Positive Feedback Defines the Timing, Magnitude, and Robustness of Angiogenesis - PubMed (original) (raw)

Positive Feedback Defines the Timing, Magnitude, and Robustness of Angiogenesis

Donna J Page et al. Cell Rep. 2019.

Abstract

Angiogenesis is driven by the coordinated collective branching of specialized leading "tip" and trailing "stalk" endothelial cells (ECs). While Notch-regulated negative feedback suppresses excessive tip selection, roles for positive feedback in EC identity decisions remain unexplored. Here, by integrating computational modeling with in vivo experimentation, we reveal that positive feedback critically modulates the magnitude, timing, and robustness of angiogenic responses. In silico modeling predicts that positive-feedback-mediated amplification of VEGF signaling generates an ultrasensitive bistable switch that underpins quick and robust tip-stalk decisions. In agreement, we define a positive-feedback loop exhibiting these properties in vivo, whereby Vegf-induced expression of the atypical tetraspanin, tm4sf18, amplifies Vegf signaling to dictate the speed and robustness of EC selection for angiogenesis. Consequently, tm4sf18 mutant zebrafish select fewer motile ECs and exhibit stunted hypocellular vessels with unstable tip identity that is severely perturbed by even subtle Vegfr attenuation. Hence, positive feedback spatiotemporally shapes the angiogenic switch to ultimately modulate vascular network topology.

Keywords: angiogenesis; endothelial cell; lateral inhibition; positive feedback; tetraspanin; tip cell.

PubMed Disclaimer

Figures

Graphical abstract

Figure 1

Positive Feedback Generates an Ultrasensitive Angiogenic Switch (A) In angiogenesis, ECs in quiescent vessels compete for VEGFR-active versus Notch-active status. VEGFR-active cells acquire motile “tip” identity and initiate branching. (B) An intercellular negative-feedback loop uses lateral inhibition (LI) to limit the number of ECs that acquire VEGFR-active status. (C) Signaling interactions underpinning construction of the two-cell ODE mathematical model. Blue arrow indicates LI. Green and red arrows indicate positive and negative feedback via VEGFR, respectively. HE refers to the combined effects of Notch-induced expression of transcriptional repressors. (D and E) Plots of DLL4 levels in two coupled cells following ODE model simulations using varying levels of positive feedback. Depending on their final level of DLL4, each coupled cell was assigned as having acquired either high VEGFR activity and stable tip identity (D; blue arrowhead indicates high levels of DLL4) or high Notch activity and repressed tip identity (E; red arrowhead indicates low levels of DLL4). Blue and red dashed lines represent maximum and minimum DLL4 thresholds for stable tip identity and repressed tip identity, respectively. (F and G) Matrix plots of tip patterning speeds in the two-cell ODE model following exposure of each coupled cell to different VEGF levels in the absence (F) or presence (G) of positive feedback. Dark gray boxes indicate the slowest rates or failure of tip patterning. Larger orange boxes indicate coupled ECs experiencing low VEGF levels (<0.05 c.u.). (H) ODE modeling of the impact of positive-feedback levels (_P_) on network bistability. Without positive feedback _(_ _P =_ 0), ECs resist switching to a VEGFR active steady state (high DLL4), even when surrounding VEGF is increased. At very high _P_ values (_P =_ 0.1), ECs remain in a VEGFR active state with changing VEGF. At intermediate _P_ values, increasing VEGF levels (>2.5) induce tip cell patterning. Moreover, this active state is retained when VEGF levels are then lowered below 2.5 to ∼1. Hence, positive feedback generates a bistable switch in EC identity that robustly maintains the active state, despite fluctuating VEGF levels. (I) Two-parameter bifurcation plot with changing VEGF and changing P values. Region inside the cusp (green shaded portion) represents values that are bistable in the EC active state. Everything outside is monostable. (J) Predicted role of positive feedback in defining the selection threshold of VEGF that drives tip identity. Data are mean.

Figure 2

Switch-like Behavior of Motile EC Selection in Angiogenesis In Vivo (A and B) Time-lapse images of EC nuclei in ISVs of control (A) and dll4 KD (B) Tg(kdrl:nlsEGFP) zf109 embryos from 19 h post-fertilization (hpf). Brackets indicate dividing cells. Nuclei are pseudocolored. (C–E) Quantification of the number of ECs that are selected to branch (C), undergo proliferation (D), or the total number of ECs per ISV (E) in control, dll4 KD, flt1 KD, and 0.3 μM SU5416-treated embryos (n = 47 ISVs from 16 control, 78 ISVs from 24 dll4 KD, 28 ISVs from 8 flt1 KD, and 81 ISVs from 23 0.3 μM SU5416-treated embryos). (F) Illustration of the biphasic nature of the selection of motile ECs in angiogenesis. Vegf signal levels define the number of ECs selected to branch, and Dll4-mediated LI prevents further selection of motile ECs. Increased Vegf (flt1 KD) or decreased Vegf (0.3 μM SU5416) signaling results in the selection of more or less motile ECs, respectively. In the absence of dll4, motile ECs continue to be selected. Data are mean ± SEM. ∗p < 0.05, two-way ANOVA test. Scale bars, 25 μm. See also Figure S1.

Figure 3

Identification of Putative Positive-Feedback Modulators of Vegf Signaling (A) Fold change in the indicated transcript levels by microarray following inhibition of Vegfr signaling (2.5 μM SU5416), Notch activity (100 μM DAPT), or both, from 22 to 30 hpf. (B and C) Fold change in tm4sf18, kdrl, flt4, and dll4 transcript levels by qPCR in embryos incubated with 2.5 μM SU5416 for the indicated times (B) and tm4sf18 and kdrl transcript levels by qPCR upon dll4 KD (C; n = 3 separate experiments). (D) Illustration of the putative transcriptional regulation of tm4sf18 by Vegf-Notch and proposed function as a positive-feedback modulator of Vegfr signaling. (E) Lateral views of sprouting ISVs in Tg(kdrl:GFP) s843 embryos (left) or WT embryos following whole-mount in situ hybridization analysis of tm4sf18 expression (right). Blue brackets indicate nascent ISVs; red brackets indicate anastomosed ISVs; arrows indicate _tm4sf18_-expressing ISVs; and arrowheads indicate tm4sf18 expression at regions of future angiogenic remodeling. (F) Whole-mount in situ hybridization analysis of tm4sf18 expression in _npas4l_s5 mutant embryos showing loss of expression, as well as upon dll4 KD showing ectopic expansion of tm4sf18 expression to the DA, consistent with de-repression of Vegfr signaling. Data are mean ± SEM. Scale bar, 100 μm. See also Figure S2.

Figure 4

TM4SF1/Tm4sf18 Expression Feeds Back to Amplify VEGF/Vegf Signaling (A) Relative expression levels of TM4SF1 by qPCR in HUVECs transfected with control or TM4SF1-targeted siRNA (n = 4 separate experiments). (B and C) Western blot analysis of pERK/ERK levels in HUVECs after VEGF-A stimulation following transfection with control or _TM4SF1_-targetting siRNA (B) and quantification of pERK/ERK ratios (C) (n = 3 separate experiments). (D) Lesions introduced into the tm4sf18 loci by TALEN and CRISPR gene editing. A 19-bp deletion of tm4sf18 exon-1 and a 16-bp deletion and 2-bp insertion of exon-2 were generated using TALENs and CRISPR/Cas9, respectively. Genomic target sites for the TALENs, gRNA target site, and PAM sequence are indicated by blue, red, and green highlights, respectively. (E) Strategy for assessing Vegfr signaling dynamics in vivo. (F–I) Lateral views of pErk immunostaining in ECs of WT (F) or _tm4sf18_−/− (H) Tg(kdrl:nlsEGFP) zf109 embryos at 0 and 2 h after inhibitor washout and quantification of pErk fluorescence intensity in WT, tm4sf18+/− (G) or _tm4sf18_−/− (I) embryos. Arrowheads in (F) indicate pErk in neuronal cells (n = at least 39 ECs from 8 WT, 129 ECs from 20 tm4sf18+/−, and 74 ECs from 13 _tm4sf18_−/− embryos at each time point). Data are mean ± SEM. ∗p < 0.05, two-tailed t test. Scale bars, 25 μm.

Figure 5

Tm4sf18 Modulates the Magnitude and Timing of the Angiogenic Response (A and B) Quantification of the number of ECs selected to branch (A) or the percentage of ECs that undergo proliferation (B) in WT, tm4sf18+/−, and _tm4sf18_−/− embryos (n = 62 ISVs from 16 WT, 58 ISVs from 15 tm4sf18+/−, and 31 ISVs from 8 _tm4sf18_−/− embryos). (C) Quantification of the distribution of ISV cellularity in WT, tm4sf18+/−, _tm4sf18_−/−, and HU/Ap-treated embryos (n = 65 ISVs from 16 WT, 62 ISVs from 15 tm4sf18+/−, 31 ISVs from 8 _tm4sf18_−/−, and 88 ISVs from 22 HU/Ap-treated embryos). (D) Quantification of the total number of ECs per ISV in WT, tm4sf18+/−, and _tm4sf18_−/− embryos. n is the same as in (A). (E) Predicted shift in the level of VEGF signaling required to achieve a selection threshold in the absence of positive feedback. (F and G) Quantification of the number of ECs selected to branch in 40 nM ZM323881-treated WT, tm4sf18+/− and _tm4sf18_−/− embryos (F) and corresponding time-lapse images of sprouting ISVs in 40 nM ZM323881-treated WT and _tm4sf18_−/− embryos from 20 hpf (G). Embryos were incubated with 40 nM ZM323881 from 18 hpf onward. Nuclei of sprouting ECs emerging from the DA are pseudocolored (n = 26 ISVs from 10 WT, 50 ISVs from 20 tm4sf18+/−, and 21 ISVs from 10 _tm4sf18_−/− embryos). Data are means ± SEM. ∗p < 0.05, two-way ANOVA or two-tailed t test. Scale bar, 25 μm. See also Figure S3.

Figure 6

Hypocellular Vessels in _tm4sf18_−/− Mutants Fail to Extend Appropriately (A) Time-lapse images of sprouting ISVs in WT and tm4sf18_−/−_Tg(kdrl:nlsEGFP) zf109 embryos from 19 hpf. Brackets indicate dividing cells. Nuclei are pseudocolored. ISVs appear shorter in the absence of tm4sf18. (B–D) Quantification of the dorsal movement of tip (cell 1) or stalk (cell 2) ECs in WT and tm4sf18+/− (B), _tm4sf18_−/− (C), or HU/Ap-treated (D) embryos (n = 71 ISVs from 16 WT, 69 ISVs from 15 tm4sf18+/−, 39 ISVs from 8 _tm4sf18_−/−, and 89 ISVs from 22 HU/Ap-treated embryos). (E) Quantification of the dorsal movement of tip ECs in non-proliferating ISVs consisting of 1, 2, and 3 or more ECs and comparison with the motility of tip ECs in _tm4sf18_−/− embryos (n = 39 ISVs from 8 _tm4sf18_−/− embryos, as well as 17 ISVs with 3 cells, 53 ISVs with 2 cells, and 16 ISVs with 1 cell from 22 embryos). (F) Quantification of the number of ECs that reach the DLAV position in WT, tm4sf18+/−, _tm4sf18_−/−, and HU/Ap-treated embryos (n = 63 ISVs from 16 WT, 56 ISVs from 15 tm4sf18+/−, 31 ISVs from 8 _tm4sf18_−/−, and 86 ISVs from 22 HU/Ap-treated embryos). (G) Illustration of the causes of vessel hypoplasia and phenotypic effect on vessel extension. Data are means ± SEM. ∗p < 0.05, two-way ANOVA or two-tailed t test. Scale bars, 25 μm. See also Figure S4.

Figure 7

Robustness of Tip Identity Is Lost in _tm4sf18_−/− Mutants (A) Lateral views of sprouting ECs in ISVs of WT and tm4sf18_−/−_Tg(kdrl:nlsEGFP) zf109 embryos immunostained for pErk. Prior to fixation, embryos were incubated with DMSO or 40 nM ZM323881 from 22 hpf for 3 h. (B) Quantification of pErk fluorescence intensity in WT, tm4sf18+/−, and _tm4sf18_−/− embryos following incubation with DMSO or increasing concentrations of ZM323881 (n = at least 32 ECs from 8 WT, 87 ECs from 22 tm4sf18+/−, and 35 cells from 8 _tm4sf18_−/− embryos at each concentration). (C) Putative role of positive-feedback-generated bistability and hysteretic dynamics in the control of VEGFR signal level robustness in angiogenesis. Bistability ensures that higher levels of VEGF are required to induce tip patterning than to reverse this active state, conferring robustness on tip identity against fluctuations in VEGF levels. (D) Impact of Tm4sf18-mediated positive feedback on the magnitude, speed, and robustness of motile EC selection during ISV branching. Tm4sf18 drives quick and robust decision making but also ensures delicate modulation of the magnitude of EC selection by Vegf levels. In the absence of Tm4sf18, the magnitude of EC selection is diminished, and both the speed and robustness of EC selection are highly variable and more dependent on Vegf level. Data are means ± SEM. ∗p < 0.05, two-way ANOVA or two-tailed t test. Scale bar, 25 μm.

References

1. Arima S., Nishiyama K., Ko T., Arima Y., Hakozaki Y., Sugihara K., Koseki H., Uchijima Y., Kurihara Y., Kurihara H. Angiogenic morphogenesis driven by dynamic and heterogeneous collective endothelial cell movement. Development. 2011;138:4763–4776. - PubMed
2. Arima, S., Nishiyama, K., Ko, T., Arima, Y., Hakozaki, Y., Sugihara, K., Koseki, H., Uchijima, Y., Kurihara, Y., and Kurihara, H. (2011). Angiogenic morphogenesis driven by dynamic and heterogeneous collective endothelial cell movement. Development 138, 4763-4776. - PubMed
1. Benedito R., Rocha S.F., Woeste M., Zamykal M., Radtke F., Casanovas O., Duarte A., Pytowski B., Adams R.H. Notch-dependent VEGFR3 upregulation allows angiogenesis without VEGF-VEGFR2 signalling. Nature. 2012;484:110–114. - PubMed
2. Benedito, R., Rocha, S.F., Woeste, M., Zamykal, M., Radtke, F., Casanovas, O., Duarte, A., Pytowski, B., and Adams, R.H. (2012). Notch-dependent VEGFR3 upregulation allows angiogenesis without VEGF-VEGFR2 signalling. Nature 484, 110-114. - PubMed
1. Bentley K., Chakravartula S. The temporal basis of angiogenesis. Philos. Trans. R. Soc. Lond. B Biol. Sci. 2017;372:20150522. - PMC - PubMed
2. Bentley, K., and Chakravartula, S. (2017). The temporal basis of angiogenesis. Philos. Trans. R. Soc. Lond. B Biol. Sci. 372, 20150522. - PMC - PubMed
1. Bentley K., Gerhardt H., Bates P.A. Agent-based simulation of notch-mediated tip cell selection in angiogenic sprout initialisation. J. Theor. Biol. 2008;250:25–36. - PubMed
2. Bentley, K., Gerhardt, H., and Bates, P.A. (2008). Agent-based simulation of notch-mediated tip cell selection in angiogenic sprout initialisation. J. Theor. Biol. 250, 25-36. - PubMed
1. Bentley K., Mariggi G., Gerhardt H., Bates P.A. Tipping the balance: robustness of tip cell selection, migration and fusion in angiogenesis. PLoS Comput. Biol. 2009;5:e1000549. - PMC - PubMed
2. Bentley, K., Mariggi, G., Gerhardt, H., and Bates, P.A. (2009). Tipping the balance: robustness of tip cell selection, migration and fusion in angiogenesis. PLoS Comput. Biol. 5, e1000549. - PMC - PubMed

Positive Feedback Defines the Timing, Magnitude, and Robustness of Angiogenesis - PubMed (original) (raw)