Characterisation of FAP-1 expression and CD95 mediated apoptosis in the
A818-6 pancreatic adenocarcinoma differentiation system
Boris J.N. Winterhoffa,b, Alexander Arltc, Angelika Duttmanna, Hendrik Ungefrorend,e, Heiner Sch¨ aferc,
Holger Kalthoffa, Marie-Luise Krusec,n
aInstitute for Experimental Cancer Research, Division Molecular Oncology, University Hospital Schleswig-Holstein Campus Kiel, Germany
bDepartment of Obstetrics and Gynecology, Mayo Clinic, Rochester, Minnesota, USA
cDepartment of General Internal Medicine , Laboratory for Molecular Gastroenterology and Hepatology, University Hospital Schleswig-Holstein, Campus Kiel, Germany
dClinic for Applied Cellular Medicine, University Hospital Schleswig-Holstein, Campus Kiel, Germany
eDepartment of Internal Medicine I, University Hospital Schleswig-Holstein, Campus L¨ ubeck, Germany
a r t i c l e i n f o
Received 18 March 2010
Received in revised form
27 September 2011
Accepted 23 November 2011
Available online 21 December 2011
a b s t r a c t
The present study investigated the expression and localisation of FAP-1 (Fas associated phosphatase-1)
and CD95 in a 3D differentiation model in comparison to 2D monolayers of the pancreatic
adenocarcinoma cell line A818-6. Under non-adherent growth conditions, A818-6 cells differentiate
into 3D highly organised polarised epithelial hollow spheres, resembling duct-like structures. A818-6
cells showed a differentiation–dependent FAP-1 localisation. Cells grown as 2D monolayers revealed
FAP-1 staining in a juxtanuclear cisternal position, as well as localisation in the nucleus. After
differentiation into hollow spheres, FAP-1 was relocated towards the actin cytoskeleton beneath the
outer plasma membrane of polarised cells and no further nuclear localisation was observed. CD95
surface staining was found only in a subset of A818-6 monolayer cells, while differentiated hollow
spheres appeared to express CD95 in all cells of a given sphere. We rarely observed co-localisation of
CD95 and FAP-1 in A818-6 monolayer cells, but strong co-localisation beneath the outer plasma
membrane in polarised cells. Analysis of surface expression by flow cytometry revealed that only a
subset (36%) of monolayer cells showed CD95 surface expression, and after induction of hollow spheres,
CD95 presentation at the outer plasma membrane was reduced to 13% of hollow spheres. Induction of
apoptosis by stimulation with agonistic anti-CD95 antibodies, resulted in increased caspase activity in
both, monolayer cells and hollow spheres. Knock down of FAP-1 mRNA in A818-6 monolayer cells did
not alter resposiveness to CD95 agonistic antibodies. These data suggested that CD95 signal transduc-
tion was not affected by FAP-1 expression in A818-6 monolayer cells. In differentiated 3D hollow
spheres, we found a polarisation-induced co-localisation of CD95 and FAP-1. A tight control of receptor
surface representation and signalling induced apoptosis ensures controlled removal of individual cells
instead of a ‘‘snowball effect’’ of apoptotic events.
& 2011 International Society of Differentiation. Published by Elsevier B.V. All rights reserved.
Pancreatic adenocarcinoma is characterised by profound resis-
tance of tumour cells towards apoptosis, induced by cytotoxic
drugs (Arlt et al., 2001; Arlt et al., 2003) or death receptor
activation (Ungefroren et al., 1998; Trauzold et al., 2001). We
have previously described a possible mechanism for evasion of
CD95 induced cell death in Panc89 pancreatic adenocarcinoma
cells, involving FAP-1 (Fas-associated phosphatase-1) as a med-
iator of resistance (Ungefroren et al., 2001). FAP-1 (Sato et al.,
1995) is also known as PTP-Bas, (phosphotyrosine phosphatase
basophil) (Maekawa et al., 1994), hPTP1E (Banville et al., 1994),
PTPL1 (Saras et al., 1994), and the mouse homologue was named
PTP-BL (phosphotyrosine phosphatase-basophil like) (Hendriks
et al., 1995). FAP-1 is a multidomain protein consisting of a
domains known as FERM (band four/ezrin/radixin/moesin), con-
ferring association with the actin cytoskeleton and 5 PDZ (PSD90/
drosophila large disc/zonula occludens) domains. PDZ domains are
known for their interaction with free C-termini of other proteins in
a sequence dependent manner (van Ham and Hendriks, 2003).
While the substrate specificity of FAP-1 is not very well defined,
more is known about its various interaction partners, e.g. CD95,
APC (Erdmann et al., 2000) and IKBa (Maekawa et al., 1999). FAP-1
Contents lists available at SciVerse ScienceDirect
journal homepage: www.elsevier.com/locate/diff
0301-4681/$-see front matter & 2011 International Society of Differentiation. Published by Elsevier B.V. All rights reserved.
Join the International Society for Differentiation (www.isdifferentiation.org)
nCorresponding author. Tel.: þ49 431 597 1491; fax: þ49 431 597 1302.
E-mail address: email@example.com (M.-L. Kruse).
Differentiation 83 (2012) 148–157
has been localised to the submembranar cytoskeleton in polarised
epithelial cells, to the Golgi complex and the nucleus. Its multi-
domain structure equips FAP-1 with features of a scaffolding
protein, directing its interaction partners to specific subcellular
locations (Erdmann, 2003; Abaan and Toresky, 2008). FAP-1
overexpression has been observed in various malignancies, i.e.
cancer of the pancreas (Ungefroren et al., 1998), ovaries
(Meinhold-Heerlein et al., 2001), head and neck (Wieckowski
et al., 2007), and has been implicated in resistance of tumour cells
towards CD95 induced apoptosis. It has been shown that a
tripeptide representing the C-terminal three amino acids in
human CD95 disrupted binding of CD95 to FAP-1 (Yanagisawa
et al., 1997). Injection of this tripeptide into pancreatic adenocar-
cinoma cells rendered these cells sensitive to CD95 induced
apoptosis, providing strong evidence for CD95/FAP-1 interactions
in resistance towards CD95 induced cell death (Ungefroren et al.,
2001). This activity appears to be restricted to the human system,
as the mouse homologue PTB-BL does not bind CD95 (Cuppen
et al., 1997). The death receptor CD95 is barely expressed in the
human pancreas, but is upregulated upon malignant transforma-
tion or during inflammation (Hasel et al., 2001). Most pancreatic
adenocarcinoma cell lines express CD95 at a high level, but are
resistant towards CD95 induced apoptosis (Ungefroren et al.,
1998; von Bernstorff et al., 1999) and the interaction of CD95
with FAP-1 provides a mechanistic explanation for resistance
(Ungefroren et al., 2001). In our reference cell line Panc89,
we found strong co-localisation of FAP-1 and CD95 in the
Golgi complex, which was even more enhanced upon stimulation
with agonistic anti-CD95 antibodies. Treatment of Panc89 cells
with brefeldin A, a drug that disrupts the Golgi complex, interfered
with FAP-1 localisation, led to enhanced surface representation of
CD95 and sensitised cells for CD95 induced cell death. Over-
expression of FAP-1 rendered a CD95 sensitive cell line (Capan-
1) resistant towards CD95 induced cell death (Ungefroren et al.,
2001). These findings were supported by data from expression
studies of FAP-1 and Fas-GFP fusion proteins in monolayer culture,
revealing interference with CD95 trafficking between intracellular
pools and the plasma membrane (Ivanov et al., 2003). Here
we report analogous experiments in a more complex model of
pancreatic carcinoma using A818-6 cells. This cell line differenti-
ates into three-dimensional (3D) polarised epithelial hollow
spheres, resembling duct-like structures upon prevention of sub-
strate adherence (Lehnert et al., 2001). In vitro differentiation
systems have gained interest lately as differentiation itself
involves apoptosis (Debnath et al., 2002) and appears to render
cell less sensitive to apoptosis by external stimuli (Weaver et al.,
2002). This has been shown in systems of human mammary
epithelial cells reorganising into polarised acini after growth in
substrate containing basement membrane proteins. Several stu-
dies have compared biologic characteristics of cell culture models
in 2D and 3D and shown significant differences when comparing
gene expression levels in melanoma cells (Ghosh et al., 2005) or
differences in protein expression in endometrial and ovarian
cancer cell lines (Grun et al., 2009). In this study we investigated
FAP-1 and CD95 localisation and CD95 induced apoptosis in
standard 2D adherent culture and after 3D differentiation.
2. Materials and methods
All chemicals were of analytical grade purity and purchased
from Sigma (Deisenhofen, Germany) Biomol (Hamburg, Germany),
or Merck (Darmstadt, Germany). All tissue culture reagents were
from Life Technologies (Karlsruhe, Germany), unless otherwise
stated. Secondary antibodies were donkey-anti-mouse and donkey
anti-rabbit coupled with AlexaFluor 488 and donkey anti-goat
coupled with AlexaFluor 546 or AlexaFlour 488 (all from Molecular
Probes, Eugene, OR), TRITC (tetramethyl rhodamine isothiocya-
nate)-coupled donkey anti-mouse, and donkey-anti-rabbit (Jackson
ImmunoResearch Laboratories, West Grove, PA). Horse radish
peroxidase coupled anti-rabbit and anti-mouse and anti-goat
secondary antibodies were from Santa Cruz Biotechnologies, (Santa
Cruz, CA) Primary antibodies used were goat anti-FAP1 (C20),
rabbit anti-CD95 (C20), anti-PARP (all from Santa Cruz), mono-
clonal mouse anti-CD95 (APO-1.3, Alexis Biochemicals, L¨ orrach,
Germany) and CH11 (mouse IgM, Immunotech, Marseille, France),
anti-caspase3 (MBL, Woburn, MA) and mouse anti-b-actin (Sigma),
mouse anti-E-cadherin (BD Biosciences Heidelberg, Germany).
2.2. Cell culture
Pancreatic adenocarcinoma cell lines A818-6 (Lehnert et al.,
2001), and Panc89 (also termed T3M4 (Moore et al., 2001; Sipos
et al., 2003)) were kept in monolayer culture in RPMI-Medium
supplemented with 10% FCS (foetal calf serum, Biochrom, Berlin,
Germany), 1% Glutamax I and 50 mg/ml gentamicin. Cells were
split every 3–6 days and incubated at 37 1C in a humid atmo-
sphere under 5% CO2. For induction of hollow spheres, A818-6 cell
were either transferred to agarose covered plates to inhibit
adherence as described before or submitted to rotating culture
vessels (Synthecon Inc., Houston, TX) at a density of 6?106cells
per 50 ml and rotated at 20 rpm at 37 1C in a humid atmosphere
under 5% CO2.
2.3. Indirect immunofluorescence
plated on glass coverslips in 12-well plates (Costar, Acton, MA).
Samples were fixed in ice cold methanol at ?20 1C for 10 min. For
indirect immunofluorescence staining, fixed cells were treated
with TRIS-buffered saline (20 mM Tris–HCl, 150 mM sodium
chloride, pH 7.6) containing 0.2% glycine and 0.1% BSA (bovine
serum albumin, blocking buffer). Incubation with primary anti-
bodies were carried out overnight at 4 1C in a moist chamber.
Secondary antibodies were incubated at 37 1C for 60 min. Samples
were mounted with slow fade mounting medium (Molecular
Probes, Eugene, OR) and analysed using a Zeiss LSM510 confocal
laser scanning microscope (Carl Zeiss Jena, Jena, Germany).
Hollow spheres were collected using transwell filters with clear
polycarbonate membranes and a pore diameter of 12 mM (Costar),
allowing single cells to pass through the membranes and retain-
ing only spheres larger than 12 mM in diameter. Cells were then
fixed as described. All further steps of washing and incubation
were carried out using transwell filters. Hollow spheres were
mounted with slow fade mounting medium and analysed by
confocal laser scanning microscopy. In order to follow the
sequence of differentiation over time in rotation culture, aliquots
of cells were removed from the culture device at different time
points and collected by centrifugation. Following fixation for
10 min in 2.5% paraformaldehyde in PBS, cells were centrifuged,
re-suspended in 1% BSA in PBS and collected by cytospins on
glass slides, which were air dried before further processing., as
2.4. Knock down experiments
For knock down of protein expression small inhibitory RNA
(siRNA) was employed. A818-6 cells were seeded at a density of
2?105per well into 6-well plates and allowed to settle overnight
in standard culture medium. The following day medium was
B.J.N. Winterhoff et al. / Differentiation 83 (2012) 148–157
replaced by serum free medium and cells were transfected with
500 ng FAP-1 siRNA or an appropriate control (SantaCruz) using
NanofectinTM(PAA, Vienna, Austria) according to the manufac-
turer’s protocol. Cells were harvested at different time points.
Monitoring of transfection efficiency was carried out using
Fluorescein-labelled control and fluorescence microscopy.
2.5. Quantitative real time PCR
RNA isolation was carried out using chaotropic homogenisation
buffers and silica matrix (NucleoSpinTMRNA/Protein, Macherey-
Nagel, Dueren, Germany), according to the manufacturer’s instruc-
tions. 200 ng of total cellular RNA were reverse transcribed using
MaximaTMfirst strand cDNA synthesis kit (Fermentas, Germany).
For qPCR quantification, cDNA was diluted and amplified using
MaximaTMSYBR Green/ROX qPCR maser mix (Fermentas). Primers
for FAP-1 and beta-actin were from RealTimePrimers (Elkins Park,
PA, USA). Quantitative real time PCR was carried out and analysed
using an iCycler and MyiQ single colour qPCR detection system
and software (BioRad, Munich, Germany).
2.6. Western blot analysis
For Western blot analysis, proteins were resolved on 4–20%
gradient gels (Invitrogen) and transferred onto ImmobilonTM-P
membrane by wet blot technique (Millipore, Eschwege, Germany).
Fig. 1. Double immunofluorescence staining for CD95 and FAP-1 in A818-6 cells. A818-6 monolayer cells (A–C) were stained for CD95 (depicted in green) and FAP-1
(depicted in red) under control conditions (A) or after CH11 stimulation for 15 min (B) and 60 min (C). A818-6 hollow spheres were stained for CD95 (D,F, depicted in
green) and FAP-1 (E,F, depicted in red) under control conditions.
B.J.N. Winterhoff et al. / Differentiation 83 (2012) 148–157
Membranes were blocked in 3% dry milk in TBS. Antibodies were
incubated overnight and immune complexes were detected using
horse-radish peroxidase (HRP)-coupled secondary antibodies and
chemiluminescence detection (Super Signal West Dura, Pierce,
Rockford, IL), signals were digitally recorded using the ChemiDoc
system and Quantity One software (BioRad, Munich, Germany).
2.7. Flow cytometry of CD95 expression
For investigation of cell surface expression of CD95 in A818-6
monolayers, cells were detached with trypsin-EDTA for 10 min,
neutralised with FCS containing medium and washed twice with
PBS. Subsequently, cells were fixed with 2.5% paraformaldehyde
in PBS for 30 min at RT. Cells were then washed in TBS and
blocked with blocking buffer (see indirect immunofluorescence
staining). FITC-coupled APO-1.3 antibody was incubated for 1 h at
37 1C and afterwards cells were washed once before FACS analysis
was performed using a DAKO Galaxy system with FloMax analysis
software (DAKO, Denmark). Controls were incubated with FITC-
coupled isotype antibodies. Hollow spheres were collected in
15 ml tubes, and washed once with PBS. All further steps were
carried out as described for monolayer cells.
2.8. Surface biotinylation and CD95 ELISA
Surface biotinylation was carried out using sulfo-LS-NHS-
biotin (Pierce) according to the manufacturer’s suggestions.
Briefly, monolayer cells were detached from the culture substrate
and left to recover in full media for 30 min. at 37 1C. Afterwards,
suspended cells as well as hollow spheres were transferred to
serum free medium and either left untreated or incubated with
500 ng/ml CH11 or 2 mg/ml cytochalasin D for 45 min, Subse-
quently, cells were washed twice with ice cold PBS and then
labelled with 1 mg/ml sulfo-NHS-LC biotin in PBS for 30 min on
ice to inhibit active transport. Labelling was stopped by addition
of 100 mM glycine in PBS, After several washes in PBS cells were
pelleted and solubilised in HEPES-buffer containing 140 mM NaCl
and 0.1% OSG (n-octyl-ß-D thioglucopyranoside). For enzyme-
linked immunosorbent assay, 2 mg/100 ml protein solution was
added to the wells of a streptavidin-coated microtiter plate and
incubated for 2 h at room temperature (RT). Plates were then
washed and blocked with 0.1% BSA in TBS for 1 h and subse-
quently incubated with rabbit anti-FAS C20 at a concentration of
0.2 mg/ml in TBS overnight at 4 1C. Plates were then washed and
incubated with secondary goat anti-rabbit-HRP-coupled antibody
for 1 h at RT and after further washing imunocomplexes were
detected by addition of TMB one reagent (Promega, Madison, WI).
The reaction was stopped after 15 min and OD was evaluated in a
microtiter plate reader.
2.9. DNA fragmentation assay (JAM-[3H]-thymidine-incorporation
JAM assay was performed as described before (Ungefroren
et al., 2001). Briefly, A818-6 monolayer cells were plated in
96-well culture dishes (Nunc, Wiesbaden, Germany) at a density
of 10,000 cells per well. After overnight incubation a total amount
of 7.4 kBq methyl-[3H]-thymidine (Amersham, Braunschweig,
Germany) per well was added for 6 h and then removed. Cells
were then treated with the indicated agents for 24 h. Cells were
harvested using a cell-harvester (Skatron, Lier, Norway) and
incorporated radioactivity was determined by liquid scintillation
counting. DNA fragmentation is expressed as the percentage of
intact DNA in cells treated with the indicated agents compared to
untreated controls. As fragmented DNA does not precipitate as
efficiently as unperturbed DNA, the assay measures the amount of
2.10. Caspase 3 and 7 activity assay
To test for caspase activity, cells were incubated for 24 h with the
indicated agents and then harvested. For the detection of caspase-
3/-7 activity, a homogeneous luminescent assay was performed
Fig. 2. Double immunofluorescence staining for FAP-1 and b-COP in A818-6 cells
(A-D). A818-6 monolayer cells were stained for FAP-1 (depicted in red) and b-COP
(depicted in green) under control conditions (A), after 15 min of BFA treatment (B),
and after 60 min of BFA treatment (C). Secondary antibody control is depicted in D.
(Original magnification 40?) Association of FAP-1 with b-actin (E–H). A818-6
monolayer cells were either left untreated (E) or incubated with cytochalasinD (F).
FAP-1 is depicted in green (E–H), b-actin staining is shown in red (E–H), A818-6
cells from 3D culture were harvested after 7 days and left untreated (G) or
subjected to cytochalasin D treatment (H). (Original magnification 40?).
B.J.N. Winterhoff et al. / Differentiation 83 (2012) 148–157
according to the manufacturer’s instructions (Promega). Activity was
monitored using a MicroLumat Plus. (Berthold Technologies, Bad
Wildbad, Germany). Luminescence signals were normalised to
Analysis of FAP-1 and CD95 expression was first carried out on
cells grown in monolayer (2D) culture. Confocal laser scanning
microscopy of double labelled cells revealed that A818-6 mono-
layer (A818-6M) cells showed a homogenous expression for FAP-1,
but a patchwork pattern of CD95 expression. A818-6M cells
exhibited a perinuclear cisternal pattern for FAP-1 in untreated
controls (Fig. 1A, depicted in red), and after CH11 stimulation no
major change occurred in localisation (Fig. 1B, C). CD95 staining
was found in some cells distinctly at the plasma membrane, while
others were negative for CD95 immunoreactivity or showed scarce
intracellular staining (Fig. 1A–C, depicted in green). After stimula-
tion with CH11 agonistic anti-CD95 antibodies for 15 min and
60 min (Fig. 1B, C) more immunoreactivity was seen for CD95
intracellularly, a finding we have made before after CD95 stimula-
tion of Panc89 cells (Ungefroren et al., 2001). This diffuse staining
throughout the cytoplasm appeared to be due to epitope uncover-
ing, as the time of stimulation was insufficient for de novo protein
synthesis. We did not observe extensive co-localisation of CD95
Fig. 3. CD95 distribution during A818-6 differentiation in 3D culture. A818-6 cells were incubated in rotary cell culture for 4, 7, 9, 11 and 18 days (panels A,B,C,D,E) and
stained for CD95 (green, lower left) and E-cadherin (red, upper left). Depicted are differential interference contrast (DIC, upper right) and overlays of immunofluorescence
and DIC (lower right). Panel F shows secondary antibody controls at day 7. E: overlay of CD95 (green) and E-cadherin (red) staining at day 18 in rotary culture. (Original
B.J.N. Winterhoff et al. / Differentiation 83 (2012) 148–157
and FAP-1 after CH11-stimulation in A818-6 monolayer cells as in
Panc89 cells after CH11 stimulation. When differentiated A818-6
hollow spheres (A818-6HS) were double-labelled for CD95 and
FAP-1, most cells were immunoreactive for CD95. Fig. 1 D-F shows
a section through differentiated hollow spheres of polarised A818-
6 cells. Staining for CD95 (Fig. 1D, F) appeared to be concentrated
beneath the outer plasma membrane. Furthermore, we found FAP-
1 staining (Fig. 1E, F) reorganised from intracellular cisternal
structures into a dense, less structured appearance associated with
the outer plasma membrane of polarised A818-6 cells, and the
nuclei were devoid of FAP-1 staining. Extensive co-localisation of
CD95 with FAP-1, as depicted by yellow colouring (Fig. 1F), was
detected in hollow spheres, most prominently in an area just below
the outer plasma membrane. CD95 appeared to be localised to the
plasma membrane throughout the hollow sphere, a phenomenon
that was further investigated by flow cytometry (see below).
To further characterise the FAP-1 staining pattern in A818-6
cells, monolayer cells were stained for FAP-1 and b-COP, a Golgi
marker protein. Reactivity of FAP-1 localisation towards brefeldin
A (BFA), a drug disrupting the architecture of the Golgi complex,
was analysed first. Double staining of BFA treated cells revealed
differences between FAP-1 localisation and reactivity in A818-6
monolayer cells compared to Panc89 cells. In untreated A818-6
cells (Fig. 2A), similar structures mostly of cisternal appearance in
a perinuclear region were positive for FAP-1 (red) and b-COP
(green), with yellow colouring revealing co-localisation. This
feature was also seen in Panc89 cells. Upon 15 min of BFA
treatment, b-COP staining was dispersed throughout the cell with
disappearance of cisternal structures (Fig. 2B, green), and this
dissociation of b-COP persisted after 60 min of BFA treatment
(Fig. 2C). In contrast, FAP-1 was still associated with elongated
structures after 15 min of BFA treatment (Fig. 2B, red) and
appeared to be located in the nucleus upon BFA treatment after
15 min and more pronounced after 60 min of BFA treatment
(Fig. 2C, red). This feature was not seen in Panc89 cells after
BFA treatment. Instead, FAP-1 re-localised to punctate, vesicular
structures close to the plasma membrane (Ungefroren et al.,
The lack of co-localisation of CD95 and FAP1 after CH11
stimulation and a different distribution of FAP-1 after BFA treat-
ment led us to investigate FAP-1 localisation in more detail. Fig. 2E
shows double immunofluorescence staining for FAP-1depicted in
green, and b-actin depicted in red in A818-6 monolayer cells. FAP-1
was found in the nucleus and in cytoplasmic vesicular structures,
while b-actin was mostly found at the cell periphery. Upon
cytochalasin D treatment, (Fig. 2F), b-actin staining appeared
coalesced into aggregates mostly associated with the plasma mem-
brane, while FAP-1 was more intense in the nucleus and some of the
vesicular staining appeared to be co-localised with b-actin aggre-
gates (see yellow staining in Fig. 2F). Furthermore, FAP-1 showed a
Fig. 4. Analysis of A818-6 monolayer cells and differentiated hollow spheres for CD95 surface representation. A 818-6 monolayer cells and hollow spheres were analysed
for APO-1.3-FITC binding at the cell surface (A) under control conditions and after 15 and 60 min of CH11 stimulation. Data represent APO 1.3-FITC positive cells (Data
represent means plus S.D., n¼3, and were tested for significance by students t-test, p¼0.05). The reduction of surface representation in monolayers after 15 min was not
significant, and 60 min values were not significantly different from controls as well. B: Surface biotinylation of untreated A818-6 monolayer cells and hollow spheres. Cells
surface was labelled by sulfo-NHS-LC biotin and the amount of CD95 biotinylation was detected by ELISA from streptavidin captured total biotin-labelled surface proteins.
The difference of surface CD95 biotinylation was significantly less (2.7-fold) in hollow spheres versus monolayer cells. (Data represent means plus S.D., n¼3, and were
tested for significance by students t-test, p¼0.05). C: Surface biotinylation of CD95 after treatment with CH11 (0.5 mg/ml) surface CD95 biotinylation was not significantly
enhanced after 45 min of CH11 stimulation in monolayer cells , but removal of CD95 from the outer surface in hollow spheres after CH11 stimulation was significant. (Data
represent means plus S.D., n¼3, and were tested for significance by students t-test, p¼0.05).
B.J.N. Winterhoff et al. / Differentiation 83 (2012) 148–157
fine line of staining at the plasma membrane, apparently trapped
there after collapse of the actin cytoskeleton. To investigate the
distribution of FAP-1 and b-actin under non-adherent growth
conditions, A818-6 cells were grown in rotary culture for 7 days,
to allow for differentiation into hollow spheres to begin (see below).
In Fig. 2G, FAP-1 was no longer localised in the nucleus or extended
cisternal structures but dispersed throughout the cytoplasm with
slightly more intense staining next to b-actin reactivity, which is
fully concentrated beneath the outer membrane of the spherical
structure. After cytochalasin D treatment, of 7 days old 3D aggre-
gates, as depicted in Fig. 2H, FAP-1 and b-actin were mostly
dispersed throughout the cells, showing intense co-localisation.
The redistribution of FAP-1 after induction of differentiation by
non-adherent growth further underlined differences between FAP-1
localisation and reactivity towards manipulation in A818-6 mono-
layer cells compared to Panc89 monolayers.
In mature hollow spheres, as depicted in Fig. 1, CD95-staining
was found co-localised with FAP-1 beneath the outer plasma
changes in CD95 immunolocalisation, we analysed the sequence
of formation of hollow spheres through the first 18 days of
rotation culture. Fig. 3 shows indirect immunofluorescence stain-
ing of cells and aggregates from day 4, 7, 9, 11 and 18. E-cadherin
staining was combined with CD95 staining to depict the plasma
membrane and the development of polarisation. After 4 days of
3D culture (Fig. 3A) E-cadherin, depicted in red, was found at the
plasma membrane of small round longitudinally arranged cell
assemblies, and in the cytoplasm of signet-ring-like cells. CD95,
depicted in green, was found intracellular and concentrated at the
cell periphery. After 7 days in rotary culture (Fig. 3B), the
association of small round cells became more prominent and
E-cadherin and CD95 appeared in a layered fashion at the plasma
membrane, with E-cadherin giving a more defined alignment
than CD95 (Fig. 3B, upper left), which appeared to be localised in
multiple small dots mostly apart from E-cadherin (Fig. 3B, lower
left). The overlay of DIC and immune label (Fig. 3B, lower right)
showed that the outline of the cell is not always congruent with
E-cadherin localisation and thus showed CD95 accumulating at a
distance from the plasma membrane, most of it beneath the
plasma membrane. The different layers of staining give an
impression of distances. After 9 days of non-adherent growth
(Fig. 3C), a first impression of basolateral versus apical membrane
domains was apparent, while E-cadherin was preferentially found
at sites of cell-cell-contact, CD95 staining appeared to shift
towards outer, free membrane areas of spherical structures. This
phenomenon was even more pronounced after 11 days of rotation
culture (Fig. 3D) and was obvious after 18 days in differentiating
spherical structures (Fig. 3F) E-cadherin was found at the sites of
cellular interaction and CD95 at the outer membrane space as
well as intracellular.
The findings from indirect immunofluorescence staining were
confirmed using flow cytometry analysis of A818-6 monolayer and
hollow spheres for CD95. Fig. 4A shows the results of APO1.3-FITC
surface staining. In untreated A818-6 monolayer cells 36% of cells
were positive for cell surface-associated CD95 (Fig. 4A). Upon
stimulation with agonistic CH11 antibodies, the surface representa-
tion of CD95 was significantly decreased by about 10% after 15 min.
After prolonged stimulation of monolayer cells surface representa-
tion increased by about 5%. Analysis of CD95 surface expression in
A818-6 hollow spheres showed reduced surface expression of CD95,
compared to monolayer cells, only about 13% of spheres were
positive for CD95 at the outer surface (Fig. 4A). When tested for
reactivity towards CH11 stimulation, surface representation of CD95
rose to higher levels within 15 min and dropped significantly
towards barely detectable levels after 60 min. These data confirmed
access of antibody to CD95 surface molecules in hollow spheres.
Furthermore, CH11 antibodies entered hollow sphere cells presum-
ably by means of receptor-mediated endocytosis (data not shown).
It has to be kept in mind that flow cytometry analysis cannot
differentiate between individual cells within a given hollow sphere,
so any positive signal recorded by the flow cytometer will score
positive the entire hollow sphere. Surface staining by immunofluor-
escence (see Fig. 1D–F) most likely did not represent plasma
Fig. 5. Analysis of induction of apoptosis in A818-6 monolayer cells and hollow
spheres. A: Intact DNA as per cent of control values in A818-6 monolayer cells
after 24 h of treatment with CH11 at concentrations of 1 mg/ml and 0.5 mg/ml,
cycloheximide (CHX) at 1 mg/ml and TNFa at 10 ng/ml and a combination of TNFa
and CHX at given concentrations, (n¼3,
significant versus untreated, all other comparisons were). B: Westernblot analysis
of PARP cleavage in A818-6 monolayer cells after 24 h of treatment as indicated.
Active caspase 3 was detected concommittanly, visible only in those lysates that
also show PARP cleavage. For loading control HSP90 was detected after stripping
of the membrane. C: Induction of caspase activity in A818-8 monolayer and
hollow spheres. Chemiluminescence signals of caspase 3 and 7 activity were
normalised to protein content and expressed as multiples of control values
(n-fold) for n¼4 individual experiments, significance was evaluated by students‘
t-test. No significance with p¼0.05 was found for TNF/CHX in monolayer versus
hollow spheres, all other comparissons met the criterion. (Depicted are means
plus standard deviation.)
nvalues for CHX and TNFa were not
B.J.N. Winterhoff et al. / Differentiation 83 (2012) 148–157
membrane staining, but submembranar trapping of CD95 mole-
cules. To confirm the findings of flow cytometry analysis, a different
biochemical approach, namely surface biotinylation and ELISA-
based quantification of biotinylated CD95 was employed. We found
the labelling of CD95 was 2.7 fold less in hollow spheres compared
to monolayer cells (Fig. 4B) which resembled the relations found by
flow cytometry (36% versus 13%). CH11 stimulation was carried out
for 45 min, adding approximately 15 min of washing in ice cold PBS
to stop the reaction and active cellular transport, stimulation values
were comparable to 60 min stimulation in FACS experiments.
A 15 min stimulation period appeared technically not sound accord-
ing to the aforementioned procedure. Fig. 4C shows that monolayer
cells did not show significant changes in surface representation, while
in hollow spheres surface CD95 biotinylation significantly decreased
by 17%, supporting the tendency found by flow cytometry.
Analysis of CD95 induced apoptosis was carried out in A818-6
monolayer cells by assessment of DNA fragmentation. The results
of the JAM assay are depicted in Fig. 5A. Cells were treated for
24 h with either 1 mg/ml or 0.5 mg/ml CH11, 1 mg/ml CHX and
10 ng/ml TNFa separately, as well as TNFa and CHX as a positive
control for apoptosis. CH11-treatment induced a significant
reduction of intact DNA by roughly 35% compared to untreated
controls, CHX and TNFa as single agents were uneffective, while
TNFa/CHX combined treatment induced roughly 50% of DNA
fragmentation. These findings were in accordance with 36% of
cells bearing CD95 receptors. Analysis of PARP cleavage showed a
caspase-3 dependent degradation as seen in Fig. 5B, also, when
probed for caspase 3, the blot showed activated caspase 3 only in
those cells which also gave a PARP cleavage signal. Analysis of
DNA fragmentation requires DNA labelling by tritiated thymidine.
Since mature A818-6 hollow spheres do not contain dividing cells,
they cannot be efficiently labelled, hence the JAM assay was not
suitable for assessment of DNA fragmentation in hollow spheres.
Instead we employed a chemiluminescent caspase activity assay
which measures activity of caspase 3 and 7. The results are shown
in Fig. 5C. CH11 treatment significantly induced caspase activity
in monolayer cells and hollow spheres 2.9- and 1.5-fold, respec-
tively, again reflecting differences in CD95 surface representation.
Combined TNFa/CHX treatment for 24 h raised caspase activity
levels 6.8- and 7.7-fold in monolayer and hollow spheres, respec-
tively. As the caspase activity assay is not exclusive for caspase 3,
we performed western blot analysis for activated caspase 3. Caspase
3 activation is induced by CD95 stimulation in A818-6 monolayer
cells, as seen in Figure in 5B by 0.5 mg/ml and 0.25 mg/ml CH11, as
well as in hollow spheres. (Data are not shown.)
For further investigation of the influence of FAP-1 on CD95-
mediated apoptosis in A818-6 cells we employed small inhibitory
RNA to knock down FAP-1 mRNA, hence protein expression. These
experiments require transient transfection of cells and, unfortu-
nately, hollow spheres are resistant to transient transfection, so
we performed these analyses in A818-6 monolayer cells. Fig. 6A
shows the result of initial experiments using 500 ng of FAP-1
siRNA for transfection over a time course of 24 h–72 h, compared
to a control oligonucleotide as recommended by the manufac-
turer. Western blot analysis revealed a reactve band at roughly
Caspase 3/7 Activity n-fold
siR 24h scr 48h siR 48h scr 72h siR 72h
Fig. 6. Inhibition of FAP-1 expression in transiently transfected A818-6 monolayer
cells. A: Effect of FAP-1 specific siRNA (siR) on FAP-1 protein expression 24, 48 and
72 h after transfection, compared to control (scr) transfections. Western blot
analysis revealed two reactive bands the lower of roughly 250 kDa. After 24 h the
lower band is slightly less than in the corresponding control, while the upper band
is almost gone. After 48 h of transfection FAP-1 appeared strongly reduced as
compared to control, while after 72 h FAP-1 protein expression re-appeared.
HSP90 loading controls revealed those effect were not due to imbalanced amounts
of protein. B: Results of qPCR analysis for FAP-1 mRNA in transiently transfected
A818-6 cells. FAP-1 amplification Ct values were normalised to beta-actin Ct
values (dCt) and subsequently expresssed as fractions of controls, which were set
as 1.0. QPCR analysis revealed that after 24h the effect of FAP-1 siRNA on FAP-1
mRNA was most pronounced and subsequently mRNA levels started to rise again.
C: Effect of FAP-1 siRNA treatment on CD95 induced apoptosis. All data were
normalised to control siRNA (con scr) values, set as 1.0. Control transfected cells
(hatched) showed a significant rise in caspase 3/7 activity after CH11 and TNFa/
cycloheximide (TNF/CHX) treatment. FAP-1 specific siRNA (siR, black) induced
capase 3/7 acivity in untreated controls, further increased levels of activity were
seen after CH11 and TNF/CHX treatment. CH11 stimulation and TNF/CHX treat-
ment were not significantly (#) different between controls and FAP-1 specific
siRNA transfected cells (n¼3).
B.J.N. Winterhoff et al. / Differentiation 83 (2012) 148–157
250 kDa and a higher molecular weight band. Both were dimin-
ished after 24 h and 48 h of transfection, but reappeared after
72 h. HSP90 was tested for loading control. We then performed
RNA analysis by means of quantitative real time PCR (qPCR) to
investigate effects at RNA level. Fig. 6B shows copy numbers
normalised to b-actin mRNA levels expressed as 2?dCtvalues, as
fractions of controls. Levels of mRNA were lowest at 24 h of
transfection and then started to rise after 48 h, but still remained
half as much as control values after 72 h. According to these
analyses cells were transfected with siRNA or controls for 24 h
and then treated with CD95 agonistic antibody CH11 or TNFa and
cycloheximide as before and after 24 h, that is 48 h of transfec-
tion, cells were harvested and subjected to caspase 3/7 activity
assays, as before. We considered this to be the time frame of
highest inhibition of FAP-1 protein expression. Fig. 6C shows the
results of apoptosis induction in A818-6 FAP-1 siRNA transfected
cells. All values were normalised to untreated control transfected
cells. Compared to these, even FAP-1 siRNA transfected untreated
cells showed a significantly higher inherent rate of caspase 3/7
activity though activity levels were generally lower than in
untransfected cells. CH11 and TNFa/CHX treatment significantly
raised caspase 3/7 acitvity versus control. Specific CD95 stimula-
tion with agonistic CH11 antibody, did not differ between control
transfected and FAP-1 siRNA transfected cells, thus knock down of
FAP-1 protein expression did not alter CH11-induced apoptosis.
The three-dimensional A818-6 differentiation system was
investigated for FAP1 expression and its possible role in CD95
induced apoptosis. All investigations were based on minimal
manipulation, as the differentiation procedure is refractory to
transient overexpression or knock down of proteins. The aim of
the study was to learn more about the effect of differentiation on
FAP-1-related resistance towards CD95 induced apoptosis. The
reference system for this investigation were Panc89 cells, in
which FAP-1 partially mediates resistance against CD95 induced
cell death. Panc89 cells grow as monolayers, these cells showed a
close spatial–temporal relationship between FAP-1 and CD95,
suggesting a blockade of CD95 trafficking from the Golgi complex
by FAP-1 as the underlying mechanism (Ungefroren et al., 2001;
Ivanov et al., 2006). A818-6 cells were grown in 2D monolayer
culture under standard conditions, for induction of 3D differen-
tiation, the cells were detached and grown on agarose or in
rotation culture to avoid adherence to tissue culture plastics.
Non-adherence (lack of integrin stimulation) is the crucial
factor for formation of 3D polarised epithelial hollow spheres
resembling duct-like structures (Lehnert et al., 2001). In the
undifferentiated state, when cells are growing as 2D monolayer
culture, FAP-1 expression was monitored by immunofluores-
cence. In untreated cells, we observed a juxtanuclear, cisternal
localisation of FAP-1 resembling the Golgi complex, which was
quite similar to the situation in Panc89 cells, but we also found
nuclear staining for FAP-1 in A818-6 monolayer cells. Closer
investigation using BFA, a drug interfering with Golgi dynamics,
revealed differences between A818-6 and Panc89 cells. In contrast
to Panc89 cells, FAP-1 in A818-6 cells was not distributed into
small vesicular structures positioned next to the plasma mem-
brane, but was found prominently in the nucleus and showed a
delayed dispersion from cisternal structures. These data provided
the first hint for fundamental differences between Panc89 cells
and A818-6 regarding FAP-1 function. Upon investigation of CD95
expression we noted another fundamental difference, namely
a lack of co-localisation between FAP-1 and CD95 even under
conditions, when co-localisation of these molecules was maximal
in Panc89 cells. Moreover, CD95 was expressed in all Panc89 cells
of a given population, but only a subset of A818-6 cells, amount-
ing to about 36%, revealed cell surface expression of CD95. The
analysis of differentiated epithelial hollow spheres derived from
A818-6 cells revealed further differences in FAP-1 localisation.
Upon induction of differentiation under non-adherent growth
conditions, the cytoskeleton of A818-6 cells is reorganised. Early
redistribution phenomena are the recruitment of E-cadherin
towards the plasma membrane, aud subsequently to the basolat-
eral aspect of the plasma membrane to sites of cell-cell-contact.
Upon formation of hollow spheres, the actin cytoskeleton resem-
bles that of polarised epithelia with a dense layer of actin found
beneath the outer plasma membrane of hollow spheres (compare
Fig. 2C). This is precisely the location of FAP-1 staining in
differentiated A818-6 hollow spheres. Upon differentiation, FAP-
1 relocates to the submembranar actin network in polarised
epithelial cells. This finding is most likely due to a functional
FERM domain in FAP-1 (Cuppen et al., 1999). More surprising was
the finding, that CD95 appeared to be expressed at higher levels
in almost every cell of a given hollow sphere, moreover, there was
a strong co-localisation between FAP-1 and CD95 in polarised
cells, located beneath the outer plasma membrane under control
conditions. FACS analysis revealed that the exposure of the
extracellular domain of CD95 at the outer plasma membrane
was scarce, with only 13% of hollow spheres being positive for
CD95 in the outer membrane. Compared to the findings of
confocal microscopy, these data suggest that CD95 is indeed
trapped in structures beneath the outer plasma membrane, but
is readily mobilised upon stimulation. The plasma membrane
representation increased during incubation with agonistic anti-
bodies and was followed by removal from the plasma membrane
within 60 min. These data suggest that FAP-1 expression, which is
found in all A818-6 cells, is functionally unrelated to CD95
expression in A818-6 monolayer cells with respect to an inhibi-
tory role in CD95 induced cell death. After differentiation FAP-1
resumed a normal localisation, as suggested by the presence of a
FERM domain, and co-localised with CD95 as has been shown
before. In A818-6 epithelial hollow spheres though, this co-
localisation appears to indicate an intracellular, readily available
reservoir for CD95, where FAP-1 acts as a scaffolding protein
(Ghiran et al., 2008). In differentiated A818-6 hollow spheres,
CD95 surface presentation is tightly controlled, but we found no
interference with receptor signalling. The analysis of CD95
induced apoptosis appears to support these data. Overall, A818-
6 cells, taken as a homogenous population, could be considered
resistant towards CD95 induced cell death, since 35% of DNA
fragmentation within 24 h of treatment in monolayer cells at a
concentration of 0.5 mg/ml CH11 was low compared to Capan-1, a
sensitive cell line. However, taken into account that only 36% of
A818-6 cells exhibit surface expression of CD95 (meaning that
64% are a priori unresponsive), a reduction of intact DNA by 35%
matches the amount of surface receptor presenting cells. PARP
analysis revealed that, indeed, A818-6 monolayer cells have
activated caspase-3. Furthermore, knock down of FAP-1 mRNA
by means of small inhibitory RNA (siRNA) did not alter sensitivity
towards CH11 induced apoptosis. If FAP-1 were involved in CD95
mediated apoptosis in A818-6 cells, inhibition of FAP-1 protein
expression should have a sensitising effect. This was not the case.
The analysis of FAP-1 in A818-6 cells revealed yet another aspect
of regulation of FAP-1 activity in pancreatic cancer cells. While
Capan-1 lacked FAP-1 expression, Panc89 cells ‘‘utilised’’ FAP-1 to
interfere with CD95 induced cell death by disruption of CD95
trafficking. Yet others, e.g. A818-6 cells appear to express FAP-1,
but do not show interference with CD95 trafficking. The 3D
differentiation model suggests that in polarised epithelial cells,
FAP-1 functions to provide a scaffold for a readily available
B.J.N. Winterhoff et al. / Differentiation 83 (2012) 148–157
reservoir of CD95 rather than acting to prevent CD95 from
reaching the plasma membrane. Data from the A818-6 differen-
tiation system provide evidence that apoptosis needs to be
investigated insystemsof higher
Weinberg, 2002). Studying three-dimensional growth and differ-
entiation in mammary epithelia (Debnath et al., 2002; Weaver
et al., 2002) benign and malignant pancreatic duct epithelial cells
(Gutierrez-Barrera et al., 2007) revealed an important role of
polarisation on cell survival as well as early carcinogenesis.
Actually, our data presented here on the role of FAP-1 in CD95
induced apoptosis and FAP-1 suggest a one-to-one action, meaning
that only those cells are condemned to die which carry the receptor
at the cell surface. The differentiation-dependent mechanism of
CD95 presentation at the cell surface points towards differential
effects of cell structure and cellular interactions in complex 3D
systems on apoptosis regulation. Results presented here indicate
that CD95 induced apoptosis is a process that is linked to
individual cells, which is mandatory for selective removal of cells
in a living organism.
This work was supported by the Deutsche Forschungsge-
meinschaft (SFB 415/A3 and EG 3/8/2-1).
Arlt, A., Vorndamm, J., Breitenbroich, M., Folsch, U.R., Kalthoff, H., Schmidt, W.E.,
Schafer, H., 2001. Inhibition of NF-kappaB sensitises human pancreatic
carcinoma cells to apoptosis induced by etoposide (VP16) or doxorubicin.
Oncogene 20, 859–868.
Arlt, A., Gehrz, A., Muerkoster, S., Vorndamm, J., Kruse, M.L., Folsch, U.R., Schafer, H.,
2003. Role of NF-kappaB and Akt/PI3K in the resistance of pancreatic carcinoma
cell lines against gemcitabine-induced cell death. Oncogene 22, 3243–3251.
Abaan, O.D., Toresky, J.A., 2008. PTPL1: a large phosphatase with a split person-
ality. Cancer and Metastasis Reviews 27 (2), 205–214.
Banville, D., Ahmad, S., Stocco, R., Shen, S.H., 1994. A novel protein-tyrosine
phosphatase with homology to both the cytoskeletal proteins of the band
4.1 family and junction-associated guanylate kinases. Journal of Biological
Chemistry 269, 22320–22327.
Cuppen, E., Nagata, S., Wieringa, B., Hendriks, W., 1997. No evidence for involve-
ment of mouse protein-tyrosine phosphatase-BAS-like Fas-associated phos-
phatase-1 in Fas-mediated apoptosis. Journal of Biological Chemistry 272,
Cuppen, E., Wijers, M., Schepens, J., Fransen, J., Wieringa, B., Hendriks, W., 1999.
A FERM domain governs apical confinement of PTP-BL in epithelial cells.
Journal of Cell Science 112, 3299–3308.
Debnath, J., Mills, K.R., Collins, N.L., Reginato, M.J., Muthuswamy, S.K., Brugge, J.S.,
2002. The role of apoptosis in creating and maintaining luminal space within
normal and oncogene-expressing mammary acini. Cell 111, 29–40.
Erdmann, K.S., Kuhlmann, J., Lessmann, V., Herrmann, L., Eulenburg, V., Muller, O.,
Heumann, R., 2000. The Adenomatous Polyposis Coli-protein (APC) interacts
with the protein tyrosine phosphatase PTP-BL via an alternatively spliced PDZ
domain. Oncogene 19, 3894–3901.
Erdmann, K.S., 2003. The protein tyrosine phosphatase PTP-Basophil/Basophil-like.
Interacting proteins and molecular functions. European Journal of Biochem-
istry 270, 4789–4798.
Ghosh, S., Spagnoli, G.C., Martin, I., Ploegert, S., Demougin, P., Heberer, M.,
Reschner, A., 2005. Three-dimensional culture of melanoma cells profoundly
affects gene expression profile: a high density oligonucleotide array study.
Journal of Cellular Physiology 204, 522–531.
Grun, B., Benjamin, E., Sinclair, J., Timms, J.F., Jacobs, I.J., Gayther, S.A., Dafou, D.,
2009. Three-dimensional in vitro cell biology models of ovarian and endome-
trial cancer. Cell Proliferation 42, 219–228.
Ghiran, I., Glodeck, A.M., Weaver, G., Klickstein, L.B., Nicholson-Weller, A., 2008.
Ligation of erythrocyte CR1 induces clustering in complex with scaffolding
protein FAP1. Blood 112, 3465–3473.
Gutierrez-Barrera, A.M., Menter, D.G., Abbruzzese, J.L., Reddy, S.A., 2007. Establish-
ment of three-dimensional cultures of human pancreatic duct epithelial cells.
Biochemical and Biophysical Research Communications 358, 698–703.
Hendriks, W., Schepens, J., Bachner, D., Rijss, J., Zeeuwen, P., Zechner, U.,
Hameister, H., Wieringa, B., 1995. Molecular cloning of a mouse epithelial
protein-tyrosine phosphatase with similarities to submembranous proteins.
Journal of Cellular Biochemistry 59, 418–430.
Hasel, C., Rau, B., Perner, S., Strater, J., Moller, P., 2001. Differential and mutually
exclusive expression of CD95 and CD95 ligand in epithelia of normal pancreas
and chronic pancreatitis. Laboratory Investigation 81, 317–326.
Ivanov, V.N., Lopez Bergami, P., Maulit, G., Sato, T.A., Sassoon, D., Ronai, Z., 2003.
FAP-1 association with Fas (Apo-1) inhibits Fas expression on the cell surface.
Molecular Cell Biology 23, 3623–3635.
Ivanov, N.V., Ronai, Z., Hei, T.K., 2006. Opposite roles of Fap-1 and dynamin in the
regulation of Fas (CD95) Translocation to the cell surface and susceptibility to Fas
ligand – mediated apoptosis. Journal of Biological Chemistry 281, 1840–1852.
Jacks, T., Weinberg, R.A., 2002. Taking the study of cancer cell survival to a new
dimension. Cell 111, 923–925.
Lehnert, L., Lerch, M.M., Hirai, Y., Kruse, M.L., Schmiegel, W., Kalthoff, H., 2001.
Autocrine stimulation of human pancreatic duct-like development by soluble
isoforms of epimorphin in vitro. Journal of Cell Biology 152, 911–922.
Maekawa, K., Imagawa, N., Nagamatsu, M., Harada, S., 1994. Molecular cloning of a
novel protein-tyrosine phosphatase containing a membrane-binding domain
and GLGF repeats. FEBS Letters 337, 200–206.
Maekawa, K., Imagawa, N., Naito, A., Harada, S., Yoshie, O., Takagi, S., 1999.
Association of protein-tyrosine phosphatase PTP-BAS with the transcription-
factor-inhibitory protein IkappaBalpha through interaction between the PDZ1
domain and ankyrin repeats. Biochemical Journal 337, 179–184.
Meinhold-Heerlein, I., Stenner-Liewen, F., Liewen, H., Kitada, S., Krajewska, M.,
Krajewski, S., Zapata, J.M., Monks, A., Scudiero, D.A., Bauknecht, T., Reed, J.C.,
2001. Expression and potential role of Fas-associated phosphatase-1 in
ovarian cancer. American Journal of Pathology 158, 1335–1344.
Moore, P.S., Orlandini, S., Zamboni, G., Capelli, P., Rigaud, G., Falconi, M., Bassi, C.,
Lemoine, N.R., Scarpa, A., 2001. Pancreatic tumours: molecular pathways
implicated in ductal cancer are involved in ampullary but not in exocrine
nonductal or endocrine tumorigenesis. British Journal of Cancer 84, 523–562.
Sato, T., Irie, S., Kitada, S., Reed, J.C., 1995. FAP-1: a protein tyrosine phosphatase
that associates with Fas. Science 268, 411–415.
Saras, J., Claesson-Welsh, L., Heldin, C.H., Gonez, L.J., 1994. Cloning and character-
isation of PTPL1, a protein tyrosine phosphatase with similarities to cytoske-
letal-associated proteins. Journal of Biological Chemistry 269, 24082–24089.
Sipos, B., Moser, S., Kalthoff, H., Torok, V., Lohr, M., Kloppel, G., 2003.
A comprehensive characterisation of pancreatic ductal carcinoma cell lines:
towards the establishment of an in vitro research platform. Virchows Archiv.
Trauzold, A., Wermann, H., Arlt, A., Schutze, S., Schafer, H., Oestern, S., Roder, C.,
Ungefroren, H., Lampe, E., Heinrich, M., Walczak, H., Kalthoff, H., 2001. CD95
and TRAIL receptor-mediated activation of protein kinase C and NF-kappaB
contributes to apoptosis resistance in ductal pancreatic adenocarcinoma cells.
Oncogene 20, 4258–4269.
Ungefroren, H., Voss, M., Jansen, M., Roeder, C., Henne-Bruns, D., Kremer, B.,
Kalthoff, H., 1998. Human pancreatic adenocarcinomas express Fas and Fas
ligand yet are resistant to Fas-mediated apoptosis. Cancer Research 58,
Ungefroren, H., Kruse, M.L., Trauzold, A., Roeschmann, S., Roeder, C., Arlt, A.,
Henne-Bruns, D., Kalthoff, H., 2001. FAP-1 in pancreatic cancer cells: func-
tional and mechanistic studies on its inhibitory role in CD95-mediated
apoptosis. Journal of Cell Science 114, 2735–2746.
Wieckowski, E., Atarashi, Y., Stanson, J., Sato, T.A., Whiteside, T.L., 2007. FAP-1-
mediated activation of NF-kappaB induces resistance of head and neck cancer
to Fas-induced apoptosis. Journal of Cellular Biochemistry 100 (1), 16–28.
Weaver, V.M., Lelievre, S., Lakins, J.N., Chrenek, M.A., Jones, J.C., Giancotti, F., Werb, Z.,
Bissell, M.J., 2002. beta4 integrin-dependent formation of polarised three-dimen-
sional architecture confers resistance to apoptosis in normal and malignant
mammary epithelium. Cancer Cell 2, 205–216.
Yanagisawa, J., Takahashi, M., Kanki, H., Yano-Yanagisawa, H., Tazunoki, T., Sawa, E.,
Nishitoba, T., Kamishohara, M., Kobayashi, E., Kataoka, S., Sato, T., 1997. The
molecular interaction of Fas and FAP-1. A tripeptide blocker of human Fas
interaction with FAP-1 promotes Fas-induced apoptosis. Journal of Biological
Chemistry 272, 8539–8545.
van Ham, M., Hendriks, W., 2003. PDZ domains-glue and guide. Molecular Biology
Reports 30, 69–82.
von Bernstorff, W., Spanjaard, R.A., Chan, A.K., Lockhart, D.C., Sadanaga, N., Wood, I.,
Peiper, M., Goedegebuure, P.S., Eberlein, T.J., 1999. Pancreatic cancer cells can
evade immune surveillance via nonfunctional Fas (APO-1/CD95) receptors and
aberrant expression of functional Fas ligand. Surgery 125, 73–84.
B.J.N. Winterhoff et al. / Differentiation 83 (2012) 148–157