Content uploaded by Imane Song
Author content
All content in this area was uploaded by Imane Song on Oct 20, 2015
Content may be subject to copyright.
RESEARCH ARTICLE
Beta Cell Mass Restoration in Alloxan-
Diabetic Mice Treated with EGF and Gastrin
Imane Song
1
*, Oelfah Patel
2
, Eddy Himpe
1
, Christo J. F. Muller
2
, Luc Bouwens
1
1Cell Differentiation Lab, Vrije Universiteit Brussel (Brussels Free University), Brussels, Belgium,
2Diabetes Discovery Platform, South African Medical Research Council, Tygerberg, South Africa
*Imane.Song@vub.ac.be
Abstract
One week of treatment with EGF and gastrin (EGF/G) was shown to restore normoglycemia
and to induce islet regeneration in mice treated with the diabetogenic agent alloxan. The
mechanisms underlying this regeneration are not fully understood. We performed genetic
lineage tracing experiments to evaluate the contribution of beta cell neogenesis in this
model. One day after alloxan administration, mice received EGF/G treatment for one week.
The treatment could not prevent the initial alloxan-induced beta cell mass destruction, how-
ever it did reverse glycemia to control levels within one day, suggesting improved peripheral
glucose uptake. In vitro experiments with C2C12 cell line showed that EGF could stimulate
glucose uptake with an efficacy comparable to that of insulin. Subsequently, EGF/G treat-
ment stimulated a 3-fold increase in beta cell mass, which was partially driven by neogen-
esis and beta cell proliferation as assessed by beta cell lineage tracing and BrdU-labeling
experiments, respectively. Acinar cell lineage tracing failed to show an important contribu-
tion of acinar cells to the newly formed beta cells. No appearance of transitional cells co-
expressing insulin and glucagon, a hallmark for alpha-to-beta cell conversion, was found,
suggesting that alpha cells did not significantly contribute to the regeneration. An important
fraction of the beta cells significantly lost insulin positivity after alloxan administration, which
was restored to normal after one week of EGF/G treatment. Alloxan-only mice showed more
pronounced beta cell neogenesis and proliferation, even though beta cell mass remained sig-
nificantly depleted, suggesting ongoing beta cell death in that group. After one week, macro-
phage infiltration was significantly reduced in EGF/G-treated group compared to the alloxan-
only group. Our results suggest that EGF/G-induced beta cell regeneration in alloxan-diabetic
mice is driven by beta cell neogenesis, proliferation and recovery of insulin. The glucose-low-
ering effect of the treatment might play an important role in the regeneration process.
Introduction
Type 1 and type 2 diabetes result from inadequate beta cell mass, which leads to persistent
hyperglycemia. Restoration of beta cell mass by pancreas or islet cell transplantation can nor-
malize blood glucose levels [1–3]. However, donor shortage and the need of immunosuppres-
sion make transplantation therapy only available to a small number of diabetic patients. A very
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 1/17
OPEN ACCESS
Citation: Song I, Patel O, Himpe E, Muller CJF,
Bouwens L (2015) Beta Cell Mass Restoration in
Alloxan-Diabetic Mice Treated with EGF and Gastrin.
PLoS ONE 10(10): e0140148. doi:10.1371/journal.
pone.0140148
Editor: Ilse Rooman, Garvan Institute of Medical
Research, AUSTRALIA
Received: June 4, 2015
Accepted: September 21, 2015
Published: October 9, 2015
Copyright: © 2015 Song et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any
medium, provided the original author and source are
credited.
Data Availability Statement: All relevant data are
within the paper and its Supporting Information files.
Funding: This work was supported by Fund for
Scientific Research-Flanders (FWO; Fonds
Wetenschappelijk Onderzoek - Vlaanderen), URL:
http://www.fwo.be/. Grant number: G003110N.
Recipient: L.B. The funders had no role in study
design, data collection and analysis, decision to
publish, or preparation of the manuscript.
Competing Interests: The authors have declared
that no competing interests exist.
attractive possibility is the restoration of a functional beta cell mass by stimulating endogenous
regeneration of beta cells within the pancreas with pharmacological agents. To this end, drugs
should be developed that stimulate beta cell neogenesis, replication and/or survival. This could
offer a much more accessible therapy for both type 1 and type 2 patients, provided that in the
former, a way can be found to prevent autoimmune destruction of the regenerated beta cells.
Several candidate growth factors, hormones or cytokines have been already studied in the
context of beta cell regeneration [4–7]. In particular, the combination of gastrin hormone and
epidermal growth factor (EGF) was among the first combination of compounds that was pro-
posed to stimulate beta cell mass increase or regeneration in beta cell-depleted or autoimmune
diabetic mice and has been incorporated in clinical trials [8]. Gastrin and EGF combination
therapy was shown to revert hyperglycemia and increase beta cell mass in rodents [9–13]. Its
mode of action was proposed to include both a stimulation of beta cell replication and neogen-
esis from progenitor cells. However, the exact contribution of these two mechanisms to beta
cell mass expansion remains unclear and controversial in these studies and in many other
experimental models. More recently a genetic lineage tracing study confirmed the antidiabetic
action of gastrin/EGF and its effect on regenerating beta cell mass in alloxan-treated mice [10];
however the study failed to find evidence for a contribution of putative ductal progenitors to
beta cell regeneration. In the present study we tried to elucidate the cellular mechanisms that
contribute to beta cell regeneration in mice, using a model of severe beta cell injury by alloxan
followed by treatment with gastrin/EGF combination. Our main aim was to evaluate the rela-
tive importance of beta cell neogenesis in this model. To this end, we used the beta cell genetic
lineage tracing method, first described by Dor et al., which is generally accepted as the only
method allowing direct and unequivocal proof of beta cell neogenesis [14,15].
Materials and Methods
Animals and treatments
Male RIP-CreER;R26-Lox-STOP-Lox-LacZ (RIP-CreER/R26-LacZ) mice, provided by Dr.
Melton [14], and Ela-CreERT;R26-Lox-STOP-Lox-YFP (Ela-CreERT/R26-YFP) mice, pro-
vided by Dr. Stoffers [16], were housed in standard conditions with free access to food and
water. Animal procedures were approved by the ethical committee of the Vrije Universiteit
Brussel (permit number: LA1230277) and performed in accordance with the national guide-
lines and regulations.
Six to eight week old mice received 50 mg of tamoxifen (Sigma Aldrich), dissolved in 0.9%
NaCl and 10% EtOH, by oral gavage in three doses over a 5-day period (Fig 1). After a wash-
out period of 2 weeks, mice were randomly divided into three groups, namely control (CTRL),
alloxan only (ALX) and alloxan plus EGF/G (ALX+EGF/G). Mice in the two latter groups were
injected intravenously with alloxan (70 mg/kg; Sigma Aldrich). One day after alloxan adminis-
tration, EGF/G treatment was started (ALX+EGF/G) as previously described [12]. Mice were
euthanized by cervical dislocation on day 3 and day 8 post-alloxan.
Metabolic analysis
Blood samples from the tail vein were collected to measure glycemia using Glucocard Memory
strips (A. Menarini Diagnostics). Only mice that reached 11.1 mmol/l or more, one day after
alloxan administration, were included. Glucose concentrations exceeding the device’s upper
detection limit were scored as 33.3 mmol/l. For plasma C-peptide analysis, blood samples of
fed and fasted mice (overnight) were collected via cardiac puncture into tubes containing
EDTA-aprotinin and plasma was obtained by centrifugation. Plasma C-peptide concentration
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 2/17
was determined by radioimmunoassay (
125
I C-peptide; Millipore) following the manufacturer’s
recommendation.
Beta cell proliferation
To study beta cell proliferation, mice were continuously treated with BrdU (5-Bromo-2’-deox-
yuridine; Sigma Aldrich) via drinking water (1 mg/ml) from day 3 to day 8 post-alloxan.
Immunodetection of BrdU incorporation was performed on pancreatic sections.
Histochemistry of pancreatic sections
To determine the labeling efficiency for lineage tracing, X-gal and insulin staining were per-
formed on cryosections as previously described [15]. For beta cell tracing analysis, the fraction
of insulin-positive cells expressing X-gal (A = X-gal
+
INS
+
/INS
+
) and the fraction of X-gal-posi-
tive cells expressing insulin (B = X-gal
+
INS
+
/X-gal
+
) were counted for each mouse. At least
5000 insulin-positive cells and 2000 X-gal-positive cells were analyzed per group. Alloxan
administration resulted into a decrease in the fraction of X-gal-positive cells expressing insulin
due to degranulation or dedifferentiation of beta cells (see results). To evaluate the contribution
of beta cell neogenesis, beta cells that lost insulin positivity after alloxan were included in the
total beta cell population. The fraction of beta cells expressing X-gal was calculated as follow:
Xgalþbeta cells ¼XgalþINSþþXgalþINS
Where X galþINS¼A
Bð1BÞ
For immunohistochemistry on paraffin sections, pancreata were fixed in 4% formaldehyde for
4 h, dehydrated and embedded in paraffin. Sections of 4 μm were cut at an interval of 10 sections
and stained. Following primary antibodies were used: guinea pig anti-insulin pAb (1:3000; Van
Schravendijk), mouse anti-glucagon mAb (1:1000; G2654; Sigma Aldrich), rabbit anti-amylase
pAb (1:500; A8273; Sigma Aldrich), mouse anti-BrdU mAb (1:10; 11200; Pro-gen), goat anti-
GFP pAb (1:100; GTX26658; Genetex) and rat anti-F4/80 mAb (1:200; MCA497; Serotec). For
immunofluorescence, species-matched FITC- and Cy3-conjugated secondary antibodies (Jack-
son) were used, and counterstained with Hoechst (Sigma Aldrich). F4/80 staining was visualized
with DAB (Vector). For acinar cell tracing analysis, at least 10000 amylase-positive and 3800
insulin-positive cells were counted per group.For all other analyses, at least 3000 cells were
counted per group. Beta cell apoptosis was assessed by TUNEL staining using the In Situ Cell
Death Detection kit (Roche). Images were acquired with a Nikon Eclipse 90i fluorescence micro-
scope or Carl Zeiss multi-photon confocal laser scanning microscope LSM710.
Morphometry
Morphometric analysis of insulin- and glucagon-stained paraffin sections was performed to
measure beta cell and alpha cell mass, respectively. An area of at least 100 mm
2
was analyzed
Fig 1. Scheme of the experimental design.
doi:10.1371/journal.pone.0140148.g001
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 3/17
for each mouse. Cell masses were calculated by multiplying the relative insulin or glucagon-
positive cell area with the corresponding pancreatic weight. The number of insulin-positive
clusters per mm
2
pancreatic area was scored to assess the relative islet density. Beta cell size
was calculated by dividing the insulin-positive cell area by the number of nuclei in that area.
In vitro glucose uptake by C2C12 myocytes
Differentiated C2C12 myocytes, purchased from the European Collection of Cell Cultures
(ECACC; Catalogue No. 91031101), were exposed for three hours to EGF or gastrin or combi-
nations of EGF and gastrin prepared in Krebs-ringer bicarbonate HEPES buffer (KRBH) con-
taining 8 mM glucose. Vehicle controls were 0.5μM acetic acid for EGF and 0.000003% DMSO
for gastrin, respectively. After the three-hour incubation, glucose uptake in C2C12 cells was
determined using pulse-labeling with
3
H-2-deoxy-D-glucose (
3
H-2-DOG) in glucose-free
KRBH buffer containing EGF and/or gastrin for fifteen minutes. Liquid scintillation counting
was used to measure intracellular
3
H-2-DOG. Exposure of C2C12 cells to insulin (1μM) was
included as a positive control.
Statistical analysis
Data are presented as mean ± SEM and analyzed with GraphPad Prism using 1-way ANOVA
with Bonferroni post-hoc and Dunnett post-hoc test or 2-way ANOVA with Bonferroni post-
hoc test. A P value of less than 0.05 was considered to be statistically significant.
Results
Combined treatment with EGF and gastrin rapidly restores
normoglycemia following alloxan administration
From one day after alloxan injection, mice were treated with the combination of EGF and gas-
trin (EGF/G) for one week. Non-fasting glycemia was measured at different time points as
shown in Fig 2A. Glycemia rose rapidly within one day following alloxan administration, and
gradually increased to severely hyperglycemic values that persisted until the end of the experi-
ment. EGF/G treatment one day after alloxan administration, reversed hyperglycemia to nor-
mal within one day, and this persisted until the end of experiment. We did not observe
significant differences in body weight changes between ALX and ALX+EGF/G groups (Fig 2B).
Combined treatment with EGF and gastrin induces beta cell mass
regeneration between day 3 and day 8 post-alloxan
Insulin-positive beta cell mass was determined on day 3 and day 8 post-alloxan. As shown in
Fig 2C beta cell mass was reduced by more than 80% on day 3 in ALX and ALX+EGF/G
groups. No statistical significance was found between ALX and ALX+EGF/G groups, which
indicates that EGF/G treatment could not attenuate the alloxan-induced beta cell mass
reduction.
Beta cell mass reduction was reflected by a significant decrease in mean number of insulin-
positive cells per cluster and islet density (number of INS
+
clusters/mm
2
) in ALX and ALX+-
EGF/G groups (Fig 2D and 2E). The reduction was not associated with beta cell atrophy.
Indeed, the mean beta cell size was even slightly increased in both groups compared to controls
(Fig 2F). No significant difference was found between ALX and ALX+EGF/G groups.
At a later stage, between day 3 and day 8, we observed a 3-fold increase in beta cell mass in
ALX+EGF/G group, while beta cell mass of ALX group tended to further decrease (Fig 2C). No
further significant changes were observed in islet density and mean beta cell size (Fig 2E and 2F).
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 4/17
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 5/17
However, we did observe a significant increase in mean number of insulin-positive cells per clus-
ter in the ALX+EGF/G group (Fig 2D). These data indicate that the EGF/G-induced beta cell
mass regeneration resulted from an absolute increase in number of beta cells, leading to growth
of pre-existing islets.
EGF treatment enhances glucose uptake in muscle cells
Although normoglycemia was already achieved in the EGF/G-treated group on day 3, while the
ALX group remained hyperglycemic, beta cell mass in both groups did not differ. To understand
this discrepancy, the plasma C-peptide level was measured on day 3. In ALX+EGF/G group, plasma
C-peptide was found to be significantly reduced in non-fastedandfastedstatessuggestingabenefi-
cial effect of EGF/G treatment on glucose uptake following alloxan administration (Fig 2G and 2H).
To test the effect of EGF/G on glucose uptake in an in vitro model, we treated C2C12 myo-
cytes with EGF, gastrin or a combination of both factors. Glucose uptake activity of the cells
was measured for each condition and results show that EGF alone was able to significantly
increase the uptake with an efficacy comparable to that of insulin (Fig 3A). No significant glu-
cose uptake activity was observed with gastrin alone, and the combination of the two factors
showed no additional effect compared to EGF alone (Fig 3A and 3B).
Combined treatment with EGF and gastrin restores insulin expression in
pre-existing beta cells following alloxan administration
To follow beta cell fate and to evaluate the importance of beta cell neogenesis during the regen-
eration process, inducible genetic lineage tracing experiments were performed. RIP-CreER/
R26-LacZ mice were used in which tamoxifen administration leads to heritable lacZ gene
expression, which encodes for beta-galactosidase, in insulin-expressing beta cells present at the
time of administration and in their progeny. By performing a pulse-chase experiment, the fate
of pre-existing X-gal-labeled beta cells can be followed over time. In control group, roughly all
X-gal-labeled cells are insulin positive (98.5 ± 0.3%), which confirms the high specificity of X-
gal for beta cells (Fig 4A and 4C). Alloxan administration, with or without EGF/G treatment,
induced a significant loss of insulin positivity of approximately 20% in the X-gal-labeled beta
cell population on day 3. No significant difference was observed between ALX and ALX+EGF/
G groups. On day 8, X-gal
+
beta cells in ALX+EGF/G group recovered their insulin positivity
to almost control values, as opposed to ALX group. These data indicate that alloxan treatment
induces loss of insulin positivity in beta cells, either by degranulation or by dedifferentiation
towards a less mature phenotype and that one week of EGF/G treatment was able to normalize
the proportion of insulin-positive beta cells.
Contribution of beta cell neogenesis to beta cell mass regeneration
Next, we investigated the contribution of beta cell neogenesis using the above-mentioned line-
age tracing system. In order to determine the fraction of beta cells expressing X-gal
+
, we first
assessed the fraction of X-gal-labeled insulin-positive cells (Fig 4C and S1 Fig). To correct for
insulin-negative cells as a result of degranulation following alloxan, we included the insulin-
Fig 2. EGF/G treatment restores glycemia and stimulates beta cell regeneration after alloxan-induced ablation. Non-fasting glycemia (A) and body
weight (B) were monitored; n = 24–37 per group. (C) Beta cell mass, (D) number of insulin-positive cells per islet, (E) islet density (number of insulin-positive
clusters per mm
2
) and (F) beta cell size were assessed in CTRL, ALX and ALX+EGF/G on d3 and d8 post-alloxan. For histological analysis n=4–7 per group
per time point. (G) Fasted and fed plasma C-peptide levels (day 3) were significantly reduced in ALX+EGF/G compared to CTRL. (H) Fasting glycemia on
day 3; n= 3 per group per time point. Symbol *represents the statistical significance of each condition compared to CTRL. The horizontalbar denotes the
significant difference between the experimental groups. *,+, P <0.05; **,++, P <0.01; ***,+++, P <0.001.
doi:10.1371/journal.pone.0140148.g002
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 6/17
negative beta cell fraction (INS
-
X-gal
+
) (Material and methods for calculations; Fig 4B). In con-
trol group, 65.8 ± 5.3% of the beta cells were X-gal-labeled. Three days following alloxan
administration, the proportion of labeled beta cells tended to decrease in ALX and ALX+EGF/
G groups, but no significance was found compared to control. However, on day 8, the propor-
tion of labeled beta cells significantly decreased to 52.5 ± 4.1% and 75.9 ± 5.4% of control in
ALX and ALX+EGF/G groups, respectively. This indicates that a significant fraction of the beta
cell population in both groups (~47% in ALX group and ~25% in ALX+EGF/G group) origi-
nated from a non-beta cell source during the 8-day period in both groups.
Analysis of the X-gal-labeled insulin-positive clusters showed no indication of neoformation
of these clusters during the regeneration period and indicates that beta cell neogenesis primar-
ily occurred within, or immediately adjacent to, pre-existing islets (Fig 4D). These data suggest
that intra-islet precursor cells might be the source of neogenic beta cells, or that extra-islet pre-
cursor cells have migrated into the pre-existing islets to give rise to insulin-expressing cells.
No important contribution of acinar cells in beta cell neogenesis
To directly assess whether neogenic beta cells originated from acinar cells, acinar cell tracing
was performed with Ela-CreERT/R26-YFP mice. In these mice, after tamoxifen administration,
Fig 3. In vitro effect of EGF and/or gastrin on glucose uptake in C2C12 myocytes. The percentage glucose uptake activity when treated with EGF alone,
gastrin alone (A) or EGF+gastrin (B). Symbol *represents the statistical significance of each condition compared to positive control (1μM Ins); **,P<0.01;
***,P<0.001.
doi:10.1371/journal.pone.0140148.g003
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 7/17
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 8/17
acinar cells are heritably labeled with YFP. The proportion of YFP-labeled amylase-expressing
acinar cells was similar in control, ALX and ALX+EGF/G groups (Fig 5A). No significant
Fig 4. Beta cell lineage tracing: Contribution of beta cell neogenesis. (A) The percentage of X-gal-labeled cells that co-express insulin. Beta cell tracing
in RIP-CreER/R26-LacZ mice showed that almost all X-gal-positive cells are insulin positive in CTRL. A decrease of insulin positivity was observed in ALX
and ALX+EGF/G on d3. Insulin positivity was recovered in ALX+EGF/G on d8. (B) X-gal-labeling of total beta cells (percentage beta cells labeled with X-gal).
Total beta cell population includes INS
-
beta cell fraction. Percentage labeled beta cells in ALX and ALX+EGF/G was significantly reduced compared to
CTRL on d8. (C) Pancreata of RIP-CreER/R26-LacZ mice were stained for insulin and X-gal. (D) The percentage insulin-positive clusters labeled with X-gal.
The percentage X-gal-labeled insulin-positive clusters remained unchanged between d3 and d8. n = 4–6 per group per time point. Symbol *represents the
statistical significance of each condition compared to CTRL. The horizontal bar denotesthe significant difference between the experimental groups. +,
P<0.05; **,P<0.01; ***,+++, P <0.001.
doi:10.1371/journal.pone.0140148.g004
Fig 5. Acinar cell lineage tracing: No important contribution of acinar cells. (A) The percentage AMY
+
acinar cells labeled with YFP. Acinar cell tracing in
Ela-CreERT/R26-YFP mice showed comparable percentages YFP-labeled AMY
+
acinar cells in CTRL, ALX and ALX+EGF/G. (B) The percentage of
INS
+
beta cells labeled with YFP. Labeling index of beta cells show no contribution of acinar cells in newly generated beta cells. (C) Pancreata of Ela-CreERT/
R26-YFP mice were stained for insulin and YFP. n=3–4 per group. No statistical difference (P 0.05) was found between the conditions.
doi:10.1371/journal.pone.0140148.g005
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 9/17
increase was observed in the proportion of YFP-labeled beta cells in ALX and ALX+EGF/G
groups compared to control (Fig 5B and 5C). This indicates that acinar cells did not signifi-
cantly contribute to the newly generated beta cells in this experimental model.
No evidence for an important contribution of alpha cells in beta cell
neogenesis
Three days after alloxan administration islet architecture was severely disturbed, which could
not be prevented by EGF/G treatment (Fig 6A). This disorganization is most likely caused by
alloxan’s rapid ablation of beta cells, which resulted into the collapse of the central beta cell
core and the “appearance”of alpha cells within the islet core. Given this histological observa-
tion together with the finding that neogenic beta cells mainly reside within pre-existing islets,
we wanted to examine whether alpha cells could have contributed to beta cell neogenesis. The
appearance of bi-hormonal cells, representing an intermediate cell type, would be indicative
for this [17]. However, no significant changes were observed in the percentage of glucagon-
insulin co-expressing cells on day 3 and day 8 in ALX and ALX+EGF/G groups compared to
the control group (Fig 6B).
Analysis of alpha cell mass and number of alpha cells per islet showed no significant
changes in ALX and ALX+EGF/G groups compared to control (Fig 6C and 6D), which sug-
gests that the alpha cells residing in the core of the islets were pre-existing alpha cells that reor-
ganized from the periphery of the islet after alloxan induced beta cell depletion.
Combined treatment with EGF and gastrin promotes beta cell
proliferation
In order to determine whether beta cell proliferation contributes to beta cell mass regeneration,
a continuous BrdU-labeling experiment from day 3 to day 8 was performed. As shown in Fig
7A, ALX+EGF/G treatment induced a 3-fold increase in beta cell proliferation. However, a
much higher beta cell proliferation (6-fold) was observed in ALX group despite the low beta
cell mass, which suggests that increased beta cell death continued to occur between day 3 and
day 8 in the ALX group, and that this cell loss was compensated by proliferation and neogen-
esis. Consequently, it appears that EGF/G treatment induces beta cell mass regeneration by
promoting beta cell survival, which allows the accumulation of cells resulting from prolifera-
tion and neogenesis. In order to measure beta cell apoptosis, TUNEL staining was performed.
In control, ALX+EGF/G day 3 and day 8 mice, no TUNEL-positive beta cells could be detected,
whereas in ALX day 3 and ALX day 8 there were 0.21 ± 0.11% and 0.07 ± 0.07% TUNEL-posi-
tive beta cells, respectively. However, percentages remained extremely low and no significant
difference was found between the groups.
Reduced macrophage infiltration after EGF/G-induced beta cell mass
regeneration
Alloxan-induced beta cell ablation was associated with a significant increase in macrophage
infiltration in the pancreas as assessed by F4/80 immunostaining (Fig 7B). On day 3, ALX and
ALX+EGF/G groups showed a significant increase in infiltration compared to control group.
In ALX group, infiltration further increased until day 8, whereas in ALX+EGF/G group further
infiltration was not observed, but rather slightly diminished. Based on histological observa-
tions, we also observed a marked peri- and intra-islet infiltrate on day 3 in both ALX and ALX
+EGF/G groups, (Fig 7C). This was less pronounced in EGF/G-treated group on day 8.
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 10 / 17
Fig 6. No evidence for an important contribution of alpha cells. (A) IHC for insulin and glucagon. (B) The percentageof glucagon-positive cells that co-
express insulin. Transitional cells co-expressing glucagon and insulin were not observed. There were no significant differences in (C) alpha cell mass and (D)
number of alpha cells per islet in ALX and ALX+EGF/G on d3 and d8 compared to CTRL; n=4–5 per group per time point. No statistical difference (P 0.05)
was found between the conditions.
doi:10.1371/journal.pone.0140148.g006
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 11 / 17
Discussion
In previous studies, co-treatment with EGF and gastrin induced beta cell regeneration and nor-
malized glycemia in rodent models of chemically induced diabetes and in NOD mouse model
[10–13]. Treatment with either EGF or gastrin alone failed to induce regeneration and normali-
zation of glycemia. In these studies, indirect evidence suggested that EGF/G treatment could
stimulate beta cell regeneration via neogenesis, but this was not confirmed by genetic lineage
tracing, which is the only method allowing to assess neogenesis in a direct way. Our present
Fig 7. Beta cell proliferation and macrophage infiltration. (A) Continuous BrdU-labeling from day 3 to day 8 showed a significantincrease in proliferating
beta cells in ALX and ALX+EGF/G compared to CTRL. (B) Alloxan-induced beta cell ablation promotes F4/80
+
macrophage infiltration (number of F4/80
+
cells
per mm
2
) in ALX and ALX+EGF/G on d3. EGF/G suppresses further infiltration on d8. (C) Pancreata stained for insulin and F4/80 showed reduced peri- and
intra-islet infiltration in ALX+EGF/G on d8. n= 3 per group per time point. Symbol *represents the statistical significance of each condition compared to
CTRL. The horizontal bar denotes the significant difference between the experimental groups. *,+, P <0.05; **,++, P <0.01; ***,P<0.001.
doi:10.1371/journal.pone.0140148.g007
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 12 / 17
results confirmed the effect of EGF/G administration on alloxan-treated mice in inducing a
rapid normalization of glycemia within the first days of treatment. Thereafter a significant beta
cell mass increase occurred between 2 and 7 days of treatment, which was mainly due to
growth of pre-existing islets. The rapid glucose-lowering effect of the treatment, despite low
beta cell mass is possibly achieved by increasing glucose uptake in peripheral tissues. The lower
fasting and non-fasting plasma C-peptide levels in EGF/G-treated mice, measured on day 3
post-alloxan, indeed indicates improved peripheral glucose uptake. In vitro, EGF was found to
directly stimulate glucose uptake by C2C12 muscle cell line. This insulin-mimicking effect of
EGF and the role of EGFR on glucose uptake were also reported previously by several studies
[18–21].
The inducible Insulin-Cre/Lox genetic tracing method [14] and continuous BrdU-labeling,
respectively, revealed that both beta cell neogenesis and replication took place during the
regeneration of the beta cell mass in EGF/G-treated animals. In alloxan-only animals, that
remained hyperglycemic, neogenesis and proliferation were even more pronounced. The pro-
liferative effects noted in treated and untreated animals can be explained by previous studies
showing that EGF, gastrin and high glucose are mitogenic for beta cells [9,22–28]. In alloxan-
only mice, increased beta cell proliferation and neogenesis did not result in increased beta cell
mass, most likely because it was compensated by increased beta cell death. TUNEL assay failed
to show significant beta cell apoptosis in alloxan-only mice, although it tended to increase. It is
notoriously difficult to reliably detect and quantify apoptosis in beta cells in situ due to the
short duration of the process and rapid elimination of apoptotic cells, especially in alloxan-
treated mice where only a small beta cell number remains. Besides apoptosis as a mechanism of
alloxan-induced beta cell death, numerous researchers reported that alloxan mainly induces
beta cell loss by necrosis [29–32].
Analysis of pancreatic and islet infiltration of F4/80-positive macrophages showed that
alloxan administration promoted inflammation, which persisted until the end of the experi-
ment. Macrophage infiltration was less pronounced after one week of EGF/G treatment. This is
not necessarily a direct effect of the treatment, as it could also have resulted from the lower gly-
cemia and hence the protection against glucotoxicity. These observations may also indicate
ongoing beta cell damage and beta cell loss during persisting hyperglycemic conditions in the
alloxan-only group.
EGF/G treatment could also have reduced a possible beta cell dedifferentiation effect result-
ing from hyperglycemic conditions or it could have stimulated redifferentiation [33,34]. Our
beta cell tracing experiments rule out an underestimation of total beta cell population since we
could still recognize insulin-negative beta cells by X-gal staining. It revealed that about 20% of
beta cells had lost insulin expression after alloxan administration, as demonstrated immuno-
histochemically in X-gal
+
cells. This degranulation or dedifferentiation effect was restored to
normal after one week of EGF/G treatment, when the animals had restored beta cell mass. It
remains possible that the regranulation resulted from the glucose-lowering effect of the
treatment.
Since we observed beta cell neogenesis both in untreated hyperglycemic and in EGF/G-
treated normoglycemic mice, we wanted to find out which type of progenitors were involved.
We previously reported by genetic lineage tracing with Hnf1ß labeling, that duct cells did not
contribute to the beta cell mass in this experimental model [10]. In the present study we exam-
ined another possibility, namely that exocrine acinar cells might act as progenitors as was pre-
viously observed following treatment with EGF and CNTF of hyperglycemic mice [35].
However, lineage tracing in inducible Ela-CreERT/R26-YFP mice showed no important contri-
bution of acinar cells to the beta cell mass in the present experimental model. A third possibil-
ity that was considered is transdifferentiation of alpha cells to beta cells. However, double
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 13 / 17
immunohistochemical staining for glucagon and insulin could not reveal the presence of
“transitional”cells that are a hallmark for this type of conversion [17,36]. Insulin-expressing
multipotent progenitors have been described in adult pancreas [37]. However, in our study,
genetic lineage tracing showed that beta cell neogenesis resulted from an important fraction
of cells that did not express insulin and therefore caused a “dilution”of the pre-labeled beta
cells.
We can conclude that EGF/G-induced beta cell regeneration is accomplished by promoting
beta cell neogenesis, proliferation and (re-)differentiation or regranulation. An estimation of
the relative contribution to the beta cell mass of pre-existing and neogenic beta cells is repre-
sented in Fig 8. An important fraction of beta cells formed by neogenesis originated from a yet
unidentified type of beta cell progenitor that is distinct from duct, acinar, alpha and insulin-
positive progenitor cells. The glucose-lowering effect of the treatment might play an important
role in the regeneration of the beta cell mass, as it could relieve beta cell stress, allowing beta
cell (re-)differentiation or regranulation and the expansion of the beta cell mass resulting from
proliferation and neogenesis, as opposed to the alloxan-only group where no beta cell mass
expansion occurred despite high proliferation and neogenesis. We did observe the attenuation
of pancreatic and islet inflammation, which could have resulted from a direct EGF/G effect or
indirectly via its glucose-lowering effect, and hence, lowering beta cell stress. On the other
hand, it could also indicate better survival of the beta cells compared to the alloxan-only group
where inflammation persisted. Interestingly, the present study not only confirms that EGF/G
stimulates beta cell neogenesis, but also demonstrates the “spontaneous”occurrence of neogen-
esis in hyperglycemic alloxan-treated mice. It remains unclear whether hyperglycemia in itself
is responsible, directly or indirectly, for triggering this effect or whether alloxan treatment may
also influence the process of neogenesis via other effects.
Fig 8. Estimation of relative contribution to the beta cell mass of pre-existing and neogenic beta cells.
doi:10.1371/journal.pone.0140148.g008
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 14 / 17
Supporting Information
S1 Fig. X-gal-labeling of insulin-expressing beta cells. Symbol represents the statistical sig-
nificance of each condition compared to CTRL. The horizontal bar denotes the significant dif-
ference between the experimental groups. ,P<0.01; ,+++, P <0.001.
(PDF)
Acknowledgments
We thank Dr. Isabelle Houbracken and Dr. Josue Mfopou for their comments and discussions,
and William Rabiot, Emmy De Blay and Iris Mathijs for their assistance in this project. We
also thank Gaby Schoonjans for performing RIA on plasma samples.
Author Contributions
Conceived and designed the experiments: IS LB OP CM. Performed the experiments: IS OP
EH. Analyzed the data: IS OP EH LB CM. Wrote the paper: IS LB OP EH CM.
References
1. McCall M, Shapiro AM. Islet cell transplantation. Semin Pediatr Surg. 2014; 23(2):83–90. Epub 2014/
06/17. S1055-8586(14)00008-0 [pii] doi: 10.1053/j.sempedsurg.2014.03.006 PMID: 24931353.
2. Weir GC, Bonner-Weir S. Islet beta cell mass in diabetes and how it relates to function, birth, and death.
Ann N Y Acad Sci. 2013; 1281:92–105. Epub 2013/02/01. doi: 10.1111/nyas.12031 PMID: 23363033;
PubMed Central PMCID: PMC3618572.
3. Meier JJ. Beta cell mass in diabetes: a realistic therapeutic target? Diabetologia. 2008; 51(5):703–13.
Epub 2008/03/05. doi: 10.1007/s00125-008-0936-9 PMID: 18317728.
4. Song I, Muller C, Louw J, Bouwens L. Regulating the Beta Cell Mass as a Strategy for Type-2 Diabetes
Treatment. Curr Drug Targets. 2015. Epub 2015/02/06. CDT-EPUB-64988 [pii]. PMID: 25654737.
5. Ding L, Gysemans C, Mathieu C. beta-Cell differentiation and regeneration in type 1 diabetes. Diabetes
Obes Metab. 2013; 15 Suppl 3:98–104. Epub 2013/09/18. doi: 10.1111/dom.12164 PMID: 24003926.
6. Vasavada RC, Gonzalez-Pertusa JA, Fujinaka Y, Fiaschi-Taesch N, Cozar-Castellano I, Garcia-Ocana
A. Growth factors and beta cell replication. Int J Biochem Cell Biol. 2006; 38(5–6):931–50. Epub 2005/
09/20. S1357-2725(05)00251-7 [pii] doi: 10.1016/j.biocel.2005.08.003 PMID: 16168703.
7. Nielsen JH, Galsgaard ED, Moldrup A, Friedrichsen BN, Billestrup N, Hansen JA, et al. Regulation of
beta-cell mass by hormones and growth factors. Diabetes. 2001; 50 Suppl 1:S25–9. Epub 2001/03/29.
PMID: 11272193.
8. van Belle TL, Coppieters KT, von Herrath MG. Type 1 diabetes: etiology, immunology, and therapeutic
strategies. Physiol Rev. 2011; 91(1):79–118. Epub 2011/01/21. 91/1/79 [pii] doi: 10.1152/physrev.
00003.2010 PMID: 21248163.
9. Wang M, Racine JJ, Song X, Li X, Nair I, Liu H, et al. Mixed chimerism and growth factors augment beta
cell regeneration and reverse late-stage type 1 diabetes. Sci Transl Med. 2012; 4(133):133ra59. Epub
2012/05/11. 4/133/133ra59 [pii] doi: 10.1126/scitranslmed.3003835 PMID: 22572882.
10. Solar M, Cardalda C, Houbracken I, Martin M, Maestro MA, De Medts N, et al. Pancreatic exocrine duct
cells give rise to insulin-producing beta cells during embryogenesis but not after birth. Dev Cell. 2009;
17(6):849–60. Epub 2010/01/12. S1534-5807(09)00477-8 [pii] doi: 10.1016/j.devcel.2009.11.003
PMID: 20059954.
11. Suarez-Pinzon WL, Yan Y, Power R, Brand SJ, Rabinovitch A. Combination therapy with epidermal
growth factor and gastrin increases beta-cell mass and reverses hyperglycemia in diabetic NOD mice.
Diabetes. 2005; 54(9):2596–601. Epub 2005/08/27. 54/9/2596 [pii]. PMID: 16123347.
12. Rooman I, Bouwens L. Combined gastrin and epidermal growth factor treatment induces islet regenera-
tion and restores normoglycaemia in C57Bl6/J mice treated with alloxan. Diabetologia. 2004; 47
(2):259–65. Epub 2003/12/11. doi: 10.1007/s00125-003-1287-1 PMID: 14666367.
13. Brand SJ, Tagerud S, Lambert P, Magil SG, Tatarkiewicz K, Doiron K, et al. Pharmacological treatment
of chronic diabetes by stimulating pancreatic beta-cell regeneration with systemic co-administration of
EGF and gastrin. Pharmacol Toxicol. 2002; 91(6):414–20. Epub 2003/04/12. PMID: 12688387.
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 15 / 17
14. Dor Y, Brown J, Martinez OI, Melton DA. Adult pancreatic beta-cells are formed by self-duplication
rather than stem-cell differentiation. Nature. 2004; 429(6987):41–6. Epub 2004/05/07. doi: 10.1038/
nature02520 [pii]. PMID: 15129273.
15. Mathijs I, Houbracken I, Bouwens L. Genetic lineage tracing of beta cell neogenesis. Methods Mol Biol.
2012; 933:317–22. Epub 2012/08/16. doi: 10.1007/978-1-62703-068-7_21 PMID: 22893417.
16. Desai BM, Oliver-Krasinski J, De Leon DD, Farzad C, Hong N, Leach SD, et al. Preexisting pancreatic
acinar cells contribute to acinar cell, but not islet beta cell, regeneration. J Clin Invest. 2007; 117
(4):971–7. Epub 2007/04/04. doi: 10.1172/JCI29988 PMID: 17404620; PubMed Central PMCID:
PMC1838936.
17. Thorel F, Nepote V, Avril I, Kohno K, Desgraz R, Chera S, et al. Conversion of adult pancreatic alpha-
cells to beta-cells after extreme beta-cell loss. Nature. 2010; 464(7292):1149–54. Epub 2010/04/07.
nature08894 [pii] doi: 10.1038/nature08894 PMID: 20364121; PubMed Central PMCID: PMC2877635.
18. Gogg S, Smith U. Epidermal growth factor and transforming growth factor alpha mimic the effects of
insulin in human fat cells and augment downstream signaling in insulin resistance. J Biol Chem. 2002;
277(39):36045–51. Epub 2002/07/26. doi: 10.1074/jbc.M200575200 [pii]. PMID: 12138086.
19. Van Epps-Fung M, Hardy RW, Williford J, Gupta K, Wells A. Epidermal growth factor induces glucose
storage in transgenic 3T3-L1 adipocytes overexpressing epidermal growth factor receptors. Diabetes.
1996; 45(11):1619–25. Epub 1996/11/01. PMID: 8866569.
20. Ishii K, Kamohara S, Hayashi H, Todaka M, Kanai F, Imanaka T, et al. Epidermal growth factor triggers
the translocation of insulin-responsive glucose transporter (GLUT4). Biochem Biophys Res Commun.
1994; 205(1):857–63. Epub 1994/11/30. S0006-291X(84)72743-4 [pii] doi: 10.1006/bbrc.1994.2743
PMID: 7999123.
21. Inman WH, Colowick SP. Stimulation of glucose uptake by transforming growth factor beta: evidence
for the requirement of epidermal growth factor-receptor activation. Proc Natl Acad Sci U S A. 1985; 82
(5):1346–9. Epub 1985/03/01. PMID: 2983339; PubMed Central PMCID: PMC397257.
22. Tellez N, Joanny G, Escoriza J, Vilaseca M, Montanya E. Gastrin treatment stimulates beta-cell regen-
eration and improves glucose tolerance in 95% pancreatectomized rats. Endocrinology. 2011; 152
(7):2580–8. Epub 2011/05/12. en.2011-0066 [pii] doi: 10.1210/en.2011-0066 PMID: 21558313.
23. Yesil P, Lammert E. Islet dynamics: a glimpse at beta cell proliferation. Histol Histopathol. 2008; 23
(7):883–95. Epub 2008/04/26. PMID: 18437687.
24. Alonso LC, Yokoe T, Zhang P, Scott DK, Kim SK, O'Donnell CP, et al. Glucose infusion in mice: a new
model to induce beta-cell replication. Diabetes. 2007; 56(7):1792–801. Epub 2007/04/03. db06-1513
[pii] doi: 10.2337/db06-1513 PMID: 17400928; PubMed Central PMCID: PMC2921922.
25. Topp BG, McArthur MD, Finegood DT. Metabolic adaptations to chronic glucose infusion in rats. Diabe-
tologia. 2004; 47(9):1602–10. Epub 2004/09/07. doi: 10.1007/s00125-004-1493-5 PMID: 15349726.
26. Miettinen PJ, Ustinov J, Ormio P, Gao R, Palgi J, Hakonen E, et al. Downregulation of EGF receptor
signaling in pancreatic islets causes diabetes due to impaired postnatal beta-cell growth. Diabetes.
2006; 55(12):3299–308. Epub 2006/11/30. 55/12/3299 [pii] doi: 10.2337/db06-0413 PMID: 17130473.
27. Cras-Meneur C, Elghazi L, Czernichow P, Scharfmann R. Epidermal growth factor increases undiffer-
entiated pancreatic embryonic cells in vitro: a balance between proliferation and differentiation. Diabe-
tes. 2001; 50(7):1571–9. Epub 2001/06/26. PMID: 11423478.
28. Bonner-Weir S, Deery D, Leahy JL, Weir GC. Compensatory growth of pancreatic beta-cells in adult
rats after short-term glucose infusion. Diabetes. 1989; 38(1):49–53. Epub 1989/01/01. PMID: 2642434.
29. Rohilla A, Ali S. Alloxan induced diabetes: Mechanisms and effect. International Journal of Research in
Pharmaceutical and Biomedical Sciences. 2012; 3(2):819–23.
30. Lenzen S. The mechanisms of alloxan- and streptozotocin-induced diabetes. Diabetologia. 2008; 51
(2):216–26. Epub 2007/12/19. doi: 10.1007/s00125-007-0886-7 PMID: 18087688.
31. Szkudelski T. The mechanism of alloxan and streptozotocin action in B cells of the rat pancreas. Phy-
siol Res. 2001; 50(6):537–46. Epub 2002/02/07. PMID: 11829314.
32. Abdel-Rahman MS, Elrakhawy FI, Iskander FA. Protection of B cells against the effect of alloxan. Toxi-
col Lett. 1992; 63(2):155–64. Epub 1992/11/01. 0378-4274(92)90007-7 [pii]. PMID: 1455447.
33. Wang Z, York NW, Nichols CG, Remedi MS. Pancreatic beta cell dedifferentiation in diabetes and redif-
ferentiation following insulin therapy. Cell Metab. 2014; 19(5):872–82. Epub 2014/04/22. S1550-4131
(14)00114-4 [pii] doi: 10.1016/j.cmet.2014.03.010 PMID: 24746806.
34. Jonas JC, Sharma A, Hasenkamp W, Ilkova H, Patane G, Laybutt R, et al. Chronic hyperglycemia trig-
gers loss of pancreatic beta cell differentiation in an animal model of diabetes. J Biol Chem. 1999; 274
(20):14112–21. Epub 1999/05/13. PMID: 10318828.
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 16 / 17
35. Baeyens L, Lemper M, Leuckx G, De Groef S, Bonfanti P, Stange G, et al. Transient cytokine treatment
induces acinar cell reprogramming and regenerates functional beta cell mass in diabetic mice. Nat Bio-
technol. 2013; 32(1):76–83. Epub 2013/11/19. nbt.2747 [pii] doi: 10.1038/nbt.2747 PMID: 24240391.
36. Chung CH, Hao E, Piran R, Keinan E, Levine F. Pancreatic beta-cell neogenesis by direct conversion
from mature alpha-cells. Stem Cells. 2010; 28(9):1630–8. Epub 2010/07/24. doi: 10.1002/stem.482
PMID: 20653050.
37. Smukler SR, Arntfield ME, Razavi R, Bikopoulos G, Karpowicz P, Seaberg R, et al. The adult mouse
and human pancreas contain rare multipotent stem cells that express insulin. Cell Stem Cell. 2011; 8
(3):281–93. Epub 2011/03/03. S1934-5909(11)00016-6 [pii] doi: 10.1016/j.stem.2011.01.015 PMID:
21362568.
Beta Cell Regeneration after EGF and Gastrin
PLOS ONE | DOI:10.1371/journal.pone.0140148 October 9, 2015 17 / 17