Soluble TL1A is sufficient for activation of death
receptor 3
€ llsack2, Maria Kurz1, Harald Wajant2 and
Sebastian Bittner1, Gertrud Knoll1, Simone Fu
1
Martin Ehrenschwender
1 Institute of Clinical Microbiology and Hygiene, University Hospital of Regensburg, Germany
€rzburg, Germany
2 Division of Molecular Internal Medicine, Department of Internal Medicine II, University Hospital Wu
Keywords
death receptor; DR3; TL1A; TNFRSF25;
TNFSF15
Correspondence
M. Ehrenschwender, Institute of Clinical
Microbiology and Hygiene, University
Hospital of Regensburg, Franz Josef Strauss
Allee 11, 93053 Regensburg, Germany
Fax: +49 941 9446402
Tel: +49 941 9446440
E-mail: [email protected]
(Received 7 August 2015, revised 22
October 2015, accepted 23 October 2015)
doi:10.1111/febs.13576
Death receptor 3 (DR3) is a typical member of the tumor necrosis factor
receptor family, and was initially identified as a T-cell co-stimulatory molecule. However, further studies revealed a more complex and partly dichotomous role for DR3 and its ligand TL1A under (patho)physiological
conditions. TL1A and DR3 are not only a driving force in the development of autoimmune and inflammatory diseases, but also play an important role in counteracting these processes through an increase in the
number of regulatory T cells. Ligands of the tumor necrosis factor family
typically occur in two forms, membrane-bound and soluble, that can differ
strikingly with respect to their efficacy in activating their corresponding
receptor(s). Ligand-based approaches to activate the TL1A–DR3 pathway
therefore require understanding of the molecular prerequisites of TL1Abased DR3 activation. To date, this has not been addressed. Here, we
show that recombinant soluble trimeric TL1A is fully sufficient to strongly
activate DR3-associated pro- and anti-apoptotic signaling pathways. In
contrast to the TRAIL death receptors, which are much better activated by
soluble TRAIL upon secondary ligand oligomerization, but similarly to the
death receptor tumor necrosis factor receptor 1, DR3 is efficiently activated
by soluble TL1A trimers. Additionally, we have measured the affinity of
TL1A–DR3 interaction in a cell-based system, and demonstrated TL1Ainduced DR3 internalization. Identification of DR3 as a tumor necrosis
factor receptor that responds to soluble ligand trimers without further
oligomerization provides a basis for therapeutic exploitation of the TL1A–
DR3 pathway.
Introduction
Death receptor 3 (DR3, also known as TNFRSF25,
Wsl-1, TRAMP and APO-3) and its ligand TL1A
(TNFSF15) are typical members of the tumor necrosis
factor (TNF) receptor superfamily (TNFRSF) and the
TNF ligand superfamily (TNFSF). DR3 belongs to
the death receptor sub-group of the TNFRSF [1], and
its highest degree of sequence similarity is with TNF
receptor 1 (TNFR1). Binding of TL1A to DR3
induces recruitment of adapter proteins such as
TNFR-associated death domain (TRADD), Fas-
Abbreviations
DR3, death receptor 3; FADD, Fas-associated death domain; GpL, Gaussia princeps luciferase; IjBa, inhibitor of jBa; IL8, interleukin 8;
MTT, 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide; NFjB, nuclear factor jB; RIP1, receptor interacting protein 1; SMAC,
second mitochondria-derived activator of caspase; TNC, chicken tenascin C; TNF, tumor necrosis factor; TNFR1, tumor necrosis factor
receptor 1; TNFRSF, tumor necrosis factor receptor superfamily; TNFSF, tumor necrosis factor ligand superfamily; TRADD, TNFR-associated
death domain; TRAF2, TNFR-associated factor 2; TRAIL, TNF-related apoptosis-inducing ligand; TWEAK, TNF-like weak inducer of apoptosis.
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
323
S. Bittner et al.
Soluble TL1A is sufficient for DR3 activation
associated death domain (FADD) and TNFR-associated factor 2 (TRAF2) to the cytoplasmic tail of the
receptor. Association of caspase-8 with FADD allows
stepwise activation of the former, and initiates a cascade of proteolytic cleavage events that culminate in
caspase-3 activation and trigger the execution phase of
apoptosis [2]. Additionally, DR3 is capable of activating the pro-inflammatory and anti-apoptotic classical
nuclear factor jB (NFjB) pathway via receptor interacting protein 1 (RIP1) [3].
TL1A and DR3 were initially identified as T-cell
co-stimulatory molecules [4], but animal models suggest a more complex, partly dichotomous role for the
TL1A–DR3 system in T-cell immunity. On the one
hand, the TL1A–DR3 system has been recognized as a
driving force in development of autoimmune and
inflammatory diseases such as asthma, inflammatory
bowel disease, rheumatoid arthritis and psoriasis [4].
On the other hand, DR3-mediated in vivo expansion
of regulatory T cells prevented allergic lung inflammation, promoted cardiac allograft survival, and reduced
graft-versus-host disease [5–7]. Increasing the frequency of regulatory T cells through, DR3 activation
may also be exploited therapeutically.
To date, attempts to activate DR3 have mainly
focused on receptor-specific agonistic antibodies [5–7].
Therapeutically, this approach may be limited by Fcc
receptor-dependent effector activities such as antibodydependent cellular cytotoxicity, but also by the requirement of many anti-TNFRSF antibodies for Fcc
receptor binding to unleash their full agonistic potential
[8]. Antibody variants with reduced or abrogated binding to the Fcc receptor are available, but this may
reduce their agonistic activity, as recently shown for
antibodies targeting the TNFRSF receptors CD40, DR5
and Fn14 [8]. Ligand-based therapeutic approaches constitute an interesting alternative, but their rational
design critically depends on defining the molecular prerequisites for ligand-mediated receptor activation.
TNFSF ligands are single membrane-spanning proteins with a characteristic extracellular C-terminal
TNF homology domain. The TNF homology domain
is essential for receptor binding and self-assembly to
form homo-trimeric ligands. Most TNFSF ligands also
occur as soluble molecules, generated either through
proteolytic cleavage or by alternative splicing. Soluble
TNFSF ligands typically also form trimers, as they
contain the TNF homology domain. Importantly,
TNFRSF receptors may differ in their ability to
respond to the membrane-bound and soluble forms of
their ligands. For example, soluble TNF binds to
TNFR1 and TNFR2, but only activates TNFR1 efficiently. In contrast, membrane-bound TNF binds to
324
and activates both TNFR1 and TNFR2 [9]. Other
examples of TNFRSF receptors that depend on the
membrane-bound ligand for full-blown activation
include CD95, TRAILR2, CD27, 41BB, TACI and
OX40 [10–13]. Notably, secondary oligomerization of
the soluble ligands targeting these TNFRSF receptors
increases specific activity by several orders of magnitude, thereby generating highly active agonists [10].
Despite the therapeutic potential of the TL1A–DR3
system, the effects of oligomerization on the ability of
soluble TL1A to activate DR3 have not yet been
addressed. Here, we report that soluble TL1A is fully
sufficient to strongly stimulate DR3. Additionally, we
characterized the TL1A–DR3 interaction in a cellbased system, and provide evidence for ligand-induced
DR3 internalization.
Results
Production and biochemical characterization of
soluble TL1A
To study TL1A-mediated DR3 activation, we generated a Flag-tagged TL1A variant (Fig. 1A, top panel)
consisting of the TNF homology domain of human
TL1A with an N-terminal Flag tag and the chicken
tenascin C (TNC) domain. The Flag tag allows affinity purification and antibody-mediated secondary
oligomerization. The TNC domain stabilizes the trimeric structure of TNFSF ligands through strong
non-covalent interactions and disulfide bonding [14].
As previously shown for CD27L [13], trimer instability may reduce ligand activity irrespective of the
oligomerization status, and this possibility therefore
had to be excluded. Flag–TNC–TL1A was expressed
in HEK293T cells, reaching concentrations of 4–
10 lgmL1 in the supernatant. Silver staining and
Western blot analysis (Fig. 1A, lower panel) revealed
that the vast majority of affinity-purified Flag–TNC–
TL1A migrated in non-reducing SDS/PAGE at a
position corresponding to the molecular weight of a
trimer. Only a small fraction of Flag–TNC–TL1A
migrated at positions corresponding to a lower molecular weight. Under reducing conditions, there was
one dominant band of approximately 28 kDa and
two somewhat faster-migrating bands, suggesting glycosylation of TL1A, similar to other TNFSF ligands.
Soluble TL1A, like soluble TNF, induces cell
death and NFjB activation
DR3 and its close relative TNFR1 trigger apoptotic as
well as non-apoptotic signaling pathways [2].
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
S. Bittner et al.
Soluble TL1A is sufficient for DR3 activation
A
B
Flag
Flag-TNC-TL1A
TNC TL1A aa 92-251
TF-1 cells
Flag-TNC-TL1A
+
–
130 kD
95 kD
72 kD
55 kD
130 kD
95 kD
72 kD
55 kD
36 kD
28 kD
36 kD
28 kD
C
αFlag
Flag-TNF(32W/86T)
Log fluorescence intensity
D
E
TNF
Flag-TNC-TL1A
Flag-TNC-TL1A
TNF
Caspase-3 activity [RLU]
125
150
125
100
75
50
100
Viability [%]
IL8 [pg·mL–1]
Isotype
αTNFR1
Isotype
αDR3
Relative cell number
–
DTT +
40
30
20
75
50
25
10
0
Ligand
[ng·mL–1]
0
500
0
Ligand
[ng·mL–1]
0
0.02
0.2
2
20
200
2000
Flag-TNC-TL1A
1x106
7.5x105
5x105
2.5x105
0
Ligand
[ng·mL–1]
0
31.25
125
500
Fig. 1. Cell death induction and NFjB activation by soluble TNF and soluble TL1A. (A) Top panel: scheme of the Flag–TNC–TL1A
variant used in the study. TL1A, human TL1A (amino acids 92-251); TNC, chicken tenascin C (amino acids 110-139). Lower panel: Flag–
TNC–TL1A was produced in HEK293T cells, purified by anti-Flag affinity chromatography, separated by SDS/PAGE under reducing and
non-reducing conditions, and visualized by silver staining (left) or Western blot analysis (right) using anti-Flag antibody. (B) Surface
expression of DR3 and TNFR1 in TF-1 cells was verified by flow cytometry. (C) TF-1 cells were challenged with the indicated
concentrations of TNFR1-specific Flag-TNF(32W/86T) (green bars) or Flag–TNC–TL1A (black bars) for 7 h. Supernatants were harvested,
and IL8 levels were measured using ELISA. Values are means SEM from three experiments. (D) TF-1 cells were treated with
cycloheximide (2.5 lgmL1) for 2 h, and subsequently challenged with the indicated concentrations of TNF (green bars) or Flag–TNC–
TL1A (black bars). Values are means SEM from three experiments. (E) TF-1 cells were treated with cycloheximide (2.5 lgmL1) for
2 h, and subsequently challenged with the indicated concentrations of TNF (green bars) and Flag–TNC–TL1A (black bars) for 3 h.
Caspase-3 activity was measured using the fluorogenic substrate (DEVD)2-R110. Values are means SEM from three experiments.
RLU, relative light units.
Interestingly, in DR3- and TNFR1-positive erythroleukemic TF-1 cells (Fig. 1B), a soluble TNFR1specific TNF mutant [Flag–TNF(32W/86T)] [15]
induced significantly higher levels of expression of the
NFjB target gene encoding interleukin 8 (IL8) compared to Flag–TNC–TL1A (Fig. 1C). A confounding
effect due to concomitant TNFR2 activation was
excluded due to use of the TNFR1-specific TNF variant. Flag–TNC–TL1A and TNF furthermore induced
cell death (Fig. 1D) and caspase-3 activation (Fig. 1E),
which is indicative of ongoing apoptosis, in a dosedependent manner. However, the respective LD50 values differed by three orders of magnitude, with a calculated LD50 for TNF of 0.54 ngmL1 (95%
confidence interval 0.38–0.71 ngmL1), and a LD50
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
for Flag–TNC–TL1A of 1740 ngmL1 (95% confidence interval 1660–1812 ngmL1).
TL1A-induced cytotoxicity is independent of
autocrine TNF secretion
The moderate cytotoxic effect of Flag–TNC–TL1A
was reminiscent of that of TNF-like weak inducer of
apoptosis (TWEAK), another TNFSF ligand. Binding
of TWEAK to its receptor Fn14 triggers cell death in
some cell lines. However, rather than relying on direct
activation of pro-apoptotic adapter proteins and caspases as in the case of TNFR1 or the TRAIL death
receptors, TWEAK utilizes NFjB-dependent induction
of TNF to exert cytotoxicity in an autocrine manner
325
S. Bittner et al.
Soluble TL1A is sufficient for DR3 activation
Flag-TNC-TL1A
0 10 30 60
– – – –
+ + + +
p-IκBα
40 kD
IκBα
40 kD
Tubulin
55 kD
Flag-TNC-TL1A + etanercept
C
100
100
75
75
Viability [%]
Time [min]
Etanercept
Flag-TNC-TL1A
0 10 30 60
Flag-TNC-TL1A
TNF
TNF + etanercept
B
Viability [%]
A
50
25
0
TNF [ng·mL–1]
0
31.25
125
500
50
25
0
Flag-TNC-TL1A 0
–1
[ng·mL ]
31.25
125
500
Fig. 2. TL1A-induced cytotoxicity and NFjB activation are independent of autocrine TNF secretion. (A) TF-1 cells were stimulated with Flag–
TNC–TL1A (200 ngmL1) for the indicated periods of time in the presence and absence of etanercept (50 lgmL1). Subsequently, cells were
washed, lysed and analyzed by Western blotting using antibodies specific for the indicated proteins. Detection of tubulin served as a loading
control. (B) TF-1 cells were treated with cycloheximide (2.5 lgmL1) for 2 h, and subsequently challenged with the indicated concentrations of
TNF in the presence and absence of etanercept (1 mgmL1). Values are means SEM from three experiments. (C) TF-1 cells were treated as
in (B), but stimulated with the indicated concentrations of Flag–TNC–TL1A in the presence and absence of etanercept (1 mgmL1).
via TNFR1 [16]. We therefore wished to determine
whether the modest cytotoxic effects observed upon
TL1A treatment were directly triggered by DR3-associated death pathways or were due to DR3-mediated
NFjB activation with subsequent TNF secretion. In
line with the data shown in Fig. 1C, Flag–TNC–TL1A
induced phosphorylation of inhibitor of jBa (IjBa), a
surrogate marker for ongoing NFjB activation
(Fig. 2A). Pre-treatment with the TNF-neutralizing
TNFR2–Fc fusion protein etanercept did not negatively affect Flag–TNC–TL1A-induced IjBa phosphorylation (Fig. 2A), but, as expected, rescued TF-1 cells
from TNF-induced cell death (Fig. 2B). Flag–TNC–
TL1A-mediated cytotoxicity remained unaffected in
the presence of etanercept (Fig. 2C). Together, these
results suggest that TL1A-induced cell death is directly
mediated through DR3.
Oligomerization of TL1A does not enhance the
efficacy of DR3 activation
Membrane-bound and oligomerized soluble TNFSF
ligand trimers can differ considerably from soluble trimeric TNFSF ligands in terms of their efficacy to activate certain TNFRSF receptors. For soluble TL1A, the
effect of ligand oligomerization on DR3 activation has
not yet been addressed. With regard to the inferior cytotoxic activity, we investigated whether oligomerization
enhances the specific activity of soluble TL1A, essentially
mimicking the activity of the membrane-bound form.
Fig. 3. Oligomerization of soluble TL1A does not enhance specific activity. (A) TF-1 cells were treated with cycloheximide (2.5 lgmL1) for
2 h, and subsequently challenged with the indicated concentrations of Flag–TNC–TL1A in the absence (green bars) and presence (black
bars) of cross-linking anti-Flag M2 antibody (M2, 1 lgmL1). Values are means SEM from three experiments. (B) Top panel: scheme of
the Fc–Flag–TNC–TL1A variant. Fc, human immunoglobulin G1 Fc fragment; TL1A, human TL1A (amino acids 92-251); TNC, chicken
tenascin C (amino acids 110-139). Lower panel: Fc–Flag–TNC–TL1A was produced in HEK293T cells, purified by anti-Flag affinity
chromatography, separated by SDS/PAGE under reducing and non-reducing conditions, and visualized by silver staining (left) or Western blot
analysis (right) using anti-Flag antibody. The migration pattern of Fc–Flag–TNC–TL1A under non-reducing conditions corresponded to a
hexameric ligand. (C) TF-1 cells were treated with cycloheximide (2.5 lgmL1) for 2 h, and subsequently challenged with the indicated
concentrations of Fc–Flag–TNC–TL1A. Values are means SEM from three experiments. (D) TF-1 cells were challenged with the indicated
concentrations of Flag–TRAIL in the presence and absence of the cross-linking antibody anti-Flag M2 (M2, 1 lgmL1). Values are
means SEM from three experiments. (E) DR3 signaling complexes were induced in TF-1 cells by stimulation with 200 ngmL1 HA-Flag–
TNC–TL1A for the indicated periods of time. Proteins associated with HA–Flag–TNC–TL1A were immunoprecipitated using anti-HA
antibodies coupled to agarose beads, and were analyzed together with the corresponding lysates by Western blotting for the presence of
DR3 and the indicated signaling complex components. (F) TF-1 cells were treated with Flag–TNC–TL1A (200 ngmL1) in the presence and
absence of anti-Flag M2 antibody (M2, 1 lgmL1) for the indicated periods of time. After washing and lysis, Western blot analyses were
performed with antibodies specific for the indicated proteins. Detection of tubulin served as a loading control. (G, H) TF-1 cells were
challenged with the indicated concentrations of Flag–TNC–TL1A or Flag–TRAIL in the presence and absence of cross-linking anti-FLAG M2
antibody (M2, 1 lgmL1). Caspase activation was blocked by addition of zVAD-fmk (100 lM) 1 h before TRAIL stimulation. After 6 h,
supernatants were collected and IL8 production was analyzed using ELISA. Values are means SEM from three experiments.
326
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
S. Bittner et al.
Soluble TL1A is sufficient for DR3 activation
A
B
100
+
DTT
–
–
+
250 kD
250 kD
130 kD
130 kD
95 kD
95 kD
D
Flag-TRAIL
E
Cell lysate
HA-Flag-TNC-TL1A –
Time [min]
100
+
5
+
15
+
45
–
0
0
IP: αHA
+
+
+
5 15 45
DR3
70 kD
TRAF2
55 kD
cIAP2
70 kD
Ub-RIP1
RIP1
70 kD
75
50
25
0
M2
only
0
31.2 62.5 125
250
Flag-TRAIL [ng·mL–1]
500
F
G
0
10
30 60
360
Time [min]
0 10 30 60
360
Flag-TNC-TL1A + M2
Flag-TNC-TL1A
IκBα
M2
10
40 kD
p-IκBα
40 kD
Tubulin
55 kD
H
20
15
IL8 (fold increase over control)
Viability [%]
αFlag
55 kD
Flag-TRAIL + M2
0
Fc-Flag-TL1A 0
–1
[ng·mL ]
55 kD
500
50
31.2 62.5 125
250
Flag-TNC-TL1A [ng·mL–1]
5
0
31
.2
5
62
.5
M2
only
25
72 kD
72 kD
0
50
25
25
75
12
50
IL8 (fold increase over control)
TL1A aa 92-251
Fc-Flag-TNC-TL1A
75
Viability [%]
TNC
Viability [%]
100
C
Flag
hIgG1 (hinge + Fc)
Flag-TNC-TL1A
Flag-TNC-TL1A + M2
5
4
Flag-TNC-TL1A
Flag-TNC-TL1A + M2
3
2
1
0
M2
only 15.6
31.2
62.5
125
250
500
Flag-TL1A [ng·mL–1]
Flag-TRAIL
+ zVAD-fmk
Flag-TRAIL + M2
10
5
4
3
2
1
0
M2
only 15.6
31.2
62.5
125
250
500
Flag-TRAIL [ng·mL–1]
Secondary oligomerization of Flag–TNC–TL1A
using an anti-Flag antibody (clone M2) did not significantly affect TL1A-induced loss of viability (Fig. 3A).
Similar observations were obtained using a hexameric
Fc–TL1A fusion protein (Fig. 3B,C). In sharp contrast, but as expected, anti-Flag antibody-mediated
oligomerization of a Flag-tagged TRAIL variant drastically increased the ability of this soluble TNFSF
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
ligand to induce cell death (Fig. 3D). This excludes the
possibility of inefficient cross-linking as a cause for
poor Flag–TNC–TL1A-mediated killing.
In line with these findings, co-immunoprecipitation
of trimeric Flag–TNC–TL1A demonstrated ligandinduced association of DR3 with TRAF2, inhibitor of
apoptosis protein 2 (cIAP2) and RIP1 (Fig. 3E). Flag–
TNC–TL1A treatment resulted in slower-migrating,
327
S. Bittner et al.
Soluble TL1A is sufficient for DR3 activation
potentially ubiquitinated, RIP1, indicative of an IjB
kinase-stimulating, DR3-associated signaling complex.
Indeed, Flag–TNC–TL1A stimulation resulted in
biphasic IjBa phosphorylation, which was slightly
more pronounced after challenge with oligomerized
Flag–TNC–TL1A+M2 (Fig. 3F). Consistently, TL1Ainduced IL8 secretion was somewhat higher in Flag–
TNC–TL1A+M2-treated cells (Fig. 3G). However, the
oligomerization-mediated enhancement of specific
Flag–TNC–TL1A activity was marginal when compared to the corresponding effect observed with Flag–
TRAIL (Fig. 3H). In TF-1 cells protected from apoptosis with the pan-caspase inhibitor zVAD-fmk, anti-Flag
monoclonal antibody-oligomerized Flag–TRAIL acted
as a potent IL8 inductor, whereas even high doses of
trimeric Flag–TRAIL failed to effectively induce IL8
secretion. In summary, oligomerization had no major
effect on the ability of soluble TL1A to activate proand anti-apoptotic DR3 signaling pathways.
Binding studies of TL1A to DR3
Affinity of a ligand for its corresponding receptor(s) is
a major determinant of ligand-induced receptor activation. To date, the biophysical properties of the TL1A–
DR3 interaction have been poorly characterized. We
generated a highly traceable TL1A probe by fusing
Flag–TNC–TL1A to the C-terminus of Gaussia princeps luciferase (GpL) yielding GpL–Flag–TNC–TL1A
(Fig. 4A, top panel). Similar to Flag–TNC–TL1A, the
majority of affinity-purified GpL–Flag–TNC–TL1A
migrated in non-reducing SDS/PAGE at a position corresponding to the molecular weight of a trimer
(Fig. 4A, lower panel). Importantly, the GpL domain
in the GpL–Flag–TNC–TL1A fusion protein did not
impede the activity of the TL1A domain, as Flag–
TNC–TL1A and GpL–Flag–TNC–TL1A were comparably effective with regard to cell death induction
(Fig. 4B). After substrate addition, light emission
increased linearly with the GpL–Flag–TNC–TL1A concentration (Fig. 4C). As expected, binding of the GpLtagged TL1A construct to DR3-positive TF-1 cells was
competed by Flag–TNC–TL1A (Fig. 4D). Increasing
the ratio of Flag–TNC–TL1A to GpL–Flag–TNC–
TL1A linearly correlated with reduced light emission,
reflecting competition of both ligands for DR3 binding.
Following functional characterization of the GpL–
Flag–TNC–TL1A construct, we next investigated the
receptor affinity of TL1A for DR3 by measuring the
equilibrium dissociation constant (KD) in equilibrium
binding studies. For this purpose, we generated a
HeLa cell line showing doxycycline-inducible expression of a DR3 construct lacking the death domain
328
(HeLa-DR3DDD). The death domain deletion in DR3
was made to avoid previously reported cytotoxic
effects due to DR3 over-expression [2]. Flow cytometry revealed significant doxycycline-mediated induction
of DR3 expression in HeLa-DR3DDD cells (Fig. 4E),
which was completely lacking in the absence of doxycycline (data not shown). Addition of doxycycline
increased binding of GpL–Flag–TNC–TL1A to HeLaDR3DDD cells approximately 300-fold compared with
untreated controls (Fig. 4F). Determination of nonspecific ligand binding was achieved by analyzing binding of GpL–Flag–TNC–TL1A to HeLa-DR3DDD
cells in the absence of doxycycline. To determine total
binding, we used doxycycline-treated, and thus
DR3DDD-expressing, cells, and GpL–Flag–TNC–
TL1A concentrations ranging from 14.7 to 7350 pM.
The specific binding data obtained (difference between
total and non-specific binding) fitted with high significance to one-site binding curves, and revealed a KD
value of 3777 2745 pM for GpL–Flag–TNC–TL1A
in HeLa-DR3DDD cells (Fig. 4G).
TL1A binding to DR3 triggers receptor
internalization
Ligand-activated death receptors depend to a varying
degree on internalization to ignite specific cellular
pathways [17]. For the TL1A–DR3 system, trafficking
events such as receptor internalization have not yet
been investigated. We therefore determined whether
association of TL1A with DR3 changes the plasma
membrane distribution of the latter. By targeting the
Flag tag of the Flag–TNC–TL1A construct using a
primary anti-Flag antibody and a fluorophore-conjugated secondary antibody, we were able to trace the
distribution of TL1A by fluorescence microscopy.
Staining of doxycycline-treated HeLa-DR3DDD
cells loaded on ice with Flag–TNC–TL1A (ligand is
shown in red) generated a fluorescent pattern with pronounced intensity at the plasma membrane (Fig. 5A).
Interestingly, the distribution of TL1A changed drastically at 37 °C (Fig. 5B). Fluorescence signals were
now barely detectable at the plasma membrane, but
accumulated in the cytoplasm. To exclude unspecific
staining, HeLa-DR3DDD cells without doxycycline
treatment, and thus without DR3 expression, were
analyzed using the same staining procedure (Fig. 5C).
As expected, no TL1A-derived fluorescence signal was
detectable. Essentially similar results were obtained
when DR3DDD-expressing cells were stimulated with
anti-Flag monoclonal antibody-oligomerized Flag–
TNC–TL1A (Fig. 5D–F), confirming our previous
results that indicated comparable activity of soluble
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
S. Bittner et al.
Soluble TL1A is sufficient for DR3 activation
A
B
C
GpL-Flag-TNC-TL1A
GpL-Flag-TNC-TL1A
Flag-TNC-TL1A
TNC TL1A aa 92-251
100
GpL-Flag-TNC-TL1A
–
+
2x105
–
180 kD
Viability [%]
DTT +
180 kD
130 kD
130 kD
95 kD
95 kD
55 kD
75
50
55 kD
1500
2.5x103
1250 1250 1250 GpL-Flag-TNC-TL1A
0
625 1250 Flag-TNC-TL1A
3x104
αDR3
2x10
RLU
5x103
0
0
1000
GpL-Flag-TNC-TL1A [ng·mL–1]
2000
F
Relative cell number
7.5x103
RLU
200
HeLa-DR3∆DD + Dox
Isotype
Specific binding (RLU)
500
0
E
1x104
G
0
αFlag
D
0
ligand
[ng·mL–1]
1x105
25
0
ligand
[ng·mL–1]
72 kD
72 kD
3x105
RLU
Flag
Gaussia luciferase aa 1-185
1x10
4
4
102
Log fluorescence intensity
0
Dox
–
+
4x104
3x104
2x10
1x10
4
4
Bmax: 69078 RLU
KD: 4576 pM
0
0
1000
3000
5000
GpL-Flag-TNC-TL1A [pM]
Fig. 4. Biophysical characterization of the TL1A–DR3 interaction. (A) Top panel: scheme of the GpL–Flag–TNC–TL1A variant used in the
study. GpL, Gaussia princeps luciferase (amino acids 1-185); TL1A, human TL1A (amino acids 92-251); TNC, chicken tenascin C (amino
acids 110-139). Lower panel: GpL–Flag–TNC–TL1A was produced in HEK293T cells, purified by anti-Flag affinity chromatography, separated
by SDS/PAGE under reducing and non-reducing conditions, and visualized by silver staining (left) or Western blot analysis (right) using antiFlag antibody. (B) TF-1 cells were treated with cycloheximide (2.5 lgmL1) for 2 h, and subsequently challenged with the indicated
concentrations of GpL–Flag–TNC–TL1A (black bars) or Flag–TNC–TL1A (green bars). Values are means SEM from three independent
experiments. (C) Luciferase substrate was added to the indicated concentrations of GpL–Flag–TNC–TL1A, and light emission was quantified
using a luminometer. RLU, relative light units. (D) TF-1 cells were incubated with the indicated concentrations of Flag–TNC–TL1A and/or
GpL–Flag–TNC–TL1A on ice. After washing with ice-cold PBS to remove unbound ligand, cells were resuspended in warm medium and
luciferase substrate was added. Light emission was quantified using a luminometer. The results for one representative experiment of three
experiments performed are shown. (E) Flow cytometric analysis of HeLa cells showing doxycycline-inducible expression of a DR3 construct
lacking the death domain (HeLa-DR3DDD). In the presence of doxycycline (Dox), the DR3 construct was detectable at the cell surface. (F)
HeLa-DR3DDD cells were incubated with GpL–Flag–TNC–TL1A (250 ngmL1) in the presence and absence of doxycycline. After washing to
remove unbound ligand, luciferase substrate was added and light emission was quantified using a luminometer. Data shown are
representative of three experiments. (G) HeLa-DR3DDD cells were seeded in 24-well plates (6 9 104 cellswell1) in the presence and
absence of doxycycline (100 ngmL1, incubation for 18 h). Subsequently, cells were challenged with the indicated concentrations of GpL–
Flag–TNC–TL1A (1 h, 37 °C). After 10 washes with ice-cold PBS to remove unbound ligand, cells were harvested and transferred into 50 lL
medium to determine cell-associated luciferase activity. Non-specific binding was determined from the samples without doxycycline, as
these cells lack DR3 expression. Total binding was determined in doxycycline-treated HeLa-DR3DDD cells challenged with GpL–FLAG–TNC–
TL1A. To calculate specific binding, non-specific binding was subtracted from the corresponding total binding values. Values for the
maximum of specifically bound GpL activity (Bmax) and the dissociation constant (KD) were calculated using GRAPHPAD PRISM5 software by nonlinear regression analysis (one site specific binding model). The results for one representative experiment of three performed are shown.
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
329
S. Bittner et al.
Soluble TL1A is sufficient for DR3 activation
Doxycycline
4 °C
37 °C
A
4 °C
B
C
HeLa-DR3∆DD
Flag-TNC-TL1A
D
E
F
HeLa-DR3∆DD
Flag-TNC-TL1A+M2
G
H
I
HeLa-DR3∆cyt
Flag-TNC-TL1A+M2
Relative cell number
αDR3
K
+ medium
+ TL1A
isotype
Log fluorescence intensity
trimeric and secondary oligomerized TL1A. Admittedly, the observed translocation of DR3DDD into the
cytoplasm upon shifting cells from 4 °C to 37 °C may
330
TF-1 cells
loss of MFI [% control]
TF-1 cells
J
100
75
50
25
0
Time [min] 30 60 90
merely reflect physiological turnover processes of the
plasma membrane rather than TL1A-induced receptor
trafficking events. We therefore generated another
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
S. Bittner et al.
Soluble TL1A is sufficient for DR3 activation
Fig. 5. TL1A binding triggers DR3 internalization. (A–I) HeLa cells showing doxycycline-inducible expression of a DR3 construct lacking the
death domain (HeLa-DR3DDD) or lacking the entire cytoplasmic domain (HeLa-DR3Dcyt) were seeded on cover slips in the presence and
absence of doxycycline (100 ngmL1). The next day, cells were challenged with Flag–TNC–TL1A (400 ngmL1) or Flag–TNC–TL1A+M2
(400 ngmL1) at the indicated temperatures for 30 min. After washing, cells were fixed and permeabilized. Cell-associated TL1A was
visualized using an anti-Flag antibody in combination with a secondary fluorophore-conjugated anti-mouse antibody (shown in red). Nuclei
were counter-stained using Hoechst 33342 (shown in blue). The results shown are representative of three experiments. (J) TF-1 cells were
incubated with Flag–TNC–TL1A (800 ngmL1, blue curve) or medium (red curve) on ice for 30 min. Subsequently, cells were shifted to
37 °C for 90 min, and surface expression of DR3 was assessed by flow cytometry. Data shown are representative of three experiments.
(K) TF-1 cells were treated as in (J), but DR3 surface expression was analyzed 30, 60 and 90 min after cells were shifted to 37 °C. Values
are means SEM from three experiments.
HeLa cell line showing doxycycline-inducible expression of a DR3 construct completely lacking all cytoplasmic residues (HeLa-DR3Dcyt, deletion of amino
acids 222-321). Doxycycline treatment induced robust
expression of this DR3 variant, again with dominant
plasma membrane localization (Fig. 5G). However, in
these cells, the TL1A-derived fluorescence signals persisted predominantly at the plasma membrane even
upon shifting the temperature to 37 °C (Fig. 5H). This
indicates disturbed DR3 trafficking due to the loss of
cytoplasmic residues. TL1A-induced DR3 internalization was also detectable by flow cytometry. Upon
TL1A treatment, we observed a significant loss of
DR3 from the surface of TF-1 cells (Fig. 5J) in a timedependent manner (Fig. 5K).
SMAC mimetics boost DR3-mediated cell death
induction
We have demonstrated that DR3 signaling, like TNFR1
signaling, is predominantly pro-inflammatory. In TNFR1
signaling, small molecules mimicking properties of the
pro-apoptotic protein second mitochondria-derived activator of caspase (SMAC) are capable of shifting the cellular response from inflammation to cell death induction
[18]. In the TL1A–DR3 system, this potentially therapeutically exploitable phenomenon has not yet been addressed.
We consequently pre-treated TF-1 cells with the SMAC
mimetics BV6 (Fig. 6A) and LCL161 (Fig. 6B). Soluble
trimeric Flag–TNC–TL1A alone displayed only marginal
cytotoxicity (green bars), whereas combinatorial treatment
with Flag–TNC–TL1A and SMAC mimetics resulted in a
drastic loss of viability (black bars). These results also
highlight the fact that trimeric soluble Flag–TNC–TL1A
is sufficient for DR3 activation.
Discussion
Surprisingly little is known about the fundamental
mechanisms that underlie DR3 activation by TL1A,
although the TL1A–DR3 system has been heavily investigated as a therapeutic target in human diseases, mainly
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
based on data from mouse models of immune pathologies [5–7]. Notably murine and human DR3 differ in
structure, isoform expression and tissue distribution
[19], which highlights the importance of our effort to
characterize the human TL1A–DR3 interaction. Additionally, many mouse models used agonistic DR3 antibodies rather than the natural ligand TL1A [5–7].
However, antibody-mediated TNFRSF receptor activation does not necessarily reflect ligand-induced activation, and may even trigger a fundamentally different
cellular response, as demonstrated for CD95 and Fn14
[20,21]. Binding of anti-TNFRSF receptor antibodies to
Fcc receptors may enhance their agonistic potential, but
may also trigger effector activities such as antibodydependent cellular cytotoxicity. TL1A-based DR3 activation may be a promising alternative to antibodymediated DR3 activation, but requires understanding of
the molecular prerequisites, as TNFRSF receptors can
differ tremendously in their capacity to respond to soluble TNFSF ligands.
Indeed, soluble TNF was strikingly more effective
than soluble TL1A with regard to cell death induction;
the LD50 for TNF was 0.54 ngmL1, compared with
1740 ngmL1 for TL1A. This difference in LD50 of
almost four orders of magnitude is astonishing, given
that the receptor signaling complexes and downstream
signaling pathways of TNFR1 and DR3 appear to be
largely overlapping [22]. Interestingly, TNFR1 and DR3
signaling involve cIAP1–TRAF2 complexes [22]. Depletion or degradation of cIAP1–TRAF2 complexes, as has
been demonstrated in response to TWEAK-induced
activation of Fn14, allows stabilization of NFjB-inducing kinase with subsequent activation of the non-canonical NFjB pathway. Subsequent secretion of the NFjB
target gene TNF finally induces apoptotic cell death in
an autocrine manner via TNFR1 [16,23,24]. However,
the modest pro-apoptotic effect of DR3 is unlikely to be
triggered by cIAP1–TRAF2 depletion and autocrine
TNF secretion, as the TNF-blocking agent etanercept
failed to rescue TL1A-treated cells, but conferred protection against exogenous TNF administration (Fig. 2).
331
S. Bittner et al.
Soluble TL1A is sufficient for DR3 activation
Flag-TNC-TL1A
Flag-TNC-TL1A + BV6
Flag-TNC-TL1A
Flag-TNC-TL1A + LCL161
B
100
100
75
75
Viability [%]
Viability [%]
A
50
25
0
50
25
0
0
31.2 62.5 125 250 500
Flag-TNC-TL1A [ng mL–1]
0
31.2 62.5 125 250 500
Flag-TNC-TL1A [ng mL–1]
Fig. 6. SMAC mimetics boost DR3-mediated cell death induction. (A, B) TF-1 cells were challenged with the indicated concentrations of
Flag–TNC–TL1A in the absence (green bars) and presence (black bars) of (A) BV6 (10 lM, added 4 h before stimulation) or (B) LCL161 (3 lM,
added 4 h before stimulation). Values are means SEM from three experiments.
The striking differences in the LD50 values for TNF
and TL1A may therefore reflect differences in the
inherent strength of receptor activation between soluble and membrane-bound (or artificially oligomerized) TL1A [10], comparable to those observed with
oligomerized and non-oligomerized Flag–TRAIL
(Fig. 3D,H). We clearly demonstrated that soluble trimeric TL1A is sufficient to induce a functional DR3
signaling complex (Fig. 3E) and to activate pro- and
anti-apoptotic
pathways
(Fig. 3A,G).
Further
oligomerization of trimeric Flag–TNC–TL1A did not
significantly enhance DR3 signaling (Fig. 3A,C). Previous studies reported secretion of interferon c from
human CD4+ T cells upon stimulation with soluble
TL1A and in co-culture assays with TL1A-positive
cells [25,26]. Additionally, binding of soluble hexameric Fc fusion proteins of murine and human TL1A to
DR3 has been demonstrated [27], and Fc–TL1A triggered proliferation of DR3-positive murine Th17 cells
[28]. However, potential differences in the specific
activity of membrane-bound versus soluble TL1A or
trimeric versus more highly oligomerized TL1A have
not been addressed in these studies. It is noteworthy
that side-by-side comparison of a murine hexameric
Fc–TL1A fusion protein and an agonistic antibody
indicated a higher capacity of the former to induce
proliferation of regulatory T cells and caspase activation [29]. Hence, the higher activity of Fc-TL1A may
be attributable to differences in the effectiveness of
ligand-based versus antibody-mediated DR3 activation, and may not be related to TL1A oligomerization,
again illustrating the importance of elucidating mechanistic aspects of DR3 activation.
To elicit a cellular response, ligand binding to
surface receptors is a prerequisite, and depends on the
affinity of the ligand for the receptor. Biophysically,
the TL1A–DR3 interaction has been poorly
332
characterized to date. Cellular binding studies involving chemically labeled TNF ligands (e.g. iodinated or
biotinylated ligands) have limitations, such as high
batch-to-batch variations in labeling efficacy and generation of mixtures of molecules with non-homogeneous activity. We circumvented this by using a GpL–
Flag–TNC–TL1A construct (Fig. 4A). This allowed
assessment of the equilibrium dissociation constant
(KD) in intact human cells. The KD of the TL1A–DR3
interaction in HeLa DR3DDD cells at 37 °C was
3777 2745 pM, which is essentially in accordance
with the results of a surface plasmon resonance-based
study that analyzed TL1A binding to an immobilized
DR3–Fc construct and reported a KD of
6450 200 pM [1]. The KD for the TL1A–DR3 interaction was comparable to that for other death ligand/
death receptor pairs (e.g. 2000–3600 pM for CD95L–
CD95, depending on cell type), but approximately two
orders of magnitude higher than the KD of the TNF–
TNFR1 interaction (19–47 pM) [30,31]. Therefore, differences in the ligand–receptor affinity may contribute
to the discrepancy in the LD50 values for TNF and
TL1A. Admittedly, cell-type dependent variations of
the KD values of TNFSF ligands have been observed
[30,31], and our measurements relied on a DR3 construct without a death domain. Moreover, the role of
other variables such as the percentage of ligand-occupied TNFR1 or DR3 molecules necessary to trigger a
cellular response, and the role of a membrane-bound
DR3 splice variant lacking CRD4 [32], remain to be
determined.
Although ligand binding to death receptors triggers
initial assembly of a signaling complex at the plasma
membrane, some downstream signaling may depend
on receptor internalization. For example, TNFR1 is
capable of initiating pro-inflammatory signaling at the
plasma membrane, but efficient cell death induction
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
S. Bittner et al.
depends on receptor internalization [33]. Due to the
high degree of similarity between TNFR1 and DR3,
we hypothesized that TL1A also initiates DR3 internalization. Trimeric Flag–TNC–TL1A (Fig. 5A,B,J,K)
and oligomerized Flag–TNC–TL1A+M2 (Fig. 5D,E)
both efficiently triggered DR3 internalization, which
was absent in DR3Dcyt-expressing cells (Fig. 5G,H).
Ligand-induced internalization of TNFR1 via clathrinmediated endocytosis critically depends on a cytoplasmic internalization motif located in the tail of the
receptor [33]. Possibly, this also holds true for DR3,
thereby explaining the predominant membrane-localized staining of HeLa DR3Dcyt cells (Fig. 5G,H).
Notably, TL1A also induced in HeLa DR3Dcyt cells a
subtle change in the staining pattern from largely coalescent areas (Fig. 5G) to distinct dots (Fig. 5H), and
a minor proportion of fluorescent spots translocated
into the cytoplasm. The latter may indicate clathrinindependent mechanisms of receptor internalization.
Most functions of the TL1A–DR3 signaling axis
in vivo are thus far thought to result from activation
of non-apoptotic pathways. DR3 signaling, similar to
TNFR1 signaling, occurs primarily through the
TRADD–TRAF axis, with subsequent activation of
NFjB and mitogen-activated protein kinase [3]. Physiological roles for non-apoptotic DR3 signaling in lymphocytes and myeloid cells and also outside the
immune system have been demonstrated [4]. In this
respect, the observed modest capability of DR3 to
induce cell death reflects the predominant pro-inflammatory DR3 function in vivo. In fact, blockade of
DR3 signaling, e.g. using neutralizing anti-TL1A antibodies, ameliorated the course of disease in several
mouse models of immune pathologies [4], and has
received attention as a therapeutic approach in
humans. Interestingly, in TNFR1 signaling, depletion
of cytoplasmic cIAP1–TRAF2 complexes (e.g. via
Fn14 or TNFR2 activation or treatment with SMAC
mimetics) may shift the cellular response from inflammation to cell death [23,24,34]. For DR3 signaling, we
showed that targeting cIAP1–TRAF2 complexes with
SMAC mimetics boosts TL1A-induced cell death
(Fig. 6). Given the restricted DR3 expression on distinct cell types, which is often dependent on activation,
e.g. in inflammatory processes, combination of SMAC
mimetics with TL1A may constitute an approach to
shut down disease-related pro-inflammatory DR3 signaling through targeted cell death induction. Taken
together, we provide evidence that the soluble trimeric
form of TL1A is fully capable of initiating signaling
events via DR3, and is therefore fully sufficient to
robustly activate DR3. Measurements of TL1A affinity
to its receptor and demonstration of TL1A-induced
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
Soluble TL1A is sufficient for DR3 activation
DR3 internalization extend our understanding of this
signaling axis, and may facilitate the development of
novel approaches for therapeutic DR3 targeting.
Experimental procedures
Cells, antibodies and reagents
TF-1 and HeLa cells were obtained from the Deutsche
Sammlung von Mikroorganismen und Zellkulturen (Braunschweig, Germany), and grown in RPMI-1640 medium
(PAN Biotech, Aidenbach, Germany) supplemented with
10% v/v fetal bovine serum (Sigma, St. Louis, MO, USA).
TF-1 cells were additionally supplemented with 5 ngmL1
human granulocyte-macrophage colony-stimulating factor
(Immunotools, Friesoythe, Germany). HeLa cells showing
doxycycline-inducible expression of DR3 constructs were
generated using the Lenti-XTM Tet-On 3G inducible
expression system (Clontech, Mountain View, CA, USA)
according to the manufacturer’s instructions. Antibodies
used in the study comprised IjBa, phospho-IjBa, cIAP2
and TRAF2 (all from Cell Signaling Technology, Beverly,
MA, USA), DR3 (R&D Systems, Wiesbaden, Germany),
tubulin (Dunnlab, Asbach, Germany) and Flag-M2
(Sigma). MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl
tetrazolium bromide) and cycloheximide were obtained
from Biomol (Hamburg, Germany). The SMAC mimetics
BV6 and LCL161 were obtained from Selleck Chemicals
(Houston, TX, USA). Human recombinant TNF was provided by D. M€annel (Institute of Immunology, University
of Regensburg, Germany). The TNFR1-specific mutant
TNF(32W/86T) has been described previously [15]. Etanercept was obtained from Pfizer (Berlin, Germany). Anti-HA
and Anti-Flag agarose was purchased from Sigma. Western
blot analysis was performed as described previously [35].
Plasmids, production and purification of
recombinant proteins
The expression vectors encoding the TL1A constructs are
derived from a variant of pCR3 (Life Technologies, Darmstadt, Germany) encoding an Ig signal peptide (provided
by P. Schneider, Department of Biochemistry, University
of Lausanne, Switzerland) followed by a Flag–TNC
domain [36]. Flag–TNC–TL1A constructs were generated
by inserting a TL1A amplicon encoding amino acids 92251, including a stop codon, in-frame 50 to the Flag–TNC
domain using standard cloning techniques. The Fc–Flag–
TNC–TL1A construct was generated by in-frame insertion
of Flag–TNC–TL1A into a pCR3 variant encoding the
IgG1 Fc domain and a linker (kind gift from P. Schneider).
Recombinant TL1A was produced in HEK293T cells by
transient transfection with polyethylenimine. Briefly,
7.5 9 106 cells were seeded in 15 cm cell culture plates and
transfected with a mixture of polyethylenimine and 25 lg
333
S. Bittner et al.
Soluble TL1A is sufficient for DR3 activation
DNA. The medium was replaced 6 h post-transfection. The
supernatant was collected after 48–72 h, cleared by centrifugation (2000 g, 5 min 4 °C), and purified by affinity
chromatography using anti-Flag antibody-conjugated agarose beads (Sigma). After elution with Tris-buffered saline
containing 100 lgmL1 Flag peptide (Sigma), fractions
were dialyzed against PBS and stored at 20 °C for further
analysis.
Co-immunoprecipitation
Co-immunoprecipitation of HA–Flag–TNC–TL1A was performed essentially as described previously for CD95L [35].
Briefly, 1 9 108 cells were stimulated with 200 ngmL1
HA–Flag–TNC–TL1A for the indicated times at 37 °C,
washed in ice-cold PBS, transferred into 1.5 mL lysis buffer
(30 mM Tris/HCl, pH 7.5, 1% Triton X-100, 10% glycerol,
120 mM NaCl) supplemented with cOmpleteTM protease
inhibitor cocktail (Roche, Mannheim, Germany), and incubated for 20 min on ice. Lysates were cleared by centrifugation (2 9 20 min, 14 000 g), and DR3 complexes were
precipitated using anti-HA antibodies coupled to agarose
beads (Sigma, 40 lL of a 50% v/v slurry) at 4 °C overnight. Lysates from non-stimulated cells supplemented with
50 ng of HA–Flag–TNC–TL1A before adding anti-HA
antibody-coupled agarose beads served as a negative control. After washing in lysis buffer, agarose-bound proteins
were eluted by incubation at 75 °C for 10 min in 49
Laemmli sample buffer.
IL8 ELISA
For the IL8 ELISA, 7 9 104 TF-1 cells were stimulated in
0.2 mL medium with the indicated concentrations of soluble TNF ligands for 7 h in triplicate. Supernatant was collected, and IL8 was quantified by ELISA (BD Biosciences,
Heidelberg, Germany).
Cell viability assay
TF-1 cells were seeded in 96-well plates (7 9 105 per well),
treated with cycloheximide (2.5 lgmL1 for 2 h), BV6
(10 lM for 4 h) or LCL161 (3 lM for 4 h), and subsequently stimulated with the indicated concentrations of the
ligands in triplicate. Cell viability was determined 18 h after
stimulation using MTT staining.
Caspase-3 activity assay
Caspase activity was determined using a caspase-3 activity
kit (AAT Bioquest, Sunnyvale, CA, USA) according to the
manufacturer’s instructions. Light emission was quantified
using a Victor3 multilabel reader (Perkin Elmer, Waltham,
MA, USA).
334
Flow cytometry
DR3 surface expression of TF-1 cells (5 9 105 per group)
was measured by staining with DR3-specific biotinylated
antibody (R&D Systems), fluorescein isothiocyanate-conjugated donkey anti-goat antibody (Jackson ImmunoResearch, Baltimore, MD, USA) and avidin-coupled
fluorescein isothiocyanate (R&D Systems). Biotinylated
goat IgG (R&D Systems) served as a specificity control.
TNFR1 surface expression was measured by staining with
phycoerythrin-conjugated TNFR1-specific antibody (R&D
Systems). Analyses were performed using a FACSCanto
flow cytometer (BD Biosciences) according to standard procedures.
Binding studies
Bioluminescent binding studies using GpL–Flag–TNC–
TL1A were essentially performed as described elsewhere
[31]. DR3 expression was triggered by adding doxycycline
(100 ngmL1) for 18 h. Untreated samples served as controls to determine non-specific binding. Induced and
untreated cells were incubated pairwise with increasing concentrations of GpL–Flag–TNC–TL1A for 1 h at 37 °C to
achieve equilibrium binding. Unbound ligand was removed
by washing the plates 10 times for 5 s in ice-cold PBS.
Next, cells were harvested using a rubber policeman, resuspended in 50 lL RMPI-1640 medium (0.5% v/v fetal
bovine serum) and transferred to black 96-well plates. GpL
activity was quantified by measuring luminescence using a
Gaussia luciferase assay kit (New England Biolabs, Frankfurt am Main, Germany) in a luminometer (AnthosLabtec
Instruments, Krefeld, Germany). Specific binding was
defined as the difference between the total binding values in
samples after doxycycline-triggered DR3 expression and the
corresponding non-specific binding values in DR3-negative
samples (without doxycycline). The maximum of specifically bound GpL activity (Bmax) and the dissociation constant (KD) were calculated using GRAPHPAD PRISM5 software
(GraphPad software, La Jolla, CA, USA) (non-linear
regression analysis, one site specific binding model).
Microscopy
For microscopy, 2.5 9 104 cells were seeded on round
12 mm cover slips. After adhesion, expression of the DR3
constructs was induced by adding doxycycline (100 ngmL1,
18 h). Untreated cells served as a negative control. The next
day, cells were washed with either cold medium (4 °C) or
warm
medium
(37 °C),
and
Flag–TNC–TL1A
(400 ngmL1) or Flag–TNC–TL1A+M2 (400 ngmL1) was
added. After incubation for 30 min at 4 °C or 37 °C,
unbound ligand was washed away by rinsing three times with
cold or warm PBS. Subsequently, cells were fixed in 4% w/v
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
S. Bittner et al.
paraformaldehyde (15 min, room temperature) and permeabilized by incubation in PBS supplemented with 0.1% v/v
Triton X-100 and 2% w/v BSA for 30 min at room temperature. To stain cell-associated TL1A, cover slips were incubated overnight with anti-Flag antibody (Sigma) diluted
1 : 2000 in 600 lL PBS + 2% w/v BSA. The next day, cover
slips were washed three times in PBS and incubated with an
Alexa Fluor 555-conjugated secondary donkey anti-mouse
antibody (Life Technologies; dilution 1 : 200 in 200 lL
PBS + 2% w/v BSA) for 2 h at room temperature protected
from light. After three washing steps with PBS, nuclei were
counter-stained with Hoechst 33342 (dilution 1:10 000, incubation for 10 min at room temperature). Cover slips were
again washed three times using PBS, mounted on microscope
slides, and analyzed using a Keyence BZ-9000 fluorescence
microscope (Keyence, Osaka, Japan).
Soluble TL1A is sufficient for DR3 activation
6
7
8
9
Acknowledgements
M.E. is supported by a grant from the German-Israeli
Foundation for Scientific Research and Development
(grant number G-2367-201.11/2014). M.E. and H.W.
are supported by the Deutsche Forschungsgemeinschaft (grants EH 465/2-1, Wa 1025/24-1 and Wa
1025/19-2).
10
11
Author contributions
ME and HW designed the experiments, analyzed the
data and wrote the paper; ME, SB, MK, SF and GK
performed the experiments.
12
References
1 Migone TS, Zhang J, Luo X, Zhuang L, Chen C, Hu
B, Hong JS, Perry JW, Chen SF, Zhou JX et al. (2002)
TL1A is a TNF-like ligand for DR3 and TR6/DcR3
and functions as a T cell costimulator. Immunity 16,
479–492.
2 Chinnaiyan AM, O’Rourke K, Yu GL, Lyons RH,
Garg M, Duan DR, Xing L, Gentz R, Ni J & Dixit
VM (1996) Signal transduction by DR3, a death
domain-containing receptor related to TNFR-1 and
CD95. Science 274, 990–992.
3 Wen L, Zhuang L, Luo X & Wei P (2003) TL1Ainduced NF-jB activation and c-IAP2 production
prevent DR3-mediated apoptosis in TF-1 cells. J Biol
Chem 278, 39251–39258.
4 Richard AC, Ferdinand JR, Meylan F, Hayes ET,
Gabay O & Siegel RM (2015) The TNF-family cytokine
TL1A: from lymphocyte costimulator to disease coconspirator. J Leukoc Biol 98, 333–345.
5 Wolf D, Schreiber TH, Tryphonopoulos P, Li S, Tzakis
AG, Ruiz P & Podack ER (2012) Tregs expanded
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS
13
14
15
16
in vivo by TNFRSF25 agonists promote cardiac
allograft survival. Transplantation 94, 569–574.
Schreiber TH, Wolf D, Tsai MS, Chirinos J, Deyev VV,
Gonzalez L, Malek TR, Levy RB & Podack ER (2010)
Therapeutic Treg expansion in mice by TNFRSF25
prevents allergic lung inflammation. J Clin Invest 120,
3629–3640.
Kim BS, Nishikii H, Baker J, Pierini A, Schneidawind
D, Pan Y, Beilhack A, Park CG & Negrin RS (2015)
Treatment with agonistic DR3 antibody results in
expansion of donor Tregs and reduced graft-versus-host
disease. Blood 126, 546–557.
Wajant H (2015) Principles of antibody-mediated TNF
receptor activation. Cell Death Differ 22, 1727–1741.
Grell M, Douni E, Wajant H, Lohden M, Clauss M,
Maxeiner B, Georgopoulos S, Lesslauer W, Kollias G,
Pfizenmaier K et al. (1995) The transmembrane form of
tumor necrosis factor is the prime activating ligand of
the 80 kDa tumor necrosis factor receptor. Cell 83,
793–802.
Schneider P, Holler N, Bodmer JL, Hahne M, Frei K,
Fontana A & Tschopp J (1998) Conversion of
membrane-bound Fas(CD95) ligand to its soluble form
is associated with downregulation of its proapoptotic
activity and loss of liver toxicity. J Exp Med 187,
1205–1213.
Wajant H, Moosmayer D, Wuest T, Bartke T, Gerlach
E, Schonherr U, Peters N, Scheurich P & Pfizenmaier
K (2001) Differential activation of TRAIL-R1 and -2
by soluble and membrane TRAIL allows selective
surface antigen-directed activation of TRAIL-R2 by a
soluble TRAIL derivative. Oncogene 20, 4101–4106.
Muller N, Wyzgol A, Munkel S, Pfizenmaier K &
Wajant H (2008) Activity of soluble OX40 ligand is
enhanced by oligomerization and cell surface
immobilization. FEBS J 275, 2296–2304.
Wyzgol A, Muller N, Fick A, Munkel S, Grigoleit GU,
Pfizenmaier K & Wajant H (2009) Trimer stabilization,
oligomerization, and antibody-mediated cell surface
immobilization improve the activity of soluble trimers
of CD27L, CD40L, 41BBL, and glucocorticoid-induced
TNF receptor ligand. J Immunol 183, 1851–1861.
Kammerer RA, Schulthess T, Landwehr R, Lustig A,
Fischer D & Engel J (1998) Tenascin-C hexabrachion
assembly is a sequential two-step process initiated by
coiled-coil a-helices. J Biol Chem 273, 10602–10608.
Loetscher H, Stueber D, Banner D, Mackay F &
Lesslauer W (1993) Human tumor necrosis factor a
(TNFa) mutants with exclusive specificity for the 55kDa or 75-kDa TNF receptors. J Biol Chem 268,
26350–26357.
Vince JE, Chau D, Callus B, Wong WW, Hawkins CJ,
Schneider P, McKinlay M, Benetatos CA, Condon SM,
Chunduru SK et al. (2008) TWEAK-FN14 signaling
induces lysosomal degradation of a cIAP1–TRAF2
335
Soluble TL1A is sufficient for DR3 activation
17
18
19
20
21
22
23
24
25
26
336
complex to sensitize tumor cells to TNFa. J Cell Biol
182, 171–184.
Schneider-Brachert W, Heigl U & Ehrenschwender M
(2013) Membrane trafficking of death receptors:
implications on signalling. Int J Mol Sci 14,
14475–14503.
Varfolomeev E, Blankenship JW, Wayson SM,
Fedorova AV, Kayagaki N, Garg P, Zobel K, Dynek
JN, Elliott LO, Wallweber HJ et al. (2007) IAP
antagonists induce autoubiquitination of c-IAPs, NFjB activation, and TNFa-dependent apoptosis. Cell
131, 669–681.
Wang EC, Kitson J, Thern A, Williamson J, Farrow
SN & Owen MJ (2001) Genomic structure, expression,
and chromosome mapping of the mouse homologue for
the WSL-1 (DR3, Apo3, TRAMP, LARD, TR3,
TNFRSF12) gene. Immunogenetics 53, 59–63.
Huang DC, Hahne M, Schroeter M, Frei K, Fontana
A, Villunger A, Newton K, Tschopp J & Strasser A
(1999) Activation of Fas by FasL induces apoptosis by
a mechanism that cannot be blocked by Bcl-2 or Bcl-x
(L). Proc Natl Acad Sci USA 96, 14871–14876.
Salzmann S, Seher A, Trebing J, Weisenberger D,
Rosenthal A, Siegmund D & Wajant H (2013)
Fibroblast growth factor inducible (Fn14)-specific
antibodies concomitantly display signaling pathwayspecific agonistic and antagonistic activity. J Biol Chem
288, 13455–13466.
Varfolomeev E, Goncharov T, Maecker H, Zobel K,
Komuves LG, Deshayes K & Vucic D (2012) Cellular
inhibitors of apoptosis are global regulators of NF-jB
and MAPK activation by members of the TNF family
of receptors. Sci Signal 5, ra22.
Wicovsky A, Salzmann S, Roos C, Ehrenschwender M,
Rosenthal T, Siegmund D, Henkler F, Gohlke F,
Kneitz C & Wajant H (2009) TNF-like weak inducer of
apoptosis inhibits proinflammatory TNF receptor-1
signaling. Cell Death Differ 16, 1445–1459.
Schneider P, Schwenzer R, Haas E, Muhlenbeck F,
Schubert G, Scheurich P, Tschopp J & Wajant H (1999)
TWEAK can induce cell death via endogenous TNF and
TNF receptor 1. Eur J Immunol 29, 1785–1792.
Biener-Ramanujan E, Gonsky R, Ko B & Targan SR
(2010) Functional signaling of membrane-bound TL1A
induces IFN-c expression. FEBS Lett 584, 2376–2380.
Jin S, Chin J, Seeber S, Niewoehner J, Weiser B,
Beaucamp N, Woods J, Murphy C, Fanning A,
Shanahan F et al. (2013) TL1A/TNFSF15 directly
induces proinflammatory cytokines, including TNFa,
from CD3+ CD161+ T cells to exacerbate gut
inflammation. Mucosal Immunol 6, 886–899.
S. Bittner et al.
27 Bossen C, Ingold K, Tardivel A, Bodmer JL, Gaide O,
Hertig S, Ambrose C, Tschopp J & Schneider P (2006)
Interactions of tumor necrosis factor (TNF) and TNF
receptor family members in the mouse and human. J
Biol Chem 281, 13964–13971.
28 Pappu BP, Borodovsky A, Zheng TS, Yang X, Wu P,
Dong X, Weng S, Browning B, Scott ML, Ma L et al.
(2008) TL1A–DR3 interaction regulates Th17 cell
function and Th17-mediated autoimmune disease. J
Exp Med 205, 1049–1062.
29 Khan SQ, Tsai MS, Schreiber TH, Wolf D, Deyev VV
& Podack ER (2013) Cloning, expression, and
functional characterization of TL1A-Ig. J Immunol 190,
1540–1550.
30 Lang I, Fick A, Schafer V, Giner T, Siegmund D &
Wajant H (2012) Signaling active CD95 receptor
molecules trigger co-translocation of inactive CD95
molecules into lipid rafts. J Biol Chem 287, 24026–24042.
31 Fick A, Lang I, Schafer V, Seher A, Trebing J,
Weisenberger D & Wajant H (2012) Studies of binding
of tumor necrosis factor (TNF)-like weak inducer of
apoptosis (TWEAK) to fibroblast growth factor
inducible 14 (Fn14). J Biol Chem 287, 484–495.
32 Screaton GR, Xu XN, Olsen AL, Cowper AE, Tan R,
McMichael AJ & Bell JI (1997) LARD: a new
lymphoid-specific death domain containing receptor
regulated by alternative pre-mRNA splicing. Proc Natl
Acad Sci USA 94, 4615–4619.
33 Schneider-Brachert W, Tchikov V, Neumeyer J, Jakob
M, Winoto-Morbach S, Held-Feindt J, Heinrich M,
Merkel O, Ehrenschwender M, Adam D et al. (2004)
Compartmentalization of TNF receptor 1 signaling:
internalized TNF receptosomes as death signaling
vesicles. Immunity 21, 415–428.
34 Fotin-Mleczek M, Henkler F, Samel D, Reichwein M,
Hausser A, Parmryd I, Scheurich P, Schmid JA &
Wajant H (2002) Apoptotic crosstalk of TNF receptors:
TNF-R2-induces depletion of TRAF2 and IAP proteins
and accelerates TNF-R1-dependent activation of
caspase-8. J Cell Sci 115, 2757–2770.
35 Ehrenschwender M, Siegmund D, Wicovsky A, Kracht
M, Dittrich-Breiholz O, Spindler V, Waschke J,
Kalthoff H, Trauzold A & Wajant H (2010) Mutant
PIK3CA licenses TRAIL and CD95L to induce nonapoptotic caspase-8-mediated ROCK activation. Cell
Death Differ 17, 1435–1447.
36 Berg D, Lehne M, Muller N, Siegmund D, Munkel S,
Sebald W, Pfizenmaier K & Wajant H (2007) Enforced
covalent trimerization increases the activity of the TNF
ligand family members TRAIL and CD95L. Cell Death
Differ 14, 2021–2034.
FEBS Journal 283 (2016) 323–336 ª 2015 FEBS