intended use is not of VEGFR-3 expression. J Clin Investig 117(11):3369–3382. permitted by statutory regulation or exceeds the permitted use, you will doi:https://doi. org/10. 1172/ JCI24 311 need to obtain permission directly from the copyright holder. To view a 13. Niessen K, Zhang G, Ridgway JB, Chen H, Kolumam G, copy of this licence, visit http://cr eativ ecommons. or g/licen ses/ b y/4.0/ . Siebel CW, Yan M (2011) The Notch1-Dll4 signaling path- way regulates mouse postnatal lymphatic development. Blood 118(7):1989 –1997. doi: https:// doi. org/ 10. 1182/ blood- 2010- 11- 319129 References 14. Fatima A, Culver A, Culver F, Liu T, Dietz WH, Thomson BR, Hadjantonakis AK, Quaggin SE, Kume T (2014) Murine Notch1 is 1. James JM, Nalbandian A, Mukouyama YS (2013) TGFbeta signal- required for lymphatic vascular morphogenesis during development. ing is required for sprouting lymphangiogenesis during lymphatic Dev Dyn 243(7):957–964. https:// doi. org/ 10. 1002/ dvdy.24129 network development in the skin. Development 140(18):3903– 15. Geng X, Yanagida K, Akwii RG, Choi D, Chen L, Ho Y, Cha 3914. doi:https://doi. org/ 10. 1242/ dev. 095026 B, Mahamud MR, Berman de Ruiz K, Ichise H, Chen H, Wythe 2. D’Souza B, Meloty-Kapella L, Weinmaster G (2010) Canonical JD, Mikelis CM, Hla T, Srinivasan RS (2020) S1PR1 regulates and non-canonical Notch ligands. Curr Top Dev Biol 92:73–129. the quiescence of lymphatic vessels by inhibiting laminar shear doi:https:// doi. org/ 10. 1016/ S0070- 2153(10) 92003-6 stress-dependent VEGF-C signaling. JCI Insight 5(14):e137652 3. Baeten JT, Lilly B (2017) Notch signaling in vascular smooth 16. Funahashi Y, Shawber CJ, Sharma A, Kanamaru E, Choi YK, muscle cells. Adv Pharmacol 78:351–382. https:// doi. org/ 10. Kitajewski J (2011) Notch modulates VEGF action in endothe- 1016/ bs. apha. 2016. 07. 002 lial cells by inducing matrix metalloprotease activity. Vasc Cell 4. Li JL, Harris AL (2009) Crosstalk of VEGF and Notch pathways 3(1):2. https:// doi. org/ 10. 1186/ 2045- 824X-3-2 in tumour angiogenesis: therapeutic implications. Front Biosci 17. Funahashi Y, Shawber CJ, Vorontchikhina M, Sharma A, Outtz 14:3094–3110 HH, Kitajewski J (2010) Notch regulates the angiogenic response 5. Shawber CJ, Kitajewski J (2004) Notch function in the vascula- via induction of VEGFR-1. J Angiogenesis Res 2(1):3. https://doi. ture: insights from zebrafish, mouse and man. BioEssays: news org/ 10. 1186/ 2040- 2384-2-3 1 3 Angiogenesis (2022) 25:205–224 223 18. Murtomaki A, Uh MK, Choi YK, Kitajewski C, Borisenko V, 33. Farnsworth RH, Karnezis T, Maciburko SJ, Mueller SN, Stacker SA Kitajewski J, Shawber CJ (2013) Notch1 functions as a nega- (2019) The interplay between lymphatic vessels and chemokines. tive regulator of lymphatic endothelial cell differentiation in Front Immunol 10:518. https://doi. or g/10. 3389/ fimmu. 2019.00518 the venous endothelium. Development 140(11):2365–2376. 34. Kanady JD, Dellinger MT, Munger SJ, Witte MH, Simon AM doi:https:// doi. org/ 10. 1242/ dev. 083865 (2011) Connexin37 and connexin43 deficiencies in mice disrupt 19. Tung JJ, Hobert O, Berryman M, Kitajewski J (2009) Chloride lymphatic valve development and result in lymphatic disorders intracellular channel 4 is involved in endothelial proliferation and including lymphedema and chylothorax. Dev Biol 354(2):253–266. morphogenesis in vitro. Angiogenesis 12(3):209–220. doi:https:// https:// doi. org/ 10. 1016/j. ydbio. 2011. 04. 004 doi. org/ 10. 1007/ s10456-009- 9139-3 35. Klein KR, Karpinich NO, Espenschied ST, Willcockson HH, Dun- 20. Uyttendaele H, Marazzi G, Wu G, Yan Q, Sassoon D, Kitajew- worth WP, Hoopes SL, Kushner EJ, Bautch VL, Caron KM (2014) ski J (1996) Notch4/Int-3, a mammary proto-oncogene, is an Decoy receptor CXCR7 modulates adrenomedullin-mediated car- endothelial cell-specific mammalian Notch gene. Development diac and lymphatic vascular development. Dev Cell 30(5):528–540. 122(7):2251–2259 https:// doi. org/ 10. 1016/j. devcel. 2014. 07. 012 21. Geback T, Schulz MM, Koumoutsakos P, Detmar M (2009) 36. Lee KM, Danuser R, Stein JV, Graham D, Nibbs RJ, Graham GJ TScratch: a novel and simple software tool for automated analysis (2014) The chemokine receptors ACKR2 and CCR2 reciprocally of monolayer wound healing assays. Biotechniques 46(4):265– regulate lymphatic vessel density. EMBO J 33(21):2564–2580. 274. doi:https://doi. org/ 10. 2144/ 00011 3083 doi:https:// doi. org/ 10. 15252/ embj. 20148 8887 22. Schmittgen TD, Livak KJ (2008) Analyzing real-time PCR data 37. Uchida Y, James JM, Suto F, Mukouyama YS (2015) Class 3 sema- by the comparative C(T) method. Nat Protoc 3(6):1101–1108. phorins negatively regulate dermal lymphatic network formation. https:// doi. org/ 10. 1038/ nprot. 2008. 73 Biol Open 4(9):1194–1205. doi:https:// doi.or g/10. 1242/bio. 012302 23. Shawber CJ, Das I, Francisco E, Kitajewski J (2003) Notch signal- 38. Zhuo W, Jia L, Song N, Lu XA, Ding Y, Wang X, Song X, Fu Y, Luo ing in primary endothelial cells. Ann N Y Acad Sci 995:162–170. Y (2012) The CXCL12-CXCR4 chemokine pathway: a novel axis doi:https:// doi. org/ 10. 1111/j. 1749- 6632. 2003. tb032 19.x regulates lymphangiogenesis. Clin Cancer Res 18(19):5387–5398. 24. Krebs LT, Xue Y, Norton CR, Shutter JR, Maguire M, Sundberg doi:https:// doi. org/ 10. 1158/ 1078- 0432. CCR- 12- 0708 JP, Gallahan D, Closson V, Kitajewski J, Callahan R, Smith GH, 39. Karpinich NO, Caron KM (2014) Apelin signaling: new G protein- Stark KL, Gridley T (2000) Notch signaling is essential for vas- coupled receptor pathway in lymphatic vascular development. Arte- cular morphogenesis in mice. Genes Dev 14(11):1343–1352 rioscler Thromb Vasc Biol 34(2):239–241. doi:https:// doi. org/ 10. 25. Srinivasan RS, Dillard ME, Lagutin OV, Lin FJ, Tsai S, Tsai MJ, 1161/ ATVBA HA. 113. 302905 Samokhvalov IM, Oliver G (2007) Lineage tracing demonstrates the 40. Lohela M, Saaristo A, Veikkola T, Alitalo K (2003) Lymphangi- venous origin of the mammalian lymphatic vasculature. Genes Dev ogenic growth factors, receptors and therapies. Thromb Haemost 21(19):2422–2432. doi:https://doi. org/10. 1101/ gad. 15884 07 90(2):167–184. doi:https://doi. org/ 10. 1160/ TH03- 04- 0200 26. Tu L, Fang TC, Artis D, Shestova O, Pross SE, Maillard I, Pear WS 41. Benedito R, Rocha SF, Woeste M, Zamykal M, Radtke F, Casano- (2005) Notch signaling is an important regulator of type 2 immunity. vas O, Duarte A, Pytowski B, Adams RH (2012) Notch-depend- J Exp Med 202(8):1037–1042. https://doi. or g/10. 1084/ jem. 20050 923 ent VEGFR3 upregulation allows angiogenesis without VEGF- 27. Nowotschin S, Xenopoulos P, Schrode N, Hadjantonakis AK (2013) VEGFR2 signalling. Nature 484(7392):110–114. doi:https:// doi. A bright single-cell resolution live imaging reporter of Notch signal- org/ 10. 1038/ natur e10908 ing in the mouse. BMC Dev Biol 13:15. doi:https:// doi.or g/10. 1186/ 42. Costa MJ, Wu X, Cuervo H, Srinivasan R, Bechis SK, Cheang E, 1471- 213X- 13- 15 Marjanovic O, Gridley T, Cvetic CA, Wang RA (2013) Notch4 is 28. Hong M, Jung E, Yang S, Jung W, Seong YJ, Park E, Bramos A, required for tumor onset and perfusion. Vasc Cell 5(1):7. https:// doi. Kim KE, Lee S, Daghlian G, Seo JI, Choi I, Choi IS, Koh CJ, Kob- org/ 10. 1186/ 2045- 824X-5-7 ielak A, Ying QL, Johnson M, Gardner D, Wong AK, Choi D, Hong 43. Zhang Y, Ulvmar MH, Stanczuk L, Martinez-Corral I, Frye M, Ali- YK (2016) Efficient assessment of developmental, surgical and talo K, Makinen T (2018) Heterogeneity in VEGFR3 levels drives pathological lymphangiogenesis using a lymphatic reporter mouse lymphatic vessel hyperplasia through cell-autonomous and non-cell- and its embryonic stem cells. PloS One 11(6):e0157126. https:// doi. autonomous mechanisms. Nat Commun 9(1):1296. doi:https:// doi. org/ 10. 1371/ journ al. pone. 01571 26 org/ 10. 1038/ s41467- 018- 03692-0 29. Cha B, Geng X, Mahamud MR, Fu J, Mukherjee A, Kim Y, Jho EH, 44. Martinez-Corral I, Zhang Y, Petkova M, Ortsater H, Sjoberg S, Cas- Kim TH, Kahn ML, Xia L, Dixon JB, Chen H, Srinivasan RS (2016) tillo SD, Brouillard P, Libbrecht L, Saur D, Graupera M, Alitalo Mechanotransduction activates canonical Wnt/beta-catenin signal- K, Boon L, Vikkula M, Makinen T (2020) Blockade of VEGF-C ing to promote lymphatic vascular patterning and the development of signaling inhibits lymphatic malformations driven by oncogenic lymphatic and lymphovenous valves. Genes Dev 30(12):1454–1469. PIK3CA mutation. Nat Commun 11(1):2869. doi:https://doi. org/ doi:https:// doi. org/ 10. 1101/ gad. 282400. 116 10. 1038/ s41467- 020- 16496-y 30. Vorontchikhina MA, Zimmermann RC, Shawber CJ, Tang H, Kita- 45. MacKenzie F, Duriez P, Wong F, Noseda M, Karsan A (2004) jewski J (2005) Unique patterns of Notch1, Notch4 and Jagged1 Notch4 inhibits endothelial apoptosis via RBP-Jkappa-dependent expression in ovarian vessels during folliculogenesis and corpus and -independent pathways. J Biol Chem 279(12):11657–11663. luteum formation. Gene Expr Patterns 5(5):701–709. doi:https:// doi:https:// doi. org/ 10. 1074/ jbc. M3121 02200 doi. org/ 10. 1016/j. modgep. 2005. 02. 001 46. Raafat A, Bargo S, McCurdy D, Callahan R (2017) The ANK 31. Du LL, Liu P (2016) CXCL12/CXCR4 axis regulates neovasculari- repeats of Notch-4/Int3 activate NF-kappaB canonical pathway in zation and lymphangiogenesis in sutured corneas in mice. Mol Med the absence of Rbpj and causes mammary tumorigenesis. Scientific Rep 13(6):4987–4994. doi:https://doi. or g/10. 3892/mmr .2016. 5179 reports 7(1):13690. doi:https://doi. or g/10. 1038/ s41598- 017- 13989-7 32. Abengozar MA, de Frutos S, Ferreiro S, Soriano J, Perez-Martinez 47. Raafat A, Lawson S, Bargo S, Klauzinska M, Strizzi L, Goldhar M, Olmeda D, Marenchino M, Canamero M, Ortega S, Megias D, AS, Buono K, Salomon D, Vonderhaar BK, Callahan R (2009) Rodriguez A, Martinez-Torrecuadrada JL (2012) Blocking ephrinB2 Rbpj conditional knockout reveals distinct functions of Notch4/ with highly specific antibodies inhibits angiogenesis, lymphangi- Int3 in mammary gland development and tumorigenesis. Oncogene ogenesis, and tumor growth. Blood 119(19):4565–4576. doi:https:// 28(2):219–230. doi:https://doi. org/ 10. 1038/ onc. 2008. 379 doi. org/ 10. 1182/ blood- 2011- 09- 380006 48. Liu X, Luo Q, Zheng Y, Liu X, Hu Y, Liu W, Luo M, Zhao Y, Zou L (2016) NOTCH4 signaling controls EFNB2-induced 1 3 224 Angiogenesis (2022) 25:205–224 endothelial progenitor cell dysfunction in preeclampsia. Reproduc- 51. Andersen P, Uosaki H, Shenje LT, Kwon C (2012) Non-canonical tion 152(1):47–55. doi:https://doi. org/10. 1530/ REP- 16- 0132 Notch signaling: emerging role and mechanism. Trends Cell Biol 49. Pedrosa AR, Trindade A, Fernandes AC, Carvalho C, Gigante J, 22(5):257–265. doi:https://doi. org/ 10. 1016/j. tcb. 2012. 02. 003 Tavares AT, Dieguez-Hurtado R, Yagita H, Adams RH, Duarte A 52. James AC, Szot JO, Iyer K, Major JA, Pursglove SE, Chapman G, (2015) Endothelial Jagged1 antagonizes Dll4 regulation of endothe- Dunwoodie SL (2014) Notch4 reveals a novel mechanism regulating lial branching and promotes vascular maturation downstream of Notch signal transduction. Biochim Biophys Acta 1843(7):1272– Dll4/Notch1. Arterioscler Thromb Vasc Biol 35(5):1134–1146. 1284. doi:https://doi. org/ 10. 1016/j. bbamcr. 2014. 03. 015 https:// doi. org/ 10. 1161/ ATVBA HA. 114. 304741 50. Murtomaki A, Uh MK, Kitajewski C, Zhao J, Nagasaki T, Shawber Publisher’s Note Springer Nature remains neutral with regard to CJ, Kitajewski J (2014) Notch signaling functions in lymphatic valve jurisdictional claims in published maps and institutional affiliations. formation. Development 141(12):2446–2451. doi:https://doi. or g/10. 1242/ dev. 101188 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Angiogenesis Springer Journals

Unique functions for Notch4 in murine embryonic lymphangiogenesis

Download PDF
20 pages

Loading next page...
 
/lp/springer-journals/unique-functions-for-notch4-in-murine-embryonic-lymphangiogenesis-21d3gIl751

References (56)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
ISSN
0969-6970
eISSN
1573-7209
DOI
10.1007/s10456-021-09822-5
Publisher site
See Article on Publisher Site

Abstract

In mice, embryonic dermal lymphatic development is well understood and used to study gene functions in lymphangiogenesis. Notch signaling is an evolutionarily conserved pathway that modulates cell fate decisions, which has been shown to both inhibit and promote dermal lymphangiogenesis. Here, we demonstrate distinct roles for Notch4 signaling versus canoni- cal Notch signaling in embryonic dermal lymphangiogenesis. Actively growing embryonic dermal lymphatics expressed NOTCH1, NOTCH4, and DLL4 which correlated with Notch activity. In lymphatic endothelial cells (LECs), DLL4 activation of Notch induced a subset of Notch effectors and lymphatic genes, which were distinctly regulated by Notch1 and Notch4 activation. Treatment of LECs with VEGF-A or VEGF-C upregulated Dll4 transcripts and differentially and temporally regulated the expression of Notch1 and Hes/Hey genes. Mice nullizygous for Notch4 had an increase in the closure of the lymphangiogenic fronts which correlated with reduced vessel caliber in the maturing lymphatic plexus at E14.5 and reduced branching at E16.5. Activation of Notch4 suppressed LEC migration in a wounding assay significantly more than Notch1, suggesting a dominant role for Notch4 in regulating LEC migration. Unlike Notch4 nulls, inhibition of canonical Notch signaling by expressing a dominant negative form of MAML1 (DNMAML) in Prox1+ LECs led to increased lymphatic density consistent with an increase in LEC proliferation, described for the loss of LEC Notch1. Moreover, loss of Notch4 did not affect LEC canonical Notch signaling. Thus, we propose that Notch4 signaling and canonical Notch signaling have distinct functions in the coordination of embryonic dermal lymphangiogenesis. Keywords Lymphangiogenesis · Notch · VEGF-C · Dermis Introduction Lymphangiogenesis is the process by which new lymphatic vessels sprout off pre-existing vessels. Sprouting of new lymphatic vessels requires coordinated lymphatic endothelial Ajit Muley and Minji Kim Uh have contributed equally to this work. * Carrie J. Shawber Wihuri Research Institute, Biomedicum Helsinki, cjs2002@cumc.columbia.edu Haartmaninkatu, 8, 00290 Helsinki, Finland Translational Cancer Medicine Program, Faculty Department of Obstetrics and Gynecology, Columbia of Medicine, Helsinki Institute of Life Science, University University Medical Center, New York, NY 10032, USA of Helsinki, FI‑00014 Helsinki, Finland Department of Pharmacology, Columbia University Medical Departments of Molecular Medicine and Experimental Center, New York, NY 10032, USA Medicine, Sapienza University, 00185 Rome, Italy Department of Physiology and Biophysics, University Department of Surgery, Columbia University Medical of Illinois Chicago, Chicago, IL 60612, USA Center, New York, NY 10032, USA Laboratory of Stem Cell and Neuro‑Vascular Biology, Cell and Developmental Biology Center, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD 20892, USA Vol.:(0123456789) 1 3 206 Angiogenesis (2022) 25:205–224 cell (LEC) proliferation, directional migration, and cell–cell lymphangiogenesis in the postnatal mouse ear [11]. In adhesion to form a properly patterned and functional net- embryonic dermal lymphangiogenesis, Notch1 deletion work. In murine dorsal skin, lymphangiogenesis begins at in LECs did not affect lymphatic branching, but increased embryonic day 12.5 (E12.5) at the side of the trunk and fol- lymphatic vessel caliber which was proposed to be second- lows dermal blood vessel development to meet at the midline ary to an increase in LEC proliferation and decreased LEC around E15.5 (Fig. S1a) [1]. Dermal lymphangiogenesis in apoptosis [14]. More recently, it was shown that loss of one mouse embryos is well characterized allowing for analysis copy of Dll4 was associated with reduced embryonic dermal of lymphatic endothelial signaling pathways, such as Notch. lymphangiogenesis in mice [15], a phenotype opposite to The Notch family of signaling proteins consists of four that seen in retinal angiogenesis [7, 8]. Additional studies cell surface receptors (NOTCH1-4) that are bound and acti- are needed to clarify the differences in the lymphangiogenic vated by membrane-bound ligands of the Delta-like (Dll1, phenotypes observed upon disruption of lymphatic endothe- 4) and Jagged (Jag1, 2) families expressed on neighboring lial Notch signaling. cells. Upon ligand activation, the extracellular domain of Here, we examined the roles for Notch4 and canonical NOTCH is released, which induces conformational changes Notch signaling in embryonic dermal lymphangiogenesis. that expose two proteolytic cleavage sites (TACE and We demonstrated that NOTCH1, NOTCH4, and DLL4 are γ-secretase/presenilin) that in turn releases the intracellular expressed, and Notch signaling active in embryonic der- cytoplasmic domain (NICD) from the cell surface [2]. In mal lymphatic endothelium. VEGF-A and VEGF-C sign- the canonical Notch signaling pathway, NICD transits to the aling differentially regulated Notch pathway gene expres- nucleus, binds the transcriptional repressor RBPjκ, where sion and activity in cultured LECs. Mice nullizygous for it recruits an activation complex including Mastermind- Notch4 displayed an embryonic dermal lymphangiogenic like (MAML) and HDACs, and activates RBPjκ-dependent phenotype characterized by increased LEC migration and transcription of Notch effectors, such as those in the HES/ reduced branching. In contrast, inhibition of canonical Notch Hey families. Notch also signals via a less well-understood signaling increased lymphatic vascular density consistent non-canonical RBPjκ-independent pathway that has been with an increase in LEC proliferation. Together, these data suggested to not require nuclear localization of NICD [2]. demonstrate that dermal lymphangiogenesis is dynami- During development of the blood vascular system, Notch cally regulated by Notch and requires both NOTCH1 and signaling is essential for arterial endothelial specification, NOTCH4 functions, as well as canonical and non-canonical vascular smooth muscle cell differentiation and viability, Notch signaling. and sprouting angiogenesis [3–5]. Studies of murine retinal angiogenesis have shown that VEGF-A, via activation of VEGFR2, upregulates DLL4 expression in the filopodia- Materials and methods extending tip cell located at the vascular front [4, 6–8]. Dll4 signals to the neighboring Notch-expressing stalk cell, where Cell culture/constructs Notch activation downregulates VEGFR2 and VEGFR3 expression and inhibits the tip cell phenotype. During reti- HeLa cells were maintained in 10% FBS DMEM. Human nal angiogenesis, inhibition of DLL4 or NOTCH1 leads to umbilical vein endothelial cells (HUVEC) were isolated a hypersprouting phenotype characterized by an increase in as previously described and maintained in EGM2 (Lonza) tip cells at the expense of the stalk cells, increased VEGFR2 [16, 17]. Neonatal human dermal lymphatic endothelial cells and VEGFR3 expression, and decreased vascular outgrowth (HdLECs) were either purchased (Promocell) or isolated as [6, 7, 9]. Although it has been shown that VEGF-C induces previously described [18] and maintained on fibronectin- DLL4 in cultured LECs [10, 11], the mechanisms by which coated plates in EGM2-MV2 (Lonza; complete medium) Notch regulates dermal lymphangiogenesis remain to be supplemented with 10 ng/mL VEGF-C (R&D). To activate elucidated. Notch signaling, HdLECs were lentivirally infected [19] We previously demonstrated that NOTCH1 and NOTCH4 using pCCL.pkg.wpre vector encoding N1IC, N4/Int-3, are expressed in the postnatal maturing dermal lymphatics or GFP. N1IC encodes the constitutively active cytoplas- [12]. Studies of postnatal lymphangiogenesis have shown mic domain of NOTCH1. N4/Int-3 encodes an activated that pharmacological inhibition or genetic manipulation Notch4 allele generated by MMTV insertion [20]. Tran- of Dll4/Notch1 signaling can result in both increased and scripts and protein expression was confirmed by quantita- decreased lymphangiogenesis [11, 13]. Neutralizing anti- tive (q)RT-PCR and Western analyses of samples collected bodies against NOTCH1 or DLL4 suppressed lymphangi- post-infection. ogenesis in the postnatal mouse ear, tail dermis, and a wounding model [13]. In contrast, an inhibitory soluble DLL4 extracellular domain fused to FC (Dll4FC) stimulated 1 3 Angiogenesis (2022) 25:205–224 207 HdLEC assays DLL4‑ligand activation assay and mRNA sequencing VEGF treatment of HdLECs: confluent HdLEC monolay - Tethered Ligand Assay: The recombinant extracellular ers were starved overnight in 1% FBS in EBM2 (Lonza) domain of the Notch ligand hDLL4FC (Sino Biologicals or in human endothelial SFM (Fisher Scientific) followed Inc.) or IgG-FC (Sino Biologicals Inc.) were coated onto by either 1 or 5  h in EBM2/SFM containing 100ng/mL a 24-well plate (Corning) on a fibronectin matrix (Sigma). VEGF-A (R&D), 100ng/mL VEGF-C (R&D), or 500ng/ Following an overnight incubation at 4 °C, primary ECs C156S mL VEGF-C (R&D) prior to RNA isolation. Assays (at 80% confluency) were trypsinized and seeded onto the were performed at least 3 times. For detection of AKT and coated plates and incubated at 37 °C with 5% CO for 6 h. ERK activity, HdLECs were serum starved overnight in Experiment was performed in triplicate. SFM containing 1% FBS and 0.1% BSA, followed by 5 h RNA was isolated using the RNEasy Mini Kit (Qiagen), in SFM alone. Cells were then switched to SFM contain- quantity and integrity measured using a Bio-analyzer (Agi- ing 0.1% BSA and either 100ng/mL VEGF-A or 100ng/mL lent TapeStation 4200, UIC Genome Research core) prior VEGF-C for 20 min prior to fixation with cold 4% PFA. to RNA sequencing. TLA HdLEC samples were sequenced Assays were performed in duplicate for two different lenti- at a ~30 million paired-end (PE) read depth with 150-base viral transductions. fragments by Novogene (https:// en. novog ene. com/). Raw Migration Assay: HdLECs were seeded in triplicate on a reads from in vitro screens were mapped to the Human data- fibronectin-coated (Thermofisher) 12-well plate in complete base (ENSEMBL/GRCh38) using STAR (version 2.5.0a) medium. The following day (0-hour time-point), a scratch and processed with Samtools (version 1.4.1). The counts through the confluent monolayer was made across each well obtained by FeatureCounts (version 1.5.2) were analyzed by using a 200 µl pipet tip, and medium was changed to EBM2 DESeq2 (version 1.18.1) to identify differentially expressed containing 100ng/ml VEGF-C. For migration assays with genes. The RNAseq datasets generated in the current study mitomycin C, confluent monolayers were pretreated with are available in the NCBI Gene Expression Omnibus reposi- 10 µg/mL mitomycin-C (Sigma) for 45 min prior to scratch- tory at https://ww w.nc bi.nlm .nih .go v/ge o (Accession Num- ing. Cells were maintained in EBM2 containing 100ng/ml ber GSE183631). VEGF-C and 0.1 µg/mL mitomycin-C while migration was assessed. Growth into the scratch was documented at 0, 4, 8, Gene expression analyses 12, and 25 h with a Zeiss Axiovert 40 CSL inverted micro- scope or at 0, 4, 8, and 24 h using an Olympus IX83 micro- RNA was isolated using the RNEasy Mini Kit (Qiagen) and scope. Cell migration rate was determined using imageJ soft- reverse transcribed using the VersoTM cDNA Synthesis ware [21] and calculated as the percentage of cell-free area Kit (Thermo Fisher) or First Strand Synthesis Kit (Invitro- at different time-points relative to the initial wound area. gen). qRT-PCR was performed in triplicate for each gene Assays were performed at least 2–3 times for two independ- (Table S1), using ABsoluteTM Blue QPCR SYBR Green ent lentivirally generated HdLEC populations. Master Mix (Thermo Fisher) or Sybr Green Master Mix (Applied Biosystems) and 7300 Real-Time PCR System Co‑culture Notch reporter assay (Applied Biosystems) or CFX96 PCR Cycler (Biorad). Gene-specific qRT-PCR standards were used to determine HdLECs (90% confluent) were lipofected (Lipofectamine transcript levels and normalized to β-actin expression [12]. 2000; Invitrogen) with the Notch reporter plasmid PCRs were set up in triplicate and performed at least 3 times. pGL3.11CSL [12] containing 11 repeats of Notch/CSL For the validation of mRNA sequencing data, qPCR was (RBPjκ) cis-elements, and phRL-SV40 renilla (Promega) to done using SYBR Green master mix (Applied Biosystems) normalize lipofection efficiency. HeLa cells were lipofected and primers specific to genes of interest (Table S2). The with pCR3 plasmids encoding either DLL4-FLAG or JAG1- mean cycle threshold (Ct) values from the triplicate run for FLAG with empty vector serving as a control. 24 h after each sample were analyzed using β-actin as the reference lipofection, HeLa and HdLECs were co-cultured together at gene. ΔΔCt method [22] was used to calculate the relative a 1:1 ratio on fibronectin-coated plates in EGM2. 24 h after expression using the following steps: (1) Normalization to co-culture, a luciferase reporter assay was performed using reference gene: ΔCt = Ct – Ct . (2) Relative expres- GOI GOI BA the Dual-Luciferase Reporter Assay System (Promega) and sion between conditions: ΔΔCt = ΔCt – ΔCt . The GOI EXP CNT a TD20/20 luminometer (Turner Designs). Luciferase val- analysis was done using Microsoft Excel and Prism. ues were normalized to Renilla values. Each condition was performed in triplicate, 4 times. 1 3 208 Angiogenesis (2022) 25:205–224 a Zeiss Axioskop2 Plus and Zeiss AxioCam MRc camera Western blotting with Zeiss Zen software, or an Olympus IX83 Inverted Sys- tem Microscope and Olympus cellSens software. Confocal NOTCH4 expression in Notch4 null mice was determined by Western blot. Fresh E14.5 hindlimbs were lysed in RIPA microscopy was performed with a Zeiss LSM 510 META Confocal Microscope and the LSM software. buffer (Invitrogen) containing protease and phosphatase inhibitors (Thermofisher) on ice and protein concentra- For cell immunochemistry, cells were fixed in 4 % PFA on ice for 15 min, permeabilized and blocked with 0.1% tion determined by BCA protein assay kit (Pierce). Equal amounts of protein were separated on an 8% SDS-PAGE Triton X-100, 2% BSA, 3% donkey serum in 1 × PBS for 1 h at room temperature. Cells were then incubated over- gel and transferred to a nitrocellulose membrane. The mem- branes were blocked with 3% nonfat milk and 3% bovine night with primary antibody at 4 °C followed by an incu- bation with Alexa Fluor-conjugated anti-donkey secondary serum albumin (Jackson ImmunoResearch) in Tris-Buff- ered Saline Tween-20 and probed with antibodies against antibodies (Invitrogen) for 1 h at room temperature. Slides were washed in 1 × PBS and mounted with Vectashield the cytoplasmic domain of NOTCH4 [23] and β-ACTIN (Abclonal). Horse radish peroxidase-conjugated secondary with DAPI mounting media. Images were captured with an Olympus IX83 Inverted System Microscope and Olympus antibodies (Life Technologies) were used, detected with SuperSignal West Pico Chemiluminescent Substrate (Pierce) CellSens software. Images were analyzed with ImageJ or Adobe Photoshop. and images captured with Biorad Chemdoc MP. Tiled 10x images were used to quantify lymphatic and blood vascular density, distance between migration fronts, fronts Mouse studies per unit length, and branch-point per unit length. 20x images −/− T2 were used to determine lymphatic vessel caliber, distance Notch4 nullizygous (N4 ) [24], Prox1CreER [25], and f l/f l DNMAML [26], CBF:H2B-Venus (NVR reporter pur- to first branch-point, number of Prox1+ cells per field, and sprout morphology. Vascular density was determined chased from Jax Labs) [27] and Prox1-tdTomato lymphatic reporter (ProxTom) [28] mice were used for these studies. as positive signal area normalized by total area. Distance between migrating fronts was determined as the mean dis- Studies were performed in mice with a mixed background, as well as a pure C57BL6j background. For studies using tance between the 2 lymphatic fronts measured at multiple T2 points (≥ 3) [1]. For analysis of migration of Notch4 mutants, Prox1CreER , tamoxifen in corn oil was administered via +/− oral gavage (10 mg/40 g) at E12.5. 3 or more independent Notch4 embryos were used to normalize between litters, as they were present in all litters analyzed and the distances litters were assessed for each analysis. Number of embryos analyzed is presented in the figure legends. between migration fronts were not statistically different than WT embryos. Sprouting fronts, defined as the sprouts that Immunohistochemistry & Imaging reside at the leading edge of the migrating front per unit length was determined as the number of sprouting fronts at E14.5 and E16.5 dorsal skin was dissected, fixed for 2 h the leading edge of the migration front normalized to the vertical length (posterior-anterior; Fig. S1a). Length of the in 4% PFA, and then immunostaining initiated. Alterna- tively, embryos were incubated overnight in 4% PFA and sprout was determined as length from tip of sprout at lym- phangiogenic front to the first branch-point. Lymphatic ves - then stored in 1 × PBS at 4 °C. E14.5 tissues were incu- bated for 2 h at room temperature in blocking buffer (10% sel caliber was determined by measuring the width of lym- phatic vessels in the maturing lymphatic plexus and adjacent donkey serum, 0.3% Triton X-100, 1 × PBS), incubated in primary antibody (Table S3) diluted in blocking buffer to the first branch-point from the migrating front. Sprout morphology at the lymphangiogenic front was determined overnight at 4 °C, and then incubated with the appropri- ate Alexa Fluor secondary antibody (Invitrogen) diluted in by counting total number of blunt-ended sprouts (rounded, lacking multiple filopodia) and spiky-ended sprouts (elon- blocking buffer overnight at 4 °C. E16.5 dermal tissues were washed in 1 × PBS containing 0.2% Triton X-100 and 20% gated with multiple filopodia) normalized to the total num- ber of sprouts assessed. Branch-points per unit length in DMSO for 4 h at room temperature and immunohistochem- istry performed as described in Cha et al. 2016 [29]. For maturing lymphatic plexus was determined as the number of branch-points per field normalized to the total length of immunostaining of sections, 5-micron sections were stained as previously described [30]. Tissue was mounted using lymphatic vessels per field. To measure Prox1+ LEC num- ber, Prox1+/LYVE1+ LECs were scored and mean number Vectashield with and without DAPI (Vector Laboratories). Images were captured using a Nikon SMZ-U Zoom 1:10 per field determined. To determine the significance between control and one experimental group, a two-tailed student’s microscope and Nikon 4500 digital camera, Nikon ECLIPSE E800 microscope and NIS Elements software, Nikon DXM t test was used. For analyses of more than two groups, one- way analysis of variance (ANOVA) was used to determine 1200 digital camera, and Image ProPlus v.4.01 software, 1 3 Angiogenesis (2022) 25:205–224 209 significance by unpaired t test. For analyses of multiple con- expressed through-out the endothelium of the sprouts at the ditions and cell populations, two-way ANOVA was used and lymphatic front which overlapped with DLL4 (Fig. 1b, c). Dunnett’s multiple comparison test performed to determine Analysis of E14.5 dermal cross-section confirmed that the significance between groups. A p < 0.05 was considered LYVE1+ dermal lymphatic endothelium expressed both significant. NOTCH1 and NOTCH4 (Fig. S1b, c). Outside of the lym- phatics, NOTCH1 expression was observed in the epider- Lymphangiography mis and blood endothelium, while NOTCH4 was expressed in the epidermis and a subset of LYVE1+ macrophages Lymphangiographies were performed as described on E17.5 (Fig. 1b, c, Fig. S1b, c) . embryos [15]. Briefly, 2 µL of 0.4% Trypan blue solution To determine where Notch is actively signaling dur- (Sigma) was injected into the dermis in periorbital region ing lymphangiogenesis, the dermal lymphatics in E14.5 with a 36G beveled needle attached to a Nano l syringe embryos carrying alleles for the Prox1-tdTomato (ProxTom) (WPI). The embryos were imaged using AMSCOPE ster- LEC reporter [28] and the Notch Venous Reporter (NVR) eomicroscope (AMSCOPE) with camera attachment 1 min [27] were assessed. Notch activity was observed throughout after injection. the lymphatic vascular plexus at both the lymphangiogenic front, defined as the LECs that make up the sprout from tip to the first branch-point, and the mature plexus where the Results vessels have begun to remodel (Fig. 2a). At the lymphangi- ogenic front, Notch activity was often observed in several Embryonic dermal lymphatics expressed NOTCH1, LECs located at the tip cell positions in spiky-ended sprouts NOTCH4, and the Notch ligand, DLL4 with filopodia (Fig.  2b), consistent with the broad expres- sion of NOTCH1, NOTCH4, and DLL4 at the front (Fig. 1). Cultured human LECs, HdLECs, express NOTCH1-4 and Notch activity was also observed in blunt-ended sprouts. In the Notch ligands, DLL4 and JAGGED1 (JAG1) [18], while the mature lymphatic plexus, Notch activity was observed NOTCH1 and NOTCH4 are expressed in the postnatal day 4 at branch-points (Fig.  2a, c). Taken together, the expres- (P4) murine dermal lymphatic vessels [12]. Notch signaling sion data suggest that DLL4 signaling via NOTCH1 and/or has been shown to be active in the E15.5 dermal lymphatics NOTCH4 has a role in regulating dermal lymphangiogenic [14], but it is not known which Notch proteins and ligands growth and maturation. are expressed in dermal LECs at this time. To study the role of Notch signaling in embryonic dermal lymphangiogenesis, Profiling of DLL4/Notch signaling in human dermal we determined the expression of NOTCH1 and NOTCH4, lymphatic endothelial cells and the angiogenic Notch ligands, DLL4 and JAG1, as well as Notch activity in E14.5 dorsal skin. This time-point is Expression studies suggested that DLL4 is the major ligand characterized by the presence of two LYVE1+ lymphatic for Notch signaling in the lymphatic endothelium during fronts migrating toward the midline which precedes a matur- dermal lymphangiogenesis. To determine if DLL4 or JAG1 ing lymphatic plexus (Fig. S1a). At this time-point, two could activate Notch in LECs, co-culture assays were per- CD31+ angiogenic fronts have fused at the midline to form formed in which endogenous Notch activation was deter- a connected blood capillary network. mined using a Notch-response CSL luciferase reporter [12]. At E14.5, DLL4 was expressed in both in the developing HeLa cells were engineered to express DLL4, JAG1, or lymphatics and blood vessels of the dermis (Fig. 1a). Unlike both (Fig. S2a) and then seeded at a 1:1 ratio with HdLECs the retina, where DLL4 expression is restricted to 1–2 tip containing a CSL-luciferase reporter. DLL4-expressing cells at the angiogenic front [6–8], high DLL4 expression HeLa cells upregulated Notch signaling nearly five-fold was observed in multiple LECs within the sprouts at the over co-cultures using parental HeLa cells, while JAG1 lymphangiogenic front. DLL4 was also expressed in the only modestly increased Notch signaling in HdLECs (Fig. arterial vessels and blood capillary network, with strongest S2b). Co-culture with HeLa cells co-expressing DLL4 and expression observed in the large arteries, consistent with JAG1 induced Notch signaling similar to the co-cultures its expression in the vasculature of the intestinal villi [10]. with DLL4 alone, suggesting that JAG1 did not interfere Unlike DLL4, JAG1 was not expressed in dermal lymphatics with DLL4 signaling. Together with the expression studies, at E14.5 (Fig. 1a). JAG1 expression was limited to the blood these data suggest that DLL4 functions as a ligand for LEC vasculature, where the highest expression was observed in NOTCH. the larger caliber arteries in a pattern consistent with vas- To further assess DLL4/Notch signaling, HdLECs were cular smooth muscle cell and endothelial cells. Staining for seeded on DLL4FC-coated or FC-coated (control) plates. NOTCH4 and NOTCH1 demonstrated that they were both After 6  h, RNA was collected and mRNA sequencing 1 3 210 Angiogenesis (2022) 25:205–224 Fig. 1 Embryonic dermal lymphatics expressed NOTCH1, NOTCH4, wild-type skin wholemount stained for LYVE1 and NOTCH4. White and DLL4. a E14.5 wild-type skin wholemounts stained for LYVE1, arrowheads mark lymphatics at the front. Yellow arrowhead marks CD31 and DLL4 or JAG1. Higher magnification of boxed areas a NOTCH4+ macrophage. Scale bars, 50  μm. c  E14.5 ProxTom presented to the right. White arrowheads mark sprouts at the lym- skin wholemount stained for DLL4 and NOTCH1. White arrow- phangiogenic front. White asterisk marks the blood vascular plexus. heads mark lymphatic sprout at the front. Yellow arrowhead marks Yellow arrowhead marks an artery. Scale bars, 100 μm. b  E14.5 NOTCH1+/DLL4+ blood vessel. Scale bars, 20 μm performed. Relative to HdLECs seeded on FC-coated plates, signaling induces genes responsible for pattern specification, HdLECs seeded on DLL4FC significantly altered the expres- neurogenesis, and chemotaxis (Fig. 3d). sion of 675 genes with a padj < 0.05 (Fig. 3a, Table S4, S5). To assess if DLL4-induced genes were downstream of 69 genes were induced 1.2-fold, while 68 genes were sup- Notch1 or Notch4 activation, HdLECs were generated to pressed 1.2-fold. Analysis of the top 30 upregulated genes express activated forms of NOTCH1 (N1IC) or NOTCH4 revealed that DLL4/Notch signaling upregulated the expres- (N4/Int-3). Quantitative RT-PCR was performed for Notch sion of known direct effectors of Notch signaling, Hey1, effectors of the Hes and Hey gene families, as well as lym- Hes4, Dll4, and Hes1, as well as key lymphangiogenic phangiogenic genes. Of the Notch effectors assessed, N1IC genes, Ackr3, Cxcr4, Ccl2, EphrinB2, Gja4 (Cx37), Gja1 and N4/Int-3 both induced the expression of Hes1, Hes4, (Cx43), and Sema3g (Fig. 3b) [31–38]. DLL4/Notch signal- Hey2, and HeyL, with the strongest induction observed for ing also downregulated lymphangiogenic genes, including Hes4 and HeyL (Fig. 3e). In contrast, only N1IC induced Apln and Adm (Fig. 3c) [35, 39]. Further analysis of the 675 the expression of Hes5. Notch1 and Notch4 activation sig- altered genes demonstrated that DLL4/Notch signaling both nificantly induced the expression of the majority of lym- upregulated and downregulated genes of the Notch pathway phangiogenic genes assessed, except for Cxcr4, Bmp2, and (Fig. S3a) and lymphangiogenesis (Fig. S3b). GO: Biologi- Tgfrb2 (Fig. 3f). Similarly, both suppressed the expression cal Pathway (BP) analysis indicates that LEC DLL4/Notch of Prox1, Podoplanin, and Lyve1 (Fig. S3b, S4b). Cxcr4 was significantly induced by N1IC, while N4/Int-3 suppressed 1 3 Angiogenesis (2022) 25:205–224 211 Fig. 2 Notch activation observed throughout the embry- onic dermal lymphatic vascular plexus. E14.5 ProxTom;NVR skin wholemounts stained for LYVE1. a Low magnifica- tion image demonstrating Notch activity throughout the developing lymphatic plexus. Blue arrowheads mark sprouts at the lymphangiogenic front with Notch activity. White arrowheads mark regions of high Notch signaling in the maturing plexus. Scale bars, 500 μm. b High magnifica- tion of spiky-ended lymphatic sprout. White arrowheads mark tip cells with Notch activity. Yellow arrowheads mark stalk cells with Notch activity. c High magnification of the maturing plexus. White arrowheads mark LECs with Notch activity. b, c Scale bars, 100 μm its expression. Neither Tgfrb2 and Bmp2 were induced. determined the effect of VEGF-A and VEGF-C on Notch While Ackr3 and Ccl2 were significantly induced by N1IC genes, ligands, and effectors in HdLECs. Serum-starved relative to the GFP controls, N4/Int-3 was a much stronger HdLECs were treated with either VEGF-A, VEGF-C or C156S inducer of both these genes (p < 0.0003, p < 0.0001 N4/Int-3 VEGF-C . In HdLECs, VEGF-A binds and activates C156S vs. N1IC, respectively). Together these data support over- VEGFR2, VEGF-C binds and activates VEGFR3, and lapping and distinct downstream signaling for Notch1 and VEGF-C activates both VEGFR2 and VEGFR3 [40]. After Notch4 in LECs. 1 h, VEGF-A and VEGF-C significantly induced Dll4 tran- scripts, which correlated with an increase in Hey1, Hey2, C156S VEGF‑C induced Dll4 expression and Notch and Hes1 transcripts (Fig. 4a, b). VEGF-C or VEGF-C activation in HdLECs both induced Dll4 expression, as well as the Notch effector, Hes1 (Fig. 4a, b). Notch1 was modestly induced by VEGF- During sprouting angiogenesis, VEGF-A/VEGFR-2 sign- A only, while Notch4 expression was unaffected (Fig.  4a). aling upregulates DLL4 in blood endothelial tip cells to To determine if VEGF-A and VEGF-C differentially activate Notch signaling in the adjacent stalk cell [6–8]. As regulate Notch signaling in BECs and LECs, HdLECs and we observed DLL4 expression and Notch activity in LECs HUVEC were serum starved and treated with either VEGF- of the sprouts located at the lymphangiogenic front, we A or VEGF-C. After 5  h, VEGF-A only induced Dll4, 1 3 212 Angiogenesis (2022) 25:205–224 Fig. 3 DLL4/Notch signaling regulated Notch and lymphangio- direct targets of Notch signaling and f  lymphangiogenic genes sig- genic genes via Notch1 and Notch4. HdLECs were seeded on either nificantly induced in the DLL4-HdLEC assay in HdLECs expressing DLL4FC- or FC-coated plates, and RNA isolated after 6 h, followed GFP, N1IC, or N4/Int-3. Data presented for two independent trans- by mRNA sequencing. Experiment was performed in triplicate. ductions done in duplicate and gene expression determined by delta a Volcano plot of genes downregulated and upregulated by DLL4FC CT method and presented relative GFP controls ± s.e.m. One-way relative to FC controls. b  Top 30 genes upregulated and c  down- ANOVA performed and significance determined by unpaired t test. e, regulated by DLL4FC in HdLECs. b, c  *mark Notch pathway and f *p < 0.05 or blue asterisks indicate p < 0.003 N4/Int-3 or N1IC rela- blue asterisks indicate lymphangiogenic genes. d  Top GO pathways tive to GFP controls. Red stars indicate  p < 0.0003 N4/Int-3 relative for biological processes for DLL4-induced genes. e  qRT-PCR for to N1IC 1 3 Angiogenesis (2022) 25:205–224 213 Fig. 4 VEGF and NOTCH signaling modulated each other in LECs. N1IC-, and N4/Int-3-HdLECS were treated with VEGF-A or VEGF- a Notch1, Notch4, and Dll4 expression, and b Hey1, Hey2, and Hes1 C for 20  min and then stained for either  e  phospho-AKT (pAKT) expression determined by qRT-PCR of HdLECs treated for 1  h or g  phospho-ERK (pERK). Scale bars, 25 μm. Quantification of C156S with no growth factor (nGF), VEGF-A, VEGF-C, or VEGF-C . mean,  f  pAKT and h  pERK expression normalized by area. Mean Experiment done in triplicate. Data presented as mean fold induc- data presented for two independent transductions and experiment tion relative to HdLECs treated with no growth factor ± s.e.m. *p < performed in duplicate. Data presented as fold expression relative to 0.05 HdLEC treated with VEGFs relative to HdLEC treated with no GFP-expressing HdLECs treated with  nGF ± s.e.m. f  pAKT: two- growth factor. c  Notch1, Notch4, and Dll4 expression, and d  Hey1, way ANOVA: p < 0.0001, Dunnett’s multiple comparison test *p < Hey2, and Hes1 expression determined by qRT-PCR of HdLECs 0.002 VEGF-A-treated N4/Int-3 vs. GFP HdLEC. h pERK: two-way treated for 5  h with nGF, VEGF-A, or VEGF-C. Data presented ANOVA: p < 0.001, Dunnett’s multiple comparison test *p < 0.03, as fold induction relative to nGF ± s.d. *p < 0.05 VEGF-treated **p < 0.0001 VEGF-C-treated N1IC or  N4/Int-3 vs. GFP HdLEC. HdLEC relative to HdLEC treated with no growth factor. e–h  GFP-, *p < 0.003 VEGF-C-treated N4/Int-3 vs. N1IC HdLEC 1 3 214 Angiogenesis (2022) 25:205–224 −/− which was associated with an increase in Hey1 and Hes1 Analysis of E14.5 Notch4 dermal wholemounts in HdLECs (Fig. 4c, d). VEGF-C induced Dll4 in HdLECs revealed that the distance between the two migrating (Fig.  4c), which correlated with an increase in Notch1 lymphatic fronts was decreased relative to wild-type and +/− transcripts (Fig. 4c), as well as Hey2 and Hes1 induction Notch4 littermates (Fig. 5a, b). This correlated with an (Fig. 4d). In HUVEC, VEGF-A induced Dll4 and Notch1 increase in the number of lymphatic fronts migrating toward −/− expression, whereas VEGF-C modestly decreased Dll4, the midline in the Notch4 dermis (Fig. 5c). Further analy- Notch1, and Notch4 transcripts (Fig. S4a). Thus, VEGF-A sis of the lymphangiogenic sprouts at the migration front and VEGF-C dynamically and temporally induced Notch revealed that the length from the front to the first branch- activity and specific Notch effectors via induction of Dll4 point did not differ between mutants and controls (Fig.  5d). in cultured LECs. We next evaluated the lymphatic vessel caliber at the lym- To determine if Notch1 or Notch4 activation altered sign- phangiogenic front and in the maturing plexus. The caliber aling downstream of the VEGFs, HdLECs expressing either of the vessel adjacent to the first branch-point at the front did N1IC, N4/Int-3 or GFP were serum starved overnight and not differ between mutant and control mice (Fig.  5e, f). How- then stimulated with VEGF-A or VEGF-C. After 20 min, ever, a significant reduction of vessel caliber was observed −/− HdLECs were assessed for AKT and ERK activation by in the maturing lymphatic plexus of Notch4 . Although immunofluorescent staining for phospho-AKT and phospho- the lymphatic vessel diameter was reduced in the maturing −/− ERK (Fig. 4e–h.). As compared to the GFP controls, activa- Notch4 plexus, branching was similar between mutants tion of AKT was reduced in N4/Int-3 HdLECs treated with and controls (Fig. S6a). To determine if the reduced dermal VEGF-A. Activation of AKT by VEGF-C was unaffected lymphatic vessel caliber was due to a change in LEC pro- −/− in HdLECs expressing either N1IC or N4/Int-3. Levels of liferation, wild-type and Notch4 E14.5 dermal skin were ERK activity were unaffected by constitutive activation of stained for the proliferation marker, KI67, and LYVE1. Der- −/− Notch1 or Notch4 in all VEGF-A-treated HdLECs (Fig. 4e, mal LEC proliferation was similar between Notch4 and f), whereas ERK activation was reduced in N1IC and N4/ control mice (Fig. S6b, c). As NOTCH4 is also expressed Int-3 HdLECs treated with VEGF-C relative to controls by the blood vasculature [20, 24], we evaluated the underly- −/− (Fig. 4 g, h). Previous studies have shown that Notch signal- ing dermal blood vascular network in E14.5 Notch4 and +/− ing alters the expression of the VEGF-A and VEGF-C recep- Notch4 embryos. Consistent with prior studies, the den- tors, VEGFR2 and VEGFR3 [9, 12, 18, 41]. The reduction sity and branching of the blood vasculature were unaffected in AKT and ERK activation downstream of VEGFs may be in the Notch4 nulls (Fig. S7) [24, 42]. secondary to Notch signaling effects on VEGFR expression. We next evaluated the dorsal dermal lymphatic phenotype −/− Therefore, the expression of Vegfr2 and Vegfr3 was deter- at E16.5 in wild-type and Notch4 embryos. The lymphatic mined in HdLECs with N1IC and N4/Int-3 by qRT-PCR. fronts had reached the midline and merged in both wild-type −/− Both N1IC and N4/Int-3 downregulated Vegfr2, while they and Notch4 embryos. Analysis of the lymphatic plexus induced Vegfr3 (Fig. S4b). Taken together, the data suggests revealed that it was disorganized with an increase in the that the decrease in AKT activation by VEGF-A and ERK distance between branch-points, while there was no differ - −/− activation by VEGF-C may be due to reduced VEGFR2 lev- ence in the mean vessel caliber in Notch4 dermis relative els, and not VEGFR3. to controls (Fig. 5g–i). Lymphatic valves were observed in both mutants and controls. A recent study has shown that Embryonic dermal lymphangiogenic defects reduced branching in the dermal lymphatic plexus was asso- in Notch4 mutant mice ciated with an increase in blunt-end sprouts due to reduced VEGFR3 signaling, which in turn led to a less branched A  prior report has shown that loss of LEC Notch1 in network [43]. In contrast, increased VEGFR3 signaling embryos leads to increased LEC proliferation and tip cells was associated with reduced blunt-ended sprouts and a [14]. To determine the role of Notch4 in embryonic lym- more densely branched network [44]. Since we observed −/− phangiogenesis, we evaluated Notch4 mice and compared a decrease in the dermal lymphatic branching in the E16.5 +/− −/− their lymphatic phenotype to that of Notch4 and wild-type Notch4 embryos, we assessed the sprout phenotype at littermates. To confirm that NOTCH4 protein is absent in E14.5 during active lymphangiogenesis. The sprouts of con- −/− the Notch4 embryos, western blots using lysates collected trol embryos uniformly expressed LYVE1 and were elon- from E14.5 embryos and staining of P4 dermal tissue with gated with numerous filopodia consistent with a lymphangi- an antibody against the intracellular domain of NOTCH4 ogenic phenotype (Fig. S8). In contrast, LECs in sprouts −/− were performed. As compared to wild-type littermates, in the Notch4 lymphatic vasculature were often rounded −/− NOTCH4 expression was absent in Notch4 tissues (Fig. with reduced and blunted filopodia. S5a–c). NOTCH1 expression determined by immunostain- Thus, Notch4 mutant mice had a distinct dermal lym- ing of P4 dermis was unaffected (Fig. S5d). phatic phenotype from that observed in mice with LEC 1 3 Angiogenesis (2022) 25:205–224 215 Fig. 5 Loss of Notch4 altered embryonic dermal lymphangiogenesis. lymphangiogenic front and in the maturing plexus. Data presented −/− a–f Dermal lymphatic phenotype was determined for E14.5 Notch4 ± s.e.m. One-way ANOVA: p = 0.019, t test: *p < 0.03, front analy- −/− +/− +/− +/− −/− (N4 ), Notch4 (N4 ) and wild-type (WT) wholemounts stained sis- WT (n = 7), N4 (n=13), N4 (n = 7), plexus analysis - WT −/− +/− +/− −/− for LYVE1. a Representative image of N4 and N4 dermal who- (n = 6), N4 (n = 9), N4 (n = 8).  f  Representative image of the −/− +/− lemounts. Red dotted line marks leading edge of lymphangiogenic E14.5 maturing lymphatic plexus in N4 and N4 littermates. front. Scale bars, 100 μm.  b  Quantification of the distance between Scale bars, 100 μm. g–i Dermal lymphatic phenotype determined for +/− −/− −/− migration fronts, normalized to the N4 controls. Data presented E16.5 Notch4 (N4 ) and wild-type (WT) wholemounts stained −4 ± s.e.m. One-way ANOVA: p = 0.014, t test: *p < 3 × 10 , **p < for VEGFR3. g Representative image of the E16.5 dermal lymphatic +/− −/− −/− 0.001. wt (n = 5), N4 (n = 13), N4 (n = 4).  c  Quantification plexus in N4 and WT littermates. Red arrowheads mark lym- of the number of lymphangiogenic sprouts normalized by length phatic valves. Scale bars, 100 μm; h quantification of mean distance of the front. Data presented ± s.e.m. One-way ANOVA: p = 0.023, between branch-points. Data presented ± s.e.m. t test: *p < 0.03 WT +/− −/− −/− t test: *p < 0.02. WT (n = 6), N4 (n = 11), N4 (n = 6).  d  Quan- (n = 4), N4 (n = 6).  i  Quantification of the average vessel caliber tification of the sprout length from the migration front to first of the dermal lymphatic plexus at E16.5. Data presented ± s.e.m. +/− −/− branch-point. Data presented ± s.e.m. WT (n = 7), N4 (n = 8), WT (n = 6), N4 (n = 8) −/− N4 (n = 6).  e  Quantification of the average vessel caliber at the 1 3 216 Angiogenesis (2022) 25:205–224 Fig. 6 Notch signaling inhibited LEC migration. a Conflu- ent N1IC-, N4/Int-3-, Hey1-, Hey2-, or GFP-expressing HdLECs were scratched and representative images for 0 and 25 h presented. Scale bars, 2.5 μm. b Quantification of percent open wound area at 0, 8, and 25 h. Data presented ± s.e.m. two-way-ANOVA: p < 0.0012, t test: *p < 0.002 N1IC or N4/Int-3 vs. GFP at 8 and 25 h, *p < 0.002 N4/ Int-3 vs. N1IC at 25 h. c Con- fluent N1IC-, N4/Int-3-, or GFP-expressing HdLECs were treated with mitomycin C and scratched and representative images for 0 and 24 h presented. Scale bars, 100 μm. d Quanti- fication of percent open wound area at 0, 4, 8, and 24 h. Data presented ± s.e.m. two-way ANOVA: p < 0.0001, Dunnett’s multiple comparison test *p < 0.0001 N1IC or N4/Int-3 vs. GFP. t test: *p < 0.0001 N4/ Int-3 vs. N1IC Notch1 deletion [14]. Rather than increased vessel diameter N1IC and N4/Int-3 expression inhibited LEC migration due to increased proliferation and branching due to increased (Fig. 6a, b). We next evaluated the effect of overexpress- sprouting lymphangiogenesis, the embryonic dermal lym- ing the downstream Notch effectors HEY1 and HEY2 phatics in Notch4 nulls had increased front closure, an early on HdLEC migration (Fig. 6a, b). Ectopic expression of decrease in the lymphatic vessel caliber, and reduced branch- HEY1 and HEY2 suppressed migration relative to con- ing without a change in LEC proliferation. trol HdLECs. Further analysis revealed that N4/Int-3 was a significantly stronger inhibitor of LEC migration than NOTCH4 activation preferentially inhibited LEC N1IC at 25 h. To insure the difference in migration was migration not due to changes in LEC proliferation, N1IC-, N4/Int- 3- and GFP-expressing HdLECs were treated with mito- To determine the effects of Notch1 and Notch4 signal acti- mycin C and LEC migration determined. Similar to the vation on LEC migration, a monolayer-wounding assay initial migration assay, both N1IC and N4/Int-3 suppressed was performed using HdLECs expressing either N1IC or HdLEC migration with N4/Int-3 suppressing migration N4/Int-3. Relative to control GFP-expressing HdLEC, both significantly more than N1IC at 24 h (Fig.  6c, d). 1 3 Angiogenesis (2022) 25:205–224 217 −/− LEC In contrast to the Notch4 phenotype, DNMAML der- Inhibition of lymphatic endothelial canonical Notch signaling increased dermal lymphatic vessel density mal lymphatics were dilated relative to control littermates LEC (Fig. 8c). One out of 5 DNMAML embryo lymphatics Notch4 has been shown to signal via RBPjκ-dependent was leaky (Fig. 8d), which was not observed in controls or −/− Notch4 lymphangiographies. These data demonstrate (canonical) and RBPjκ-independent (non-canonical) downstream pathways [45–47]. To determine the effects of that the Notch4 null and mice with a loss of LEC RBPjκ- dependent Notch signaling have distinct phenotypes at LEC specific loss of canonical Notch signaling on dermal T2 lymphangiogenesis, we used the inducible Prox1CreER E17.5, as well as at E14.5 (Figs. 5 and 7). driver to induce expression of a DNMAML transgene [26]. −/− DNMAML encodes a dominant negative form of Mammalian Canonical Notch signaling is unaffected in Notch4 embryonic dermal lymphangiogenesis Mastermind-like 1 (MAML1) that binds the NOTCH/RBPjκ complex to form an inactive complex and blocks the recruit- T2 As we observed a difference between the embryonic der - ment of transcriptional co-activators. Prox1CreER mice fl/fl −/− LEC were crossed with DNMAML mice to generate Prox1CreE mal lymphatic phenotypes of Notch4 and DNMAML T2 fl/+ LEC fl/+ mutants, we evaluated canonical Notch signaling by intro- R ;DNMAML embryos (DNMAML ) and DNMAML control littermates. To circumvent effects on early lymphatic ducing the NVR and ProxTom alleles into the Notch4 null background. Loss of Notch4 did not change canonical Notch specification caused by loss of Notch signaling in LECs [18], tamoxifen was administered to pregnant females at E12.5, signaling at the lymphangiogenic vascular front, nor the maturing lymphatic plexus (Fig. 9, S6d). This data suggested just as sprouting lymphangiogenesis begins, and the dermal lymphatic phenotype analyzed at E14.5. Unlike the Notch4 that Notch4 is not necessary for canonical Notch signaling in the embryonic dermal lymphatics. nulls, the closure of the migration fronts was the same LEC between DNMAML and control (Fig. 7a, b). The number of sprouts along the migrating front and the length of the sprout to the first branch-point were similar between mutants Discussion and controls (Fig. 7c, d). In contrast, the lymphatic density LEC was nearly 25% greater in the DNMAML compared to NOTCH1 and NOTCH4 are expressed and Notch signal- LEC ing active in the embryonic and early postnatal dermal lym- controls (Fig. 7e). The increase in the DNMAML dermal lymphatic density correlated with an enlargement of the phatic vasculature [12, 14], suggesting a role for both Notch proteins in embryonic lymphangiogenesis. Loss of Notch4 lymphatic vessel caliber at the lymphangiogenic front and in the maturing plexus (Fig. 7f, g). As compared to controls, was shown to exacerbate the Notch1 null embryonic blood LEC vascular phenotype, suggesting that Notch1 and Notch4 DNMAML dermal lymphatics had an increase in the num- ber Prox1+/LYVE1+ LECs (Fig. 7h), while branching in the have overlapping functions in the blood endothelium [24]. In contrast to the blood endothelium, we found that loss of mature plexus was unaffected (Fig.  7i). The increase in vas- cular density was specific to the lymphatics as blood vessel Notch4 led to a distinct embryonic dermal lymphangiogenic LEC phenotype, than that observed in mice with LEC deletion density was unchanged in DNMAML mutants (Fig. S9). Thus, we found that inhibition of RBPjκ-dependent Notch of Notch1 [14], or inhibition of canonical Notch signaling, presented here. At E14.5, Notch4 null embryos displayed an signaling resulted in increased lymphatic vessel density and caliber associated with an increase in LECs, suggesting that increase in the closure of the lymphangiogenic fronts to the midline, reduced vessel caliber in the maturing plexus, and canonical Notch signaling suppressed LEC proliferation in the embryonic dermal lymphatics. an increase in blunt-ended sprouts, while LEC proliferation was unaffected. By E16.5, the dermal lymphatic plexus in −/− LEC −/− Notch4 and DNMAML embryos display distinct Notch4 has reduced branching and tortuous lymphatic vessels, which may be secondary to the increase in blunt- lymphatic phenotypes at E17.5 ended sprouts at E14.5. In cultured LECs, constitutive acti- vation of Notch4 was a stronger inhibitor of migration than To assess the functionality and patterning of the dermal lymphatics, lymphangiographies were performed on E17.5 Notch1 activation and induced a subset of lymphangiogenic −/− LEC genes. In contrast, loss of LEC Notch1 at E10.5 increased Notch4 , DNMAML , and control littermates. Dye was injected within the dermis in the periorbital region embryonic dermal lymphatic density, due to increased LEC proliferation and decreased LEC apoptosis [14]. Similar and uptake by the lymphatics assessed after 1  min. The −/− Notch4 lymphatic plexus had reduced branching with to the Notch1 LEC knockout, we demonstrate that LEC expression of DNMAML, which inhibits canonical Notch/ tortuous vessels relative to the more uniform lymphatics of −/− controls (Fig.  8a). One of the 8 Notch4 embryos ana- RbpJκ signaling, increased the dermal lymphatic vascular density consistent with an increase in LEC proliferation and lyzed had blood-filled dermal lymphatics at E17.5 (Fig.  8b). 1 3 218 Angiogenesis (2022) 25:205–224 viability. Distinct functions for Notch1 and Notch4 have promote vessel maturation while having no effect on vascu- been described for endothelial progenitor cells, where Dll4 lar density [49]. Taken together, we propose that Notch1 and signaling via Notch4 specifically induced EphrinB2 and Notch4 signal dynamically to regulate lymphangiogenesis increased proliferation and migration of cultured cells [48]. and control migration and branching, versus proliferation More recently, it was proposed that endothelial Dll4/Notch1 and cell viability by distinct mechanisms. signaling induces Hey2 to suppress proliferation and tip cell Our studies suggest that Dll4 signaling via Notch1 and formation, while Jag1 activates Notch4 to induce Hey1 and Notch4 have overlapping and unique transcriptional targets 1 3 Angiogenesis (2022) 25:205–224 219 ◂Fig. 7 Loss of canonical Notch signaling in LECs increased der- further studies are necessary to understand the complexity T2 fl/fl mal lymphatic density. Prox1CreER and DNMAML mice were of Notch1 and Notch4 signaling in LECs. crossed, tamoxifen administered at E12.5 and dorsal dermis analyzed Prior studies have shown that VEGF-C induces DLL4 in at E14.5. a  Representative images of LYVE1 staining of Prox1Cre T2 fl/+ LEC fl/+ LECs leading to Notch activation [10]. We expanded these ER ;DNMAML (DNMAML ) and DNMAML (control) der- mis. Red dotted line marks leading edge of lymphatic fronts. Scale studies to understand the role of time and specific VEGFRs C156S bars, 1000 μm.  b  Quantification of the distance between migration in this process. Using VEGF-C which specifically binds fl/  + fronts, normalized to the  DNMAML controls. Data presented VEGFR3 and VEGF-A which binds VEGFR2, we found LEC ± s.e.m. Control (n = 7), DNMAML (n = 9).  c  Quantification that VEGFR2 signaling was a stronger and faster inducer of the number of lymphangiogenic sprouts normalized by length LEC of the front. Data presented ± s.e.m. Control (n = 8), DNMAML of Dll4/Notch signaling than VEGFR3 signaling. We also (n = 6).  d  Quantification of the distance between migration fronts, discovered that the induction of the Hes and Hey genes was fl/  + normalized to the DNMAML controls. Data presented ± s.e.m. time- and VEGF-dependent in HdLECs. VEGF-A induced LEC Control (n = 8), DNMAML (n = 6).  e  Quantification of average Hes1 at 1 h which persisted until 5 h, while significant Hey1 LYVE1+ vessel density normalized by area. Data presented rela- tive to control ± s.e.m. t test *p < 0.002, control (n = 7), DNMAML- upregulation was not observed until 5 h. In LECs, VEGF- LEC LEC (n = 9).  f  LYVE1, CD31, and PROX1 staining of DNMAML A was a stronger inducer of Hey1, while VEGF-C induced mutant and control dermal wholemounts. Images represent low Hey2. This differential response of Hes and Hey genes to (left) and high (middle) magnification of the lymphangiogenic front VEGF-A and VEGF-C was specific to LECs, as VEGF-C and the maturing plexus (right). Scale bars, 500 μm (left), 100 μm (middle, right).  g  Quantification of the average vessel caliber at the had no effect on Dll4, or Notch gene expression in HUVEC. lymphangiogenic front and in the maturing plexus. Data presented Our studies also suggest that additional Hes and Hey gene LEC ± s.e.m. t test: *p < 0.04, **p < 0.01, Control (n = 8), DNMAML family members than those studied in the blood vasculature, (n = 6).  h  Quantification of the number of PROX1+/LYVE1+ LECs Hes4 and HeyL, have a role in transmitting Notch signaling per field (pf). Data presented as ± s.d. ***p < 0.001. Control (n = 8), LEC DNMAML (n = 6). i) Quantification of the average number of in LECs. branch-points normalized to unit of vessel length. Data presented ± Our studies also revealed that Notch1 and Notch4 differ - LEC s.e.m. control (n = 3), DNMAML (n = 5) entially altered signaling downstream of VEGF-A/VEGFR and VEGF-C/VEGFR. Constitutive activation of Notch4 in HdLECs. Constitutive Notch1 and Notch4 activation blocked AKT activation by VEGF-A, and ERK activation by in LECs both induced the expression of Notch effectors VEGF-C, whereas Notch1 activity only modestly suppressed (Hes1, Hes4, Hey1, Hey2, HeyL) and lymphangiogenic VEGF-C activation of ERK signaling. Together, these data genes, such as EphrinB2, and Cx37 [50], as well as down- suggest that Notch4 has a role in modulating VEGFR2 and regulated essential genes in lymphangiogenesis, Podopla- VEGFR3 signaling down the PI3K/AKT and RAS/MAPK nin, and Prox1. In contrast, Hes5 expression was induced pathway. by N1IC, and not N4/Int-3. DLL4/Notch-induced genes The dermal lymphatic phenotypes were distinct between −/− LEC involved in chemokine signaling were also differentially Notch4 and DNMAML , suggesting Notch4 signals at regulated by Notch1 and Notch4. Expression of Cxcr4 was least in part via a non-canonical pathway. Notch4 has been induced by Notch1 activation, but suppressed by Notch4 shown to signal via canonical (RBPjκ-dependent) and non- signaling. In LECs, CXCR4 signaling promoted wound- canonical (RBPjκ-independent) Notch pathways in multiple induced and VEGF-C driven lymphangiogenesis in vivo, cells types [45–47]. In endothelial cells, Notch4 activation while in vitro it induced chemokine-driven LEC migration blocked LPS-induced apoptosis via RBPjκ-independent [31, 38]. Thus, it is possible that loss of Notch4 led to an upregulation of Bcl2 [45]. In mice, NOTCH4 activation increase in CXCR4 expression, which in turn increased the in the ductal epithelium required RBPjκ for physiological LEC migration toward the midline. N4/Int-3 was also a sig- alveolar development, but not for breast cancer development, nificantly stronger inducer of Ccl2 and Ackr3, than N1IC. suggesting Notch4 functions via both canonical and non- LEC-derived CCL2 has been shown to promote the recruit- canonical pathway in the breast endothelium [46, 47]. We ment of monocytes and macrophages to sites of lymphangi- observed that canonical Notch signaling was unchanged in ogenesis, where they deliver VEGF-A and VEGF-C [33, the embryonic dermal LECs in Notch4 nulls suggesting that 36]. In murine lymphatic development, ACKR3 functions to the Notch4 dermal lymphatic phenotype did not occur via a suppress LEC growth by scavenging adrenomedullin (ADM) RBPjκ-dependent mechanism. However, it is possible that [35]. Interestingly, we found that Dll4/Notch signaling also the variable phenotypes are due to differences in the pen- suppressed Adm expression, suggesting that Dll4/Notch4 etrance of global Notch4 loss versus a tamoxifen-induced signaling suppresses ADM signaling to regulate lymphatic cell mosaic expression of DNMAML in LECs. development. While our gene expression studies begin to An increase in the closure of the two lymphangiogenic elucidate some of the mechanisms by which Dll4/Notch1 fronts was observed in Notch4 mutants that correlated with and Dll4/Notch4 signaling regulates lymphangiogenesis, reduced vessel caliber in the absence of a change in LEC proliferation. This phenotype is consistent with an increase 1 3 220 Angiogenesis (2022) 25:205–224 Fig. 8 Loss of Notch4 and dele- tion of LEC canonical Notch signaling resulted in distinct lymphatic phenotypes at E17.5. a, b Lymphangiography of E17.5 wild-type (n = 6) and −/− Notch4 (n = 8) embryos. Representative images of wild- −/− type and Notch4 embryos. Boxed area enlarged to the −/− right. b Notch4 embryos with blood-filled dermal lymphatics (white arrowheads). Boxed area enlarged to the T2 right. c, d Prox1CreER and fl/fl DNMAML mice were crossed and tamoxifen administered at E12.5 and lymphangiog- raphy performed at E17.5. Control (n = 12), DNMAML- LEC (n = 5). c Representative fl/− images of DNMAML control LEC and DNMAML embryos. Boxed area enlarged to the LEC right. d DNMAML embryo with leaking dermal lymphatic vessels (red arrowheads). Boxed area enlarged below in LEC migration toward the midline. In HdLECs, ectopic NOTCH4 peptide that suppresses Notch1 signaling by Notch4 activation inhibited LEC migration significantly functioning as a ligand trap [52]. However, a loss of canoni- more than Notch1 activation. This inhibition of LEC migra- cal Notch signaling was not observed in the lymphatics of tion by Notch4 may occur via non-canonical Notch signal- Notch4 mutant mice, which would be predicted if Notch1 ing, as expression of DNMAML, an inhibitor of canonical signaling was inhibited in the model. Moreover, the Notch4 Notch signaling, did not affect the closure of the lymphangi- mutant dermal lymphatic phenotype is distinct from that ogenic fronts. Notch4 may suppress LEC migration via its observed in mice with Notch1 deleted in the LECs [14], as LEC interactions with Wnt/β-catenin signaling. Non-canonical well as the DNMAML mice. The dermal lymphatic phe- Notch4 signaling has been shown to antagonize Wnt/β- notype however may be due to loss of Notch4 in non-LECs, catenin signaling in stem and progenitor cells [29, 51]. In such as macrophages, and a conditional Notch4 allele needs LECs, loss of β-catenin signaling reduced LEC migration to be developed to better understand the cell type specific toward the midline and increased dermal lymphatic vessel requirement for NOTCH4 in lymphatic development. caliber [29, 51], phenotypes opposite to that observed in Together with published data, our studies suggest that −/− Notch4 embryos, suggesting that Notch4 via a non-canon- Notch1 and Notch4 function distinctly in embryonic dermal ical signaling suppresses LEC migration. lymphangiogenesis via a RBPjκ-dependent and -independ- Western analysis of embryo lysates and immunostaining ent pathways. We propose that Dll4/Notch1 signaling via of tissue sections using an antibody against the cytoplas- a canonical pathway suppresses LEC proliferation, while mic domain of NOTCH4 demonstrated a loss of NOTCH4 Notch4 signaling suppresses LEC migration and branching, expression in the Notch4 nulls. It has been suggested that possibly via a RBPjκ-dependent mechanism. Further studies this Notch4 null line expresses a truncated extracellular into the mechanistic interaction between Notch1 and Notch4 1 3 Angiogenesis (2022) 25:205–224 221 Fig. 9 Canonical Notch signaling in LECs is unaltered in the Notch4 mutant dermal lymphatics. +/− +/ +/− Notch4 ;ProxTOM ;NVR males were bred with +/− +/ +/− Notch4 ;ProxTom ;NVR females, and E14.5 ProxTom; NVR and −/− Notch4 ;ProxTom;NVR dermal wholemounts stained for LYVE1. a Representative images of blunted-ended and spiky-ended sprouts at the lym- phangiogenic front of wild-type −/− (WT) and Notch4 dermis. b Representative images of branch-points with Notch sign- −/− aling in WT and Notch4 in the maturing lymphatic vascular plexus. Scale bars, 50 μm Acknowledgements The authors thank Valeriya Borisenko and Marina in LECs and lymphatic development and homeostasis are Vorontchikhina for technical assistance, June Wu for critical reading necessary, as a number of therapeutics that are pan-Notch fl/fl of the manuscript, and Warren Pear (DNMAML ), Tom Gridley inhibitors or target specific receptors or ligands are currently −/− T2 (Notch4 ), Guillermo Oliver (Prox1CreER ), and Hong Young in clinical trials or the research pipeline for use in the clinic. Kwon (Prox1-tdTomato) for providing mice. Supplementary Information The online version contains supplemen- Author contributions AM and MKU share first authorship. AM, MKU, tary material available at https://doi. or g/10. 1007/ s10456- 021- 09822-5 . YM, JKK, CJS contributed to the study conception and design. Mate- rial preparation, data collection, and analysis were performed by AM, 1 3 222 Angiogenesis (2022) 25:205–224 MKU, GSD, BS, JMJ, AM, SWY, JDM, CK, MG, GR, CJS. The first and reviews in molecular. Cell Dev Biol 26(3):225–234. https:// draft of the manuscript was written by AM, MKU, and CJS and revised doi. org/ 10. 1002/ bies. 20004 by CJS. 6. Hellstrom M, Phng LK, Hofmann JJ, Wallgard E, Coultas L, Lindblom P, Alva J, Nilsson AK, Karlsson L, Gaiano N, Yoon K, Rossant J, Iruela-Arispe ML, Kalen M, Gerhardt H, Betsholtz C Funding This study was funded by the NIH/NCI (R01CA136673; (2007) Dll4 signalling through Notch1 regulates formation of tip CJS, JKK), NIH/NIDDK (R01 R01DK107633; CJS), NIH/ cells during angiogenesis. Nature 445(7129):776–780. doi:https:// NHLBI (RO1HL112626; JKK), the DOD pre-doctoral fellowship doi. org/ 10. 1038/ natur e05571 (W81XWH-10-1-0304; MKU), and the Lipedema Foundation (CJS). 7. Lobov IB, Renard RA, Papadopoulos N, Gale NW, Thurston G, These studies used the resources of the Herbert Irving Comprehen- Yancopoulos GD, Wiegand SJ (2007) Delta-like ligand 4 (Dll4) is sive Cancer Center Flow Cytometry Shared Resources funded in part induced by VEGF as a negative regulator of angiogenic sprouting. through Center Grant P30CA013696. Proc Natl Acad Sci USA 104(9):3219–3224. doi:https:// doi. org/ 10. 1073/ pnas. 06112 06104 Declarations 8. Suchting S, Freitas C, le Noble F, Benedito R, Breant C, Duarte A, Eichmann A (2007) The Notch ligand Delta-like 4 negatively Conflict of interest Jan Kitajewski has received research funding from regulates endothelial tip cell formation and vessel branching. Eisai Pharmaceuticals (CU12-3625 and UICID#084028 Eisai Ltd. Re- Proc Natl Acad Sci USA 104(9):3225–3230. doi:https:// doi. org/ search Collaborative Agreements). All other authors declare that they 10. 1073/ pnas. 06111 77104 have no conflict of interests. 9. Tammela T, Zarkada G, Wallgard E, Murtomaki A, Suchting S, Wirzenius M, Waltari M, Hellstrom M, Schomber T, Peltonen Ethical approval Isolation of HUVEC and HdLEC from anonymous R, Freitas C, Duarte A, Isoniemi H, Laakkonen P, Christofori G, discarded specimens and received IRB exempt status by Columbia Yla-Herttuala S, Shibuya M, Pytowski B, Eichmann A, Betsholtz University IRB (AAAA7338). All procedures performed in studies C, Alitalo K (2008) Blocking VEGFR-3 suppresses angiogenic involving human participants were in accordance with the ethical stand- sprouting and vascular network formation. Nature 454(7204):656– ards of the institutional and/or national research committee and with 660. doi:https://doi. org/ 10. 1038/ natur e07083 the 1964 Helsinki declaration and its later amendments or comparable 10. Bernier-Latmani J, Cisarovsky C, Demir CS, Bruand M, Jaquet ethical standards. Mouse studies were approved by Columbia Univer- M, Davanture S, Ragusa S, Siegert S, Dormond O, Benedito sity IACUC (AC-AAAE2653, AC-AAAD0577, AC-AAAP9603, AC- R, Radtke F, Luther SA, Petrova TV (2015) DLL4 promotes AAAP0452, AC-AABB9551). All procedures performed in studies continuous adult intestinal lacteal regeneration and dietary fat involving animals were in accordance with the ethical standards of the transport. J Clin Investig 125(12):4572–4586. doi:https:// doi. institution or practice at which the studies were conducted. org/10. 1172/ JCI82 045 11. Zheng W, Tammela T, Yamamoto M, Anisimov A, Holopainen T, Kaijalainen S, Karpanen T, Lehti K, Yla-Herttuala S, Alitalo Open Access This article is licensed under a Creative Commons Attri- K (2011) Notch restricts lymphatic vessel sprouting induced by bution 4.0 International License, which permits use, sharing, adapta- vascular endothelial growth factor. Blood 118(4):1154–1162. tion, distribution and reproduction in any medium or format, as long doi:https://doi. org/10. 1182/ blood- 2010- 11- 317800 as you give appropriate credit to the original author(s) and the source, 12. Shawber CJ, Funahashi Y, Francisco E, Vorontchikhina M, provide a link to the Creative Commons licence, and indicate if changes Kitamura Y, Stowell SA, Borisenko V, Feirt N, Podgrabinska were made. The images or other third party material in this article are S, Shiraishi K, Chawengsaksophak K, Rossant J, Accili D, included in the article's Creative Commons licence, unless indicated Skobe M, Kitajewski J (2007) Notch alters VEGF responsive- otherwise in a credit line to the material. If material is not included in ness in human and murine endothelial cells by direct regulation the article's Creative Commons licence and your intended use is not of VEGFR-3 expression. J Clin Investig 117(11):3369–3382. permitted by statutory regulation or exceeds the permitted use, you will doi:https://doi. org/10. 1172/ JCI24 311 need to obtain permission directly from the copyright holder. To view a 13. Niessen K, Zhang G, Ridgway JB, Chen H, Kolumam G, copy of this licence, visit http://cr eativ ecommons. or g/licen ses/ b y/4.0/ . Siebel CW, Yan M (2011) The Notch1-Dll4 signaling path- way regulates mouse postnatal lymphatic development. Blood 118(7):1989 –1997. doi: https:// doi. org/ 10. 1182/ blood- 2010- 11- 319129 References 14. Fatima A, Culver A, Culver F, Liu T, Dietz WH, Thomson BR, Hadjantonakis AK, Quaggin SE, Kume T (2014) Murine Notch1 is 1. James JM, Nalbandian A, Mukouyama YS (2013) TGFbeta signal- required for lymphatic vascular morphogenesis during development. ing is required for sprouting lymphangiogenesis during lymphatic Dev Dyn 243(7):957–964. https:// doi. org/ 10. 1002/ dvdy.24129 network development in the skin. Development 140(18):3903– 15. Geng X, Yanagida K, Akwii RG, Choi D, Chen L, Ho Y, Cha 3914. doi:https://doi. org/ 10. 1242/ dev. 095026 B, Mahamud MR, Berman de Ruiz K, Ichise H, Chen H, Wythe 2. D’Souza B, Meloty-Kapella L, Weinmaster G (2010) Canonical JD, Mikelis CM, Hla T, Srinivasan RS (2020) S1PR1 regulates and non-canonical Notch ligands. Curr Top Dev Biol 92:73–129. the quiescence of lymphatic vessels by inhibiting laminar shear doi:https:// doi. org/ 10. 1016/ S0070- 2153(10) 92003-6 stress-dependent VEGF-C signaling. JCI Insight 5(14):e137652 3. Baeten JT, Lilly B (2017) Notch signaling in vascular smooth 16. Funahashi Y, Shawber CJ, Sharma A, Kanamaru E, Choi YK, muscle cells. Adv Pharmacol 78:351–382. https:// doi. org/ 10. Kitajewski J (2011) Notch modulates VEGF action in endothe- 1016/ bs. apha. 2016. 07. 002 lial cells by inducing matrix metalloprotease activity. Vasc Cell 4. Li JL, Harris AL (2009) Crosstalk of VEGF and Notch pathways 3(1):2. https:// doi. org/ 10. 1186/ 2045- 824X-3-2 in tumour angiogenesis: therapeutic implications. Front Biosci 17. Funahashi Y, Shawber CJ, Vorontchikhina M, Sharma A, Outtz 14:3094–3110 HH, Kitajewski J (2010) Notch regulates the angiogenic response 5. Shawber CJ, Kitajewski J (2004) Notch function in the vascula- via induction of VEGFR-1. J Angiogenesis Res 2(1):3. https://doi. ture: insights from zebrafish, mouse and man. BioEssays: news org/ 10. 1186/ 2040- 2384-2-3 1 3 Angiogenesis (2022) 25:205–224 223 18. Murtomaki A, Uh MK, Choi YK, Kitajewski C, Borisenko V, 33. Farnsworth RH, Karnezis T, Maciburko SJ, Mueller SN, Stacker SA Kitajewski J, Shawber CJ (2013) Notch1 functions as a nega- (2019) The interplay between lymphatic vessels and chemokines. tive regulator of lymphatic endothelial cell differentiation in Front Immunol 10:518. https://doi. or g/10. 3389/ fimmu. 2019.00518 the venous endothelium. Development 140(11):2365–2376. 34. Kanady JD, Dellinger MT, Munger SJ, Witte MH, Simon AM doi:https:// doi. org/ 10. 1242/ dev. 083865 (2011) Connexin37 and connexin43 deficiencies in mice disrupt 19. Tung JJ, Hobert O, Berryman M, Kitajewski J (2009) Chloride lymphatic valve development and result in lymphatic disorders intracellular channel 4 is involved in endothelial proliferation and including lymphedema and chylothorax. Dev Biol 354(2):253–266. morphogenesis in vitro. Angiogenesis 12(3):209–220. doi:https:// https:// doi. org/ 10. 1016/j. ydbio. 2011. 04. 004 doi. org/ 10. 1007/ s10456-009- 9139-3 35. Klein KR, Karpinich NO, Espenschied ST, Willcockson HH, Dun- 20. Uyttendaele H, Marazzi G, Wu G, Yan Q, Sassoon D, Kitajew- worth WP, Hoopes SL, Kushner EJ, Bautch VL, Caron KM (2014) ski J (1996) Notch4/Int-3, a mammary proto-oncogene, is an Decoy receptor CXCR7 modulates adrenomedullin-mediated car- endothelial cell-specific mammalian Notch gene. Development diac and lymphatic vascular development. Dev Cell 30(5):528–540. 122(7):2251–2259 https:// doi. org/ 10. 1016/j. devcel. 2014. 07. 012 21. Geback T, Schulz MM, Koumoutsakos P, Detmar M (2009) 36. Lee KM, Danuser R, Stein JV, Graham D, Nibbs RJ, Graham GJ TScratch: a novel and simple software tool for automated analysis (2014) The chemokine receptors ACKR2 and CCR2 reciprocally of monolayer wound healing assays. Biotechniques 46(4):265– regulate lymphatic vessel density. EMBO J 33(21):2564–2580. 274. doi:https://doi. org/ 10. 2144/ 00011 3083 doi:https:// doi. org/ 10. 15252/ embj. 20148 8887 22. Schmittgen TD, Livak KJ (2008) Analyzing real-time PCR data 37. Uchida Y, James JM, Suto F, Mukouyama YS (2015) Class 3 sema- by the comparative C(T) method. Nat Protoc 3(6):1101–1108. phorins negatively regulate dermal lymphatic network formation. https:// doi. org/ 10. 1038/ nprot. 2008. 73 Biol Open 4(9):1194–1205. doi:https:// doi.or g/10. 1242/bio. 012302 23. Shawber CJ, Das I, Francisco E, Kitajewski J (2003) Notch signal- 38. Zhuo W, Jia L, Song N, Lu XA, Ding Y, Wang X, Song X, Fu Y, Luo ing in primary endothelial cells. Ann N Y Acad Sci 995:162–170. Y (2012) The CXCL12-CXCR4 chemokine pathway: a novel axis doi:https:// doi. org/ 10. 1111/j. 1749- 6632. 2003. tb032 19.x regulates lymphangiogenesis. Clin Cancer Res 18(19):5387–5398. 24. Krebs LT, Xue Y, Norton CR, Shutter JR, Maguire M, Sundberg doi:https:// doi. org/ 10. 1158/ 1078- 0432. CCR- 12- 0708 JP, Gallahan D, Closson V, Kitajewski J, Callahan R, Smith GH, 39. Karpinich NO, Caron KM (2014) Apelin signaling: new G protein- Stark KL, Gridley T (2000) Notch signaling is essential for vas- coupled receptor pathway in lymphatic vascular development. Arte- cular morphogenesis in mice. Genes Dev 14(11):1343–1352 rioscler Thromb Vasc Biol 34(2):239–241. doi:https:// doi. org/ 10. 25. Srinivasan RS, Dillard ME, Lagutin OV, Lin FJ, Tsai S, Tsai MJ, 1161/ ATVBA HA. 113. 302905 Samokhvalov IM, Oliver G (2007) Lineage tracing demonstrates the 40. Lohela M, Saaristo A, Veikkola T, Alitalo K (2003) Lymphangi- venous origin of the mammalian lymphatic vasculature. Genes Dev ogenic growth factors, receptors and therapies. Thromb Haemost 21(19):2422–2432. doi:https://doi. org/10. 1101/ gad. 15884 07 90(2):167–184. doi:https://doi. org/ 10. 1160/ TH03- 04- 0200 26. Tu L, Fang TC, Artis D, Shestova O, Pross SE, Maillard I, Pear WS 41. Benedito R, Rocha SF, Woeste M, Zamykal M, Radtke F, Casano- (2005) Notch signaling is an important regulator of type 2 immunity. vas O, Duarte A, Pytowski B, Adams RH (2012) Notch-depend- J Exp Med 202(8):1037–1042. https://doi. or g/10. 1084/ jem. 20050 923 ent VEGFR3 upregulation allows angiogenesis without VEGF- 27. Nowotschin S, Xenopoulos P, Schrode N, Hadjantonakis AK (2013) VEGFR2 signalling. Nature 484(7392):110–114. doi:https:// doi. A bright single-cell resolution live imaging reporter of Notch signal- org/ 10. 1038/ natur e10908 ing in the mouse. BMC Dev Biol 13:15. doi:https:// doi.or g/10. 1186/ 42. Costa MJ, Wu X, Cuervo H, Srinivasan R, Bechis SK, Cheang E, 1471- 213X- 13- 15 Marjanovic O, Gridley T, Cvetic CA, Wang RA (2013) Notch4 is 28. Hong M, Jung E, Yang S, Jung W, Seong YJ, Park E, Bramos A, required for tumor onset and perfusion. Vasc Cell 5(1):7. https:// doi. Kim KE, Lee S, Daghlian G, Seo JI, Choi I, Choi IS, Koh CJ, Kob- org/ 10. 1186/ 2045- 824X-5-7 ielak A, Ying QL, Johnson M, Gardner D, Wong AK, Choi D, Hong 43. Zhang Y, Ulvmar MH, Stanczuk L, Martinez-Corral I, Frye M, Ali- YK (2016) Efficient assessment of developmental, surgical and talo K, Makinen T (2018) Heterogeneity in VEGFR3 levels drives pathological lymphangiogenesis using a lymphatic reporter mouse lymphatic vessel hyperplasia through cell-autonomous and non-cell- and its embryonic stem cells. PloS One 11(6):e0157126. https:// doi. autonomous mechanisms. Nat Commun 9(1):1296. doi:https:// doi. org/ 10. 1371/ journ al. pone. 01571 26 org/ 10. 1038/ s41467- 018- 03692-0 29. Cha B, Geng X, Mahamud MR, Fu J, Mukherjee A, Kim Y, Jho EH, 44. Martinez-Corral I, Zhang Y, Petkova M, Ortsater H, Sjoberg S, Cas- Kim TH, Kahn ML, Xia L, Dixon JB, Chen H, Srinivasan RS (2016) tillo SD, Brouillard P, Libbrecht L, Saur D, Graupera M, Alitalo Mechanotransduction activates canonical Wnt/beta-catenin signal- K, Boon L, Vikkula M, Makinen T (2020) Blockade of VEGF-C ing to promote lymphatic vascular patterning and the development of signaling inhibits lymphatic malformations driven by oncogenic lymphatic and lymphovenous valves. Genes Dev 30(12):1454–1469. PIK3CA mutation. Nat Commun 11(1):2869. doi:https://doi. org/ doi:https:// doi. org/ 10. 1101/ gad. 282400. 116 10. 1038/ s41467- 020- 16496-y 30. Vorontchikhina MA, Zimmermann RC, Shawber CJ, Tang H, Kita- 45. MacKenzie F, Duriez P, Wong F, Noseda M, Karsan A (2004) jewski J (2005) Unique patterns of Notch1, Notch4 and Jagged1 Notch4 inhibits endothelial apoptosis via RBP-Jkappa-dependent expression in ovarian vessels during folliculogenesis and corpus and -independent pathways. J Biol Chem 279(12):11657–11663. luteum formation. Gene Expr Patterns 5(5):701–709. doi:https:// doi:https:// doi. org/ 10. 1074/ jbc. M3121 02200 doi. org/ 10. 1016/j. modgep. 2005. 02. 001 46. Raafat A, Bargo S, McCurdy D, Callahan R (2017) The ANK 31. Du LL, Liu P (2016) CXCL12/CXCR4 axis regulates neovasculari- repeats of Notch-4/Int3 activate NF-kappaB canonical pathway in zation and lymphangiogenesis in sutured corneas in mice. Mol Med the absence of Rbpj and causes mammary tumorigenesis. Scientific Rep 13(6):4987–4994. doi:https://doi. or g/10. 3892/mmr .2016. 5179 reports 7(1):13690. doi:https://doi. or g/10. 1038/ s41598- 017- 13989-7 32. Abengozar MA, de Frutos S, Ferreiro S, Soriano J, Perez-Martinez 47. Raafat A, Lawson S, Bargo S, Klauzinska M, Strizzi L, Goldhar M, Olmeda D, Marenchino M, Canamero M, Ortega S, Megias D, AS, Buono K, Salomon D, Vonderhaar BK, Callahan R (2009) Rodriguez A, Martinez-Torrecuadrada JL (2012) Blocking ephrinB2 Rbpj conditional knockout reveals distinct functions of Notch4/ with highly specific antibodies inhibits angiogenesis, lymphangi- Int3 in mammary gland development and tumorigenesis. Oncogene ogenesis, and tumor growth. Blood 119(19):4565–4576. doi:https:// 28(2):219–230. doi:https://doi. org/ 10. 1038/ onc. 2008. 379 doi. org/ 10. 1182/ blood- 2011- 09- 380006 48. Liu X, Luo Q, Zheng Y, Liu X, Hu Y, Liu W, Luo M, Zhao Y, Zou L (2016) NOTCH4 signaling controls EFNB2-induced 1 3 224 Angiogenesis (2022) 25:205–224 endothelial progenitor cell dysfunction in preeclampsia. Reproduc- 51. Andersen P, Uosaki H, Shenje LT, Kwon C (2012) Non-canonical tion 152(1):47–55. doi:https://doi. org/10. 1530/ REP- 16- 0132 Notch signaling: emerging role and mechanism. Trends Cell Biol 49. Pedrosa AR, Trindade A, Fernandes AC, Carvalho C, Gigante J, 22(5):257–265. doi:https://doi. org/ 10. 1016/j. tcb. 2012. 02. 003 Tavares AT, Dieguez-Hurtado R, Yagita H, Adams RH, Duarte A 52. James AC, Szot JO, Iyer K, Major JA, Pursglove SE, Chapman G, (2015) Endothelial Jagged1 antagonizes Dll4 regulation of endothe- Dunwoodie SL (2014) Notch4 reveals a novel mechanism regulating lial branching and promotes vascular maturation downstream of Notch signal transduction. Biochim Biophys Acta 1843(7):1272– Dll4/Notch1. Arterioscler Thromb Vasc Biol 35(5):1134–1146. 1284. doi:https://doi. org/ 10. 1016/j. bbamcr. 2014. 03. 015 https:// doi. org/ 10. 1161/ ATVBA HA. 114. 304741 50. Murtomaki A, Uh MK, Kitajewski C, Zhao J, Nagasaki T, Shawber Publisher’s Note Springer Nature remains neutral with regard to CJ, Kitajewski J (2014) Notch signaling functions in lymphatic valve jurisdictional claims in published maps and institutional affiliations. formation. Development 141(12):2446–2451. doi:https://doi. or g/10. 1242/ dev. 101188 1 3

Journal

Angiogenesis – Springer Journals

Published: May 1, 2022

Keywords: Lymphangiogenesis; Notch; VEGF-C; Dermis

There are no references for this article.