Differential effects of EGFR ligands on endocytic sorting of the receptor.

Abstract

Endocytic downregulation is a pivotal mechanism turning off signalling from the EGF receptor (EGFR). It is well established that whereas EGF binding leads to lysosomal degradation of EGFR, transforming growth factor (TGF)-alpha causes receptor recycling. TGF-alpha therefore leads to continuous signalling and is a more potent mitogen than EGF. In addition to EGF and TGF-alpha, five EGFR ligands have been identified. Although many of these ligands are upregulated in cancers, very little is known about their effect on EGFR trafficking. We have compared the effect of six different ligands on endocytic trafficking of EGFR. We find that, whereas they all stimulate receptor internalization, they have very diverse effects on endocytic sorting. Heparin-binding EGF-like growth factor and Betacellulin target all EGFRs for lysosomal degradation. In contrast, TGF-alpha and epiregulin lead to complete receptor recycling. EGF leads to lysosomal degradation of the majority but not all EGFRs. Amphiregulin does not target EGFR for lysosomal degradation but causes fast as well as slow EGFR recycling. The Cbl ubiquitin ligases, especially c-Cbl, are responsible for EGFR ubiquitination after stimulation with all ligands, and persistent EGFR phosphorylation and ubiquitination largely correlate with receptor degradation.

Full text

Page 1

Traffic 2009; 10: 1115–1127
© 2009 The Authors
Journal Compilation © 2009 John Wiley & Sons A/S
doi: 10.1111/j.1600-0854.2009.00943.x
Differential Effects of EGFR Ligands on Endocytic
Sorting of the Receptor
Kirstine Roepstorff1,†, Michael Vibo Grandal1,†,
Lasse Henriksen1, Stine Louise Jeppe
Knudsen1, Mads Lerdrup1, Lene Grøvdal1,
Berthe Marie Willumsen2and Bo van Deurs1,*
1Department of Cellular and Molecular Medicine, The
Panum Building, Faculty of Health Sciences, University of
Copenhagen, Blegdamsvej 3, DK-2200 Copenhagen,
Denmark
2Department of Biology, University of Copenhagen, Ole
Maaløes Vej 5, DK-2200 Copenhagen, Denmark
∗Corresponding author: Bo van Deurs,
bvd@sund.ku.dk
†These authors contributed equally to this work.
Endocytic downregulation
turning off signalling from the EGF receptor (EGFR). It
is well established that whereas EGF binding leads to
lysosomal degradation of EGFR, transforming growth
factor (TGF)-α causes receptor recycling. TGF-α therefore
leads to continuous signalling and is a more potent
mitogen than EGF. In addition to EGF and TGF-α, five
EGFR ligands have been identified. Although many of
these ligands are upregulated in cancers, very little is
known about their effect on EGFR trafficking.
is a pivotal mechanism
We have compared the effect of six different ligands on
endocytic trafficking of EGFR. We find that, whereas
they all stimulate receptor internalization, they have
very diverse effects on endocytic sorting. Heparin-
binding EGF-like growth factor and Betacellulin target
all EGFRs for lysosomal degradation. In contrast, TGF-
α and epiregulin lead to complete receptor recycling.
EGF leads to lysosomal degradation of the majority
but not all EGFRs. Amphiregulin does not target EGFR
for lysosomal degradation but causes fast as well
as slow EGFR recycling. The Cbl ubiquitin ligases,
especially c-Cbl, are responsible for EGFR ubiquitination
after stimulation with all ligands, and persistent EGFR
phosphorylation and ubiquitination largely correlate
with receptor degradation.
Key words: amphiregulin, betacellulin, endocytic traffick-
ing, epidermal growth factor, epidermal growth factor
receptor, epiregulin, heparin-binding EGF-like growth
factor, transforming growth factor-α
Received 23 January 2009, revised and accepted for pub-
lication 11 May 2009, uncorrected manuscript published
online 19 May 2009 , published online 17 June 2009
Re-use of this article is permitted in accordance with the
Terms and Conditions set out at http://www3.interscience.
wiley.com/authorresources/onlineopen.html
The epidermal growth factor receptor (EGFR) is a receptor
tyrosine kinase involved in normal cellular growth and
differentiation as well as in the pathogenesis of cancer.
Seven different EGFR ligands have been identified: EGF,
transforming growth factor-α (TGF-α), heparin-binding
EGF-like growth factor (HB-EGF), betacellulin (BTC),
amphiregulin (AR), epiregulin (EPI), and epigen (1). Several
of the ligands are found in increased concentrations in
human cancers, where they engage in auto- and paracrine
signalling stimulating tumour progression (2–4).
Endocytic sorting is an important mechanism regulating
EGFR signalling. Upon ligand-binding, EGFR is internalized
and trafficked to endosomes. From endosomes, the
receptor is either recycled to the cell surface, or it
is transported to lysosomes for degradation (recently
reviewed in (5)). The sorting to lysosomes is regulated
by ubiquitination of the receptor, which in turn is
dependent on the pH sensitivity of the ligand-binding.
The EGF binding to EGFR is relatively stable at the
pH of endosomes, so upon EGF binding EGFR is
continuously ubiquitinated by the Cbl ubiquitin ligases (6)
and transported to lysosomes (7,8). In contrast, TGF-α
rapidly dissociates from the receptor when exposed to
the low pH of endosomes (9), and the receptor becomes
de-ubiquitinated and is therefore recycled back to the
plasma membrane (7,10).
Whether EGFR is degraded in lysosomes or recycled to
the plasma membrane is vital for the duration of the
EGFR signal. Following receptor degradation, the EGFR
signal will be diminished until the number of EGFRs at
the plasma membrane has been reestablished by protein
synthesis. In contrast, following receptor recycling, the
cell is immediately able to undergo an additional round of
full EGFR activation. In accordance with this, TGF-α is a
more potent mitogen than EGF (11). At present, very little
is known about how the remaining EGFR ligands affect
endocytic sorting and degradation of EGFR, although
several studies have shown that they differ substantially
in their oncogenic potential (3,4,12,13).
In this study we have compared the effects of six different
EGFR ligands on EGFR endocytosis, phosphorylation,
ubiquitination,degradation,andrecycling.Alloftheligands
stimulate EGFR endocytosis. However, their effects on
intracellular trafficking of EGFR vary significantly, from
complete recycling to complete lysosomal degradation.
Whether EGFR is recycled or degraded after stimulation
with the various ligands correlates well with persistent
receptor phosphorylation and also with persistent Cbl
ubiquitin ligase-dependent receptor ubiquitination and pH
sensitivity of ligand-binding.
www.traffic.dk
1115

Page 2

Roepstorff et al.
Based on our results and the current knowledge about the
role of EGFR ligands in the pathogenesis of cancer, we
propose that the oncogenic potential of the various EGFR
ligands in part depends on their ability to induce receptor
recycling rather than degradation.
Results
EGFR ligands differentially induce receptor
internalization and recycling
To compare the ability of the different EGFR ligands to
induce internalization of EGFR, we used the common
approach of pre-binding ligand on ice, thereby looking at
a synchronized wave of receptor internalization (7,14,15).
We initially determined the incubation time necessary
to obtain maximal ligand-binding. First,125I-EGF binding
to HEp2 cells as a function of time was tested, and
in Figure S1A it is seen that the equilibrium is reached
after only 15 min. To investigate the binding kinetics of
the other ligands to EGFR, cells were incubated with
unlabelledligandoniceforupto2 h.Thereafter,cellswere
briefly washed followed by binding of125I-EGF for 30 min.
The level of125I-EGF binding reflects unbound EGFRs.
In Figure S1B it is seen that all ligands have reached
maximal binding well before 1 h. Thus, to investigate
EGFR internalization, HEp2 cells were incubated with
increasing ligand concentrations on ice for 1 h followed
by washing and 15 min chase at 37◦C. Subsequently,
the amount of EGFR present at the cell surface was
determined by fluorescence-activated cell sorter (FACS)
analysis.AsseeninFigure 1A,allsixligandsinducedEGFR
internalization. However, the degree of internalization
varied between the ligands. HB-EGF and BTC were very
efficient at inducing EGFR internalization. At saturating
concentrations of ligand, 70–80% of cell surface EGFR
was internalized upon the 15 min chase. EGF and TGF-α
both induced internalization of approximately 50% of the
receptors at saturating concentrations of ligand. In case
of EPI and AR, the internalization did not reach saturation
with the ligand concentrations used. This is in accordance
with the fact that these two ligands have a lower affinity
for EGFR compared with the other ligands (16,17).
The varying capability of the ligands to clear EGFR from
the cell surface during a 15 min chase could be because
of different trafficking properties. If a ligand induces a
substantial amount of EGFR recycling to the cell surface,
Figure 1:
affect EGFR endocytosis and recycling.
A) EGFR internalization following stimulation
with increasing concentrations of various
EGFR ligands. HEp2 cells were incubated on
ice with increasing concentrations of ligand
for 1 h, washed, and incubated at 37◦C for
15 min. Subsequently, the amount of EGFR
at the cell surface was determined by FACS
analysis. Data points represent mean +/−
SEM for three independent experiments.
B) Time–course of EGFR internalization and
recycling following stimulation with different
EGFR ligands. Cells were incubated on ice
with 10 or 100 nM of ligands as indicated,
washed, and incubated at 37◦C for different
time periods. The amount of EGFR present
at the cell surface was determined by FACS
analysis. Data points represent mean +/−
SEM for four independent experiments.
EGFR ligands differentially
1116
Traffic 2009; 10: 1115–1127

Page 3

EGFR Ligands
it will clear the cell surface for EGFR less efficiently
than a ligand that does not induce receptor recycling.
Such recycling can best be detected with pulse-chase
experiments, because continuous incubation with ligand
will cause recycled receptors to repeatedly undergo
endocytosis (Figure S2). Thus, to investigate recycling,
HEp2 cells were incubated with a fixed concentration of
ligand on ice for 1 h, washed, and subsequently chased for
the indicated period of time (Figure 1B). At the end of the
chase, the amount of cell surface EGFR was quantified by
FACS analysis. In case of EGF, TGF-α, HB-EGF, and BTC,
10 nM of ligand was used because this concentration
induced a close to maximal EGFR internalization in the
experiment shown in Figure 1A. For the lower affinity
ligands EPI and AR, a concentration of 100 nM was
used. Although the in vivo ligand concentrations have
not been determined for all of the investigated ligands,
it is conceivable that the concentrations used here
are physiologically and pathophysiologically relevant (see
discussion).
Ascan be seenfrom Figure 1B, verylittle EGFRis recycled
to the cell surface following stimulation with HB-EGF
or BTC. In contrast, close to 100% of the receptors
is recycled following stimulation with either TGF-α or
EPI. EGF and AR give intermediary responses, and
induce recycling of approximately 50% of the internalized
receptors. Thus, the six EGFR ligands have very different
effects on EGFR trafficking.
All ligands induce EGFR transport to EEA1-positive
endosomes
It is known that EGFR is transported through EEA1
positive endosomes after stimulation with EGF (18). To
furthertest traffickingand intracellularlocalization ofEGFR
after stimulation with the other ligands, we investigated
the association of EGFR with EEA1 positive endosomes.
Cells were incubated with ligand on ice for 1 h, washed,
and incubated at 37◦C for different time periods. They
were subsequently fixed and labelled for EGFR and EEA1.
Figure 2A shows images of EGFR and EEA1 following
15 min of internalization. As can be seen, all six EGFR
ligands target EGFR to early EEA1-positive endosomes.
Figure 2B shows image quantification of the amount of
cellular EGFR associated with EEA1 positive endosomes
at different time-points. All of the tested EGFR ligands
induce colocalization of EGFR with EEA1, peaking after
15–30 min of internalization. However, AR is slightly less
efficient at targeting EGFR to EEA1-positive endosomes
than the other ligands.
EGFR ligands vary in their potential to stimulate
EGFR degradation
To test how the various EGFR ligands affect receptor
degradation, two different methods were applied. Cells
were stimulated with ligand for 1 h on ice, washed, and
chased for 0–8 h in the presence of cycloheximide to
inhibit de novo EGFR synthesis. The cells were subse-
quently lysed, and the amount of EGFR determined by
ELISA (Figure S3). Alternatively, to avoid the use of cyclo-
heximide, cells were pulse-labelled with35S-methionine,
stimulated with ligand for 1 h on ice, washed, and incu-
bated for 2 or 6 h at 37◦C. EGFR was subsequently
immunoprecipitated and the amount of35S-labelled EGFR
quantified by PhosphorImaging of an SDS-PAGE gel
(Figure 3). As can be seen, stimulation with TGF-α, EPI,
or AR does not lead to significant degradation of EGFR.
Stimulation with either EGF or HB-EGF leads to degrada-
tion of 40–60% of the cellular EGFR, whereas stimulation
with BTC leads to degradation of approximately 70% of
the cellular EGFR.
The degradation data following stimulation with EPI,
TGF-α, EGF, and BTC all correlate well with the degree of
EGFR recyclinginduced bytheligands (compareFigure 1B
and Figure 3). Interestingly, stimulation with AR does not
lead to significant degradation for up to 6 h of EGFR in
spite of the fact that only half of the internalized receptors
are recycled within the time of the recycling experiment.
One possible explanation for this is that EGFR is recycled
more slowly after stimulation with AR compared with
other ligands. To investigate this, we incubated cells with
eitherEGF orARfor1 honice,washed,andchasedfor6 h
at 37◦C, 3 h longer than the recycling experiments shown
in Figure 1B. We subsequently measured the amount of
EGFR present at the cell surface by FACS analysis. Six
hours after EGF stimulation, 60% of the initial amount of
EGFR was present at the cell surface whereas 6 h after
AR stimulation, the amount of EGFR on the cell surface
was 80% of the initial level. This indicates that there may
be some additional, slow recycling after stimulation with
AR compared with EGF.
EGFR ligands differentially induce transport of EGFR
to Lamp1-positive endosomes
We next investigated which ligands that target EGFR
to lysosomes. Cells were incubated with ligand on ice
for 1 h, washed, and further incubated at 37◦C for
different time periods. They were subsequently fixed
and labelled for EGFR and Lamp1, and imaged by
confocal microscopy. As can be seen from Figure 4A,
after 60 min of stimulation with TGF-α
substantial amount of EGFR is present at the cell surface,
which correlates well with the lack of degradation and
complete recycling seen after stimulation with these
ligands. Stimulation with HB-EGF, BTC, or EGF for
60 min results in some colocalization of EGFR with
lysosomes. However, because the receptor is rapidly
degraded once it has reached lysosomes, it is difficult
to quantify the amount of EGFR trafficked to lysosomes.
Indeed, when lysosomal function was inhibited by
bafilomycin, mostEGFR could be found in lysosomesafter
120 min of stimulation with HB-EGF or BTC (not shown
and Figure 4B). Following 60 min of AR stimulation, a
substantial amount of EGFR appears to be located to
non-lysosomal vesicles (Figure 4A). Taking into account
the relatively lower colocalization with EEA1-positive
compartments (Figure 2B) and the slow recycling of
or EPI a
Traffic 2009; 10: 1115–1127
1117

Page 4

Roepstorff et al.
Figure 2: EGFR localization to early endosomes following ligand stimulation. HEp2 cells were incubated on ice with 10 nM (EGF,
TGF-α, HB-EGF, and BTC) or 100 nM (AR and EPI) of ligand, washed, and incubated at 37◦C for different time periods. Cells were fixed
and labelled for EGFR and the early endosome marker EEA1. (A) shows confocal microscopy images of representative cells after 15 min
of EGFR internalization. The lower right panel shows a magnified field of the area boxed in the panel to the left. Bars, 10 μm. (B) shows
a quantification of the amount of EGFR colocalizing with EEA1 in an average of 50–58 cells for each time-point +/− SEM.
EGFR after AR stimulation (see above), these data
indicate that AR leads to sorting of a small fraction of
EGFR to an intracellular compartment which is distinct
from EEA1- and Lamp-1-positive compartments and from
which EGFR slowly recycles. To further characterize this
compartment, we investigated colocalization of EGFR
with Rab-4 and Rab-11, which are markers of recycling
endosomes (19). Following 60 min of AR stimulation,
EGFR partially colocalizes with both Rab-4 and Rab-11
(Figure S4). This is in agreement with a recent paper
showing colocalization between EGFR and Rab-11 after
AR stimulation (20).
Based on the above results, it is clear that the different
EGFR ligands lead to very different trafficking and down-
regulation of EGFR. We therefore further investigated the
molecular mechanisms underlying these differences.
AR, HB-EGF, and BTC bind to EGFR even at very
low pH
It is generally accepted that TGF-α leads to EGFR recycling
because the ligand dissociates from the receptor in the
acidic environment in endosomes. In contrast, EGF bind-
ingtoEGFRismoreresistanttolowpHandthusEGFleads
to degradation of the receptor (9). To investigate whether
1118
Traffic 2009; 10: 1115–1127

Page 5

EGFR Ligands
Figure 3:
degradation. Cells were incubated with35S-methionine/cysteine
for 1–2 h followed by unlabelled medium for 3 h. The cells were
subsequently incubated on ice with 10 nM (EGF, TGF-α, HB-EGF,
andBTC)or100nM(ARandEPI)ofligand,washed,andincubated
at 37◦C for 2 or 6 h. Cells were lysed, immunoprecipitated EGFR
was separated by gel electrophoresis (upper image), and the
amount of radioactive EGFR was quantified by PhosphorImager.
The column bar graph shows mean +/− SEM for quantification
of four independent experiments. Statistical difference from the
0 h control as determined by Student’s t-test is indicated by stars
(*: p < 0.05; ** p < 0.01).
EGFR ligands differentially stimulate EGFR
there is a general correlation between receptor recy-
cling/degradation and pH dependence of ligand-binding,
cells were incubated with ligand on ice for 4 h and then
briefly washed with buffers of varying pH followed by
binding of125I-EGF. The amount of125I-EGF bound to the
cells reflects the number of EGFRs from which ligand
has been removed during the acid wash (Figure 5). As
expected, TGF-α dissociates from the receptor at pH 6.5,
whereas EGF dissociates at a slightly lower pH 5.5 (9).
AR, BTC, and HB-EGF are highly acid-resistant and remain
bound to the receptor at physiologically relevant pH val-
ues. It was not possible to measure the pH dependence
of EPI because of the low affinity of the ligand. The data
show that for EGF, TGF-α, HB-EGF, and BTC the pH sen-
sitivity of ligand-binding correlates well with the level of
receptor recycling/degradation. Surprisingly, AR binding
showed high acid resistance despite the fact that this lig-
and induces recycling and negligible degradation of EGFR.
EGFR ubiquitination largely correlates with receptor
degradation
Ubiquitination of the EGFR has been shown to target
the receptor for lysosomal degradation (7). We therefore
investigated to what degree the different EGFR ligands
induce EGFR ubiquitination. Cells were incubated with
ligand on ice for 1 h, washed, and chased for 5 min
at 37◦C. Subsequently, EGFR was immunoprecipitated
and ubiquitination detected by western blotting. As can
be seen in Figure 6A, all ligands induce ubiquitination of
EGFR,albeit toavaryingextent.Toinvestigatethekinetics
of EGFR ubiquitination, cells were chased for 0 to 20 min
at 37◦C. In Figure S5 it is seen that all ligands induce
85–100% of maximal ubiquitination after 5 min. After
this time-point, ubiquitination persists to different degrees
dependingontheligand.HB-EGFandinparticularBTCgive
strong and/or persistent EGFR ubiquitination. In contrast,
TGFα and EPI give low ubiquitination that is quickly
lost. AR gives a strong initial EGFR ubiquitination which,
however, is quickly lost. Compared with the ubiquitination
seen after HB-EGF and BTC, the level of EGF-stimulated
EGFR ubiquitination is low, but it is more persistent than
seen after TGFα, AR and EPI (Figure 6A and Figure S5).
After 60 min, the level of EGFR ubiquitination was back to
background levels for all ligands (data not shown). When
correlating the level and kinetics of EGFR ubiquitination
(Figures 6A and S5) with that of receptor degradation, it
is seen that the overall pattern of ubiquitination correlates
well with the ability of the various ligands to target EGFR
to lysosomal degradation.
AR does not stimulate persistent recruitment of c-Cbl
to EGFR
c-Cbl is an important ubiquitin ligase responsible for
ubiquitination of EGFR (21). It can bind to EGFR either
directly via phosphorylated Tyr1045 or indirectly via
Grb2 (22–24). To investigate how the different EGFR
ligands affect recruitment of c-Cbl to activated EGFR,
cells were incubated with ligand on ice for 1 h,
washed, and chased at 37◦C for 0–15 min. Cells
were lysed in co-immunoprecipitation (IP) buffer, EGFR
immunoprecipitated, and the immunoprecipitate analysed
by western blotting for EGFR, c-Cbl, and Grb2. As can be
seen from Figure 6B,C, all ligands stimulate c-Cbl binding
to EGFR, but the association is more prolonged after
EGF and BTC stimulation than the remaining ligands. In
particular, AR stimulates a transient c-Cbl binding to EGFR
(Figure 6B). Based on the high ubiquitination level seen
after HB-EGF, it is surprising that we found a low c-Cbl
binding to EGFR after this ligand (see discussion). Like
c-Cbl, Grb2 bound EGFR more efficiently following EGF
or BTC stimulation for 15 min compared with the other
EGFR ligands (Figure 6C). This fits well with the reports
that c-Cbl primarily binds EGFR via Grb2 (25).
Recruitment of c-Cbl to the EGFR and ubiquitination of
the receptor does not correlate after stimulation with
AR and HB-EGF, and this could indicate that c-Cbl is not
decisive for the ubiquitination of EGFR after stimulation
with all ligands.
Cbl ubiquitin ligases are important for EGFR
ubiquitination after stimulation with all ligands
tested
To further investigate the role of Cbl ubiquitin ligases
in EGFR ubiquitination, we transfected cells with small
interfering RNA (siRNA) against c-Cbl and/or Cbl-b,
another ubiquitin ligase reported to ubiquitinate EGFR
after EGF stimulation (26), and tested how this affected
Traffic 2009; 10: 1115–1127
1119