J. Biomedical Science and Engineering, 2009, 2, 216-226
doi: 10.4236/jbise.2009.24035 Published Online August 2009 (http://www.SciRP.org/journal/jbise/
JBiSE
).
Published Online August 2009 in SciRes. http://www.scirp.org/journal/jbise
Effects of extracellular matrix proteins on expansion,
proliferation and insulin-producing-cell differentiation of
ARIP cells
Gary G. Adams1, Yu-Xin Cui2
1Insulin Diabetes Experimental Research Group (IDER), Faculty of Medicine and Health Science, University of Nottingham, Clifton
Boulevard, Nottingham, NG7 2UH UK; 2Department of Physiology, Development and Neuroscience, University of Cambridge,
Downing Street, Cambridge, CB2 3EG, UK.
Email: Gary.Adams@nottingham.ac.uk
Received 17 April 2009; revised 25 April 2009; accepted 30 April 2009.
ABSTRACT
Regeneration of transplantable pancreatic islet
cells has been considered to be a promising
alternative therapy for type 1 diabetes. Re-
search has suggested that adult pancreatic
stem and progenitor cells can be derived into
insulin-producing cells or cultivated islet-like
clusters given appropriate stimulating condi-
tions. In this study we explored the effect of
selective extracellular matrix (ECM) proteins on
the potential of insulin-producing cell differen-
tiation using ARIP cells, an adult rat pancreatic
ductal epithelial cell line, as a model in vitro.
Quantitative single cell morphology analysis
indicated that all the four ECM proteins we
have used (type I collagen, laminin, fibronectin
and vitronectin) increased the single cell area
and diameter of ARIP cells. In addition, se-
rum-free cell cultivation was dependent on cell
density and particular components; and serum
could be replaced when systematic optimisa-
tion could be performed. Surface treated with
laminin was shown to be able to enhance
overall cell expansion in the presence of de-
fined serum-free medium conditions. Collagen
treated surfaces enhanced insulin production
in the presence of GLP-1 although the insulin
gene expression was however weak accord-
ingly. Our results suggest that selective ECM
proteins have effects on single cell morphol-
ogy, adhesion and proliferation of ARIP cells.
These ECM molecules however do not have a
potent effect on the insulin-producing cell dif-
ferentiation potential of ARIP cells even com-
bining with GLP-1.
Keywords: Extracellular Matrix; Proliferation; Dif-
ferentiation; ARIP Cells; Incretin GLP-1
1. INTRODUCTION
The pluripotent cells that develop into pancreatic β-cells
are initially derived from pancreatic ductal epithelium.
Ductal cell proliferation is subsequently followed by the
budding of endocrine cells but little is known about the
intrinsic and extrinsic factors which surround different-
iation.
Although a number of studies have been carried out to
discover the factors that may be responsible for β-cell pro-
liferation and differentiation, no conclusive evidence is
forthcoming [2,34,35,40,45]. Some of the known con-
tributory factors reported to be involved in β-cell acquisi-
tion in vitro, include islet neogenesis- associated protein
(INGAP), nicotinamide, retinoic acid, glycogen
like-peptide (GLP-1) and the pancreatic regenerating gene
(Reg) [6, 8,25, 29,30,33,41,43].
ARIP cells, an adult rat pancreatic ductal epithelial
cell line, are derived from an azaserine induced rat non-
tumourigenic pancreatic carcinoma [17]. ARIP cells
express ductal cell markers and carbonic anhydrase un-
der basal cultivation conditions. GLP-1, a peptide which
is originally produced from small intestine, has been
demonstrated to induce the insulin-producing cell dif-
ferentiation of ARIP cells; and this effect may be ac-
companied by the overexpression of PDX-1, a transcrip-
tion factor which is crucial for pancreatic embryogenesis
and insulin secretion [15]. Thus, it can be speculated that
ARIP cell is a useful in vitro model with which to study
the mechanism of the insulin-producing cell regenera-
tion/neogenesis mediated by corresponding stimuli. On
the other hand, cell-ECM and cell-cell interaction is be-
lieved an important cellular process in regulating cellular
proliferation and differentiation. Particularly, extracellu-
lar matrix (ECM) and soluble factors may be involved in
the facilitation or induction of cell differentiation, al-
though the exact signalling pathway is not well known
[17,19,26]. For instance, a previous study has proved
that omitting serum is essential for insulin-producing
G. G. Adams et al. / J. Biomedical Science and Engineering 2 (2009) 216-226 217
SciRes Copyright © 2009 JBiSE
cell differentiation, as medium containing serum is
found to decrease insulin production and inhibit cyst
development, whilst, a Matrigel™ overlay procedure is
found to be vital to encourage the formation of
three-dimensional structure from primarily isolated pan-
creatic ductal cells [13]. It is suggested that ECM adhe-
sion molecules reorganize cytoskeletal architecture and
mediate expression of genes responsible for cell differ-
entiation regulation by binding with specific cell surface
integrins, and triggering certain downstream signal
transduction, [9,20]. Within the pancreas, integrin dis-
tribution is distinctive on the surfaces of different pan-
creatic cell types. Thus, by manipulating the interaction
of ECM integrin, it may provide a promising approach to
inducing insulin-producing cell differentiation. Laminin
is found, for example, to promote β-cell differentiation
during cell cultivation [18].
We hypothesised, therefore, that in the presence of
certain serum-free media conditions and ECM proteins,
ARIP cells could provide an appropriate model for pan-
creatic insulin-producing differentiation. Herein, we
report the use of ARIP cells to explore the possibility of
insulin-producing cell acquisition in vitro. Single cell
cytogeometry analysis was carried out to find out
whether major ECM proteins are capable of modulating
ARIP cell morphology in the first instance. Subsequently,
ATP-luciferase based assay was performed to see if cell
proliferation is dependent on seeding density and par-
ticular serum-free media components, when compared
with medium containing foetal calf serum (FCS). Sur-
face coverage of ARIP cells was estimated to determine
the potential effect of ECM-protein-coated surfaces in
the presence of defined serum-free medium. ARIP cells
were eventually encouraged to differentiate into insu-
lin-producing cells given ECM proteins, GLP-1 and se-
rum-free medium conditions.
2. MATERIALS AND METHOD
2.1. Reagents
All chemical reagents and media components were ob-
tained from Sigma (UK) unless mentioned otherwise.
2.2. Tissue Culture
ARIP cells were obtained from ATCC (USA) and main-
tained in F12K medium supplemented with 10% FCS,
100 μg/ml streptomycin and 100 units/ml penicillin at 37
˚C in a humidified, 5% CO2 atmosphere. For experi-
mental use, ARIP cells were passaged from tissue cul-
ture flasks using a 0.25% trypsin/0.02% EDTA solution.
In order to determine the optimal serum-free tissue cul-
ture conditions, F12K medium was supplemented with 1
g/L ITS supplement (containing 5 mg/L insulin + 5
mg/L transferrin + 5 mg/L selenium), 2 g/L BSA, 10
mM nicotinamide, 10ng/ml keratinocyte growth factor
(KGF) and 10nM Glucagon-Like Peptide-1 (GLP-1)
together or separately.
2.3. Surface Coating and Cell Treatment
Multiwell plates with tissue culture (TC) surface (Nunc,
Denmark) were coated with Collagen Type I (CN. Up-
state, USA), Fibronectin (FN), Laminin (LN) and Vi-
tronectin (VN) respectively according to product in-
structions. The concentrations of the four ECM proteins
were 0.1 μg/cm2, 1 μg/cm2 and 10μg/cm2 (not for VN)
respectively. Following surface treatment, plates were
washed for 5 min in PBS and blocked with 1% bovine
serum albumin (BSA) in PBS for 1 h at 37°C. ARIP
cells with 80% confluence were disattached with tryp-
sin/EDTA and resuspended in F12K containing 10%
FCS. Cells were spun down at 250 xg for 5 min and re-
suspended in serum-free F12K. Cells were seeded at a
density of 3000 cells/cm2 and cultured for 6 hours at
37°C in a humidified, 5% CO2 atmosphere.
2.4. Single Cell Morphology Analysis
The cultured cells were stained with 2uM Calcein AM
(Invitrogen, UK), a membranepermeant fluorescence
viable cell indicator, for 45 minutes at room temperature.
Images were captured from 4 random areas surrounding
each centre of a well using an inverted fluorescent Leica
DM IRB microscope at X200 magnification.
Fluorescence images were subsequently analysed us-
ing a highly standardised macro program language,
QUIPS written in the Leica Qwin imaging software
((Leica, Bensheim, Germany). Briefly, the image analy-
sis program was delicately designed to identify half
maximal colour intensity so as to recognise the body of
the single cell but not any background automatically.
The intensity threshold was manually altered to reach the
cell membrane. A binary editing function enabled the
manual removal of artefacts that may have contributed to
a false measurement. Once the area of choice had been
accepted, single cell features (e.g. area, perimeter, num-
ber and roundness) were measured simultaneously ac-
cording to the calibration of microscopic lens.
2.5. ATP Luminescence Assay
The determination of cell proliferation was carried out
on the basis of a luminometric ATP measurement by
means of the ATPlite™-M Luminescence Assay System
(PerkinElmer, USA) according to the manufacturer's
instruction. Luminescence was measured using a LUCY
1 luminometer (Anthos Labtech Instruments, Austria).
Standard curves for individual culture of mammalian
cell number were performed to confirm both linearity
and sensitivity of the method.
2.6. Surface Coverage Measurement
Similar to the surface coating procedures described
above, multiwell plates with tissue culture surface was
218 G. G. Adams et al. / J. Biomedical Science and Engineering 2 (2009) 216-226
SciRes Copyright © 2009 JBiSE
coated with CN (10μg/cm2), LN (10μg/cm2), CN 50%+
LN 50%, Matrigel matrix (BD Biosciences, UK), Poly
L-Lysine (PLL; 0.01% solution), PLL followed by CN,
PLL followed by LN, PLL followed by CN 50%+ LN
50%, PLL followed by Matrigel, respectively. ARIP
cells at a concentration of 2.5X104 cells/cm2 were cul-
tured on the above surfaces with F12K medium contain-
ing 10% FCS for 4 hours, and sequentially with Se-
rum-free F12K medium supplemented with 2g/l BSA,
1g/l ITS supplement, 10mM nicotinamide and P/S. Im-
ages were captured from the centre region of each well
with the magnification of X100. Cell surface coverage
analysis was similar to the measurement of single cell
morphology mentioned above.
2.7. Measurement of Insulin Content by
Enzyme-Linked Immunosorbent Assay
(ELISA)
Cultured cells were washed three times for 10 min each
in prewarmed fresh Krebs-Ringerbicarbonate buffer
(KRBB; 115 mM NaCl, 24 mM NaHCO3, 5 mM KCl, 1
mM MgCl2, 2.5 mM CaCl2, 10 mM HEPES, pH 7.4)
with 0.1% BSA. Cells were then lysed directly by add-
ing 0.5 ml CelLyticTM-M mammalian cell lysis/ extrac-
tion reagent. After incubated for 15 min on a shaker, cell
lysate was collected and centrifuged for 15 min at 1200
xg to pellet the cellular debris. Theprotein-containing
supernatant was transferred to a chilled test tube and
stored at –80°C until being assayed for the presence of
insulin using an Ultra Sensitive Rat Insulin ELISA Kit
(Crystal Chem Inc, USA) according to the manufac-
turer’s instruction.
2.8. Measurement of Insulin Release
Stimulated by Glucose
Cells were washed three times with KRBB buffer. After
pre-incubation with fresh KRBB buffer for another 30
minutes at 37 ºC, cells were subjected to 5.5 mM and
then 16.7 mM glucose within the KRBB buffer for 1
hour. The amount of insulin released in supernatant was
determined by rat insulin ELISA as mentioned above.
2.9. Quantification of Cellular DNA Content
Cultured cells in 96-well plates were washed with PBS
and treated with 0.02% SDS in salinesodium citrate
buffer at 37°C for 1h. Equal volume of 2μg/ml Hoechst
33258 working solution was added in the above wells.
100 μl of the above mixture was transferred to a fresh
96-well assay plate. Measurement of fluorescence was
performed with a plate reader (Dynex technologies, UK)
at excitation λ 360 nm and emission λ 460 nm. Deoxy-
ribonucleic acid sodium salt from calf thymus (Sigma,
UK) was used as a DNA standard.
2.10. Data Analysis
The differences at various time points and experimental
conditions were analysed by one-way ANOVA at 95%
significance level for multiple comparisons using SPSS
11.5 (SPSS Inc., USA).
3. RESULTS
3.1. Single Cell Morphology on ECM
Protein-Coated Surfaces
The first step in this research was to establish that the
chosen blocking agent, 1% BSA, could decrease single
cell area and perimeter when coated on to tissue culture
treated plastic surfaces. Figure 1 confirmed the blocking
action of BSA. Subsequently, effects of the four ECM
protein-coated surfaces at a range of concentrations were
compared. These ECM protein-coated surfaces altered
both single cell features (P<0.01) when compared with
BSA-coated surfaces. The treatment effects on both sin-
gle cell area and single cell perimeter were in the order
of CN>FN>LN>VN. For single cell area, among CN
groups, there was a significant difference between the
coating concentrations of 1 μg/cm2 and 0.1ug/cm2
(P<0.01); and the single cell area difference between
them was approximately 200 μm2. No difference was
found between 10 μg/cm2 and 1 μg/cm2 of CN groups.
Among FN groups, the three coating concentrations
showed similar results. Among LN groups, 10 μg/cm2 of
LN-coated surface resulted in higher single cell area
compared with the other two lower concentrations
(P<0.01); no difference was found between 0.1 ug/cm2
and 1 ug/cm2. Among VN groups, the results of 0.1 ug/
cm2 and 1 ug/cm2 concentrations were similar. The fea-
ture of single cell perimeter produced similar patterns of
treatment effects on single cell area. When single cell
roundness was estimated, surfaces coated with three
concentrations of CN increased this feature, and 10
μg/cm2 of CN-coated surface appeared to induce the
highest (P<0.01 vs BSA group).
Morphologically, on the four ECM protein-coated
surfaces, ARIP cells showed increased cell membrane
spreading and cell protuberance (Figure 2). It appeared
that the effect of CN, FN and LN was stronger than VN
when they were used to coat tissue culture polystyrene
surface. These results indicated that four types of ECM
protein-coated surfaces could expand the morphological
structures of single ARIP cell in vitro.
In a follow-up experiment, ARIP cells at a density of
1.5×104 cells/cm2 were cultured under the equivalent
surface-coated conditions described above and in se-
rum-free medium for 96 hours. Results showed that the
viability of cells decreased severely, and there were just
few viable cells left surrounded by large amounts of a
cellular debris (data not shown). It appeared difficult for
cells to reach confluence. As a control, cells in medium
G. G. Adams et al. / J. Biomedical Science and Engineering 2 (2009) 216-226 219
SciRes Copyright © 2009 JBiSE
0
200
400
600
800
1000
1200
1400
A
**
Figure 1. Effect of ECM protein-coated surfaces on the single cell area (A) and perime-
ter (B) of ARIP cells. ARIP cells were seeded on surfaces as indicated at a density of
3000 cells/cm2 and cultured for 6 hours in F12K medium respectively. Each bar was
given as means±SD. Significant differences are indicated as **P<0.01 vs. BSA.
containing 10% FCS showed normal growth. Therefore,
for the long-term culture, it was not possible to maintain
growth of ARIP cells on the ECM-protein coated sur-
faces in absolute serum-free medium without providing
any extra essential component.
3.2. Growth of ARIP Cells in Defined
Serum-free Medium
In order to provide a suitable serum-free culture envi-
ronment for ARIP cells without using FCS, this work
investigated the possibility to defining and optimising
serum-free medium with known functional supplements
including ITS, BSA, nicotinamide, KGF and GLP-1. As
shown in Figure 3, when ARIP cells were cultured at a
low seeding density (1.5×104 cells/cm2), and cultured
ver 96 hours, there were significant differences in cell
proliferation between normal culture medium containing
10% FCS (named as FCS group) and all serum-free cul-
tivation groups (P<0.01). This suggested that it was dif-
ficult to establish cell proliferation within serum-free
media if the cell density was too low. At a higher density
studied (3×104 cells/cm2), none of the serum freegroups
supported cell proliferation as sufficiently as FCS group.
However, cell proliferation in the FI F12K medium sup-
plemented with BSA, ITS and nicotinamide (named as FI)
or F12K medium supplemented with BSA, ITS, nicoti-
namide and KGF (named as FII) appeared to behigher
than other serum-free groups. The addition of KGF in FII
group did not show any additiveeffect. At the highest
density studied, 6×104 cells/cm2, after tissue culture for
96 hours, there was no difference among FCS, FI and FII
groups, suggesting that FI and FII provided a tissue
o
Su
A ea
rface
verage cell ar
**
**
Average cell area
0
20
40
60
80
100
120
140
160
Sur e
Av er
fac
erage cell perimet
**
**
**
**
**
**
**
**
*
*
*
*
*
*
*
*
*
*
*
0.1 μg/cm2 VN
0.1 μg/cm2 CN
0.1 μg/cm2 LN
0.1 μg/cm2 FN
10 μg/cm2 CN
10 μg/cm2 LN
10 μg/cm2 FN
1 μg/cm2 VN
1 μg/cm2 CN
1 μg/cm2 LN
1 μg/cm2 FN
BSA
TC
Surface
B
Average cell perimeter
0.1 μg/cm2 VN
0.1 μg/cm2 CN
0.1 μg/cm2 LN
0.1 μg/cm2 FN
10 μg/cm2 CN
10 μg/cm2 LN
10 μg/cm2 FN
1 μg/cm2 VN
1 μg/cm2 CN
1 μg/cm2 LN
1 μg/cm2 FN
BSA
TC
Surface
220 G. G. Adams et al. / J. Biomedical Science and Engineering 2 (2009) 216-226
SciRes Copyright © 2009 JBiSE
 
  
Figure 2. Morphology of ARIP cells cultured on different surfaces with F12K medium plus P/S for 6 hours. Magnifica-
tion: X200. Cells were stained with 2uM of Calcein AM. A. TC surface. B. 1% BSA surface. C. 10μg/cm2 CN surface. D.
10μg/cm2 FN surface. E. 10μg/cm2 LN surface. F. 10μg/cm2 VN surface. The scale bar was 50 m.
A B
0
20
40
60
80
100
120
140
FCSFFIFIIFIII FIV
M edium
Relative Light Unit (RLU)
25000 cells/m l50000 cells/m l100000 cells/ ml
0
20
40
60
80
100
120
140
160
FCSFFIFIIFIII FIV
M edium
Relative Light Unit (RLU)
25000 cells/m l50000 cells/m l100000 cells/m l
Figure 3. Relative light unit (RLU) by ATP assay which quantitatively indicates relative number of ARIP cells. FCS:
F12K+10% FCS; F: F12K; FI: F12K+BSA+ITS+nicotinamide; FII: FI+ 10ng/ml KGF; FIII: FI+10nM GLP-1; FIV:
FI+KGF+GLP-1. Each value was give as mean±SD from 6 observations. A. 48 hours. B. 96 hours. Significant differences
are indicated as * P<0.05; **P<0.01 and ***P<0.001vs. FCS.
culture environment equivalent to culture medium con-
taining FCS. The group of F12K medium supplemented
with 10nM GLP-1 (named as FIII) did not improve cell
proliferation even compared with the group of absolute
serum-free medium (F group). Moreover, in the group of
serum-free medium supplemented with both KGF and
GLP-1 (named as FIV), ARIP cells showed the lowest
proliferation rate compared with other serum-free culture
conditions. These results implied that GLP-1 might play
an inhibitive role in proliferation of ARIP cells espe-
cially in the presence of KGF. Overall, during the
96-hour period studied, reproduction or proliferation of
ARIP cells under serum-free conditions was density-
dependent. The supplements of BSA, ITS and nicotina-
mide together were found to maintain cell proliferation
at equivalent levels to FCS.
3.3. Estimation of Cell Coverage in the
Presence of ECM Protein-Coated Sur-
faces and Defined Serum-Free Medium
This study investigated the overall cell growth of ARIP
cells in the presence of ECM proteins and defined se-
rum-free medium by estimating the surface coverage of
culture at three time points (4, 48 and 96 hours). As
shown in Figure 3, after monolayer culture for 96 hours,
the surfaces treated with CN and LN produced cell cov
G. G. Adams et al. / J. Biomedical Science and Engineering 2 (2009) 216-226 221
SciRes Copyright © 2009 JBiSE
Figure 4. Surface coverage of ARIP cells. Cells were cultured in F12K supplemented with BSA,
ITS, nicotinamide and P/S. Each value was given as means±SD from 3 individual observations.
TC: Tissue culture surface. CN: collagen I coated surface. LN: laminin coated surface. CN+LN:
50% collagen I and 50% laminin coated surface. Matrigel: Matrigel coated surface. PLL: 0.01%
ploy-l-lysine coated surface. PLL+CN: poly-l-lysine surface was further coated with collagen I.
PLL+LN: poly-l-lysine surface was further coated with laminin. PLL+CN+LN: poly-l-lysine
surface was further coated with 50% collagen I and 50% laminin. PLL+Matrigel: poly-l-lysine
surface was further coated with Matrigel. Significant differences are indicated as * P<0.05;
**P<0.01 or **P<0.001.
0.0
0
4 hr
erage of approximately 60%, while the percentage was
approximately 46% on tissue culture surface. Thus, CN
and LN coated surfaces improved the cell surface cov-
erage (P<0.05). PLL (0.01%) enhanced cell coverage
slightly after culture for 96 hours no matter whether sur-
faces were coated with ECM proteins. Matrigel inhibited
cell coverage either on its own or on PLL pre-coated
surface. Moreover, the majority of ARIP cells appeared
rounder in morphology and difficult to spread upon Ma-
trigel-coated surfaces. In addition, distribution of ARIP
cells on Matrigel-coated surfaces was patchy and uneven
(Figure 3). In some of the captured images in the Ma-
trigel-coated-surface group, a few of cells could be
found. The addition of PLL did not alter the effect of
Matrigel.
On the laminin-coated surfaces, the surface coverage
increased by approximately 16% in the absence of PLL
and 23% in the presence of PLL compared with non-
treated tissue culture surface, suggesting the positive role
of LN in the overall status of ARIP cells in defined se-
rum-freemedium. The surface treated with CN and LN
together appeared to show an effect after 48 hours, but
after incubation for 96 hours, it was similar to the sur-
faces coated with CN or LN separately. Furthermore, the
equivalent experiments were repeated using a lower cell
density, 3×104 cell/cm2, and the results also demon-
strated that LN significantly promoted cell surface cov-
erage, andMatrigel inhibited cell growth and surface
coverage.
3.4. Insulin Production and Gene Expression
under Differentiation Conditions
Immunofluorescence was carried out to identify insu-
lin-producing ARIP cells induced by GLP-1. As shown
in Figure 53, in the presence of 10 nM GLP-1 within
serum-free medium, approximately 20-30% of ARIP
cells were insulin positive, suggesting that GLP-1 was
able to direct insulin-positive cell formation.
.2
0.4
0.6
0.8
1.0
TC
CN
LN
N
M trigel
PLL
P L-CN
PLL-LN
PLL-CN-LN
PLL-Matrigel
S
Perc age
CN-L
a
L
urface
entage of cell cover
48 hr96 hr
Surface
4 hr48 hr96 hr
1.0
0.8
Percentage of cell coverage
0.6
0.4
0.2
0.0
PLL-CN
PLL-LN
PLL-CN-LN
PLL-Matrigel
Matrigel
CN-LN
PLL
CN
LN
TC
222 G. G. Adams et al. / J. Biomedical Science and Engineering 2 (2009) 216-226
SciRes Copyright © 2009 JBiSE
In order to investigate whether these insulin-positive
cells can produce insulin, the insulin content within
ARIP cells was carried out. Meanwhile, total cellular
DNA content was applied to normalize insulin content
and indicate cell proliferation. As shown in Figure 5A,
after incubation for 72 hours, surfaces coated with ECM
proteins studied resulted in large increases in total DNA
content. The cell culture upon laminin-coated surface
resulted in the highest DNA level, which was approxi-
mately 3-fold higher than that upon normal tissue culture
plastic surface. Because the total DNA content reflects
the cell number in each group, the data indicated that
ECM proteins play a role in increasing cell number over
the period studied. Furthermore, as shown in Figure 5B,
all the ECM protein-coated surfaces appeared to im-
prove the total insulin content in ARIP cells over a
72-hour period studied. Collagen-coated surface ap-
peared to have the highest impact on insulin production
when compared with surfaces coated with either fi-
bronectin or laminin. After insulin production is nor-
malisation by DNA content as shown in Figure 5C, it
appeared that only the collagen-coated surface had a
stimulative effect on insulin production given the same
size of cell population. The collagen-coated surface
showed approximately 1.8-fold higher normalised insu-
lin content in ARIP cells than the uncoated tissue culture
surface.
The presence of insulin gene expression in ARIP cells
was measured using RT-PCR. As shown in Figure 6,
despite the strong expression of β-actin as an endoge-
nous control, insulin gene expression was weak in the
entire groups studied. It appeared that ARIP cells upon
collagen coated surface resulted in slightly higher insulin
gene expression. Expressions of the other genes such as
PDX-1 were not identified. Experiments were repeated
at least three times and similar results were obtained. To
ensure the gene primers used in RT-PCR were specific,
RNA samples were isolated from RIN-m5F cells, a rat
insulin-producing cell line, and used as controls. Results
showed that RIN-m5F cells strongly expressed insulin
and PDX-1 genes under identical RT-PCR conditions
thus confirming the efficacy of the technique employed.
Given the combination conditions including collagen-
coated surface, GLP-1 and defined serum-free medium,
and after incubation for 72 hours, the ARIP cells were
challenged with 5.5 mM and 16.7 mM of glucose. As
shown in Figure 7, there appeared to be little difference
in insulin secretion response to the two concentrations of
glucose.
4. DISCUSSION
The work demonstrates that tissue culture polystyrene
surfaces coated with different types of ECM proteins
have effects on the morphological features of single
ARIP cell. These changes can be detected by 5 hours.
0
1
2
3
4
5
TC CN FNLN
DNA (ug/mL)
0.0
0.1
0.2
0.3
0.4
0.5
TC CNFNLN
Insulin (ng/mL)
0
50
100
150
TCCN FNLN
Figure 5. Insulin released by ARIP cells with response to ECM
-protein-treated surfaces. Cells were cultured for 48 hours in
F12K medium supplemented with ITS, BSA nicotinamide and
GLP-1. TC: TC surface. CN: 10μg/cm2 CN coated surface. FN:
10μg/cm2 FN coated surface. LN: 10μg/cm2 LN coated surface.
Cell density: 5x104 cells/cm2. Each value was given as
mean±SD from 3 individual observations.
Their treatment effects on both single cell area and
single cell perimeter were in the order of collagen Type I
>fibronectin>laminin>vitronectin. Of the four ECM pro-
teins studied, collagen-coated surface achieves the high-
est level, for example, the increase of single cell area is
2.7-3.8 fold compared with the BSA-coated surface as a
control. Previous studies have identified the tripeptide
RGD is located in all the four ECM protein types, and
the RGD domain in the ECM proteins serves as a cell
Insulin (ng/mg DNA)
G. G. Adams et al. / J. Biomedical Science and Engineering 2 (2009) 216-226 223
SciRes Copyright © 2009 JBiSE
1 2 3 4 5
Insulin
PDX-1
β-actin
Figure 6. Expression of insulin and PDX-1 genes in ARIP
cells following incubation for 72 hours in defined serum-free
medium plus 10 nM GLP-1 and upon surfaces indicated. Lanes
from 1 to 4 are RT-PCR products from ARIP cells. Lane 1:TC
surface. Lane 2: collagen I coated surface. Lane 3: fibronectin
coated surface. Lane 4: laminin coated surface. Lane 5:gene
expression in RIN-m5F cells (an insulin-producing cell line
from insulinoma) and used to validate the RT-PCR assay of
each gene.
Figure 7. Insulin secretion in ARIP cells with response to glu-
cose stimulation. Cells were cultivated for 72 hours in defined
serum-free medium plus 10 nM GLP-1 and upon collagen
coated surface. Each value was given as mean ± SD from 5
individual observations.
recognition site and plays a role in cell spreading
through its binding to integrins [35]. However, as these
four ECM proteins bind to different integrin patterns,
this may explain their different effects on ARIP cells.
Vitronectin has the weakest effects according to our data,
it may be possible that vitronectin mainly binds to αvβ3
and αvβ5 integrins, which are not as effective or abun-
dant as the other integrins [31].
Because without FCS, serum-free medium is not able
to maintain the survival of ARIP cells upon ECM-pro-
tein coated surfaces, the work presents an optimal se-
rum-free culture environment for the purpose of using
ECM proteins effectively. Moreover, as proliferation of
ARIP cells is density-dependent. The initial seeding
density of cells should be high enough (e.g. 6×104 cells/
cm2) so as to maintain cells under serum-free medium
conditions. This sufficient cell density may encourage
cells interact effectively, thus enhancing survival of the
total population. The impact of KGF on proliferation
appears to be minimal when ARIP cells are cultured
under serum-free conditions. However, KGF has been
identified as an important growth factor of epithelial
cells [31,43]. A previous study also suggests that KGF
has a potential to induce pancreatic ductal cell prolifera-
tion in vivo [45]. The reason KGF doesn’t not play a role
in ARIP cells proliferation may be because ARIP are
distinctive from primary pancreatic epithelium in terms
of their origin and cellular characteristics. For example,
the pancreatic regenerating gene proteins are mitogenic
to primary cultures of ductal cells, but compared with
their effects on the primary cultured ductal cells, these
proteins were 100-fold less potent on ARIP cells [46].
Our results show that GLP-1 inhibits the proliferation
of ARIP cells. This may be the result of transdifferen-
tiation of pancreatic ductal cells induced by GLP-1
[4,5,15]. A previous study also provides evidence that
GLP-1 promotes islet-cell neogenesis from pancreatic
ductal cells in a type 2 diabetic rodent model [30]. Due
to the inhibitory effect of GLP-1 on ductal cell prolifera-
tion, GLP-1 should be supplied at differentiation but not
proliferation stage.
0.0
0.2
0.4
0.6
5.5 mM glucose16.7 mM glucose
Insulin (ng/mL/hr)
Surface coverage pattern of adhesive cells is deter-
mined by total cell number, individual cell features and
surface properties in theory, thus it can reflect the dy-
namics of the overall cell performance under particular
tissue culture conditions [1,12]. The ECM provides a
backbone to influence the attachment of other proteins or
influence cell adhesion directly via embedded cell sig-
nalling. Therefore, the coverage of monolayer cells on
ECM protein-coated surfaces can indicate the overall
patterns of cell status as consequence of adhesion, mi-
gration, proliferation and differentiation. Our results
demonstrates that when monolayer ARIP cells are cul-
tured for 4 days in optimised serum-free medium, sur-
face coated with laminin has a significant supportive
effect on coverage of ARIP cells. Because laminin has
been demonstrated to improve the insulin accumulation
and preservation of pancreatic endocrine cells [17,18,19],
it can be employed when cells with pancreatic ductal
characteristics are differentiated into an insu-
lin-producing phenotype.
Collagen-coated surface appeared to increase the per-
centage of cell coverage especially in the presence of
poly-L-lysine. A previous study has indicated that Col-
lagen is useful for the culture of adult primary pancreatic
epithelial cells from the main duct [27]. However, in
embryonic pancreatic epithelia, laminin but not collagen
has been demonstrated to induce duct formation [23].
Therefore, the exact signalling pathways of collagen and
laminin for cell differentiation should be further investi-
gated.
224 G. G. Adams et al. / J. Biomedical Science and Engineering 2 (2009) 216-226
SciRes Copyright © 2009 JBiSE
Poly-lysine can enhance electrostatic interaction be-
tween negatively-charged ions of the cell membrane and
positively-charged surface ions of attachment factors on
tissue culture plastic surface. When adsorbed to the cul-
ture surface, it may increase the number of positively
charged sites available for cell binding. It has been
known to be useful for enhancing effects of surface
coating with ECM proteins. Our results confirm that
Poly-L-lysine can improve the percentage of cell cover-
age on culture surfaces coated with selective ECM pro-
teins.
In the presence of defined serum-free medium, Ma-
trigel-coated surface inhibits the coverage of ARIP cells
significantly. Indeed this inhibition can be detected even
after 4 hours. In order to avoid the potential effects of
growth factors in Matrigel, growth factor-reduced Ma-
trigel has also been used, however, the inhibitive effect
induced by Matrigel cannot disappear. Matrigel has been
found important to promote the formation of islet-like
clusters from primary pancreatic ductal epithelial cells
[3]. It is possible that Matrigel induces the differentia-
tion of pancreatic ductal cells leading to a decreased
proliferation. However, due to the complex formulation
of Matrigel, the exact signalling pathway needs to be
further investigated.
In the presence of GLP-1, ARIP cells show different
morphological structures and a loss of their typical cob-
blestone-like morphology. This morphological change
may be due to the regulation of metabolism activities or
induced differentiation. GLP-1 has been reported to in-
duce insulin-producing cell differentiation from pancre-
atic acinar AR42J and ductal ARIP cells [15,47]. For
example, exposure of rat AR42J cells to GLP-1 over 2
days result in an initial increase in levels of cyclic
adenosine monophosphate and cellular proliferation,
followed by cessation of proliferation and expression of
the islet-associated hormones, insulin, glucagon, and
somatostatin, in up to 50% of cells [47]. Furthermore,
GLP-1 treatment also induces expression of glucose
transporter 2 and glucokinase genes, in association with
the capacity to secrete insulin in a glucose-dependent
manner [47]. Therefore, the morphological change found
here may be associated with transdifferentiation of ARIP
cells. Interestingly, defined serum-fee medium also re-
sulted in a slight change in morphology of ARIP cells.
As nicotinamide has been considered capable of stimu-
lating insulin-producing cell differentiation in other
progenitor cell types [7,29], this may partly explain the
morphological change of ARIP cells in defined se-
rum-free medium. Immunofluorescence data in this
study indicate that 20-30% of ARIP cells can obtain an
insulin-positive phenotype after exposure to GLP-1. The
insulin content of ARIP cells in the presence of GLP-1
can also be detected. Thus, it confirms that GLP-1 has
the capability of converting ARIP cells into insu-
lin-producing cells to some extent. The combination of
the defined serum-free medium and GLP-1 leads to the
maximum insulin production rate. This is as expected,
because the defined serum-free medium is able to sustain
the maximum growth of ARIP cells in a serum-free en-
vironment.
After the treatment with GLP-1 and/or defined se-
rum-free medium, the gene expression of insulin in
ARIP cells is observed at a low level. This indicates that
the insulin content observed by ELISA is not from po-
tential contamination or uptake that has been described
in a previous study on the insulin-producing cell differ-
entiation of embryonic stem cells [37]. The result ob-
tained here appears similar to a previous study, which
demonstrates the weak expression of insulin gene in
GLP-1-induced ARIP cells by northern blot analysis
[15]. However, no insulin gene expression was detected
by another similar study [21]. Moreover, the expression
of other relevant genes such as PDX-1 is not detected
here. Therefore, although the insulin gene, not normally
present in ARIP cells, is expressed, it is difficult to relate
this study to the observations by other relevant studies.
In addition, although the role of GLP-1 in insulin-
producing cell formation is confirmed, it remains unclear
as to how GLP-1 triggers the insulin production in ARIP
cells. Previous studies suggests that the initial step of
GLP-1 action requires identification and binding of the
peptide to the GLP-1 receptor, which is in the su-
per-family of G protein-coupled receptors [41]. Activa-
tion of the GLP-1 receptor results in the stimulation of
adenylyl cyclase, leading to an increase in intracellular
cyclic adenosine monophosphate (cAMP) and the
cAMP-dependent protein kinase A (PKA) activation in
the pancreatic tissue [10,11,14,22]. However, the gene
expression of GLP-1 receptor remains controversial as
described in previous studies [15,21]. It is also reported
that the action of GLP-1 depends on the expression of
PDX-1. However, PDX-1 gene expression is not de-
tected in this study, which is consistent with two of the
previous studies [21,36]. Relevant studies on other adult
non-β cells may provide additional clues. For example,
the PDX-1 expression in the duodenal tissue is stronger
than in intestinal epithelial cells, but GLP-1 could induce
insulin production in intestinal epithelial cells but not in
duodenum cells [38].
On the other hand, GLP-1 has been suggested to up-
regulate expression of some other transcription factors
such as Neurogenin 3 but not PDX-1 during insulin ex-
pression [38]. A study using adult hepatic progenitor
cells has shown that the acute insulin secretion of
non-PDX-1- expressing cells can be enhanced by GLP-1
in a time-dependent manner, though the effect of GLP-1
is strengthened in the presence of PDX-1 [25]. Indeed,
there is no sufficient evidence to show that GLP-1 in-
creases the PDX-1 gene expression during insulin pro-
ducing cell conversion from other adult cell types. Taken
G. G. Adams et al. / J. Biomedical Science and Engineering 2 (2009) 216-226 225
SciRes Copyright © 2009 JBiSE
together, during the induction of insulin-producing-cell
generation using adult progenitor cells, the role of
GLP-1 may be enhanced by PDX-1 gene expression, but
the exact signalling pathway of GLP-1 remains unclear.
5. CONCLUSIONS
In the presence of GLP-1 and defined serum-free me-
dium, the total amount of insulin content in ARIP cells
significantly increases upon the surfaces coated with
ECM proteins including collagen, fibronectin and
laminin. Meanwhile, comparison of total DNA content
indicates that ECM-protein coated surfaces play a role in
cell duplication, possibly through improved single cell
spreading. After normalisation by DNA content, only
collagen-coated surface appears to stimulate the insulin
production given the same cell population size. The data
obtained from gene expression also show that the insulin
gene expression in ARIP cells upon collagen-coated
surface appears slightly stronger. To date, there is still
no sufficient information to determine how important
ECM proteins are to insulin-producing cell formation
from other cell types. Therefore, more experiments need
to be carried out before the roles of ECM proteins can be
fully illustrated. In particular, the distribution and activ-
ition of integrins may be important in revealing the pro-
found function of these proteins.
6. ACKNOWLEDGEMENT
This work was funded by the EPSRC. We thank Professor Kevin
Shakesheff, Dr Sue Chan, Dr Richard Pearson and Mr Darryl Jackson
for their critical advice and technical assistance.
REFERENCES
[1] B. T. M. Baharloo and D. Brunette, (2005) Substratum
roughness alters the growth, area, and focal adhesions of
epithelial cells, and their proximity to titanium surfaces, J
Biomed Mater Res A, 74, 12-22.
[2] S. B. Weir, L. A. Baxter, G. T. Schuppin, and F. E. Smith,
(1993) A second pathway for regeneration of adult exo-
crine and endocrine pancreas: A possible recapitulation
of embryonic development, Diabetes, 42, 1715-1720.
[3] S. T. B. Weir, M. Weir, G. C. Tatarkiewicz, K. Song, K.
H. Sharma, and A. O'Neil, (2000) In vitro cultivation of
human islets from expanded ductal tissue, Proc Natl
Acad Sci, 97, 7999-8004.
[4] A. H. Bulotta, H. Anastasi, E. Bertolotto, C. Boros, L. G.
Di Mario, U. Perfetti, R. Silver, K. Yao, F. Zenilman, M.
E. Chen, and T. H. Magnuson, (2002) Cultured pancre-
atic ductal cells undergo cell cycle re-distribution and
beta-cell-like differentiation in response to glucagon-like
peptide-1, J Mol Endocrinol, 29, 347-360.
[5] A. P. Bulotta, H. R. Hui, and L. G. Boros, (2003) GLP-1
stimulates glucose-derived de novo fatty acid synthesis
and chain elongation during cell differentiation and insu-
lin release, J. Lipid Res, 44, 1559-1565.
[6] L. A. Chen, M. C. Alam, T. Miyaura, C. Sestak, A.
O'Neil, J. Unger, and C. B. Newgard, (1992) Factors
regulating islet regeneration in the post-insulinoma
NEDH rat, Adv Exp.Med Biol, 321, 71-80, 81-74.
[7] L. B. Chen, X. B. Jiang, and L. Yang, (2004) Differentia-
tion of rat marrow mesenchymal stem cells into pancre-
atic islet beta-cells, World J Gastroenterol, 10, 3016-
3020.
[8] B. S. Chertow, N. Q. Goking, and H. K. Driscoll, (1997)
Effects of all-trans-retinoic acid (ATRA) and retinoic
acid receptor (RAR) expression on secretion, growth, and
apoptosis of insulin-secreting RINm5F cells, Pancreas,
15, 122-31.
[9] E. A. Clark and J. S. Brugge, (1995) Integrins and signal
transduction pathways: The road taken, Science, 268,
233-239.
[10] W. G. R. Ding, E. Rorsman, P. Buschard, and J. K. Gro-
mada, (1997) Glucagon-like peptide I and glucose- de-
pendent insulinotropic polypeptide stimulate Ca2+ in-
duced secretion in rat alpha-cells by a protein kinase
A-mediated mechanism, Diabetes, 46, 792-800.
[11] H. C. S. Fehmann, M. and Goke, B. (1994) Interaction of
glucagon-like peptide-I (7-37) and somatostatin-14 on
signal transduction and proinsulin gene expression in
beta TC-1 cells, Metabolism, 43, 787-792.
[12] P. Feugier, R. A. Black, J. A. Hunt, and T. V. How,
(2005) Attachment, morphology and adherence of human
endothelial cells to vascular prosthesis materials under
the action of shear stress, Biomaterials, 26, 1457-1466.
[13] R. U. Gao, J. Pulkkinen, M. A. Lundin, K. Korsgren, and
T. Otonkoski, (2007) Characterization of endocrine pro-
genitor cells and critical factors for their differentiation
in human adult pancreatic cell culture, Diabetes, 52,
2003-2015.
[14] D. H. Gefel, G. K. Mojsov, S. Habener, and G. C. Weir,
(1990) Glucagon-like peptide-I analogs: effects on insu-
lin secretion and adenosine 3’,5’-monophosphate forma-
tion, Endocrinology, 126, 2164-2168.
[15] H. Hui, C. Wright, and R. Perfetti, (2001) Glucagon-like
peptide 1 induces differentiation of islet duodenal ho-
meobox-1-positive pancreatic ductal cells into insu-
lin-secreting cells, Diabetes, 50, 785-96.
[16] N. W. Jessop, (1980) Characteristics of two rat pancre-
atic exocrine cell lines derived from transplantable tu-
mors, In vitro, 16, 212.
[17] F. X. Jiang, (2005) Laminin-1 and epidermal growth
factor family members costimulate fetal pancreas cell
proliferation and colony formation, Differentiation, 73,
45-49.
[18] F. X. C. Jiang, D. S. DeAizpurua, and L. C. Harrison,
(1999) Laminin-1 promotes differentiation of fetal mouse
pancreatic beta-cells, Diabetes, 48, 722-730.
[19] F. X. Jiang and L. C. Harrison, (2001) Regulation of
laminin 1-induced pancreatic beta-cell differentiation by
alpha6 integrin and alpha-dystroglycan, Mol Med, 7,
107-114.
[20] C. Q. Lin, (1993) Multi-faceted regulation of cell differ-
entiation by extracellular matrix, Faseb, 7, 737-743.
[21] H. K. G. Liu, B. D. Gault, V. A. McCluskey, J. T.
McClenaghan, N. H. O’Harte, and P. R. Flatt, (2004)
N-acetyl-GLP-1: A DPP IV-resistant analogue of gluca-
gon-like peptide-1 (GLP-1) with improved effects on
226 G. G. Adams et al. / J. Biomedical Science and Engineering 2 (2009) 216-226
SciRes Copyright © 2009 JBiSE
pancreatic beta-cell-associated gene expression, Cell Biol
Int., 28, 69-73.
[22] M. W. Lu, M. B. Leng, and A. E. Boyd, (1993 The role
of the free cytosolic calcium level in beta-cell signal
transduction by gastric inhibitory polypeptide and glu-
cagonlike peptide I (7-37), Endocrinology, 132, 94-100.
[23] T. S. C. Maldonado, C. A. Kadison, A. S. Alkasab, S. L.
Longaker, and G. K. Gittes, (2000) Basement membrane
exposure defines a critical window of competence for
pancreatic duct differentiation from undifferentiated
pancreatic precursor cells, Pancreas, 21, 93-96.
[24] C. Miyaura, L. Chen, M. Appel, T. Alam, L. Inman, S.
Hughes, D. Milburn, J. L. Unger, and C. B. Newgard,
(1991) Expression of reg/PSP, a pancreatic exocrine gene:
relationshipto changes in islet beta-cell mass, Mol Endo-
crinol, 5, 226-234.
[25] N. S. N. Nagata, T. Mitaka, T. Katakai, T. Yamato, E.
Miyazaki, J. Tabata, Y. Sugai, and A. Shimizu, (2004) In
vitro induction of adult hepatic progenitor cells into insu-
linproducing cells, Biochem Biophys Res Commun, 318,
625-630.
[26] C. O. Welsh, (2001) Long-term culture in matrigel en-
hances the insulin secretion of fetal porcine islet-like cell
clusters in vitro, Pancreas, 22, 157-163.
[27] D. S. Oda, C. E. Nguyen, T. D. Swenson, and S. P. Lee,
(1998) Culture of human main pancreatic duct epithelial
cells, In Vitro Cell Dev Biol Anim, 34, 211-216.
[28] T. Otonkoski, G. M. Beattie, M. I. Mally, C. Ricordi, and
A. Hayek, (1993) Nicotinamide is a potent inducer of
endocrine differentiation in cultured human fetal pancre-
atic cells, J Clin Invest, 92, 1459-1466.
[29] T. Otonkoski, J. Ustinov, M. A. Huotari, E. Kallio, and P.
Hayry, (1997) Nicotinamide and sodium butyrate for the
induction of fetal porcine beta-cell differentiation prior to
transplantation, Transplant Proc,29, 2045.
[30] R. A. H. Perfetti, (2004) The role of GLP-1 in the life
and death of pancreatic beta cells, Horm Metab Res, 36,
804-810.
[31] K. A. S. D. Preissner, (1998) Role of vitrGonectin and its
receptors in haemostasis and vascular remodeling,
Thromb Res, 89, 1-21.
[32] R. P. Rafaeloff, G. L. Barlow, S. W. Qin, X. F. Yan, B.
Rosenberg, L. Duguid, and A. I. Vinik, (1997) Cloning
and sequencing of the pancreatic islet neogenesis associ-
ated protein (INGAP) gene and its expression in islet
neogenesis in hamsters, J Clin Invest, 99, 2100-2109.
[33] L. Rosenberg, (1995) In vivo cell transformation: neo-
genesis of beta cells from pancreatic ductal cells, Cell
Transplant, 4, 371-383.
[34] L. Rosenberg, R. Rafaeloff, , D. Kakugawa, Y. Pittenger,
A. I. G. Vinik, and W. P. Duguid, (1996) Induction of is-
let cell differentiation and new islet formation in the
hamster-further support for a ductular origin, Pancreas,
13, 38-46.
[35] E. A. P. Ruoslahti, (1987) New perspectives in cell adhe-
sion: RGD and integrins, Science, 238, 491-497.
[36] K. A. Y. F. Silver, (2001) ARIP cells as a model for pan-
creatic beta cell growth and development, Pancreas, 22,
141-147.
[37] S. E. Sipione, A. Lyon, J. G. Korbutt, and R. C. Bleack-
ley, (2004) Insulin expressing cells from differentiated
embryonic stem cells are not beta cells, Diabetologia.
[38] A. N. H. Suzuki and H. Taniguchi, (2003) Glucagon-like
peptide 1 (1-37) converts intestinal epithelial cells into
insulin-producing cells, Proc Natl Acad Sci U S A, 100,
5034-5039.
[39] G. L. Teitelman and D. J. Reis, (1987) Differentiation of
prospective mouse pancreatic islet cells during develop-
ment in vitro and during regeneration, Dev Biol, 120,
425-433.
[40] K. Terazono, H. Yamamoto, S. Takasawa, K. Shiga, Y.
Yonemura, Y. Tochino, and H. Okamoto, (1988) A novel
gene activated in regenerating islets, J Biol Chem, 263,
2111-2114.
[41] B. Thorens, (1992) Expression cloning of the pancreatic
beta cell receptor for the gluco-incretin hormone gluca-
gon-like peptide 1, Proc Natl Acad Sci U S A, 89,
8641-8645.
[42] B. E. A. O. Tuch, (1990) Maturation of insulinogenic
response to glucose in human fetal pancreas with retinoic
acid, Horm Metab Res Suppl, 25, 233-238.
[43] T. R. Y. Ulich, E. S. Cardiff, R. Yin, S. Bikhazi, N. Biltz,
R. Morris, and G. F. Pierce, (1994) Keratinocyte growth
factor is a growth factor for mammary epithelium in vivo:
The mammary epithelium of lactating rats is resistant to
the proliferative action of keratinocyte growth factor, Am
J Pathol, 144, 862-868.
[44] R. N. K. Wang, and L. Bouwens, (1995) Ductto islet-cell
differentiation and islet growth in the pancreas of
duct-ligated adult rats, Diabetologia, 38, 1405-1411.
[45] E. S. Y. Yi, S. Harclerode, D. L. Bedoya, A. Bikhazi, N.
B. Housley, R. M. Aukerman, S. L.Morris, C. F. Pierce,
and T. R. Ulich, (1994) Keratinocyte growth factor in-
duces pancreatic ductal epithelial proliferation, Am J
Pathol, 145, 80-85.
[46] M. E. C. Zenilman and T. H. Magnuson, (1998) Effect of
reg protein on rat pancreatic ductal cells, Pancreas, 17,
256-261.
[47] J. W. Zhou, X. Pineyro, and J. M. Egan, (1999) Gluca-
gon-like peptide 1 and exendin-4 convert pancreatic
AR42J cells into glucagon and insulin-producing cells,
Diabetes, 48, 2358-2366.