1. Cell Biology
  2. Human Biology and Medicine
Download icon

Vascular remodeling is governed by a VEGFR3-dependent fluid shear stress set point

  1. Nicolas Baeyens
  2. Stefania Nicoli
  3. Brian G Coon
  4. Tyler D Ross
  5. Koen Van den Dries
  6. Jinah Han
  7. Holly M Lauridsen
  8. Cecile O Mejean
  9. Anne Eichmann
  10. Jean-Leon Thomas
  11. Jay D Humphrey
  12. Martin A Schwartz Is a corresponding author
  1. Yale University School of Medicine, United States
  2. Yale University School of Engineering and Applied Science, United States
  3. Université Pierre et Marie Curie, France
  4. INSERM, CNRS U-1127, UMR-7225, France
  5. Groupe Hospitalier Pitié Salpêtrière, France
Research Article
Cited
26
Views
3,294
Comments
0
Cite as: eLife 2015;4:e04645 doi: 10.7554/eLife.04645

Abstract

Vascular remodeling under conditions of growth or exercise, or during recovery from arterial restriction or blockage is essential for health, but mechanisms are poorly understood. It has been proposed that endothelial cells have a preferred level of fluid shear stress, or ‘set point’, that determines remodeling. We show that human umbilical vein endothelial cells respond optimally within a range of fluid shear stress that approximate physiological shear. Lymphatic endothelial cells, which experience much lower flow in vivo, show similar effects but at lower value of shear stress. VEGFR3 levels, a component of a junctional mechanosensory complex, mediate these differences. Experiments in mice and zebrafish demonstrate that changing levels of VEGFR3/Flt4 modulates aortic lumen diameter consistent with flow-dependent remodeling. These data provide direct evidence for a fluid shear stress set point, identify a mechanism for varying the set point, and demonstrate its relevance to vessel remodeling in vivo.

https://doi.org/10.7554/eLife.04645.001

eLife digest

Blood and lymphatic vessels remodel their shape, diameter and connections during development, and throughout life in response to growth, exercise and disease. This process is called vascular remodeling.

The endothelial cells that line the inside of blood and lymphatic vessels are constantly exposed to the frictional force from flowing blood, termed fluid shear stress. Changes in shear stress are sensed by the endothelial cells, which trigger vascular remodeling to return the stress to the original level. It has been proposed that remodeling is governed by a preferred level of fluid shear stress, or set point, against which deviations in the shear stress are compared. Thus, changing the fluid flow through a blood vessel increases or decreases shear stress, which results in the vessel remodeling to restore the original level of shear stress. Like all remodeling, this process involves inflammation to recruit white blood cells, which assist with the process.

Baeyens et al. investigated whether such a shear stress set point exists and what its biological basis might be using cultured endothelial cells from human umbilical veins. These cells remained stable and in a resting state when a particular level of shear stress was applied to them; above or below this shear stress level, the cells produced an inflammatory response like that seen during vascular remodeling. This suggests that these cells do indeed have a set point for shear stress. The same response occurred in human lymphatic endothelial cells, although in these cells the shear stress set point was much lower, correlating with the low flow in lymphatic vessels.

Baeyens et al. then discovered that the shear stress set point is related to the level of a protein called VEGFR3 in the cells, which was recently found to participate in shear stress sensing. Endothelial cells from lymphatic vessels normally produce much greater quantities of VEGFR3 than those from blood vessels. Reducing the amount of VEGFR3 in lymphatic endothelial cells increased the set point shear stress, while increasing the levels in blood vessel cells decreased the set point. This suggests that the levels of this protein account for the difference in the response of these two cell types. Baeyens et al. then tested this pathway by reducing the levels of VEGFR3 in zebrafish embryos and in adult mice. In both animals, this caused arteries to narrow, showing that VEGFR3 levels also control sensitivity to shear stress—and hence vascular remodeling—inside living creatures.

Understanding in detail how vascular remodeling is regulated could help improve treatments for a wide range of cardiovascular conditions. To do so, further work will be needed to develop methods to control the sensitivity of endothelial cells to shear stress and to identify other proteins that might specifically control the narrowing or the expansion of vessels in human patients.

https://doi.org/10.7554/eLife.04645.002

Introduction

Homeostasis, one of the central concepts in physiology (Cannon, 1929), posits that physiological variables have an optimum value or set point such that deviations from that set point activate responses that return those variables toward their original value. For example, changes in central body temperature trigger sweating, altered blood flow to the skin or shivering to restore normal temperature. In the vasculature, arteries remodel under sustained changes in blood flow, with increased or decreased flow triggering outward or inward remodeling, respectively, to adjust lumen diameters accordingly (Thoma, 1893; Kamiya and Togawa, 1980; Kamiya et al., 1984; Langille and O'Donnell, 1986; Langille et al., 1989; Langille, 1996; Tronc et al., 1996; Tuttle et al., 2001). These studies have given rise to the concept that the endothelium encodes a fluid shear stress set point that governs remodeling responses (Rodbard, 1975; Cardamone and Humphrey, 2012) (Figure 1A). While appealing, there is no direct evidence for such a mechanism. Moreover, if it exists, the set point must itself be variable, since different types of vessels, for example, arteries, veins and lymphatics, generally have very different magnitudes of fluid shear stress (Lipowsky et al., 1980; Dixon et al., 2006; Suo et al., 2007).

Testing the set point hypothesis.

(A) Definition of the ‘shear stress set point’. (B) Picture of a silicone gasket used in the gradient flow chamber with the corresponding calculation of the theoretical shear stress level across the channel with two different conditions of gasket thickness and flow rate.

https://doi.org/10.7554/eLife.04645.003

Arterial remodeling is crucial in normal physiological adaptation to growth and exercise, and is a major determinant of outcomes in cardiovascular disease (Kohler et al., 1991; Corti et al., 2011; Padilla et al., 2011). Outward remodeling of atherosclerotic vessels helps to maintain lumen diameter and blood flow, whereas inward remodeling leads to ischemia associated with angina and peripheral vascular disease (Ward et al., 2000). Additionally, flow-dependent remodeling of small blood vessels near sites of myocardial infarction provides collateral circulation that plays a major role in restoring cardiac function (Heil and Schaper, 2004), whereas failure to remodel is a major factor in progression to heart failure.

Flow-dependent remodeling is initiated by inflammatory activation of the endothelium, leading to recruitment of leukocytes that assist with remodeling in several ways including secretion of matrix metalloproteinases, cytokines and extracellular matrix proteins (Silvestre et al., 2008; Schaper, 2009; Silvestre et al., 2013). Once the remodeling phase is completed, inflammation is resolved and the vascular wall stabilized. NF-κB plays a major role in the initial inflammatory activation (Castier et al., 2009; Sweet et al., 2013), whereas signaling through TGF-β is critical in the anti-inflammatory, stabilization phase (Walshe et al., 2009) .

These considerations led us to investigate the existence of a fluid shear stress set point and its relevance to vascular remodeling. Our results provide strong evidence for a fluid shear stress set point in vascular endothelium. They also show that vascular and lymphatic endothelium have different set points, that this difference is mediated by differences in expression of VEGFR3, and provide evidence that this pathway controls artery remodeling in vivo.

Results

Is there a set point for fluid shear stress?

To test the existence of a shear stress set point, we built a flow chamber that creates a gradient of shear stress along a single culture slide. Following a previous design (Usami et al., 1993), the width of the chamber progressively decreases to yield a linear gradient (Figure 1B). We then measured several biological responses associated with fluid shear stress and vascular remodeling. To assay responses as a function of shear stress, we took successive microscopic images along the chamber. Depending on localization, these responses correlated with calculated values of shear stress. Changing the gasket thickness and flow rate allowed us to control the range of shear stress for each experiment (Figure 1B).

We first measured endothelial cell alignment in flow, which is a well-studied response associated with vessel stabilization and suppression of inflammatory pathways (Levesque and Nerem, 1985; Wang et al., 2012; Baeyens et al., 2014). Alignment was quantified by measuring the angle between the major axis of the nucleus and the flow direction (Baeyens et al., 2014). Human umbilical vein endothelial cells (HUVECs) were subjected to 16 hr of laminar shear stress ranging from 2 to 60 dynes.cm−2. HUVECs aligned in the direction of the flow, between approximately 10 and 20 dynes.cm−2, but were misaligned or oriented perpendicularly, against the flow direction, outside this range (Figure 2A, Figure 2—figure supplement 1). This result agrees with previous studies showing perpendicular alignment of endothelial cells under very high shear stress (Viggers et al., 1986; Dolan et al., 2011; Dolan et al., 2012).

Figure 2 with 1 supplement see all
Set point for shear stress.

(A) Cell orientation: the average orientation of HUVEC nuclei was measured in each picture, to obtain average orientation at a given shear stress. (n = 16, Mean ± SEM, ANOVA: F = 15.02, p < 0.0001). With no flow, cell orientation was random (average = 45°). (B) NF-κB activation: p65 nuclear translocation in HUVEC was measured either in no flow (dotted line: average) or after 16 hr of flow in the gradient chamber (n = 6, Mean ± SEM, ANOVA: F = 10.97, p < 0.0001). (C) Smad1 activation: Smad1 nuclear translocation in HUVECs was measured without flow (dotted line: average) or after 16 hr of flow in the gradient chamber (n = 6, Mean ± SEM, ANOVA: F = 13.47, p < 0.0001).

https://doi.org/10.7554/eLife.04645.004

Next, to assess NF-κB activation, we measured the nuclear translocation of the p65 subunit of NF-κB. NF-κB showed baseline activation in cells without flow, which decreased between approximately 10 and 25 dynes.cm−2, and dramatically increased at very high shear (Figure 2B, Figure 2—figure supplement 1). The suppression of NF-κB translocation in this range is consistent with previous observations that sustained laminar flow is anti-inflammatory (Mohan et al., 1997; Berk, 2008). Lastly, we measured the activation of TGFβ/SMAD signaling by assaying nuclear translocation of Smad1. Strikingly, flow induced Smad translocation with a sharp maximum between 10 and 20 dynes.cm−2 and repressed translocation at higher values (Figure 2C, Figure 2—figure supplement 1). The results obtained with the gradient chamber were validated by examining 2, 12 and 50 dynes.cm−2 using normal parallel flow chambers (Figure 2—figure supplement 1).

Taken together, these results show that HUVECs have a biphasic response to shear stress such that anti-inflammatory, stabilization pathways are activated between approximately 10 and 20 dynes.cm−2, while lower or higher shear stress is pro-inflammatory. This behavior is consistent with a shear stress set point within the range of 10 and 20 dynes.cm−2 for these cells.

Analysis of lymphatic endothelial cells

An essential aspect of the set point hypothesis is that it must differ between different types of vessels. In vivo, average shear stress in lymphatic vessels is much lower than in arteries or veins (Lipowsky et al., 1980; Dixon et al., 2006; Suo et al., 2007). We therefore examined the behavior of human dermal lymphatic endothelial cells (HDLEC), using modified chamber parameters to obtain values of shear stress from 0.5 to 20 dynes.cm−2 (Figure 1). In these experiments, HUVECs aligned between 8 and 20 dynes.cm−2, (Figure 2A and Figure 3A) whereas HDLEC aligned maximally between 4 and 6 dynes.cm−2 (Figure 3A, Figure 3—figure supplement 1). The minimum for NF-κB translocation also shifted to between 4 and 10 dynes.cm−2 (Figure 3B, Figure 3—figure supplement 1), which corresponds well to in vivo measurements (Dixon et al., 2006). These results indicate that lymphatics have a higher sensitivity to shear stress compared to HUVECs, consistent with the set point concept.

Figure 3 with 1 supplement see all
Set point in HUVECs vs lymphatic endothelial cells.

(A) The average orientation of venous cell (HUVEC) or lymphatic cell (HDLEC) nuclei across the slide was measured as in Figure 2A. (n = 11, Mean ± SEM). The difference between HUVECs and HDLECs is statistically significant (ANOVA Two-way, p < 0.0001). (B) NF-κB activation: p65 nuclear translocation in HDLEC was measured either in no flow (dotted line: average) or after 16 hr of flow in the gradient chamber (n = 4, Mean ± SEM, ANOVA: F = 34.32, p < 0.0001). (C) Expression of VE-cadherin, PECAM-1, VEGFR2 and VEGFR3, proteins involved in the shear stress mechanotransduction through the junctional complex. Actin was used as a loading control.

https://doi.org/10.7554/eLife.04645.006

VEGFR3 expression regulates the set point for shear stress in vitro

A number of shear stress responses, including cell alignment and NF-κB activation, require mechanotransduction via VEGFR2, whose ligand-independent transactivation by flow requires PECAM-1 and VE-cadherin (Tzima et al., 2005). We therefore considered whether differences in expression of these proteins might account for the difference in flow sensitivity between HUVECs and HDLECs. However, no major differences in levels of these proteins were observed (Figure 3C). VEGFR3, a close homolog of VEGFR2, is highly expressed in lymphatic cells (Kaipainen et al. 1995) and recent work in our lab showed that it is activated by flow in vascular endothelial cells similarly to VEGFR2 (Coon et al., 2015). These considerations prompted us to examine levels of this receptor as well, which showed approximately 10-fold higher expression in lymphatic ECs compared to HUVECs (Figure 3C). We therefore considered whether VEGFR3 levels might be responsible for the higher flow sensitivity of lymphatic ECs.

HDLECs were therefore transfected with VEGFR3 siRNA, which reduced its expression to approximate the level in HUVECs (Figure 4A). We also transduced HUVECs with adenovirus coding for hVEGFR3-GFP (Figure 4A), which increased levels by ∼10-fold and infected >90% of the cells (Figure 4—figure supplement 1). Cell alignment in flow was then analyzed. Depletion of VEGFR3 in HDLECs shifted the optimal alignment to between 10 to 20 dynes.cm−2 (Figure 4B, Figure 4—figure supplement 2), similar to HUVECs. Conversely, over-expression of VEGFR3 in HUVECs decreased the optimal response toward the lower shear stress levels seen with lymphatic ECs (Figure 4C, Figure 4—figure supplement 2). Taken together, these results show that VEGFR3 levels are a major determinant of the difference in shear stress sensitivity between HUVECs and HDLECs.

Figure 4 with 2 supplements see all
VEGFR3 expression controls the shear stress set point.

(A) Western Blot of VEGFR3 and GFP in HDLECs with and without VEGFR3 siRNA (10 nM), and in HUVECs with and without adenoviral expression of hVEGFR3-GFP. Actin serves as a loading control. (B) Effect of VEGFR3 siRNA in HDLECs on set point. Cell alignment was assayed after shear stress for 16 hr (n = 6). Data were smoothed with a LOWESS fit to improve visualization (mean ± SEM; HDLEC vs HDLC + VEGFR3 siRNA: p = 0.004; HDLEC + VEGFR3 siRNA vs HUVEC: p = 0.45). (C) Effect of VEGFR3 over-expression on set point. Alignment after 16 hr flow was assayed in HUVECs infected with adenovirus expressing mCherry or hVEGFR3-GFP as before. Data were smoothed with a LOWESS fit to improve visualization (n = 10, values are means ± SEM; HUVEC + mCherry vs HUVEC + VEGFR3-GFP: p < 0.0001).

https://doi.org/10.7554/eLife.04645.008

We also confirmed VEGFR3 activation by flow in lymphatic endothelial cells. Onset of flow stimulated VEGFR3 phosphorylation maximally at 6 dynes.cm−2 in HDLEC (Figure 5), which corresponds well to the set point of around 5 dynes.cm−2 in these cells. HUVECs, by contrast, exhibited a weaker response that was shifted to higher shear, consistent with their higher set.

VEGFR3 activation by shear stress.

HDLECs (left) and HUVECs (right) were stimulated for 15 min with shear stress at the indicated levels. VEGFR3 transactivation was assayed by phosphorylation on Y1230, detected by Western blotting with pY1230 antibody (n = 5 independent experiments; *: p < 0.05, **: p < 0.01, ****: p < 0.0001).

https://doi.org/10.7554/eLife.04645.011

VEGFR3 controls blood vessels diameter in zebrafish, in a VEGF-C-independent manner

To test whether VEGFR3 levels control sensitivity to shear stress and vascular remodeling in vivo, we examined Danio rerio (zebrafish). This system has the advantage that development proceeds normally without blood flow, thus, fluid shear stress can be altered or even stopped without affecting viability (Langheinrich et al., 2003). The notion that levels of VEGFR3 (Flt4 in zebrafish) determine the shear stress set point predicts that reducing VEGFR3 expression will induce inward remodeling of the vessels in order to increase shear stress and restore normal signaling. We used a strain in which blood vessels are labeled by expression of kdrl:mCherry (VEGFR2) and flt4:Citrine (VEGFR3) reporters. kdrl:mCherry was highly visible in the dorsal aorta and the posterior cardinal vein, whereas flt4:Citrine was low (though detectable) in the dorsal aorta and higher in the cardinal posterior vein and the developing thoracic duct (Figure 6, Figure 6—figure supplement 1). Flt4/VEGFR3 and its ligand, VEGF-C, are associated with development of lymphatic vasculature and segmental arteries in zebrafish (Covassin et al., 2006; Kuchler et al., 2006). To assay the effect of FLT4 and VEGFC dosage on vessels diameter, we injected zebrafish embryos at the one cell stage with previously validated VEGFC and FLT4 morpholinos at two different concentrations. These antisense oligos target the respective mRNAs and induce a dose dependent loss of function (Nicoli et al., 2012; Villefranc et al., 2013). At 72 hr post fertilization (hpf), the progressive inhibition of VEGFC did not perturb the remodeling of blood vessel or vessel diameter but as expected inhibited the development of the thoracic duct, the first zebrafish lymphatic vessel (Yaniv et al., 2006) (Figure 6, white stars). By contrast, progressive inhibition of FLT4 reduced the diameter of the dorsal aorta with loss of thoracic duct evident at a higher dose of FLT4 morpholino (Figure 6). These results suggested that VEGF-C-independent Flt4 activation is required for artery diameter and exclude an indirect effect of lymphatic development on the artery development. Interestingly, a similar decrease of the dorsal aorta diameter can be observed in a recent paper (Kwon et al., 2013). Although these authors focused on the growth of motoneurons, the dorsal aorta is readily visible in images of Flt1 mCherry reporter embryos; its diameter is obviously smaller in expando embryos expressing a kinase dead Flt4, as well as in wildtype embryos treated with Flt4 morpholino or VEGFR3 inhibitors but not after injection with VEGFC morpholino, in accordance with our own observations.

Figure 6 with 1 supplement see all
VEGFR3 (Flt4) controls blood vessel caliber in zebrafish.

Representative pictures of the dorsal aorta (DA), posterior cardinal vein (PCV) and thoracic duct (white *) at 72 hr post-fertilization (hpf) in wild type zebrafish embryos or embryos injected with Flt4 (VEGFR3) morpholino at 0.06 or 0.1 mM, or with VEGF-C morpholino at 0.06 or 0.1 mM. The mCherry reporter driven by the KDR (VEGFR2) promoter (kdrl:mCherry) is depicted in red and the citrine reporter driven by the Flt4 promoter (flt4:citrine) is depicted in green. Scale = 20 µm and applies to all pictures. n = 6-15 fishes for each condition, whiskers represents the minimum and maximum data point (NS: non-significant, ***: p < 0.001 and ****: p < 0.0001, ANOVA).

https://doi.org/10.7554/eLife.04645.012

To test the role of flow in this process, embryos were treated with 40 μM nifedipine, a voltage-dependent calcium channel blocker that stops the heart and thus blood flow (Langheinrich et al., 2003). Blocking flow led to a decreased vessel diameter (Figure 6, Figure 6—figure supplement 1), supporting the role of shear stress in determining lumen size. Interestingly, lumen diameter was similar in embryos treated with high dose Flt4 morpholino and with nifedipine. To test whether Flt4 acts on a flow pathway, we then combined these treatments. Strikingly, in the absence of flow, neither Flt4 nor VEGF-C morpholinos caused further changes in vessel size. Taken together, these results support the conclusion that VEGF-C-independent activation of VEGFR3 by flow may determine the endothelial cell sensitivity to flow and vessel remodeling, consistent with the existence of a fluid shear stress set point.

Interestingly, ligand-independent responses for VEGFR3 are consistent with developmental mouse phenotypes: deletion of VEGF-C and VEGF-D does not affect the development and maturation of blood vessels during mice development, while deletion of VEGFR3 does (Haiko et al., 2008). Ligand-dependent responses are thus required for lymphangiogenesis but probably not for flow responses.

VEGFR3 and artery remodeling in mice

Lastly, we investigated whether VEGFR3 controls artery remodeling in mice in a similar manner. Expression of VEGFR3 in adult arteries has been reported to be low (Gu et al., 2001; Witmer et al., 2002; Tammela et al., 2008), thus, we first verified its transcription in the thoracic aorta. Using a transgenic Vegfr3::YFP reporter mouse (Calvo et al., 2011), expression of YFP was readily detected, confirming Vegfr3 expression in adult arteries (Figure 7A). We confirmed this observation by staining a longitudinal section of the thoracic aorta with an anti-VEGFR-3 antibody (Figure 7B). Interestingly, VEGFR3 expression was not uniform: weaker expression was detected in the outer curvature or some portions of the carotid artery, associated with higher shear stress, while stronger expression was observed in the inner curvature, associated with low shear stress (Figure 7—figure supplement 1).

Figure 7 with 1 supplement see all
Transient vascular remodeling in EC iΔR3 mice.

(A) Longitudinal paraffin section of the thoracic aorta of a Vegfr3::YFP (VEGFR3 reporter) mouse. YFP was detected with an anti-GFP antibody. Scale bar = 50 µm. (B) VEGFR3 and DAPI staining of a longitudinal section of the thoracic aorta of Vegfr3fl/fl (WT) or EC iΔR3 mice, 3 weeks after Tx injection. Scale bar: 50 µm. (C) Lyve1 and VEGFR3 staining of the lymphatic network in ear skin from Vegfr3fl/fl (WT) or EC iΔR3 mice. Pictures were taken 1 week after Tx injection. Scale bar = 50 µm. (D) Aorta from oil injected-EC iΔR3 (WT) or Tx-injected EC iΔR3 mice, 2 weeks after Tx injection. Scale bar = 1 mm. (E) Diameters (graph on right) were measured in thoracic aortas (images on left) from Vegfr3fl/fl (WT) or EC iΔR3 mice, 3 weeks (WT: n = 8 and EC iΔR3: n = 7) and 7 weeks (WT: n = 6 and EC iΔR3: n = 5) after Tx treament (whiskers indicate the minimum and maximum data point, ***: p < 0.001, ANOVA). The measurement was performed right after the curvature, 1 mm below the subclavian artery bifurcation. (F) Longitudinal paraffin sections of the thoracic aorta from Vegfr3fl/fl (WT) or EC iΔR3 mice or EC iΔR3 mice, probed for MMP9 (blue) and nuclei (red) after injection of Tx for the indicated time (WT is 1 week post-injection). Distribution of the area under the curve of MMP9 fluorescence from the media is plotted on the left (n ≥ 3 mice for each condition, whiskers are 10–90%, cross is the arithmetic mean).

https://doi.org/10.7554/eLife.04645.014

Because deletion of Vegfr3 in mice leads to major cardiovascular defects and embryonic lethality (Dumont et al., 1998), we used an inducible knock out strategy in adult Vegfr3fl/fl mice (Haiko et al., 2008) that also contain an endothelium-specific, tamoxifen-inducible Cre (Cdh5-CreERT2) allele (Wang et al., 2010). Cdh5 CreERT2, Vegfr3fl/fl mice, referred as EC iΔR3, grow normally without any defect prior to tamoxifen injection. Two month old Vegfr3fl/fl (wild-type, WT) and EC iΔR3 mice were injected with tamoxifen and examined at 1, 2, 3 or 7 weeks. 1 week after tamoxifen injection, no VEGFR3 expression was visible in the thoracic aorta (Figure 7B) and in the ear skin lymphatics of EC iΔR3 mice (Figure 7C). 3 weeks after deletion of Vegfr3, the dermal lymphatic network in the skin was completely intact but vessel diameter was dramatically decreased (WT: 38 ± 5 µm and EC iΔR3: 22 ± 2 µm, n = 4, p < 0.001). We also observed a ∼15% reduction of the diameter of the descending aorta (Figure 7D,E). No further change was observed when mice were examined at 7 weeks (Figure 7E), indicating that vessels remodeled and then stabilized. No change in body weight was observed 3 weeks after injection (28.4 g ± 2 for WT and 28.3 g ± 2.7 for EC iΔR3 mice). The curvature of the aortic arch was also reproducibly decreased after excision, an unexpected result that we have not further investigated.

To investigate the role of remodeling pathways, we stained longitudinal sections of the thoracic aorta for MMP9, a matrix metalloprotease involved in flow-dependent vascular remodeling (Bond et al., 1998; Godin et al., 2000; Magid et al., 2003). Following Vegfr3 deletion, MMP9 in the thoracic aorta was highly elevated at 1 week but decreased to baseline at later times (Figure 7F). This observation strongly supports the notion that Vegfr3 deletion induces inward remodeling of the thoracic aorta which is followed by stabilization. Increased MMP9 expression may be induced through NF-κB (Sun et al., 2007). We hypothesize that elevating the set point causes the endothelium to signal low shear, which induces inward remodeling. Together, these data support the concept that vessel lumen diameter is controlled by a VEGFR3-dependent shear stress set point.

Discussion

Living organisms have developed an extensive repertoire of mechanisms to adapt to stresses and maintain homeostasis. For more than a century, investigators have observed effects suggesting that the blood flow controls vascular diameter (Thoma, 1893; Langille and O'Donnell, 1986; Langille et al., 1989; Langille, 1996), a mechanism that would optimize perfusion by adjusting vascular morphology in response to tissue demand. It has been hypothesized that, as for thermoregulation, there is an optimal value of flow which is maintained through feedback mechanisms to prevent deviation from this value. This is what we term the ‘shear stress set point’ theory (Rodbard, 1975). The current data show that HUVECs align in the direction of flow, inhibit NF-κB and activate Smads within a narrow range of fluid shear stress magnitudes. This range corresponds to the physiological flow within the umbilical vein estimated at around 8.4 to 12.5 dynes.cm−2 ((Kiserud and Rasmussen, 1998; Boito et al., 2002; Christensen et al., 2014); shear stress = 8 × viscosity (velocity/diameter), with viscosity = 0.06–0.09 poisse, velocity = 7.1 cm.s−1 and diameter = 4.1 mm). These results imply that physiological flow inhibits inflammatory pathways and activates anti-inflammatory/stabilization pathways. By contrast, cells in low or high flow fail to align, activate NF-κB and suppress Smads. We propose that these responses are involved in the vessel remodeling that reestablishes optimal blood flow.

It is known that inflammation is a critical component of flow-dependent as well as other forms of vessel remodeling (Silvestre et al., 2008; Schaper, 2009; Silvestre et al., 2013). It has been recently demonstrated that inhibiting NF-κB impairs outward remodeling associated with increased shear stress as well as aneurysm formation (Saito et al., 2013). On the other hand, defective Smad1 signaling in the endothelium is associated with hereditary haemorrhagic telengiectasia (HHT), which is characterized by the development of unstable, arteriovenous malformations (Dupuis-Girod et al., 2010). Interestingly, these malformations are preceded by increased vascular lumen diameter, which occurs in a flow dependent manner (Corti et al., 2011). These observations, combined with ours, suggest that these two signaling pathways contribute to balanced control of the vessel caliber.

Fluid shear stress varies among different types of vessels, and to some extent even within the same vessel, suggesting that different cells must have different set points for shear stress, depending on their location. Relevant to our experiments, the shear stress in the human umbilical vein is estimated at around 8.4–12.5 dynes/cm−2 whereas lymphatic vessels have highly pulsatile flow with peaks values at around 4–8 dynes.cm−2 and averages that are much lower (Dixon et al., 2006). The shear stress set point model therefore predicts that these cell types will have different set points, which was borne out in our studies. Furthermore, we found that this difference can be largely accounted for by differences in VEGFR3 expression. This receptor, a close homolog of VEGFR2, is also activated in response to flow. Both expression levels in vivo (Witmer et al., 2002) and our functional experiments in vitro lead to the conclusion that high expression of VEGFR3 increases sensitivity to shear to give a low shear stress set point, while low expression of VEGFR3 is associated with higher set points. However, it is highly likely that other mechanotransducers or mediators influence set point values. While we did not observe any major difference in PECAM-1 and VE-cadherin expression between HDLEC and HUVEC, these two proteins can vary between different vascular beds (Pusztaszeri et al., 2006; Herwig et al., 2008), which might also affect the set point. We used HUVECs as a model for blood endothelial cells because they are readily available and their response to shear stress is well characterized. However, it has been recently showed that arterial and venous markers greatly diminish in culture (Aranguren et al., 2013), thus, whether they fully represent typical venous cells in vivo should be treated with caution. Comparing fresh primary cells from veins and arteries will be an interesting direction for future work. Mechanotransducers apart from the junctional complex are also likely to be important. There must also be pathways that distinguish high and low shear to initiate outward vs inward remodeling. Future work will be required to explore these pathways in more detail and their relevance to vascular remodeling.

The notion that vascular remodeling is governed by a shear stress set point, which is itself set by activation of various receptors and signaling pathways, may be relevant to a number of applications. Recovery from atherosclerotic luminal narrowing or myocardial infarction is thought to proceed in part via flow-dependent vessel remodeling (Heil and Schaper, 2004). Vascular graft adaptation also requires activation of signaling pathways activated by high shear stress to promote healing of the graft by preventing intimal proliferation (Kohler et al., 1991). Arteriovenous malformations are also thought to have a flow-dependent component (Corti et al., 2011). Thus, further understanding of the molecular sensors and downstream signaling pathways that control flow-dependent remodeling is relevant to a broad range of vascular dysfunction.

Materials and methods

Cell culture

Human Umbilical Vein Endothelial Cells (HUVECs) pooled from three different donors were obtained from the Yale Vascular Biology and Therapeutics program and cultured in M199 medium supplemented with 20% Fetal Bovine Serum, 50 µg.ml−1 of Endothelial Cell growth Supplement (ECGS) prepared from bovine hypothalamus, 100 μg.ml−1 heparin, 100 U.ml−1 penicillin and 100 μg.ml−1 streptomycin. They were used between passage 3 and 7. Human Dermal Lymphatic Endothelial Cells (HDLECs) were obtained from Lonza (Basel, Switzerland) and cultured in EGM-2 MV medium and used from passage 5 to 7. Cells were starved in M199 medium supplemented with 5% FBS and 100 U.ml−1 penicillin and 100 μg.ml−1 streptomycin for a minimum of 4 hr before further treatments.

Shear stress

Cells were seeded on tissue culture plastic slides cut from 150 mm tissue culture dishes (Falcon), coated with 20 μg.ml−1 fibronectin. Confluent cells were subjected to steady laminar shear stress in a modified parallel plate flow chamber (Figure 1) in which the gasket was a silicon sheet of either 0.8 or 1.6 mm height (Grace Bio-Labs, Bend, OR, #664172 and #664283) cut to generate a linear gradient of shear stress, calculated from (Usami et al., 1993). Flow was applied for 16 hr in starvation medium. Cells were then fixed with 4% formaldehyde in PBS for 10 min, permeabilized with 0.5% Triton x-100 in PBS for 10 min, blocked with Startingblock buffer (ThermoScientific) for 30 min at room temperature and probed overnight at 4°C with a primary antibody diluted in Startingblock buffer. Slides were stained with Hoechst 33342 to label nuclei, with rabbit anti-p65 antibody (Cell Signaling) to label NF-κB, and with rabbit anti-Smad1 antibody (Cell Signaling).

Image analysis

Images were acquired with a Perkin Elmer spinning disk confocal microscope equipped with an automated stage which was used to take successive pictures along the chamber channel. Masks of the images were made using a combination of an adaptive histogram equalization algorithm with intensity and size thresholding. Cell orientation was calculated by taking the masks of the cell nuclei, fitting to an ellipse, and finding the angle between the flow direction and the major-axis of the ellipse. Nuclear translocation was computed by taking the mask of the nucleus and determining the integrated intensity of the transcription factor stain (Smad1 or p65) in the nucleus and in the whole cell. The ‘translocation factor’ (TF) was calculated by dividing the integrated intensity in the nucleus by the value for the whole cell. If the entire signal is localized to the nucleus, TF = 1, while if the entire signal is cytoplasmic, TF = 0.

siRNA transfection and adenoviral expression

Depletion of VEGFR3 was achieved by transfecting 10 nM siRNA (L-003138-00 OnTarget Smartpool Human FLT4, ThermoScientific) with Lipofectamine RNAi Max (Invitrogen), following the manufacturer's instructions. Transfection efficiency was assessed by Western-blot. Human VEGFR3-GFP was cloned in adenoviral (pAd) expression vector. Cells were infected with the virus in medium with polybrene (5 mg/ml) overnight and used 48 hr later.

FACS

GFP expression in HUVEC or HUVEC infected with VEGFR3-GFP was assayed on a Stratedigm S1000EX (Stratedigm, San Jose, CA). Data were analyzed with the FlowJo software (TreeStar, Ashland, OR).

Western blotting

Cells were washed with cold PBS and proteins extracted with Laemmli's buffer. Samples were run on 10 or 12% SDS-PAGE and transferred onto nitrocellulose membranes. The membranes were blocked with StartingBlock buffer (ThermoScientific) and probed with primary antibodies overnight at 4°C: VEGFR3 (R&D systems), phospho-VEGFR3 (Cell Applications), VEGFR2 (Cell Signaling), PECAM-1 (Abcam), VE-cadherin (Santa Cruz), GFP (Invitrogen) and actin (Santa Cruz). DyLight conjugated fluorescent secondary antibodies (680 nm and 800 nm, Thermoscientific) or HRP-conjugated antibodies were used to detect primary antibodies. Bands were detected and quantified with an Odyssey infrared imaging system for DyLight antibodies (Li-Cor) or a BioRad western blot imaging system (Bio Rad).

Zebrafish

Zebrafish were grown and maintained according to protocols approved by the Yale University Animal Care. The Tg(kdrl:mCherry; flt4:citrine) was used (Bussmann and Schulte-Merker, 2011). Morpholinos (Nicoli et al., 2012) were injected at the indicated concentrations and morphants were observed in a confocal microscope (SP5 Leica Microsystems). Images captured using Leica application suite software. Chemical treatment with nifedipine 40 µM was performed as previously described, 4 hr prior imaging (Bussmann et al., 2011).

Mice

All animal experiments were approved by the Institutional Care and Use Committee of Yale University. The Vegfr3::YFP (Calvo et al., 2011), Cdh5CreERT2 (Pitulescu et al., 2010; Wang et al., 2010), Vegfr3flox/flox (Haiko et al., 2008) mice were described previously. Cdh5CreERT2 mice were crossed with Vegfr3flox/flox mice to generate endothelial-specific inducible Vegfr3 mutant mice. 6–8 weeks old Vegfr3flox/flox mice, with or without the Cre recombinase, were injected intra-peritoneally with 2 mg tamoxifen (TX; at 20 mg/ml in peanut oil (Sigma) with 10% Ethanol) once per day for 5 consecutive days (induction period). Mice were euthanized then fixed by perfusion with 3.7% formaldehyde 1, 2, 3 or 7 weeks after induction. Ear tissue was fixed overnight in 3.7% formaldehyde. The ear skin was removed, cleaned of connective tissue and cartilage, and permeabilized for 4 hr in permeabilization buffer (1% BSA, 1% NGS, 0.5% Tween in PBS). The skin was then incubated with antibody against LYVE-1 or VEGFR3, in 50% permeabilization buffer/50% PBS for 2 days at 4°C. After washing, the skin was incubated with secondary antibodies overnight 4°C in the same buffer. The skin was then flat mounted in Fluoromount G (Southern Biotech) and imaged with a Perkin Elmer spinning disk confocal microscope with a 20× objective. The aorta was removed, cleaned of all connective tissue, fixed overnight in 3.7% formaldehyde at 4°C and embedded in paraffin. Paraffin embedding and sectioning were performed by the Yale Pathology department, in the Tissue Microarray facility. Aortas were cut longitudinally, paraffin was removed in xylene baths and sections progressively rehydrated before antigen retrieval for 30 min at 95°C in citrate buffer (10 mM sodium citrate, 0.05% Tween, pH = 6). Sections were blocked for 30 min in StartingBlock blocking buffer (ThermoScientific) and probed either with anti-MMP9 antibody (Abcam, 1/400), anti-VEGFR3 antibody (R&D) or anti-GFP antibody (Invitrogen, 1/400). Slides were then washed 3× in PBS-Tween and once in PBS, then incubated with donkey-anti rabbit AlexaFluor 647 secondary antibody (Molecular Probes, 1 hr at RT, 1/500). Slides were washed 3× in PBS-Tween and once in PBS, then mounted in Fluoromount G (Southern Biotech). Slides were imaged with a Nikon Eclipse 80i epifluorescence microscope. Image analysis of MMP9 staining was performed by measuring the area under the curve of the fluorescence signal coming from the media in 4 different 20× pictures for each individual aorta. The fluorescence profile was obtained with MeasureEndo, an ImageJ macro.

Statistics

Values indicated in the text are mean ± SD. At least three independent experiments were performed for each condition. Statistical tests were performed by using either analysis of variance tests (ANOVA) or unpaired Student's t-tests. The ANOVA test performed on Figures 2 and 3 tested the null hypothesis that shear stress magnitude does not have an effect on either cell orientation or p65 and Smad1 nuclear translocation.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
    Organization for physiological homeostasis
    1. WB Cannon
    (1929)
    Physiological Reviews 9:399–431.
  10. 10
    Modeling of physiological flows
    1. L Cardamone
    2. JD Humphrey
    (2012)
    187–203, Modeling of physiological flows, Springer.
  11. 11
  12. 12
  13. 13
    Intra-membrane binding of VE-cadherin to VEGFR2 and VEGFR3 assembles the endothelial mechanosensory complex
    1. BG Coon
    2. N Baeyens
    et al. (2015)
    Journal of Cell Biology In press.
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
    Vascular endothelial growth factor-A and -C protein up-regulation and early angiogenesis in a rat photothrombotic ring stroke model with spontaneous reperfusion
    1. W Gu
    2. T Brännström
    3. W Jiang
    4. A Bergh
    5. P Wester
    (2001)
    Acta Neuropathologica 102:216–226.
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
    Adaptive regulation of wall shear stress to flow change in the canine carotid artery
    1. A Kamiya
    2. T Togawa
    (1980)
    The American Journal of Physiology 239:H14–H21.
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
    Arterial remodeling: relation to hemodynamics
    1. BL Langille
    (1996)
    Canadian Journal of Physiology and Pharmacology 74:834–841.
    https://doi.org/10.1139/y96-082
  35. 35
    Adaptations of carotid arteries of young and mature rabbits to reduced carotid blood flow
    1. BL Langille
    2. MP Bendeck
    3. FW Keeley
    (1989)
    The American Journal of Physiology 256:H931–H939.
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
    Differential activation of NF-kappa B in human aortic endothelial cells conditioned to specific flow environments
    1. S Mohan
    2. N Mohan
    3. EA Sprague
    (1997)
    The American Journal of Physiology 273:C572–C578.
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
    Untersuchungen über die Histogenese und Histomechanik des Gefäßsystems
    1. R Thoma
    (1893)
    Stuttgart: Enke Verlag.
  55. 55
  56. 56
    Shear level influences resistance artery remodeling: wall dimensions, cell density, and eNOS expression
    1. JL Tuttle
    2. RD Nachreiner
    3. AS Bhuller
    4. KW Condict
    5. BA Connors
    6. BP Herring
    7. MC Dalsing
    8. JL Unthank
    (2001)
    American Journal of Physiology Heart and Circulatory Physiology 281:H1380–H1389.
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64
  65. 65
  66. 66

Decision letter

  1. Fiona M Watt
    Reviewing Editor; King's College London, United Kingdom

eLife posts the editorial decision letter and author response on a selection of the published articles (subject to the approval of the authors). An edited version of the letter sent to the authors after peer review is shown, indicating the substantive concerns or comments; minor concerns are not usually shown. Reviewers have the opportunity to discuss the decision before the letter is sent (see review process). Similarly, the author response typically shows only responses to the major concerns raised by the reviewers.

Thank you for sending your work entitled “Vascular remodeling is governed by a VEGFR3-dependent fluid shear stress set point” for consideration at eLife. Your article has been favorably evaluated by Fiona Watt (Senior editor) and three reviewers.

The Senior editor and the reviewers discussed their comments before we reached this decision, and the Senior editor has assembled the following comments to help you prepare a revised submission. All the reviewers considered your work to be interesting and important. Although there is a long list of comments for you to address, many can be dealt with in the text and by strengthening the Discussion.

However, two critical issues are as follows:

First, the cited unpublished results (Coon et al.) that VEGFR3 is a part of the mechanosensory complex consisting of PECAM–1, VEC and VEGFR2 should be presented in this manuscript. It is impossible to accept the authors' statements without reviewing this evidence.

Second, there does not seem to be any detectable Flt4 signal in the zebrafish aorta. Previously published results indicate that Flt4 is expressed in the PCV, not aorta, and in segmental arteries, and Vegfc is expressed in the zebrafish aorta (Covassin et al., PNAS, 2006; Siekman and Lawson, Nature, 2007). How could Flt4 MO have any effect on aorta width in this case? High MO concentrations could have off-target effects in aorta development, or on blood flow. In fact, recent results have indicated that few phenotypes based on morpholinos are similar to those based on corresponding gene-edited deletions, casting doubt on experiments that employ morpholinos (see Schulte-Merker and Stainier, 2014, Development, 3103-3104).

Specific points:

1) Flow chamber:

a Please explain what is the actual flow profile. It is unclear if the linear dependence shown in Figure 1 was obtained experimentally or calculated using laminar flow models.

b. Are there any downstream effects of the cells exposed to low shear to those exposed to higher shear?

c. Please clarify how reproducible are data shown in Figure 2 (all three panels). Could statistical differences be shown?

2) It is stated that HUVECs aligned between 8 and 20 dynes/cm2 (Figures 2A and 3A), lymphatic endothelial cells aligned between 4 and 6 dynes/cm2, (Figure 3A), and the minimum NF-kB translocation shifted to 4 and 10 dynes/cm2 (Figure 3B). This is not obvious from the data shown. It would be helpful to clarify how these ranges were determined, and if they correspond to statistically significant differences relatively to the other regions. In Figure 2 representative pictures of cell alignment (junctional staining or bright field images or at least the one in supplemental figure) should be included for 3 values of flow (under 10, between 10 and 20, above 20 dynes/cm2).

3) Figure 4: Again, it would be instructive to know what is the reproducibility of these data (which look like single data sets), and important to determine which differences are significant.

4) The zebrafish loss-of-function experiments are not temporally controlled and flt4 will have additional roles that are difficult to separate from the alleged mechanosensory role. Therefore the authors should avoid overstating these results in saying that they show that ligand independent activation of flt4 is involved in this process. The authors do not provide any data on flt4 activation at all. So it remains unclear how, mechanistically, VEGFR3 is involved.

5) The authors should include representative images of cell alignment, p65 and Smad1 under different shear stress conditions. It should be noted in the text that the quantification shown in the figure uses nuclear anisotropy as a proxy for cell alignment. This may well be justifiable but needs to be made clear. The authors assess alignment and NFκB–SMAD translocation at 16 hr for both HUVECs and HDLEC. Could these cells have also a different time-response to shear stress to alignment and NFκB–SMAD activation?

6) Is MMP9 part of the set-point by affecting “stiffness”, or just a downstream effector involved in remodelling? If not possible to tie down experimentally, this should at least be discussed. Are p65 and Smad1 affected in vivo? This should be tested to better link the parts of the manuscript.

7) Regarding the quantification program for the translocation factor, how do the authors differentiate between cells with very low level of NFkB everywhere (cytoplasm and nucleus with the same staining intensity) and cells with very high level of NFkB everywhere? Biologically these two situations are different because NFκB increases in the nucleus but the ratio would give the same value. In other words, are total levels affected by shear levels and is this relevant? Can it be excluded? Same comment for Smad1.

8) The authors mention in the text that more than 90% of cells expressed VEGFR3-GFP. Based on the images provided in Figure 4–figure supplement 1, we can see approximately half of the cell GFP+ (or it is not a confluent monolayer). The authors should assess the transduction efficiency with a more accurate method.

9) Please comment on what happens to the thoracic duct in nifedipine treated fish.

10) The authors nicely show that VEGFR3 is expressed in EC of thoracic Aorta. As the inner part of the curvature of the aorta is exposed to lower shear stress, do the authors see any differential expression of VEGFR3 in this area? Is there any regulation of VEGFR3 expression or phosphorylation by the shear stress level in the in vitro chamber?

11) The authors observe a strange effect on aorta curvature in EC iDR3 mice. Could it be linked with the increased production of MMPs potentially softening the surrounding matrix. Could the authors comment on this?

12) How is MMP9 expression modulated in the flow chamber model?

13) Do the authors have any explanation on how VEGFR3 expression regulates MMP9 expression?

14) How are p65 and SMAD modified in the aorta of EC iDR3 mice or in the different morpholino fish? Is the alignment of EC modified?

15) Could it be that VEGFR3 deletion has a short time effect on vessel diameter by modulating the vasomotor tone in addition to the remodelling effect of MMPs production?

16) The YFP fluorescent reporter seems to be based on a large recombinant genomic clone in an unknown chromosomal location, where it can be influenced by the chromosomal context. Does the YFP expression reflect reliably the expression of VEGFR3 in the adult aorta? What is the half-life of the YFP used in Figure 6A? The YFP expression appears “patchy” in Figure 6A; does this mean that ECs in the aorta expresses variable amounts of VEGFR3? The authors should obtain convincing data showing that the VEGFR3 protein is indeed expressed at a significant level in the aortic endothelial cells in vivo.

17) The authors use HUVECs in their in vitro studies. These cells do not correspond to the endothelial cells in the mouse or zebrafish aorta. Although the flow conditions in umbilical vein may be similar to those in an artery, venous endothelial cells express higher levels of VEGFR3 than arterial endothelial cells (for example, see Hogan et al., Development, 2009). Somehow, based on in vitro analysis, the shear stress “setpoints” are handled rather indiscriminately, considering that arteries and veins as well as primary lymph vessels and collecting vessels also have obvious differences in shear stress in vivo. It would be useful and relevant to perform the gradient flow chamber experiments using freshly isolated arterial endothelial cells, or even better, aortic endothelial cells. If the authors are not willing to do this, then at least the limitation of using the cell lines should be discussed.

18) The completeness of recombination achieved in the mouse aorta using the Cdh5-Cre line produced by Wang et al. is not reported. How does the dermal lymphatic data shown in Figure 6B connect with the rest of the manuscript? It certainly does not provide surrogate evidence for VEGFR3 deletion, and should be replaced by the assessment of VEGFR3 expression levels in mouse aortas before and after Cre-mediated deletion by IHC, WB and qPCR.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled “Vascular remodeling is governed by a VEGFR3-dependent fluid shear stress set point” for further consideration at eLife. Your revised article has been evaluated by Fiona Watt (Senior editor) and the original reviewers. The manuscript has been improved but there are some remaining issues that need to be addressed before acceptance.

Reviewer #1 requests a number of specific clarifications, as listed below. In addition Reviewers #2 and #3 flag up the problem of how to avoid duplication of data in your two manuscripts, while nevertheless giving eLife readers sufficient data to enable them to evaluate your findings without having to refer to the Coon et al paper. We have copied verbatim some of Reviewer #3's comments on this issue.

We believe that you can address the reviewers' concerns without further experimentation. It is harder to see a solution to the problem of overlapping findings in the two papers and would appreciate your comments.

Reviewer #1:

1) Flow chamber:

a) Critique: Please explain what is the actual flow profile It is unclear if the linear dependence shown in Figure 1 was obtained experimentally or calculated using laminar flow models.

Response: “The linear dependence of the flow profile was calculated using flow models from the original manuscript describing this chamber (Usami, Chen et al. 1993). To confirm the results obtained with the gradient chamber, we performed additional experiments with a conventional parallel plate chamber under uniform shear stress of 2, 12 or 50 dynes.cm-2.”

There is still no clear and direct validation of model predictions. The authors should either validate or remove the model predictions.

b) Critique: Are there any downstream effects of the cells exposed to low shear to those exposed to higher shear?

Response: “The results of these experiments (Figure 2–figure supplement) support the results obtained with the gradient chamber, ruling out downstream effect from cells exposed to low shear on the responses measured at higher shear stress.”

Again, it is not obvious how the responses to uniform shear at three different levels rule out the downstream effects from cells exposed to low shear to cells exposed to higher shear.

c) Critique: Please clarify how reproducible are data shown in Figure 2 (all three panels). Could statistical differences be shown?

Response: “Lastly, the statistical significance of the results obtained with the gradient chamber has been tested with an ANOVA test and the result of the test for each variable is now included in the legends of the Figure 2 and Figure 2–figure supplement 1.”

It remains unclear how the statistics was done and what it tells us about the reproducibility of the data. I see the response unrelated to the question.

Also, the responses to points 12 and 13 about the MMP 9 expression are vague. The authors state that they could not observe any change in the MMP 9 intensity in HUVECs in the gradient chamber and conclude that “MMP9 analysis in vitro does not appear to be a fruitful line of investigation”, whereas there is a body of literature reporting in vitro studies of MMP 9 expression in various cell types. Similarly, they do not offer an explanation on how VEGFR3 expression regulates MMP9 expression.

Reviewer #3:

The Coon et al. manuscript that has apparently been sent to another journal presents some complications:

1) There is overlap as in both manuscripts induction of VEGFR3 phosphorylation by FSS is now presented as a novel finding. The new Figure 5 in the revised eLife manuscript is not of the necessary standard as it fails to show total VEGFR3 protein in the lysates, only P-VEGFR3 is displayed. The total VEGFR3 protein is analysed in a similar Figure 4 of the other manuscript, but one cannot ask the eLife readers to consult another paper for such basic control.

2) The eLife manuscript avoids clearly mentioning the VEC;VEGFR2 complex, whereas the other manuscript claims VEGFR2;VEGFR3;VEC mechanosensory complex is involved, and speculates that a critical component may actually be a phosphatase.

3) Whereas the eLife manuscript emphasizes VEGFR3 alone as the FSS set point, the other manuscript says in the abstract that “VEGFR2 and VEGFR3 signal redundantly downstream of VE-cadherin”.

4) Both manuscripts also show evidence of “significant” VEGFR3 expression in aortic endothelium. It is difficult to know if this amount is biologically significant.

5) The eLife abstract says: “VEGFR3 modulates arterial lumen diameter consistent with flow-dependent remodeling” (it should probably be “aortic” rather than “arterial” here, as other arteriae were not studied). The abstract of the other manuscript has a similar result: “VEGFR3 expression is observed in the aortic endothelium where it contributes to flow responses in vivo”.

https://doi.org/10.7554/eLife.04645.016

Author response

First, the cited unpublished results (Coon et al.) that VEGFR3 is a part of the mechanosensory complex consisting of PECAM–1, VEC and VEGFR2 should be presented in this manuscript. It is impossible to accept the authors' statements without reviewing this evidence.

The main point that is relevant to the current study is that VEGFR3 contributes to flow signaling similarly to VEGFR2. The mechanism by which it is activated by flow is a separate issue that is better left for the other paper where it is studied in detail. We have therefore added new data demonstrating activation of VEGFR3 by flow in both HUVECs and lymphatic ECs (Figure 5) and provide the unpublished manuscript for the reviewers’ inspection.

Second, there does not seem to be any detectable Flt4 signal in the zebrafish aorta. Previously published results indicate that Flt4 is expressed in the PCV, not aorta, and in segmental arteries, and VEGFC is expressed in the zebrafish aorta (Covassin et al., PNAS, 2006; Siekman and Lawson, Nature, 2007). How could Flt4 MO have any effect on aorta width in this case? High MO concentrations could have off-target effects in aorta development, or on blood flow. In fact, recent results have indicated that few phenotypes based on morpholinos are similar to those based on corresponding gene-edited deletions, casting doubt on experiments that employ morpholinos (see Schulte-Merker and Stainier, 2014, Development, 3103-3104).

We have now included a full-size, detailed picture of the Flt4 reporter embryo (Figure 6–figure supplement 1), where we observe a gradient of Flt4 transcription: high expression in the developing thoracic duct, lower in the posterior cardinal vein and lower but still readily detectable in the dorsal aorta. This observation is very much in accordance with our paradigm that expression of VEGFR3 inversely correlates with shear stress magnitude in vasculature. We also would like to stress that these former studies by Nathan Lawson’s laboratory were performed at 24 hpf, during the initial phases of vasculature development, while our observation is done at 72 hpf, when the vasculature is more mature.

Concerning the potential off-target effect of the morpholinos, the concentration dependent effect on dorsal aorta diameter without affecting the whole vasculature of the embryo or its viability, supports the specificity of the effect. More critically, the kinase dead expando mutant for Flt4 (Hogan, Herpers et al., 2009) yielded similar effects on the diameter of the dorsal aorta, though these were not noted by the authors (Kwon, Fukuhara et al., 2013). This effect is not observable with a VEGFC morpholino, in accordance with our own observations. These results, plus the similar observation in mice, strongly support the conclusion that the decrease of the dorsal aorta diameter is related to the decreased levels of Flt4 rather than an off-target effect. All of these issues are now discussed in the manuscript.

Specific points:

1) Flow chamber:

a. Please explain what is the actual flow profile. It is unclear if the linear dependence shown in Figure 1 was obtained experimentally or calculated using laminar flow models.

b. Are there any downstream effects of the cells exposed to low shear to those exposed to higher shear?

c. Please clarify how reproducible are data shown in Figure 2 (all three panels). Could statistical differences be shown?

The linear dependence of the flow profile was calculated using flow models from the original manuscript describing this chamber (Usami, Chen et al., 1993). To confirm the results obtained with the gradient chamber, we performed additional experiments with a conventional parallel plate chamber under uniform shear stress of 2, 12 or 50 dynes.cm-2. The results of these experiments (Figure 2–figure supplement 1) support the results obtained with the gradient chamber, ruling out downstream effect from cells exposed to low shear on the responses measured at higher shear stress. Lastly, the statistical significance of the results obtained with the gradient chamber has been tested with an ANOVA test and the result of the test for each variable is now included in the legends of the Figure 2 and Figure 2–figure supplement 1.

2) It is stated that HUVECs aligned between 8 and 20 dynes/cm2 (Figures 2A and 3A), lymphatic endothelial cells aligned between 4 and 6 dynes/cm2, (Figure 3A), and the minimum NF-κB translocation shifted to 4 and 10 dynes/cm2 (Figure 3B). This is not obvious from the data shown. It would be helpful to clarify how these ranges were determined, and if they correspond to statistically significant differences relatively to the other regions. In Figure 2 representative pictures of cell alignment (junctional staining or bright field images or at least the one in supplemental figure) should be included for 3 values of flow (under 10, between 10 and 20, above 20 dynes/cm2).

The set point ranges given in the test were determined by visual inspection of the data and, as indicated in the text, are approximate. As requested, representative pictures for different flow magnitudes have been included, including phalloidin staining to highlight cell alignment. Statistics were determined with an ANOVA test and the result of the test for each variable is included in the legend of the Figures 2, Figure 2–figure supplement 1 and Figure 3.

3) Figure 4: Again, it would be instructive to know what is the reproducibility of these data (which look like single data sets), and important to determine which differences are significant.

4) The zebrafish loss-of-function experiments are not temporally controlled and flt4 will have additional roles that are difficult to separate from the alleged mechanosensory role. Therefore the authors should avoid overstating these results in saying that they show that ligand independent activation of flt4 is involved in this process. The authors do not provide any data on flt4 activation at all. So it remains unclear how, mechanistically, VEGFR3 is involved.

We performed an ANOVA test to compare the different situations and the results are now included in the legend for Figure 4. We also included the raw data in Figure 4–figure supplement 2.

5) The authors should include representative images of cell alignment, p65 and Smad1 under different shear stress conditions. It should be noted in the text that the quantification shown in the figure uses nuclear anisotropy as a proxy for cell alignment. This may well be justifiable but needs to be made clear. The authors assess alignment and NFkB–SMAD translocation at 16h for both HUVECs and HDLEC. Could these cells have also a different time-response to shear stress to alignment and NFkB–SMAD activation?

We have added representative pictures and have highlighted that cell alignment was quantified by measuring nuclei orientation. We previously validated that nuclear alignment is a suitable metric that is highly correlated with cell alignment and that reference is now cited (Baeyens, Mulligan-Kehoe et al., 2014). Regarding the time dependence of the responses, our opinion is that this may not be relevant to the set point theory: lymphatic cells can align at a magnitude where HUVEC cannot, after 16 hours of flow, indicating clearly that the shear sensing mechanisms are optimized for lower values of shear stress in lymphatic cells.

6) Is MMP9 part of the set-point by affecting “stiffness”, or just a downstream effector involved in remodelling? If not possible to tie down experimentally, this should at least be discussed. Are p65 and Smad1 affected in vivo? This should be tested to better link the parts of the manuscript.

We examined MMP9 as a well-known downstream effector in arterial remodeling primarily to assess the time course of remodeling. The key point from the results is that following MMP9 expression after VEGFR3 deletion, it then resolves, indicating that vessels have reached a new steady state, thus, are not undergoing regression or some other continuing process. Its expression is thought to be controlled in part by NFkB but also by other factors whose identification is not important here. We have been unable to reliably quantify p65 and SMAD nuclear translocation in longitudinal sections of the thoracic aorta.

7) Regarding the quantification program for the translocation factor, how do the authors differentiate between cells with very low level of NFkB everywhere (cytoplasm and nucleus with the same staining intensity) and cells with very high level of NFkB everywhere? Biologically these two situations are different because NFkB increases in the nucleus but the ratio would give the same value. In other words, are total levels affected by shear levels and is this relevant? Can it be excluded? Same comment for Smad1.

While this is in principle a valid point, there are in fact no discernable changes in staining intensity but rather in staining localization (Figure 2–figure supplement 1). Western blotting also shows little change in total p65 expression under flow (Orr, Sanders et al., 2005). This is why we developed the high-throughput quantification method, allowing us to assess this very specific feature of nuclear translocation, which is a well-established feature of NFκB and Smad1 activation.

8) The authors mention in the text that more than 90% of cells expressed VEGFR3-GFP. Based on the images provided in Figure 4–figure supplement 1, we can see approximately half of the cell GFP+ (or it is not a confluent monolayer). The authors should assess the transduction efficiency with a more accurate method.

While this is in principle a valid point, there are in fact no discernable changes in staining intensity but rather in staining localization (Figure 2–figure supplement 1). Western blotting also shows little change in total p65 expression under flow (Orr, Sanders et al. 2005). This is why we developed the high-throughput quantification method, allowing us to assess this very specific feature of nuclear translocation, which is a well-established feature of NFκ B and Smad1 activation.

9) Please comment on what happens to the thoracic duct in nifedipine treated fish.

Although the thoracic duct is still developing at 3dpf (Yaniv, Isogai et al., 2006), we observed very clearly that it still developed in the presence of nifedipine (Figure 6–figure supplement 1).

10) The authors nicely show that VEGFR3 is expressed in EC of thoracic Aorta. As the inner part of the curvature of the aorta is exposed to lower shear stress, do the authors see any differential expression of VEGFR3 in this area? Is there any regulation of VEGFR3 expression or phosphorylation by the shear stress level in the in vitro chamber?

We do observe higher VEGFR3 expression in the inner curvature and some portions of the carotids, and lower expression on the outer curvature of the aorta (Figure 7–figure supplement 1), suggesting that VEGFR3 expression is higher in areas of low shear stress. However, within each vessel, expression appears nonuniform, suggesting local regulation. We also have in vitro data suggesting that VEGFR3 is slightly reduced under flow mainly due to internalization and degradation. The factors that govern VEGFR3 expression in arterial endothelial cells therefore appear to be complex and represent a distinct question separate from the one we address in this manuscript. We feel this issue is best left for further work.

11) The authors observe a strange effect on aorta curvature in EC iDR3 mice. Could it be linked with the increased production of MMPs potentially softening the surrounding matrix. Could the authors comment on this?

Yes, it is plausible that asymmetric remodeling of the aorta involving MMP9 resulted in the change in aorta curvature. Such asymmetric remodeling may be the consequence of a non-uniform VEGFR3 expression along the aorta. Or it could be a feature of the complex mechanics of the ascending aorta, which moves axially during each cardiac cycle and would be predicted to differentially load the inner and outer curvature (Kassab, 2006). We thought it appropriate to make a note of the observation but feel that anything we could say would be largely speculative as well as being unrelated to the main point of the paper and to any known remodeling process in vivo.

12) How is MMP9 expression modulated in the flow chamber model?

We stained for MMP9 in HUVECs in the gradient chamber but failed to observe any consistent change in intensity. However, MMP9 is a determinant of 3-dimensional tissue remodeling in vivo; ECs on a flat coverslip without a smooth muscle layer cannot undergo any sort of comparable process and may not express or capture secreted MMP9 in the same way. MMP9 analysis in vitro does not appear to be a fruitful line of investigation.

13) Do the authors have any explanation on how VEGFR3 expression regulates MMP9 expression?

Our current explanation is that reduction of VEGFR3 expression induces a change in the set point of the endothelial cells, which results in altered flow signaling, which induces remodeling. In essence, VEGFR3 deletion results in ECs sensing that shear levels are too low, which induces inward remodeling to increase shear. MMP9 may be a fairly indirect downstream consequence of these events.

14) How are p65 and SMAD modified in the aorta of EC iDR3 mice or in the different morpholino fish? Is the alignment of EC modified?

We tried to measure nuclear translocation in longitudinal sections of the aorta but have been unable to reliably quantify the effects in these sections. Expression levels of Smad1 and p65 seemed unaffected. Also, it is impossible, to our knowledge, to assess properly endothelial cell alignment in zebrafish embryos or in longitudinal sections of the aorta.

15) Could it be that VEGFR3 deletion has a short time effect on vessel diameter by modulating the vasomotor tone in addition to the remodelling effect of MMPs production?

Vasomotor tone is currently thought to be a critical component of vascular remodeling. The idea is that high or low shear induce, respectively, vasorelaxation or constriction, which over long times is entrained by tissue remodeling (Humphrey, Dufresne et al., 2014). It is also well established that production of nitric oxide in response to flow is mediated in part by the PECAM/VE-cadherin-VEGFR2 pathway (Jin, Ueba et al., 2003; Fleming, Fisslthaler et al., 2005). Thus, a similar role for VEGFR3, which leads to vessel remodeling, is quite reasonable.

16) The YFP fluorescent reporter seems to be based on a large recombinant genomic clone in an unknown chromosomal location, where it can be influenced by the chromosomal context. Does the YFP expression reflect reliably the expression of VEGFR3 in the adult aorta? What is the half-life of the YFP used in Figure 6A? The YFP expression appears “patchy”in Figure 6A; does this mean that ECs in the aorta expresses variable amounts of VEGFR3? The authors should obtain convincing data showing that the VEGFR3 protein is indeed expressed at a significant level in the aortic endothelial cells in vivo.

To address this problem, we performed an immunohistological staining for VEGFR3 in the thoracic aorta and the ear skin (Figure 7). This staining matches our observation with the YFP fluorescent reporter.

17) The authors use HUVECs in their in vitro studies. These cells do not correspond to the endothelial cells in the mouse or zebrafish aorta. Although the flow conditions in umbilical vein may be similar to those in an artery, venous endothelial cells express higher levels of VEGFR3 than arterial endothelial cells (for example, see Hogan et al., Development, 2009). Somehow, based on in vitro analysis, the shear stress “setpoints” are handled rather indiscriminately, considering that arteries and veins as well as primary lymph vessels and collecting vessels also have obvious differences in shear stress in vivo. It would be useful and relevant to perform the gradient flow chamber experiments using freshly isolated arterial endothelial cells, or even better, aortic endothelial cells. If the authors are not willing to do this, then at least the limitation of using the cell lines should be discussed.

Our fundamental reason for using HUVECs is that since most of the venous vs arterial gene expression signature is lost when ECs are cultured (Aranguren, Agirre et al., 2013), we felt it was better to use cells that are easy to obtain and use, and that respond well to flow. The feature we are studying appears to be conserved among different types of ECs, thus, HUVECs provide a suitable in vitro model. We tried freshly isolated mouse arterial ECs but they failed to show any alignment in response to flow. This is very likely a protocol problem but in practical terms is a significant issue. We now include a statement about the limitation of using these two cells lines and invite the reader to consider that different cells lines may have different set points.

18) The completeness of recombination achieved in the mouse aorta using the Cdh5-Cre line produced by Wang et al. is not reported. How does the dermal lymphatic data shown in Figure 6B connect with the rest of the manuscript? It certainly does not provide surrogate evidence for VEGFR3 deletion, and should be replaced by the assessment of VEGFR3 expression levels in mouse aortas before and after Cre-mediated deletion by IHC, WB and qPCR.

As mentioned earlier, we have performed an immunohistological staining of VEGFR3 in the thoracic aorta and the skin ear in both WT and VEGFR3 iKO. Figure 7 shows that VEGFR3 expression is drastically reduced after recombination.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Reviewer #1:

1) Flow chamber:

a) Critique: Please explain what is the actual flow profile It is unclear if the linear dependence shown in Figure 1 was obtained experimentally or calculated using laminar flow models.

Response: “The linear dependence of the flow profile was calculated using flow models from the original manuscript describing this chamber (Usami, Chen et al. 1993). To confirm the results obtained with the gradient chamber, we performed additional experiments with a conventional parallel plate chamber under uniform shear stress of 2, 12 or 50 dynes.cm-2.

There is still no clear and direct validation of model predictions. The authors should either validate or remove the model predictions.

We have not carried out any additional validation of the flow profile per se beyond what was done in the original publication (Usami et al, 1993). This point is now clarified in the Methods section. What we did do was validate the biological conclusions from the gradient chamber by using standard flow chambers. The new data that we added fully supported those conclusions, which we feel is the main issue.

b) Critique: Are there any downstream effects of the cells exposed to low shear to those exposed to higher shear?

Response: “The results of these experiments (Figure 2–figure supplement 1) support the results obtained with the gradient chamber, ruling out downstream effect from cells exposed to low shear on the responses measured at higher shear stress.

Again, it is not obvious how the responses to uniform shear at three different levels rule out the downstream effects from cells exposed to low shear to cells exposed to higher shear.

The reviewer was concerned that medium flowing from the region of low shear stress could influence the cells in the downstream region under high shear stress. We therefore carried out experiments where each slide was under uniform shear. The effects confirmed the results with the gradient chamber. In that setting, there is no medium flowing from low to high shear, thus, we believe that this addresses the issue. Please also note that our hydraulic circuit comprises a circulating volume of 150 ml of culture media flowing at 1.5 ml by second, which should quickly dilute any soluble factor produced by the cells. Lastly, we also carried out experiments with standard flow chambers where the medium flowed in the opposite direction (high shear stress chamber to low shear stress chamber); these gave identical results.

c) Critique: Please clarify how reproducible are data shown in Figure 2 (all three panels). Could statistical differences be shown?

Response: “Lastly, the statistical significance of the results obtained with the gradient chamber has been tested with an ANOVA test and the result of the test for each variable is now included in the legends of the Figure 2 and Figure 2–figure supplement 1.

It remains unclear how the statistics was done and what it tells us about the reproducibility of the data. I see the response unrelated to the question.

We now provide additional clarifications about the way the statistics were done. The one way ANOVA test performed on the data in Figure 2 tested the null hypothesis that shear stress magnitude has no effect on endothelial alignment, p65 and Smad1 nuclear translocation. This analysis showed very clearly that it is not the case. This clarification is mentioned in the methods. This analysis treated the effect of flow on these variables as a trend; we discussed this issue with a biostatistician, who felt that there are no other statistical tests appropriate for this situation. We also provide the standard error on the mean for each curve, to address the reproducibility of the data. Additionally, we tested the difference between specific shear stress magnitudes in Figure 2–figure supplement 1, we performed a 2 way ANOVA test to compare the 3 conditions. The results validate the conclusion that fluid shear stress at different magnitudes induces statistically different responses.

Also, the responses to points 12 and 13 about the MMP 9 expression are vague. The authors state that they could not observe any change in the MMP 9 intensity in HUVECs in the gradient chamber and conclude that “MMP9 analysis in vitro does not appear to be a fruitful line of investigation”, whereas there is a body of literature reporting in vitro studies of MMP 9 expression in various cell types. Similarly, they do not offer an explanation on how VEGFR3 expression regulates MMP9 expression.

Regarding MMP9, available methods for analyzing its expression are limited to biochemistry (QPCR, zymography, etc.).These are not readily applicable to the gradient chamber where staining is the only relevant method for visualizing differences across the chamber.

Regarding VEGFR3, we do not think the relationship between VEGFR3 and MMP9 expression is direct. Changing the set point results in changes in many signaling pathways. Elucidating all of the consequences is clearly beyond the scope of this paper.

Moreover, we do not feel that elucidating the mechanisms that govern MMP9 expression in vitro is an essential part of this story as endothelial cells in vitro cannot remodel in a meaningful way.

Reviewer #3:

The Coon et al. manuscript that has apparently been sent to another journal presents some complications:

1) There is overlap as in both manuscripts induction of VEGFR3 phosphorylation by FSS is now presented as a novel finding. The new Figure 5 in the revised eLife manuscript is not of the necessary standard as it fails to show total VEGFR3 protein in the lysates, only P-VEGFR3 is displayed. The total VEGFR3 protein is analysed in a similar Figure 4 of the other manuscript, but one cannot ask the eLife readers to consult another paper for such basic control.

The other manuscript is now accepted and is cited to address this issue.

2) The eLife manuscript avoids clearly mentioning the VEC;VEGFR2 complex, whereas the other manuscript claims VEGFR2;VEGFR3;VEC mechanosensory complex is involved, and speculates that a critical component may actually be a phosphatase

3) Whereas the eLife manuscript emphasizes VEGFR3 alone as the FSS set point, the other manuscript says in the abstract that “VEGFR2 and VEGFR3 signal redundantly downstream of VE-cadherin”.

4) Both manuscripts also show evidence of “significant” VEGFR3 expression in aortic endothelium. It is difficult to know if this amount is biologically significant.

5) The eLife abstract says: “VEGFR3 modulates arterial lumen diameter consistent with flow-dependent remodeling” (it should probably be “aortic”rather than “arterial” here, as other arteriae were not studied). The abstract of the other manuscript has a similar result: “VEGFR3 expression is observed in the aortic endothelium where it contributes to flow responses in vivo”.

We do not understand what is the concern behind these comments that compare the two manuscripts. None of the data overlap and the main conclusions are completely distinct. The major issue flagged by the reviewers was that we invoked VEGFR3 as a component of the mechanosensory mechanism without providing the evidence. This work is now in press and cited in the current manuscript.

Regarding the biological significance of VEGFR3 in arterial endothelial cells, we demonstrated that its deletion results in dramatic remodeling of the aorta in two different organisms. It seems to us that this demonstrates biological significance. We also changed “arterial” by “aortic” in the new abstract, to reflect the reviewer’s concern.

https://doi.org/10.7554/eLife.04645.017

Article and author information

Author details

  1. Nicolas Baeyens

    Department of Internal Medicine, Yale Cardiovascular Research Center, Yale University School of Medicine, New Haven, United States
    Contribution
    NB, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article
    Competing interests
    The authors declare that no competing interests exist.
  2. Stefania Nicoli

    Department of Internal Medicine, Yale Cardiovascular Research Center, Yale University School of Medicine, New Haven, United States
    Contribution
    SN, Acquisition of data, Analysis and interpretation of data
    Competing interests
    The authors declare that no competing interests exist.
  3. Brian G Coon

    Department of Internal Medicine, Yale Cardiovascular Research Center, Yale University School of Medicine, New Haven, United States
    Contribution
    BGC, Helped with experimental design, Analysis and interpretation of data
    Competing interests
    The authors declare that no competing interests exist.
  4. Tyler D Ross

    Department of Internal Medicine, Yale Cardiovascular Research Center, Yale University School of Medicine, New Haven, United States
    Contribution
    TDR, Development of analytical tools
    Competing interests
    The authors declare that no competing interests exist.
  5. Koen Van den Dries

    Department of Internal Medicine, Yale Cardiovascular Research Center, Yale University School of Medicine, New Haven, United States
    Contribution
    KVD, Development of analytical tools
    Competing interests
    The authors declare that no competing interests exist.
  6. Jinah Han

    Department of Internal Medicine, Yale Cardiovascular Research Center, Yale University School of Medicine, New Haven, United States
    Contribution
    JH, Acquisition of data, Contributed unpublished essential data or reagents
    Competing interests
    The authors declare that no competing interests exist.
  7. Holly M Lauridsen

    Department of Biomedical Engineering, Yale University School of Engineering and Applied Science, New Haven, United States
    Contribution
    HML, Design of the flow chamber
    Competing interests
    The authors declare that no competing interests exist.
  8. Cecile O Mejean

    Department of Internal Medicine, Yale Cardiovascular Research Center, Yale University School of Medicine, New Haven, United States
    Contribution
    COM, Development of analytical tools
    Competing interests
    The authors declare that no competing interests exist.
  9. Anne Eichmann

    Department of Internal Medicine, Yale Cardiovascular Research Center, Yale University School of Medicine, New Haven, United States
    Contribution
    AE, Provided critical ideas about the use of the VEGFR3 endothelial inducible KO mice, Contributed unpublished essential data or reagents
    Competing interests
    The authors declare that no competing interests exist.
  10. Jean-Leon Thomas

    1. Department of Neurology, Yale University School of Medicine, New Haven, United States
    2. Université Pierre et Marie Curie, Paris, France
    3. INSERM, CNRS U-1127, UMR-7225, Paris, France
    4. PHP, Groupe Hospitalier Pitié Salpêtrière, Paris, France
    Contribution
    J-LT, Provided critical ideas about the use of the VEGFR3 endothelial inducible KO mice, Contributed unpublished essential data or reagents
    Competing interests
    The authors declare that no competing interests exist.
  11. Jay D Humphrey

    Department of Biomedical Engineering, Yale University School of Engineering and Applied Science, New Haven, United States
    Contribution
    JDH, Provided the original idea, Contributed unpublished essential data or reagents
    Competing interests
    The authors declare that no competing interests exist.
  12. Martin A Schwartz

    1. Department of Internal Medicine, Yale Cardiovascular Research Center, Yale University School of Medicine, New Haven, United States
    2. Department of Biomedical Engineering, Yale University School of Engineering and Applied Science, New Haven, United States
    3. Department of Cell Biology, Yale University School of Medicine, New Haven, United States
    Contribution
    MAS, Provided the original idea, Conception and design, Analysis and interpretation of data, Drafting or revising the article
    For correspondence
    martin.schwartz@yale.edu
    Competing interests
    The authors declare that no competing interests exist.

Funding

U.S. Public Health Service (P01 HL107205)

  • Martin A Schwartz

American Heart Association (AHA) (14POST19020010)

  • Nicolas Baeyens

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Dr Yingdi Wang for her help with the ear skin dissection and lymphatic vessels labeling, David D Simon for his help with the gradient chamber design and Dr Marleen Ansems for the acquisition of the FACS data. NB was supported by BAEF fellowship, WBI. World excellence scholarship and American Heart Association postdoctoral fellowship (14POST19020010). The work was supported by USPHS grant PO1 HL107205 to MAS.

Ethics

Animal experimentation: All animal experiments were performed in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and approved by the Institutional Care and Use Committee of Yale University (protocol #11406).

Reviewing Editor

  1. Fiona M Watt, Reviewing Editor, King's College London, United Kingdom

Publication history

  1. Received: September 6, 2014
  2. Accepted: February 1, 2015
  3. Accepted Manuscript published: February 2, 2015 (version 1)
  4. Version of Record published: February 18, 2015 (version 2)

Copyright

© 2015, Baeyens et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 3,294
    Page views
  • 553
    Downloads
  • 26
    Citations

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

Comments

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)