(341c) VEGFR1 Mediates Cell Migration through Activation of PI3K and PLCɣ

VEGFR1 mediates
cell migration through activation of PI3K and PLC_ɣ

Introduction:

VEGF-VEGFR  regulate neovascularization;
despite this, efforts to control this signaling axis towards directing
neovascularization have yet to be achieved [1].  This challenge underlies a continuing
need to increase our understanding of signal-to-response. Indeed, there is
significantly less knowledge of the VEGFR1 role, which is often described as a
decoy receptor due to its high VEGF affinity but low kinase activity [2], relative
to the strongly pro-angiogenic VEGFR2. However, our recent computational models
have identified that high-VEGFR1 levels may negate the effect of a VEGF
inhibitor [3]. Thus, to understand the VEGFR1 signaling role, we develop computational
models of the initial intracellular signaling mechanisms connected to VEGFR1:
tyrosine site-specific VEGFR1-adapter interactions. Furthermore, we examine how
cooperativity in adapter binding to specific VEGFR1 tyrosine sites, and
adapter-adapter interactions, direct VEGFR1 signaling. We introduce a novel
weighing scheme that correlates adapter-VEGFR1 binding to a cell response
(e.g., proliferation, migration, or ubiquitination). We validate our
computational predictions using wound healing assays, in vitro.

Materials and Methods:

Computational methods: We develop and compare four endothelial VEGFR1-adapter
interaction models in MATLAB Simbiology: each increases in complexity to gauge
how each mechanistic component affects the physiologic response (Fig. 1A). The
four models are as follows: Model 1, simplest: adapters bind to one C-terminal
Tyr site; Model 2: Model 1 + adapter-adapter binding: adapter-adapter
interactions can occur, with binding to one Tyr site; Model 3: Model 1 + Tyr site
specificity: adapters bind to physiologically known Tyr sites; Model 4, most
physiologically accurate: adapter-adapter interactions occur with binding to
known Tyr sites. Therefore, the following comparisons provide insight into
VEGFR1 signaling: Model 1 vs. Model 3 = site-specific binding, Model 1 vs. Model
2 = adapter cooperativity, and Model 1 vs. 4 = site-specific binding +
cooperativity. We take a data-driven approach towards model development: our
experimental qFlow (quantitative) cytometry provides membrane-receptor concentrations
while literature mining provides adapter concentrations via western blots, kinetics
via surface plasmon resonance, and protein-protein interactions via co-immunoprecipitation.
We mine literature to determine how the cell response (migration, proliferation,
or ubiquitination) changes when each adapter is inhibited. We quantify these
changes in migration, proliferation, and ubiquitination, and assign a
corresponding weight to each adapter and linearly correlate adapter
contributions to cell response.

Experimental methods: Wound healing assays are performed using mouse RAW
264.7 macrophages: their high VEGFR1 levels (4,820 ± 112 VEGFR1/cell) relative
to VEGFR2 (1,770 ± 132 VEGFR2/cell) allows us to focus on VEGF-VEGFR1 response.
RAW macrophages are seeded into a 12-well plate and cultured as monolayer to
~90% confluence. The cells are serum starved overnight with DMEM supplemented
with 0.5% FBS and 1% PS. The monolayer is scratched with a 100 µL pipette tip
and washed once with PBS to remove floating cells. Wells are treated with 750
µL of the serum starved growth factor media, VEGF-A₁₆₄ (50 ng/mL),
10 µM Wortmannin (Anti-PI3K, IC50 = 3 nM), 10 µM U73122 (Anti-PLCɣ, IC50 =
1 µM), 10 µM Imatinib Mesylate (Anti-Abl, IC50 = 600 nM), or a combination of
VEGF-A₁₆₄ and an inhibitor. Images of the wounded cell monolayer are
taken using a brightfield microsocope at 0 h and 24 h after scratching. All
experiments are independently carried out in triplicate. Cell migration is quantified
as the number of cells contained within the total gap area relative to the
number of cells immediately after the scratch, using Image J.

Results and Discussion:

Adapter negative cooperativity decreases
VEGFR1-mediated migration. We find
that VEGFR1 regulates HUVEC migration: in Model 1, the integrated migratory response
is 2.4-fold greater than the proliferation, and 4.6-fold greater than the
ubiquitination, integrated responses. Examining how adapter cooperativity
(Model 2) affects VEGFR1-mediated cell responses reveals two key findings.
First, VEGFR1 still primarily promotes migration when accounting for adapter
cooperativity, evidenced by the migration pathway having the greatest
integrated response, 2.1-fold and 3.1-fold greater than proliferation and ubiquitination,
respectively. Second, accounting for adapter cooperativity significantly alters
the phosphorylation of multiple adapters. Specifically, the integrated response
and phosphorylation amplitude of the adapter Nck decrease 88% and 86%,
respectively, when adapter-adapter interactions occur (Model 2) compared to no
adapter cooperativity (Model 1). Thus, cooperativity in adapter binding
mechanistically preventing VEGFR1-mediated Nck activation, implying that Nck
negatively cooperates with the other adapters.

VEGFR1 Tyr sites preferentially activate
the PI3K and PLCɣ adapters. Accounting
for VEGFR1 tyrosine site specificity directs the preferentiality of
adapter-VEGFR1 binding and subsequent adapter phosphorylation. Specifically, we
find that accounting for VEGFR1 tyrosine sites (Model 3) increases PI3K and PLCɣ
signaling; we observe a 1.7-fold and 1.5-fold increase in PI3K and PLCɣ
signaling, respectfully, when accounting for VEGFR1 tyrosine sites (Model 3)
compared to no tyrosine site specificity (Model 1). Thus, when accounting for
specific VEGFR1 tyrosine sites (Model 3), we identify that VEGFR1 is structured
to preferentially activate PI3K and PLCɣ.

Tyr site specificity with adapter
cooperativity amplifies VEGFR1 through increased PLCɣ activation. By comparing VEGFR1 signaling in the simplest (Model
1) and most physiologically relevant (Model 4) cases, we find that VEGFR1.
First, the cell response is unchanged: VEGFR1 signaling primarily promotes cell
migration, followed by proliferation, with ubiquitination having the lowest activation.
Second, both adapter cooperativity and binding specificity are necessary to
enable signal amplification. We arrive at this conclusion because all
integrated responses increase in Model 4 compared to Model 1: 6.5-fold for
proliferation and 3.5-fold for both migration and degradation. These increased
cell responses do not occur with only Tyr site specificity (Model 3) or
adapter-adapter interactions (Model 2). Furthermore, we find that signal
amplification primarily results from the preferential activation of PLCɣ;
the PLCɣ integrated response is 2.5-fold greater than the next most
activated adapter. We believe this high PLCɣ activation is due to coordination
between VEGFR1 Tyr sites and adapter cooperativity. Indeed, PLCɣ has the
highest number of available Tyr binding sites on VEGFR1, in addition to
undergoing cooperative binding with

Wound healing assays validate VEGFR1-induced
migration predictions. Our results
indicate that PI3K and PLCɣ are the two adapters primarily regulating cell
migration within the VEGF-VEGFR1 signaling axis. To test the robustness of this
prediction, we extend our model to macrophages, changing model parameters to
match literature-minded adapter and receptor concentrations (e.g., [VEGFR1] >>
[VEGFR2]).  Again, we predict that PI3K and PLCɣ regulate macrophage
migration. Moreover, we rank-predict that PLCɣ inhibition (max=58%/min=24%)
> PI3K inhibition (max=51%/min=15%) >> Abl inhibition (< 1%
decrease) (Fig. 1B). We validate these predictions to wound healing assays (Fig.
1C).

Conclusions:

Our findings further define VEGFR1 as an
active controller in angiogenesis-rather than a mere decoy receptor. Moreover, we define the key proteins mediating this
migration role, as PI3K and PLC_ɣ. Since these pathways are
linked to Ca²⁺signaling [4], these models offer a path towards
further understanding VEGFR1-migration via calcium signaling. Future studies can
also apply this modeling methodology to understand and control VEGFR1-signaling
in diseases where VEGFR1 concentrations are high, such as in breast cancer [3] and
ischemia [5].

References:

[1] Eichmann A. Curr Opin Cell Biol
(2012) 24:188. [2] Zhang Z. Cell Death Differ (2010) 17:499-512. [3]
Weddell JC. PLOS ONE (2014) 9:e97271. [4] Rameh LE. J Biol Chem (1998)
273:23750. [5] Imoukhuede PI. Am J Physiol Heart Circ Physiol (2013)
304:H1085.