Animal Models of DM2 With Reduced Pancreatic Β-cell Mass - 2006

8/10/2019 Animal Models of DM2 With Reduced Pancreatic -cell Mass - 2006

1/21

The International Journal of Biochemistry & Cell Biology 38 (2006) 873893

Review

Animal models of type 2 diabetes withreduced pancreatic-cell mass

Pellegrino Masiello

Dipartimento di Patologia Sperimentale, Biotecnologie Mediche, Infettivologia ed Epidemiologia,

University of Pisa, Via Roma 55, Scuola Medica, 56126 Pisa, Italy

Available online 4 October 2005

Abstract

Type 2 diabetes is increasingly viewed as a disease of insulin deficiency due not only to intrinsic pancreatic -cell dysfunction

but also to reduction of-cell mass. It is likely that, in diabetes-prone subjects, the regulated -cell turnover that adapts cell mass

to bodys insulin requirements is impaired, presumably on a genetic basis. We still have a limited knowledge of how and when

this derangement occurs and what might be the most effective therapeutic strategy to preserve -cell mass. The animal models of

type 2 diabetes with reduced -cell mass described in this review can be extremely helpful (a) to have insight into the mechanisms

underlying the defective growth or accelerated loss of-cells leading to the -cell mass reduction; (b) to investigate in prospective

studies the mechanisms of compensatory adaptation and subsequent failure of a reduced -cell mass. Furthermore, these models

are of invaluable importance to test the effectiveness of potential therapeutic agents that either stimulate -cell growth or inhibit

-cell death.

2005 Elsevier Ltd. All rights reserved.

Keywords: Type 2 diabetes; Animal models;-cell mass;-cell growth;-cell apoptosis

Contents

0. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 874

1. Animal models of type 2 diabetes with reduced-cell mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 876

1.1. Spontaneous or transgenic animal models of reduced-cell mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 876

1.1.1. Transgenic mice deficient in factors involved in pancreas development. . . . . . . . . . . . . . . . . . . . . . . . . . . . 876

1.1.2. Transgenic mice deficient in factors involved in-cell growth and/or survival . . . . . . . . . . . . . . . . . . . . . 877

1.1.3. Animal models with increased-cell apoptosis due to endoplasmic reticulum (ER) stress . . . . . . . . . . . 879

1.1.4. Animal models with increased-cell apoptosis due to islet amyloid production . . . . . . . . . . . . . . . . . . . . 880

1.1.5. Animal models with increased-cell apoptosis due to gluco- and/or lipotoxicity. . . . . . . . . . . . . . . . . . . 8801.1.6. Spontaneous animal syndrome of non-obese type 2 diabetes with reduction of-cell mass . . . . . . . . . . 881

1.2. Animal models of type 2 diabetes with experimentally induced reduction of-cell mass . . . . . . . . . . . . . . . . . . . . 881

1.2.1. Models with reduction of-cell mass induced by foetal growth retardation. . . . . . . . . . . . . . . . . . . . . . . . 881

1.2.2. Models with surgically induced reduction of-cell mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 882

Tel.: +39 050 221 8571; fax: +39 050 221 8557.

E-mail addresses:[email protected],

[email protected].

1357-2725/$ see front matter 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.biocel.2005.09.007


2/21

874 P. Masiello / The International Journal of Biochemistry & Cell Biology 38 (2006) 873893

1.2.3. Models with chemically-induced reduction of-cell mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 884

1.2.4. Effect of treatment with exendin-4 on animal models of type 2 diabetes with reduced-cell mass . . . 885

1.2.5. Advantages and limitations of the most commonly used animal models of type 2 diabetes with-cell

mass reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 887

2. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 887

Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 887

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 888

0. Introduction

Insulin insufficiency is a key feature in both type 1 and

type 2 diabetes mellitus. Type 1 diabetes is well recog-

nized as a condition of absolute insulin deficiency due to

massive autoimmune destruction of pancreatic -cells

(Mathis, Vence, & Benoist, 2001). Type 2 diabetes is

characterized by severalmetabolicdefects, among which

-cell secretory dysfunction and peripheral insulin resis-tance are considered as hallmarks of the disease in

humans (Kahn, 2003;Rizza and Butler, 1990; Weyer,

Bogardus, Mott, & Pratley, 1999).Usually, the disease

arises because of the progressive failure of endocrine

pancreas to adequately cope with the increased insulin

demand in insulin-resistant states, in particular obesity.

The results of the United Kingdom Prospective Diabetes

Study (UKPDS) clearly demonstrate that the progressive

nature of diabetes is an ongoing decline in -cell func-

tion without a change in insulin sensitivity (Kahn, 2001;

Matthews, Cull, Stratton, Holman, & Turner, 1998).Itis still debated whether this functional impairment is

due to reduced -cell mass or to an intrinsic secretory

defect of the -cells or both. Actually, it has been now

clearly established that most patients with type 2 dia-

betes, whether lean or obese, show a net decrease in

-cell mass (Butler et al., 2003a; Sakuraba et al., 2002;

Yoon, Ko, Cho, & Lee, 2003),but it remains uncertain

whether such reduction per se can account for inade-

quate insulin secretion. It is worth reminding that the

evaluation of-cell mass in man is done at autopsy and

does not derive from prospective studies. Thus, we do

not really know what is the -cell mass before the onsetof the disease.

Another intriguing aspect related to the pathogene-

sis of type 2 diabetes is the fact that albeit obesity is a

well known major risk factor for the development of the

disease (Burke et al., 1999;Center for Disease Control

and Prevention, 1997),two-thirds of obese subjects do

not actually develop diabetes (Mokdad et al., 2001),

likely because their-cell mass and insulin secretion can

permanently increase to compensate for the increased

metabolic load and insulin resistance. Conversely, in the

one-third of obese patients that progresses to type 2 dia-

betes, the same metabolic conditions finally result into-cell dysfunction and decrease in-cell mass (Lingohr,

Buettner, & Rhodes, 2002).The reasons why the same

environmental factors, i.e. nutritional excess with subse-

quentobesity andinsulin resistance, maylead to opposite

changes in the -cell mass and function in large popula-

tion groups are still unclear. In any case, the regulation

of-cell mass appears to play a pivotal role in the patho-

genesis of type 2 diabetes.Theoretically, as depicted in Fig. 1, -cell mass is

the result of the overall balance of-cell growth and-cell loss, depending on at least four mechanisms

(Bonner-Weir, 2000a,b;Rhodes, 2005): (1) replication of

existing differentiated-cells; (2) neogenesis of-cells

from precursors usually located in the pancreatic ductal

epithelium; (3)-cell size; (4) -cell death (by apopto-

sis and/or necrosis). The relative contribution of each

mechanism is variable, depending on various factors,

such as species, age, metabolic load, exposure to cyto-

toxic factors (Rhodes, 2005).Studies mainly conductedin rodents indicate that under normal circumstances -

cell growth and remodelling (this latter being marked by

apoptoticwaves), occurringduring foetal life (Boujendar

et al., 2002; Hanke, 2000),continue in the early postna-

tal period. Indeed, there is a burst of-cell replication

just after birth, followed by a rise in -cell neogene-

sis and a transient wave of apoptosis between 1 and 3

Fig. 1. Factors regulating -cell mass. On the left are the mechanisms

that expand the system (replication, i.e. proliferation of pre-existing

differentiated -cells; neogenesis, i.e. production of new -cells from

undifferentiated precursors located in the pancreatic duct epithelium

or scattered in the exocrine pancreas; hypertrophy, i.e. increased cell

size). On the right are the mechanisms that reduce the system (-cell

death by apoptosis or necrosis; hypotrophy, i.e. reduced cell size).


3/21

P. Masiello / The International Journal of Biochemistry & Cell Biology 38 (2006) 873893 875

weeks of age, during which-cell mass does not increase

despite ongoing enhanced rates of -cell replication

and neogenesis (Bonner-Weir, 2000c; Finegood, Scaglia,

& Bonner-Weir, 1995; Scaglia, Cahill, Finegood, &

Bonner-Weir, 1997).This physiological wave of devel-

opmental-cell apoptosis is coincident with a decline in

the islet expression of insulin-like growth factor-II (IGF-II), likely acting as an essential -cell survival factor

during foetal and neonatal life (Hill et al., 2000;Petrik,

Arany, McDonald, & Hill, 1998; Petrik et al., 1999), and

should mark for -cells a change from a proliferative,

poorly functional foetal phenotype to a fully differenti-

ated, highly functional adult phenotype (Hellerstrom &

Swenne, 1985). In any case, the overall net effect, at least

in rodents, is a marked increase in-cell growth in the

early postnatal period and during weaning. Thereafter,

the rates of-cell replication, neogenesis and apoptosis

stabilize at low levels. In humans, although it is obvi-ously more difficult to study in details the physiological-cell development and turnover than in rodents, there

are nevertheless enough indications thatsimilar phenom-

ena occur (Bonner-Weir, 2000c; Butler et al., 2003a;

Kloppel, Lohr, Habich, Oberholzer, & Heitz, 1985),

including waves of-cell developmental apoptosis, as

reported in the third trimester of foetal life (Tornehave

& Larsson, 1997).

In the adult, -cells have an estimated life span of 60

days (Bonner-Weir, 2000b),with a very slow turnover

involving about 0.5% of the-cell population undergo-ing mainly self-replication (Dor, Brown, Martinez, &

Melton, 2004)and a corresponding 0.5% entering apop-

tosis (Bonner-Weir, 2000c; Lingohr et al., 2002).Thus,

the-cell mass remains relatively constant under phys-

iological conditions during adult life.

A reduction in -cell mass may originate by either

an impairment of -cell replication/neogenesis or an

increased -cell apoptosis or both combined, induced

by genetic or acquired factors acting prenatally and/or

postnatally (see below). On the other hand, the plas-

ticity of-cell mass is well known (Bernard-Kargar &

Ktorza, 2001).As a remarkable example, during preg-nancy, -cell hyperplasia and hypertrophy occur, driven

by the pregnancy hormonesprolactinand placentallacto-

gen (Sorenson & Brelje, 1997),to ensure compensation

for the additional metabolic requirements of the grow-

ing foetus, while in the post-partum a decreased -cell

replication and a concomitant increase in -cell apopto-

sis result in prompt normalization of-cell population

(Scaglia, Smith, & Bonner-Weir, 1995).

Compensatory increases in -cell mass, depend-

ing on both replication/neogenesis and hypertrophy of

-cells, are also observed in obesity and other con-

ditions of insulin resistance in experimental animals,

such as Zucker fatty rats (Pick et al., 1998), and in

humans as well (Butler et al., 2003a; Kloppel et al.,

1985),likely driven by mild hyperglycaemic excursions

(Bonner-Weir, 2000c;Bernard-Kargar & Ktorza, 2001).

However, as reminded above, in one third of obese

patients, a subsequent failure of compensation occurs,mainly due to a markedly increased frequency of-cell

apoptosis, ultimately resulting in a progressive relative

decrease of-cell mass and development of diabetes

(Butler et al., 2003a; Lingohr et al., 2002).Among the

factors which might favour such cell loss, prolonged

overstimulation of -cells may play a relevant role,

promoting apoptosis by various mechanisms, such as

protein overload and consequent increased endoplasmic

reticulum(ER)stress( Araki,Oyadomari,&Mori,2003 ),

amyloid deposition (Butler et al., 2003a)and also long-

term increases in cytosolic Ca2+ (Grill & Bjorklund,2001). Furthermore, hypertrophic -cells are consid-

ered more vulnerable to stress (Bonner-Weir, 2001).It is

also well established that -cell dysfunction and loss, at

least in obesity-linked diabetes, can be due to the toxic

effects of prolonged hyperglycaemia, i.e. glucotoxic-

ity (Donath, Gross, Cerasi, & Kaiser, 1999; Jonas et al.,

1999;Kahn, 2001),or hyperlipidaemia, i.e. lipotoxic-

ity (McGarry & Dobbins, 1999;Shimabukuro, Zhou,

Levi, & Unger, 1998),or both, i.e. glucolipotoxicity

(El-Assaad et al., 2003; Poitout & Robertson, 2002;

Prentki & Corkey, 1996).Various downstream mecha-nisms driven by exposure to chronic high glucose and/or

lipid levels have been suggested to alter -cell gene

expression and metabolism and increase -cell apopto-

sis. These mechanisms include enhanced generation of

reactive oxygen species (Ihara et al., 1999; Laybutt et al.,

2002); -cell production of pro-apoptotic interleukin-1

with NF-kB activation and Fas signalling (Maedler et

al., 2002);FFA- or lipoprotein-induced apoptotic path-

ways, mediated by ceramide accumulation (Unger &

Orci, 2002)and activation of c-Jun N-terminal kinase

JNK (Roehrich et al., 2003), respectively. An intrigu-

ing link between ER stress and glucolipotoxicity, viaactivation of the lipogenic transcription factor SREBP-

1c (steroid-response-element-binding-protein), has been

very recently reported in a -cell line and in rat islets

exposed to chronic high glucose levels (Wang, Kouri, &

Wollheim, 2005).

Alternatively or additionally to glucolipotoxicity, it

can be also hypothesized that an early reduction in -

cell mass, deriving from a defective prenatal or post-

natal developmental -cell growth, might be a major

predisposing factor for type 2 diabetes. A simple yet

interesting possibility to explain the different final out-


4/21


comes of metabolic overload in different groups of obese

patients may be a different pre-existing amount of-cell

mass. In humans, a low-birth weight, most likely depen-

dent on unfavourable intrauterine environment, has been

indeed associated with proneness to type 2 diabetes later

on in life (Barker et al., 1993;Hales & Barker, 2001).

In a smaller neonate, a proportionally smaller -cellmass may grow, thereby having less capacity to expand

in response to increased insulin demand, as in obesity

or pregnancy, with final outcome in obesity-associated

type 2 diabetes or gestational diabetes (Hales & Barker,

2001).Furthermore, an early loss of-cell mass might

subsequently favour dysfunction of the residual -cells,

possibly due to overstimulation or toxic effects of even

mild chronic hyperglycaemia and/or hyperlipidaemia

(Donath & Halban, 2004).Of course, it would be very

helpful to monitor longitudinally the changes in -cell

mass throughout life and in various pathophysiologi-cal conditions. Unfortunately, whereas several tests exist

for the determination of insulin secretory capacity, in

vivo measurement of-cell mass in humans is currently

not possible. Target-specific imaging probes have been

recently developed in rodents by using -cell-specific

monoclonal antibodies modified for nuclear imaging,

which may allow non-invasive assessmentof-cellmass

(Moore, Bonner-Weir, & Weissleder, 2001). The use

of tritiated d-mannoheptulose, administered in vivo to

preferentially label hepatocytes and insulin-producing

cells, has been also proposed, taking advantage of theselective GLUT-2 mediated transport of this heptose

into hepatocytes and -cells, but not other cell types

(Malaisse, 2001, 2005).While waiting for such or other

methodologies to be fully established, at present we can

only rely on functional measures that indirectly reflect

-cell mass. In animal studies, for instance, correla-

tions between in vivo functional tests (e.g., acute insulin

response) and actual -cell mass have been reported

in primates (McCulloch, Koerker, Kahn, Bonner-Weir,

& Palmer, 1991)and minipigs (Larsen, Rolin, Wilken,

Carr, & Gotfredsen, 2003b).Also in humans, functional

insulin secretory reserve as assessed by intravenous glu-cose and/or arginine could be a possible tool for predict-

ing -cell mass, as reported in transplantation studies

(Teuscher et al., 1998).

In this context, animal models of type 2 diabetes

with reduced -cell mass can be extremely helpful for

a number of reasons: (a) they can reveal pathogenic

mechanisms of the -cell mass reduction; (b) they can

contribute to clarify if a reduced -cell mass is a pre-

requisite for type 2 diabetes development; (c) they can

contribute to assess mechanisms of compensation and

subsequent failure in the residual -cell mass; (d) they

offer opportunities to explore new approaches to treat-

ment of diabetes.

In any case, we should be aware that none of the

animal syndromes of type 2 diabetes can reproduce the

complexity of the human disease; nevertheless, each one

can be helpful for understanding at least some aspects

of the pathogenesis and evolution of the disease. Actu-ally, we should consider these experimental syndromes

as models of diabetogenesis rather than models of

diabetes.

This consideration appears particularly appropriate

for the diabetic syndromes with reduced beta cell mass

described in this review, as most of them lack one of

the major features of human type 2 diabetes, i.e. insulin

resistance, but at the same time provide important clues

to establish the impact of reduced-cell mass as a rele-

vant pathogenic mechanism of the disease.

1. Animal models of type 2 diabetes with

reduced-cell mass

1.1. Spontaneous or transgenic animal models of

reduced-cell mass

These models are particularly interesting to provide

insight into the pathogenic mechanisms of-cell mass

reduction.

1.1.1. Transgenic mice deficient in factors involvedin pancreas development

1.1.1.1. Pdx-1 +/ mice; -cell-specific Pdx-1 /

mice. Pancreatic duodenal homeobox factor 1 (Pdx-1),

also known as Ipf-1 in humans, is a homeodomain-

containing transcription factor required for the develop-

ment of the pancreas andother foregutstructures (Offield

et al., 1996).Humans and mice that do not express the

Pdx-1 gene exhibit pancreatic agenesis and congenital

diabetes (Jonsson, Carlsson, Edlund, & Edlund, 1994;

Stoffers, Zinkin, Stanojevic, Clarke, & Habener, 1997b).

Mutations of Pdx-1/Ipf-1 in the heterozygous state are

associated with one of the six known genetic forms ofmaturity onset diabetes in the young (MODY), namely

MODY-4 (Stoffers, Ferrer, Clarke, & Habener, 1997a),

or with a small fraction of patients with typical adult-

onset type 2 diabetes (Hani et al., 1999; MacFarlane et

al., 1999).Located preferentially but not exclusively in-cells, Pdx-1 has also a relevant role in the later dif-

ferentiation of-cells, as it transcriptionally regulates

a number of-cell genes, including insulin, glucoki-

nase, GLUT-2, prohormone convertases PC 1/3 and PC

2, as well as fibroblast growth factor receptor-1 (FGFR-

1) (Edlund, 1998, 2001a, 2001b). Pdx-1 mediates in


5/21


particular glucose-stimulated insulin gene transcription

(Marshak, Totary, Cerasi, & Melloul, 1996)and FGRF-

1 signalling that appears to be necessary for the nor-

mal expression of GLUT-2 and prohormone convertases

(Hart et al., 2000).

Transgenic mice with haploid insufficiency of Pdx-1

(Pdx-1 +/) have worsening glucose tolerance with ageandreduced insulin release in vitro. Both-cellmassand

islet number are similar to controls in young animals, but

fail to increase with age and are approximately 50% less

with respect to controls by 1 year (Johnson et al., 2003).

Pdx +/ mice show increased islet apoptosis, associated

with abnormal islet architecture and lymphocyte infil-

tration, whereas isolated islets and dispersed -cells are

functionally normal (Johnson et al., 2003).

The mechanism by which Pdx-1 haploinsufficiency

increases -cell apoptosis is unclear, but it might be

related to the prosurvival role of insulin receptor sub-strate 2 (IRS-2) signalling (see below). It has also been

suggested (Kulkarni et al., 2004) that cell death might be

secondary to a failure of Pdx-defective-cells to expand

in response to appropriate stimuli (e.g., during compen-

sation for insulin resistance), similarly to the apoptotic

death triggered in differentiated neuronal cells following

an abortive attempt at entering the cell cycle (Becker &

Bonni, 2004).

The relevance of Pdx-1 for-cellgrowth and function

is confirmed by the finding that -cell-specific inacti-

vation of Pdx-1 gene in mice results in 40% reductionin the -cell number, impaired expression of insulin,

glucokinase and GLUT-2 in -cells with development

of diabetes at 1015 weeks of age (Ahlgren, Jonsson,

Jonsson, Simu, & Edlund, 1998; Hart, Baeza, Apelqvist,

& Edlund, 2000).In these mice, Pdx-1 / -cells also

show downregulation of FGFR-1 and prohormone con-

vertases, leading to impaired insulin processing (Hart et

al., 2000). Interestingly, the phenotype of-cell-specific

Pdx-1 / mice could be reproduced in a transgenic

mouse line expressing a dominant-negative version of

FGRF1, in which diabetes develops at 15 weeks of age,

with reduced-cell number and impaired expression ofGLUT-2 and PC 1/3 but not glucokinase (Hart et al.,

2000).

Another transgenic mouse, with disruption of the

gene for the hepatocyte nuclear factor-1 (HNF-1),

mimics MODY-3, a rare form of MODY caused by

mutations of this gene (Fajans, Bell, & Polonsky, 2001;

Owen & Hattersley, 2001).In HFN-1 /mice, like

in MODY-3, there is a defective secretory response to

glucose and arginine, but the -cell mass, adjusted for

the reduced body weight of these animals, is notdifferent

from wild type controls (Pontoglio et al., 1998).Thus,

HFN-1 appears to be mainly involved in the regulation

of-cell differentiation rather than-cell mass.

1.1.2. Transgenic mice deficient in factors involved

in-cell growth and/or survival

1.1.2.1. Insulin receptor substrate-2 (IRS-2) / mice.

Disruption of murine insulin receptor substrate-1 (IRS-1) or insulin receptor substrate-2 (IRS-2) causes severe

insulin resistance(Witherset al., 1998), as expectedfrom

the role of these key adaptor molecules for insulin sig-

nalling in insulin target tissues (Rhodes & White, 2002).

However, the IRS-1/mice do not become diabetic,

because the -cell mass increases to compensate for

insulin resistance (Withers et al., 1998).In contrast, the

IRS-2/ mice (C57Bl-6/129sv hybrid genetic back-

ground) become profoundly diabetic with fasting hyper-

glycaemia worsening progressively, and attaining more

than 400 mg/dl at the age of 1216 weeks (Withers etal., 1998). Disruption of IRS-2 on a different hybrid

genetic background (C57Bl-6/CBA) confirms these fea-

tures, although the severity of diabetes is less striking

(Kubota, Tobe, Terauchi, & Eto, 2000).In IRS-2 /

mice, the -cell mass fails to expand in compensation

for the insulin resistance and undergoes a progressively

increasing frequency of apoptosis (Withers et al., 1998,

1999).

It is worthwhile to note that Pdx-1 expression is

unchanged in IRS-1 /mice (Kulkarni et al., 2004),

whereas it is reduced in IRS-2 / mice (Kushner et al.,2002),likely due to the inhibiting effect of the transcrip-

tion factor Foxo-1, that in the absence of the IRS-2 sig-

nalling is not phosphorylated and is largely localized in

the nucleus (Kitamura et al., 2002).Foxo-1 is ordinarily

phosphorylated by protein kinase B (PKB, also known

as Akt) (downstream of IRS-2/phosphatidylinositol 3-

kinase (PI3K) pathway), thereby being retained in the

cytosol (Brazil & Hemmings, 2001). Therelevanceof the

link between IRS-2 and Pdx-1 is supported by the reports

that either transgenic Pdx-1 over-expression (Kushner

et al., 2002) or disruption of a single Foxo-1 allele

(Kitamura et al., 2002)rescues -cell mass and functionin IRS-2 / mice, normalizing glucose homeostasis.

On the other hand, very interestingly, the marked com-

pensatory islet hyperplasia occurring in insulin-resistant

IRS-1 / or double IR/IRS-1 / mice is severely

restricted by Pdx-1 haploinsufficiency and an increased-cell apoptosis is instead observed (Kulkarni et al.,

2004).

The absence of IRS-2 expression in cultured -cells

also causes marked spontaneous apoptosis (often asso-

ciated with increased levels or activities of pro-apoptotic

factors like BAD), and reduction of -cell survival


6/21


(Lingohr et al., 2003; Withers et al., 1999). This is

consistent with the involvement of IRS-2/PI3K/PKB sig-

nalling in the phosphorylation and induction of several

anti-apoptotic substrates (activation of the ubiquitin lig-

ase Mdm2 and subsequent degradation of p53; Mayo

& Donner, 2001),enhanced efficiency of the apoptotic

inhibitor XIAP (Dan et al., 2004), inactivation of the pro-apoptotic factor BAD (Chan, Rittenhouse, & Tsichlis,

1999).

Thus, IRS-2 and its downstream elements, activated

in response to not only insulin and IGF-1 (Withers et al.,

1998, 1999)but also GLP-1 signalling through cAMP-

response element binding protein (CREB)-mediated

pathways (Jhala et al., 2003),are critically important for

regulating -cell mass in adaptation to metabolic home-

ostasis, especially by promoting-cellsurvival(Dickson

& Rhodes, 2004; Hennige et al., 2003). An acquired

defect of IRS-2 signalling pathway in -cells could ulti-mately result into -cell loss and onset or progression of

type 2 diabetes (Rhodes & White, 2002;White, 2002).

A possible mechanismof IRS-2 acquireddefect might

be an enhancement of IRS-2 ubiquitination and protea-

somal degradation induced by chronic hyperglycaemia

and/or hyperlipidaemia, or by cytokines produced in

islets (Maedler et al., 2002) or derived from the expanded

adipose tissue in obese patients (Trayhurn & Wood,

2004), as recently reviewed by Dickson and Rhodes

(2004)and byRhodes (2005).

1.1.2.2. Transgenic mice with dominant-negative CREB

in -cells. As CREB phosphorylation, induced by

glucagon-like peptide 1 (GLP-1) or other cAMP ago-

nists, promotes insulin and IRS-2 gene expression

(Dumonteil & Philippe, 1996; Jhala et al., 2003), another

mechanism of poor IRS-2 signalling might be linked to

the impairment of this pathway. Indeed, in transgenic

mice expressing the dominant-negative CREB inhibitor

A-CREB in -cells, expression of IRS-2 is severely

blunted and a marked reduction in-cell mass secondary

to -cell apoptosis develops, with progressively increas-

ing serum glucose and decreasing serum insulin levels(Jhala et al., 2003).

1.1.2.3. PKB- / mice. As already mentioned, there

are evidences that of the two major signalling path-

ways downstream of IRS-2, PKB activation plays a

crucial role in -cell survival (Lingohr et al., 2002;

Wang et al., 2004; Wrede et al., 2002), with negli-

gible contribution from ERK1/2 activation (Lingohr

et al., 2003). Thus, transgenic expression of a con-

stitutively active variant of PKB in -cells prevents

FFA-induced apoptosis (Wrede et al., 2002)and exerts

a protective effect against streptozotocin-induced dia-

betes by enhancing -cell mass mainly through increase

of-cell survival and -cell size, without significant

effect on -cell replication or neogenesis (Tuttle et

al., 2001). Surprisingly, ablation of the PKB- iso-

form has not consistently reproduced the phenotype

of IRS-2 / mice with regard to -cell mass. Twolines of PKB- / mice, established on different

genetic backgrounds, both show peripheral insulin resis-

tance, mild/moderate hyperglycaemia and glucose intol-

erance, associated with basal hyperinsulinaemia (Cho

et al., 2001; Garofalo et al., 2003).However, in one of

them (C57Bl-6/129 hybrid genetic background), pancre-

atic -cell mass is increased, suggesting -cell com-

pensation for insulin resistance (Cho et al., 2001),

whereas in the other (DBA/1lacJ inbred genetic back-

ground) it is decreased due to enhanced -cell apop-

tosis, although not as much as in IRS-2 / mice(Garofalo et al., 2003).These data underline the com-

plexity of the factors regulating-cell mass and might

also be indicative of the involvement of other isoforms

of PKB in -cell growth and survival (Tuttle et al.,

2001).

Many PKB substrates can affect -cell size, replica-

tion, differentiation and survival, as reviewed by Dickson

and Rhodes (2004). Here I briefly make reference to

some of them, whose ablation results in animal models

of diabetes or impaired glucose tolerance with reduction

of-cell mass.

1.1.2.4. p70S6K1 / mice. PKB is involved in

enhancement of protein synthesis in -cells through

activation of mTOR (mammalian target of rapamycin),

which in turn phosphorylates and activates two fac-

tors regulating protein synthesis, i.e. 4E-BPI (eukary-

otic initiation factor-binding protein-1 or PHAS-1) and

p70S6K (the70-kDa ribosomalsubunitS6 protein kinase)

(McDaniel, Marshall, Pappan, & Kwon, 2002). PKB can

also increase protein synthesis by inactivating GSK3

and hence relieving inhibition of the protein synthesis

initiation factor eIF2B (Cohen & Frame, 2001). Therelevance of this pathway for regulation of-cell size

is confirmed by the phenotypes of two types of trans-

genic mice: (a) mice expressing a constitutively active

PKB specifically in -cells which actually show an

increase in-cell size (Tuttle et al., 2001);(b) p70S6K1

null mice, that have a decrease in -cell mass due to

a selective reduction of-cell size (-cells and other

endocrine cells are not affected) and not of-cell num-

ber (Diehl, Cheng, Roussel, & Sherr, 1998;Pende et al.,

2000). The p70S6K1 / mice are glucose-intolerant and

hypoinsulinaemic (Diehl et al., 1998; Pende et al., 2000);


7/21


no increase in -cell apoptosis is observed (Pende et al.,

2000).

1.1.2.5. Cyclin D2/mice and Cdk-4/mice. -

cell growth might also be mediated by PKB-dependent

inhibition of GSK3 and subsequent poor phosphoryla-

tion of cyclin D that would prevent its degradation andpromote mitogenesis (Dickson & Rhodes, 2004).Actu-

ally, transgenic disruptionof either cyclin D, in particular

cyclin D2, or partner cyclin-dependent protein kinase

Cdk-4, causes insulin-deficient diabetes by a marked

decrease in-cell mass (Rane & Reddy, 2000).

Cyclin D2 / mice, obtained by homologous

recombination in embryonic stem cells, show nor-

mal prenatal development of endocrine pancreas, but

impaired post-natal replication of- and -cells, so that

by postnatal day 14, their total -cell mass is about 30%

of that of controls (Georgia & Bhushan, 2004). Ani-mals size, body and pancreas weights and replication of

exocrine and ductal cells are not affected. The selective

replication defect in the endocrine pancreas is surpris-

ing because in most cell types upregulation of the other

D-cyclins can compensate for the loss of any particular

D-cyclin (Ciemerych et al., 2002). The failure to upregu-

late other D-cyclins in the-cells of cyclin D2 / mice

is limited, however, to the early postnatal period since

cyclin D1 expression is readily observed in islet cells

2 weeks after birth. The metabolic phenotype of cyclin

D2/mice is not severe: moderate fasting hypergly-caemia, glucose intolerance and reduced post-loading

insulin peak are observed in 12-week-old animals. Thus,

impairment in cyclin D2 function might well determine

an early defect in-cell mass.

Cdk-4 / mice also display a defect in-cell prolif-

eration, indicating that Cdk-4 acts as the requisite partner

for cyclin D2 in controlling cell cycle progression in -

cells (Rane et al., 1999).However, the Cdk-4 / mice,

while showing a similar insulin-deficient diabetes with a

marked decrease in -cell mass as cyclin D2 / mice,

are 40% smaller than wild type mice, suggesting that

Cdk-4 plays a more general role and is required for nor-mal cell growth in most tissues and organs (Martin et al.,

2003).Re-expression of endogenous Cdk-4 in -cells of

Cdk-4 / mice restores -cell proliferation and nor-

moglycaemia (Martin et al., 2003).

Independently on the lack of these physiological

factors implicated in pancreas development and -cell

growth, differentiation and survival, reduction of-cell

mass can be also the result of a marked enhancement of

-cell apoptosis due to other pathogenic mechanisms,

as indicated by a number of other interesting animal

syndromes.

1.1.3. Animal models with increased-cell

apoptosis due to endoplasmic reticulum (ER) stress

1.1.3.1. Akita mice. The Akita mouse harbors a sponta-

neous mutation in the INS2 gene that leads to the produc-

tion of a mutant form of proinsulin 2 and causes early-

onset non-obese diabetes with a decreased -cell mass.

This decrease occurs progressively from birth and hasbeen suggested to be due to ER stress induced by reten-

tion of the mutant malfolded proinsulin 2 in the ER and

subsequent-cell dysfunction and death (Oyadomari et

al., 2002; Wang et al., 1999).The ER of affected -cells

distends and contains elevated levels of stress markers,

such as the molecular chaperone Bip/Grp78 and acti-

vated CHOP, which may induce apoptosis (Oyadomari

et al., 2002). Hyperglycaemia and falling insulin pro-

duction correlate with a progressive decrease in -cell

mass. When the Akita mutation was introduced into

a CHOP / background, islet cell destruction andhyperglycaemia were delayed in onset (Oyadomari et

al., 2002).

A similar condition is supposed to occur in the Wol-

fram syndrome, a rare genetic form of human diabetes,

due to a mutant ER resident protein that might alter ER

homeostasis and-cell integrity (Inoue et al., 1998).

1.1.3.2. PERK / mice. The WolcottRallison syn-

drome of infantile diabetes with early destruction of

pancreatic-cells, pancreatic hypoplasia and osteodys-

trophy is caused by homozygous mutation in theEIF2AK3/PERK gene (Delepine et al., 2000), which

encodes for pancreatic ER kinase (PERK), a major ER

stress transducer in mammalian cells, involved in protec-

tive inhibition of protein synthesis. Disruption of PERK

produces the same phenotype in mice. In PERK /

mice, islets are normal at birth, but afterwards -cells

undergo a progressive destruction. It is supposed that

the absence of PERK interferes with the tightly regulated

balance of protein synthesis and folding ER capacity in-cells, thereby favouring protein overload through the

ER (Harding et al., 2001).

Various apoptotic pathways (e.g. CHOP induction,JNK and caspase 12 activation;Nakagawa et al., 2000;

Oyadomari et al., 2002; Urano et al., 2000),could be

triggered by ER stress and lead to -cell loss. Thus,

as reviewed by Harding and Ron (2002), ER-stress-

induced apoptosis could be a major mechanism by

which prolonged stimulation and overload of -cells

under conditions of increased insulin demand (obesity,

insulin resistance, long-term treatment with sulphony-

lureas) might determine the decompensation phase often

termed pancreatic -cell exhaustion (Araki et al.,

2003).


8/21



apoptosis due to islet amyloid production

1.1.4.1. Mice and rats with overexpression of human

islet amyloid polypeptide (h-IAPP); h-IAPP/Agouti

mice. Amyloid deposits, derived from the physiolog-

ical islet amyloid polypeptide (IAPP), co-stored and

co-released with insulin, are present in almost all indi-viduals with type 2 diabetes and are associated with

decreased -cell mass (Butler et al., 2003a; Clark et

al., 1988; Hull, Westermark, Westermark, & Kahn,

2004), but the role of this lesion in the pathogenesis

of type 2 diabetes remains controversial (Hull et al.,

2004).Several spontaneous or genetically manipulated

animal models of islet amyloid formation are known.

Two longitudinal studies performed in primates, such

as Macaca nigra (Howard, 1986)and Macaca mulatta

(de Koning, Bodkin, Hansen, & Clark, 1993)show that

overtly diabetic animals develop extensive islet amy-loid deposits, which precede elevation of blood glucose

and are associated with islet -cell loss (de Koning et

al., 1993).Domestic cats also develop diabetes associ-

ated with an early deposition of IAPP as islet amyloid

(Ma, Westermark, Johnson, OBrien, & Westermark,

1998).

Several lines of transgenic mice with targeted expres-

sion of human IAPP (rodent IAPP is not amiloidogenic;

OBrien, Butler, Westermark, & Johnson, 1993) in -

cells have been established as models of islet amyloid

formation (DAlessio et al., 1994; Fox et al., 1993; Yaguiet al., 1995). These mice produce, store and release

human IAPP normally (Verchere, DAlessio, Palmiter,

& Kahn, 1994), but develop islet amyloid deposits

only in the presence of altered metabolic environment.

Indeed, amyloid formation associated with reduction

of -cell mass, fasting hyperglycaemia and impaired

insulin secretion are observed in human IAPP transgenic

mice either fed for 1 year with high-fat diet (Butler,

Janson, Soeller, & Butler, 2003b) or intercrossed with

genetic murine models of obesity, insulin resistance and

-cell dysfunction, such as ob/ob mice (Hoppener et al.,

1999)and Agouti viable yellow (Avy/a) mice (Soeller etal., 1998). The reduction of-cell mass is due, at least in

one of these models (Butler et al., 2003b),to a striking

increase in-cell apoptosis induced by soluble cytotoxic

IAPP oligomers, that outweighs the obesity-induced-

cell proliferation.

Recently, a human IAPP transgenic rat (the HIP rat)

has been obtained that spontaneously develops diabetes

characterized by islet amyloid formation and decreased

-cell mass (Butler et al., 2004).Normal until 5 months

of age, the -cell mass decreases by 80% in 18-month-

old HIP rats (in controls it increases by 60%), due to

an increased frequency in -cell apoptosis prior to the

development of hyperglycaemia.

Taken together, these data suggest that islet amyloid

formation can be indeed responsible for -cell mass

loss secondary to increased apoptosis. Conditions of

enhanced insulin demand and intrinsic -cell dysfunc-

tion would favour this process, at least in mice.


apoptosis due to gluco- and/or lipotoxicity

In the two models of obesity-linked type 2 diabetes

that I briefly describe here, the-cell mass is not actually

reduced below that of lean controls, but is neverthe-

less largely inadequate to compensate for the associated

insulin resistance.

1.1.5.1. Zucker diabetic fatty rats. The diabetes-prone

Zucker fatty rat model (Finegood et al., 2001; Pick et al.,1998; Tokuyama et al., 1995)develops extreme obesity

because of a genetic defect in the leptin receptor (Phillips

et al., 1996; Takaya et al., 1996).Whereas the original

Zucker fatty (ZF) rats compensate for the obesity-linked

insulin resistance by increasing-cell mass and insulin

secretion, selective breeding has generated a colony of

male (notfemale) diabetes-prone Zucker fatty(ZDF) rats

that develop diabetes because a marked enhancement in-cell apoptosis counteracts the increase of-cell mass

(Finegood et al., 2001; Pick et al., 1998).This colony is

supposed to carry additional genetic alterations becausein the prediabetic state (at 6 weeks of age), ZDF rats

already show an intrinsic -cell secretory defect with

respect to ZF (Pick et al., 1998). Interestingly, while

in ZF animals pancreatic Pdx-1 expression as well as

Akt activity are enhanced (Jetton et al., 2005), ZDF

rats show defective -cell Pdx-1 expression after the

development of hyperglycaemia, but not in the predi-

abetic state (Harmon et al., 1999). Thus, Pdx-1 defi-

ciency in ZDF rats is not genetically determined but

most likely secondary to glucotoxicity and would exac-

erbate-cell dysfunction by down-regulating essential

functional genes such as insulin and GLUT-2. The mech-anisms underlying the increased-cell apoptosis in ZDF

rats have been attributed to lipotoxicity caused by lipid

accumulation, possibly through activation of the pro-

apoptotic ceramide pathway (Lee et al., 1994, 1997;

Shimabukuro et al., 1998).

1.1.5.2. Psammomys obesus gerbil. Another relevant

model is the gerbilP. obesusfed in captivity with a rela-

tively high-energy (HE) diet (2.93 kcal/g), replacing its

low-calorie natural diet of desert saltbush plant (Adler,

Kalman, Lazarovici, Bar-On, & Ziv, 1991; Donath et


9/21


al., 1999).Upon HE diet, these rodents become obese

and, similarly to the Zucker fatty rat, selective breed-

ing has generated two lines of animals: a diabetes-prone

(DP) line that develops overt diabetes in a few days, and

a partially diabetes-resistant (DR) line, in which most

animals exhibit normoglycaemia and persistent hyperin-

sulinaemia (Kalderon, Gutman, Levy, Shafrir, & Adler,1986;Nesher, Gross, Donath, Cerasi, & Kaiser, 1999).

The diabetes-prone P. obesus gerbils show defec-

tive glucose-stimulated insulin secretion, inability to

undergo adaptive changes to keep pace with increased

insulin demand and progressive loss of -cell mass

because of an impressive increase in -cell apopto-

sis, mainly attributed to glucose toxicity (Donath et al.,

1999; Leibowitz et al., 2001b).Interestingly, no func-

tional Pdx-1 gene product, at least corresponding to the

highly conserved mammalian homeodomain, has been

found in P. obesus, with consequent impaired glucose-induced insulin gene expression, that could be restored

upon Pdx-1 gene transfer into islets (Leibowitz et al.,

2001a). However, the lack of Pdx-1 also occurs in the DR

line (Leibowitz et al., 2001a),suggesting that additional

factors are required for full development of diabetes in

this model.

1.1.6. Spontaneous animal syndrome of non-obese

type 2 diabetes with reduction of-cell mass

1.1.6.1. Goto-Kakizaki (GK) rats. The spontaneously

diabetic Goto-Kakizaki ratis a genetic lean model of type2 diabetes originating from selective breeding over many

generations of glucose-intolerant non-diabetic Wistar

rats (Goto, Suzuki, Sasaki, Ono, & Abe, 1998).GK rats

show a prenatal reduction of-cell proliferation asso-

ciated to an abnormal wave of apoptosis (Miralles &

Portha, 2001). This results in a deficit in -cell mass

from birth that becomes progressively larger as a con-

sequence of impaired new islet formation and -cell

replication rather than increased apoptosis (Movassat,

Saulnier, Serradas, & Portha, 1997;Plachot, Movassat,

& Portha, 2001;Serradas, Gangnerau, Giroix, Saulnier,

& Portha, 1998). Interestingly, this deficit appears tobe due to impaired IGF-2 production (Serradas et al.,

2002).Despite congenital reduction of-cell mass, in

the first 34 weeks of life (i.e. the period of maximal

physiological growth of beta cells), GK rats maintain

normoglycaemia (Movassat, Saulnier, & Portha, 1995;

Tourrel et al., 2002). At an adult age, GK rats have

50% depletion of total-cell mass, stable moderate non-

fasting hyperglycaemia, impaired glucose tolerance and

markedly defective insulin response to glucose in vivo

and in vitro (Briaud, Kelpe, Johnson, Tran, & Poitout,

2002; Movassat et al., 1995; Portha et al., 1994).GK

rats also have genetically induced alterations of insulin

secretion independent of the reduction of-cell mass

(Portha, 2003).

A synopsis of the mechanisms leading to reduction

of-cell mass in spontaneous or transgenic models of

type 2 diabetes is shown inTable 1.

1.2. Animal models of type 2 diabetes with

experimentally induced reduction of-cell mass

Reduction of-cell mass has been experimentally

induced in several animal species by nutritional, sur-

gical or chemical means to reproduce at least some

of the features of human type 2 diabetes. Such mod-

els are particularly helpful to understand the mecha-

nisms of compensationand decompensation of a reduced

-cell mass and assess the effectiveness of potential

therapeutics acting by enhancing -cell growth and/or

survival.

1.2.1. Models with reduction of-cell mass induced

by foetal growth retardation

1.2.1.1. Intrauterine growth retardation by uteroplacen-

tal insufficiency in the rat. A rat model of uteroplacental

insufficiency has been developed, designated as IUGR

for intrauterine growth retardation, induced by bilateral

uterine artery ligation at 19 days of gestation, i.e. 3 days

before term (Simmons, Templeton, & Gertz, 2001). Thismodel intends to mimic the unfavourable intrauterine

environment that in humans leads to low-birth weight

and is supposed to confer high risk of development of

diabetes in adult age (Barker et al., 1993; Rich-Edwards

et al., 1999). IUGR rats have lower birth weightthan con-

trols until 7 weeks of age, but later on, they surpass the

weight of controls and by 26 weeks they become obese.

At an early age, IUGR rats are glucose-intolerant and

insulin resistant; by 7 weeks they develop mild fasting

hyperglycaemia and hyperinsulinaemia. At 26 weeks,

they are markedly hyperglycaemic.

In IUGR rats, -cell mass is normal during the firstfew weeks of life, despite the 50% reduction of Pdx-1

expression observed in IUGR foetuses. However, by 7

weeks of age, -cell mass becomes lower than in con-

trols, is reduced to 50% at 15 weeks and to one-third at

26 weeks of age (Simmons et al., 2001).It is not known

if this age-related decline is caused by decreased prolif-

eration or increased apoptosis of-cells.

This model supports the hypothesis that an abnormal

intrauterine environment can induce permanent alter-

ations in glucose homeostasis after birth and lead to type

2 diabetes in the adulthood.


10/21


Table 1

Synopsis of the mechanisms leading to reduced -cell mass in spontaneous or transgenic animal syndromes of type 2 diabetes

Animals -Cell

replication

-Cell

neogenesis

-Cell

size

-Cell

apoptosis

Obesity or insulin

resistance

Transgenic mice deficient in factors involved in -cell development, growth and/or survival

Pdx-1 +/ N N N No

IRS-2/ N N N Yes

RIP A-CREB transgenic mouse N N N No

p70S6K / N N N No

Cyclin D2/ N N N No

CdK-4/ N N N No

Mice with increased-cell apoptosis due to ER stress

Akita mouse (INS-2 mutation) N N N No

PERK/ N N N No

Rodents with increased-cell apoptosis due to islet amyloid production

Mouse with overexpression of h-IAPP (upon HFD) N Yes

Rat with overexpression of h-IAPP (HIP rat) N No

h-IAPP/Agouti mousea Yes

Rodents with increased-cell apoptosis due to gluco- and/or lipotoxicityZucker diabetic rata Yes

Psammomys obesusgerbila (upon HED) Yes

Spontaneous model of decreased-cell mass

GK rat N N No

N, normal; , decrease; , increase;, marked increase; ER, endoplasmic reticulum; HFD, high-fat diet; HED, high energy diet; RIP A-CREB,

dominant-negative CREB transgene under control of rat insulin promoter-1.a Relative decrease of-cell mass.

1.2.1.2. Offsprings of pregnant rats fed with low-protein

diet. That foetal malnutritioncan be a relevantrisk factor

for the later development of type 2 diabetes is confirmed

by the observation that offsprings of pregnant rats sub-

jected to a low-protein (LP) isocaloric diet (8% versus

20%), show, both during foetal life and in the neona-

tal period, reduced body weight, altered insulin secre-

tory capacity and reduced -cell mass resulting from

reduced-cell proliferation rate and increased apopto-

sis (Bertin et al., 2002; Boujendar et al., 2002; Petrik

et al., 1999;Snoeck, Remacle, Reusens, & Hoet, 1990).

This is associated with reduced foetal and neonatal islet

expression of IGF-II (Petrik et al., 1999), which actsas-cell mitogen and protects against apoptosis (Petrik

et al., 1998). Expression of Pdx-1 is also reduced in

islets from pups of LP mothers (Arantes et al., 2002).

Interestingly, taurine supplementation is able to par-

tially prevent -cell mass reduction in the LP model

by normalization of proliferation and apoptosis rates

in concomitance with restoration of IGF-II expression

(Boujendar et al., 2002).In adulthood, rats born from

LP mothers still have reductions in -cell mass and

insulin secretion and show glucose intolerance, but usu-

ally not overt diabetes (Dahri, Reusens, Remacle, &

Hoet, 1995;Dahri, Snoeck, Reusens, Remacle, & Hoet,

1991;Reusens & Remacle, 2000).At old age, LP off-

springs may develop fasting hyperglycaemia, associated

with insulin resistance (Petry, Dorling, Pawlak, Ozanne,

& Hales, 2001).

Several variants of LP diet protocol have been used,

differing for duration, severity of protein deprivation,

association with calorie restriction, as recently reviewed

by Armitage, Khan, Taylor, Nathanielsz, and Poston

(2004). The above-described endocrine and metabolic

features of the LP model are present in most of the vari-

ants, although to a differentextent. It should be reminded

that in the LP isocaloric diets, calories are balanced byaddition of fat or carbohydrates, and this makes more

complex the interpretation of the results (seeArmitage

et al., 2004).

1.2.2. Models with surgically induced reduction of

-cell mass

1.2.2.1. Partially pancreatectomized rats. Partial pan-

createctomy has been largely used in various animal

species as a tool to directly reduce-cell mass in healthy

animals, without any genetic background and/or previ-

ous alteration of insulin-producing cells.


11/21


In several species, quite large pancreatectomies are

required to determine mild/moderate hyperglycaemia

and insulin secretory defect (5075% in man: Kendall,

Sutherland, Najarian, Goetz, & Robertson, 1990;Porte,

1991;6070% in the pig: Stump, Swindle, Saudek, &

Strandberg, 1988; more than 80% in the rat: Jonas et

al., 1999).It is intriguing that if a small additional pan-creatic mass is removed (8090%) in pigs (Lohr et al.,

1989)and humans (Gepts & Lecompte, 1981;Kloppel

et al., 1985;Saito, Yaginuma, & Takahashi, 1979);95%

in rats (Laybutt et al., 2003),overt diabetes with severe

insulin deficiency develops.

The different outcome of surprisingly small differ-

ences in the percentage of subtotal pancreatectomy in

rats has been recently studied in details by Laybutt

et al. (2003). Tissue removal is performed by gentle

abrasion with cotton applicators leaving the pancreas

within 12 mm of the common bile duct and extend-ing from the duct to the first part of the duodenum.

Control rats undergo laparatomy and gentle rubbing of

the pancreas between fingers. By varying the propor-

tion of gastric lob removed, it is possible to obtain

either 85 or 95% pancreatectomy. This slight varia-

tion resulted initially in non-fasting blood glucose lev-

els of variable severity, confirming previous results

(Jonas et al., 1999). However, over time, glycaemia

clustered into two distinct groups. Some rats main-

tained nearly normal glucose levels, indicating long-

lasting compensatory adaptation. The other rats becamehighly hyperglycaemic, with no rats in the middle range

(140250 mg/dl). These observations suggest that below

a critical value of -cell mass, when compensatory

mechanisms fail, a dysfunction in the residual -cells

is likely to occur, possibly due to overload-induced ER

stress or toxic effects of mounting hyperglycaemia and

hyperlipidaemia (Donath & Halban, 2004; Laybutt et al.,

2003).

Pancreatectomized rats appear to be a good model to

investigate the adaptive mechanisms of residual-cells

in adult animals. In 60% pancreatectomized (60% Px)

rats, that are normoglycaemic and glucose-intolerant, alimited regenerationof the residual-cellsoccurs, so that

the -cell mass increases from 40 to 55% of normal, sev-

eral weeks after surgery (Leahy, Bonner-Weir, & Weir,

1988).Moreover, an increase in insulin responsiveness

to intermediate glucose concentrations is observed, due

to up-regulation of-cell glycolytic flux driven by an

increased activity of glucokinase (Liu, Nevin, & Leahy,

2000),the main regulator of-cell glucose utilization

(Matschinsky et al., 1993).Similar up-regulation of islet

glucose metabolismalso occurs in normoglycaemic 60%

Px mice (Martn et al., 1999).Interestingly, in hypergly-

caemic 90% Px rats, an increase in hexokinase, and not

glucokinase, activity is the prevalent change affecting

islet glucose metabolism (Hosokawa, Corkey, & Leahy,

1997),associated with an enhanced basal insulin secre-

tion (Leahy, Bumbalo, & Chen, 1993).

In 90% Px rats, an initial burst of -cell regener-

ation has been observed in the first 710 days aftersurgery (Bonner-Weir, Baxter, Schuppin, & Smith,

1993). Furthermore, -cell hypertrophy is apparent in

these hyperglycaemic pancreatectomized rats 4 weeks

after surgery (Jonas et al., 1999;Xu, Stoffers, Habener,

& Bonner-Weir, 1999)and is maintained at 14 weeks

(Laybutt et al., 2003).-Cell hypertrophy is also found

in prediabetic Zucker diabetic rats with impaired glu-

cose tolerance (Pick et al., 1998), and in pregnancy

(Scaglia et al., 1995). Thus, -cell hypertrophy could

represent a major mechanism of compensatory response

to increased demand in terminally differentiated -cellsthat may prevent, at least for a while, more serious

metabolic impairment. The up-regulation of c-Myc that

occurs in 8595% Px rats more or less rapidly, depending

on the circulating glucose levels, may play an important

role in thecompensatory growth of-cells, since this fac-

tor can lead to hypertrophy in the absence of cell division

(Schuhmacher et al., 1999).

Laybutt et al. (2003) have also studied the time course

of the changes in -cell phenotypes in 8595% pancrea-

tectomized (8595% Px) rats, by evaluating the expres-

sion of a number of-cell genes coding for transductionfactors regulating -cell differentiation (Pdx-1, HNF-

1, NeuroD/BETA2, Nkx61) and for key metabolic fac-

tors (GLUT-2, glucokinase, pyruvate carboxylase, mito-

chondrial glycerophosphate dehydrogenase). In clearly

hyperglycaemic 95% Px rats, the expression of these

genes was decreased by 50% 4 weeks after pancreatec-

tomy and further reduced at 14 weeks, whereas in 85%

Px rats with mild hyperglycaemia, a 3040% decrease in

transcription factors was found at 14 weeks only. Correc-

tion of hyperglycaemia by a 2-week phlorizin treatment

fully reverses the changes at 4 weeks but only partially

at 14 weeks. Thus, chronic hyperglycaemia induces adeterioration of-cell phenotype, whose extent, timing

and reversibility depend on both severity and duration of

hyperglycaemia. However, it is noteworthy that at 4 and

14 weeks, 85% Px rats, despite significant differences

in-cell phenotype, have similar mild hyperglycaemia,

likely related to a concomitant two-fold increase in -

cell size (Laybutt et al., 2003). This suggests that, at least

in some circumstances, a reduced-cell mass is able to

maintain successful compensation for a long time, even

in the presence of mild hyperglycaemia and-cell dys-

function.


12/21


1.2.3. Models with chemically-induced reduction of

-cell mass

Differently from pancreatectomy, chemical induction

of diabetes offers the advantage of preservation of both

exocrine and endocrine cell populations other than -

cells, thus resembling the situation in human diabetes.

Furthermore, the good conditions of the animals afterchemical induction of diabetes do not require particular

welfaremeasuresandallowstudiesontheeffectsofhigh-

fat diet that cannot be carried out in pancreatectomized

animals.

1.2.3.1. Neonatal rats treated with streptozotocin

(n-STZ rats). Neonatal Wistar rats treated with

90100 mg/kg b.w. STZ (n-STZ) on the day of birth

(Portha, Levacher, Picon, & Rosselin, 1974)or 2 days

after birth (Weir, Clore, Zmachinski, & Bonner-Weir,

1981)undergo a transient hyperglycaemia followed bya rapid spontaneous remission until about 68 weeks

of age. Thereafter, non-fasting chronic hyperglycaemia

develops with plasma glucose concentrations usually

ranging 170200 mg/dl for rats injected at birth and

200350 mg/dl for rats injected at 2 days of age. The

remission of initial STZ-induced hyperglycaemia is

accompanied by partial-celland insulin stores reconsti-

tution (Bonner-Weir, Trent, Zmachinski, Clore, & Weir,

1981;Portha et al., 1989),most of new -cells originat-

ing from undifferentiated duct cells (Dutrillaux, Portha,

Roze, & Hollande, 1982;Wang, Bouwens, & Kloppel,1994). The residual -cell mass, which is reduced to

approximately 20% upon STZ administration, is finally

recovering up to 50% as a consequence of-cell regen-

eration and/or neogenesis. The moderate diabetic state

of adult n-STZ rats does not affect body weight gain nor

requires insulin treatment. However, in these rats, insulin

responsiveness to glucose and tolbutamide is lacking

(Dutrillaux et al., 1982;Giroix, Portha, Kergoat, Bailbe,

& Picon, 1983),that is not justified by the 50% reduc-

tion of-cell mass per se, but might be dependent on an

incomplete differentiation of the newly formed -cells

after the initial STZ-induced loss (Portha et al., 1989;Weir, Leahy, & Bonner-Weir, 1986).

1.2.3.2. Streptozotocin-nicotinamide-treated adult rats

(STZ-NA rats). On the basis of previous knowledge

that suitable doses of nicotinamide (NA) could exert

a partial protection against the -cytotoxic effect of

streptozotocin (STZ), we have established a new exper-

imental diabetic syndrome in adult rats that appears

closer to human type 2 diabetes than other available

models (e.g. neonatally STZ-injected rats; GK rats),

with regard to insulin responsiveness to glucose and

sulphonylureas (Masiello et al., 1998).In 2-month-old

Wistar rats, a dose of 200230 mg/kg b.w. nicotinamide,

given intraperitoneally 15 min before streptozotocin

administration (60 mg/kg i.v.), yields a maximum of

animals (7580%) with 40% reduction of pancreatic

-cell mass (and no change in-cell mass) and moderate

stable non-fasting hyperglycaemia (150180 mg/dl).Interestingly, the remaining 2025% treated animals

either become severely diabetic within 23 weeks

or remain normoglycaemic, yet glucose-intolerant

(Masiello, unpublished data). Thus, as reported for

8595% Px rats (Laybutt et al., 2003),we cannot obtain

STZ-NA rats with stable hyperglycaemic levels in the

middle range (200300 mg/kg).

Two to 3 weeks after diabetes induction, intravenous

glucose tolerance tests reveal clear abnormalities in glu-

cose tolerance and insulin responsiveness, which are

reversed by tolbutamide administration. These featuresremain unchanged for a long time after induction of

diabetes. In the isolated perfused pancreas of STZ-NA

rats, insulin response to glucose elevation is clearly

present, although significantly reduced with respect to

controls. Moreover, the insulin response to tolbutamide

is similar to that observed in normal pancreases. Such

in vivo and in vitro partial insulin responsiveness to

glucose and sulphonylureas is missing in other animal

models, such as n-STZ, GK and partially pancreatec-

tomized rats (Giroix et al., 1983;Leahy, Bonner-Weir,

& Weir, 1984; Portha et al., 1991). In islets isolated fromSTZ-NA rats, compensatory adaptations occur. Indeed,

glucose oxidation and utilization are increased when

expressed per islet DNA content (Novelli, Fabregat,

Fernandez-Alvarez, Gomis, & Masiello, 2001) and

insulin release is stimulatedby intermediate glucose con-

centrations and potentiated in presence of free fatty acids

more than in controls (manuscript in preparation). A

moderate increase of-cell neogenesis in the pancreas of

STZ-NArats without evidence of significant-cell repli-

cation, has been observed within a few weeks after treat-

ment (Novelli et al., 2001),but extensive studies on this

topic have not been conducted yet in this model. Thus,the STZ-NA-induced diabetic syndrome with decreased

-cell mass and preserved insulin responsiveness to glu-

cose and tolbutamide, may provide a particularly advan-

tageous tool for pharmacological investigations of not

only new insulinotropic agents (Broca et al., 1999),but

also factors stimulating -cell growth, such as GLP-

1/exendin-4 (see below). These studies would clarify

whether such promising growth-stimulating factors that

are currently tested in neonate or very young animals,

are also able to promote-cell growth in adult animals

with reduced -cell mass. Furthermore, the STZ-NA


13/21


model appears very suitable for longitudinalstudies aim-

ing at assessing whether a reduced -cell mass is able to

cope with increased insulin demand induced by high-fat

diet/obesity/insulin resistance and what are the mecha-

nisms underlying the compensation and the subsequent

expected failure of insulin secretory function.

We have initiated studying the effect of high-fat diet(HFD) (40% of calories provided as saturated and mono-

unsaturated fat) given for 3 months to STZ-NA rats.

The results show that HFD-fed STZ-NA rats become

obese, insulin resistant and slightly hyperlipidaemic, but

maintain unchanged the mild hyperglycaemia typical of

this model, while developing a significant and stable

hyperinsulinaemia. Glucose tolerance tests and in vitro

stimulation of isolated islets confirm that STZ-NA rats

develop successful adaptive responses to the increased

metabolic load, at least for 3 months (manuscript in

preparation). Experiments are currently in progress todetermine whether this adaptation continues for a longer

time and which are the underlying mechanisms.

It is of interest that in the genetic model of GK rats

(see above), the effects of a high-fat diet were markedly

differentthaninSTZ-NArats.Infact,inGKrats,high-fat

feeding for 6 weeks induced no compensatory response

in vivo and resulted instead in a further impairment

of glucose-stimulated insulin release in vitro, associ-

ated with UCP-2 expression (Briaud et al., 2002).The

discrepancy between the two models is likely to be

dependent on the higher basal glycaemic levels and thepresence of genetic-cell dysfunction in GK rats com-

pared to STZ-NA rats. Both of these factors could lead

to overt decompensation when associated with hyperlip-

idaemia (Poitout & Robertson, 2002).

1.2.3.3. Streptozotocin-nicotinamide-treated minipigs

(STZ-NA minipigs). The rat STZ-NA model has been

recently reproduced in the Gottingen minipig, in which

the combined administration of streptozotocin and

nicotinamide results in a mild diabetes characterized by

reduced -cell mass, fasting and postprandial hypergly-

caemia and mildly impaired insulin secretion (Larsen etal., 2002).Since pigs are more resistant to the diabeto-

genic effect of STZ than rats, doses of 125 mg/kg of STZ

i.v. and 67 mg/kg of NA were found to be the most suit-

able for diabetes induction. -Cell mass was reduced

by approximately 70% (Larsen et al., 2003a).A com-

pensatory increase in insulin secretion from the residual-cell population was observed both in vivo and in vitro

(Larsen, Rolin, Gotfredsen, Carr, & Holst, 2004).The

STZ-NA minipig model has also been used to validate

the measurement of insulin secretory capacity and glu-

cose tolerance to predict pancreatic-cell mass in vivo

(Larsen et al., 2003b),as well as to study the effect of

reduced -cell mass on the pulsatile insulin secretion

(Larsen et al., 2003a)that is impaired in type 2 diabetes

(Hollingdal et al., 2000).

Several of the animal models with reduced -cell

mass have been recently used to test the possibilities

of recovery of-cell mass and function upon treatmentwith pharmacological agents potentially capable of stim-

ulating -cell replication or neogenesis. As a relevant

example, I summarize here the results of studies on the

effects of exendin-4, a long-acting agonist of GLP-1

receptor.

1.2.4. Effect of treatment with exendin-4 on animal

models of type 2 diabetes with reduced-cell mass

A number of studies have examined the ability of

exendin-4 to increase pancreatic-cell mass in vivo, asreviewed inNielsen, Young, and Parkes (2004).

In 9095% pancreatectomized rats, the daily admin-

istration of exendin-4 for 10 days post-pancreatectomy

stimulated a 40% expansion of-cell mass and cell pro-

liferation, without affecting -cell size (Xu et al., 1999).

This resulted in attenuation but not normalization of

hyperglycaemia.

In GK rats, neonatal treatment with exendin-4 for a

few days enhanced pancreatic insulin content and total-cell mass by stimulation of both proliferation and neo-

genesis of-cells, without effect on -cell size (Tourrelet al., 2002).Two months after exendin-4 treatment, -

cell mass in GK rats achieved 63% of the -cell mass

in the Wistar group (untreated GK have 40% of the

normal -cell mass) and plasma glucose levels were

decreased compared to untreated GK rats, but not nor-

malized (Tourrel et al., 2002).

In n-STZ rats, the response to early exendin-4 expo-

sure for 4 days was a rapid three-fold increase in -cell

mass, attributable to stimulation of-cell neogenesis,

that was maintained in adulthood, but resulted only in a

partial decrease of hyperglycaemia and no improvement

of insulin secretory function in vivo and in vitro (Tourrel,Bailbe, Meile, Kergoat, & Portha, 2001).

In IUGR rats, neonatal treatment with exendin-4

for 6 days prevented the development of diabetes by

completely rescuing -cell mass from the progressive

reductionobserved in untreated animals (Stoffers, Desai,

DeLeon, & Simmons, 2003). This effect was due to

enhanced -cell proliferation and Pdx-1 expression.

Interestingly, it has been reported that exendin-4 fails

to increase-cell proliferation and reduce-cell apop-

tosis in mice with -cell-specific inactivation of Pdx-1

(Li et al., 2005).


14/21


Table 2

Scheme of the main advantages and limitations of various rodent models of type 2 diabetes with reduction of pancreatic -cell mass and some

suggestions for their use in diabetes research

Model Associated

insulin

resistance

Advantages Disadvantages and limitations Suggestions for use

ZDF rats Yes Genetic model associated with

obesity and insulin resistance;

occurrence of a prediabetic phase

Specific genetic background

(alteration of leptin receptor)

affecting appetite; intrinsic

-cell dysfunction on unknown

genetic basis; markedly

abnormal lipid metabolism

Studies on the mechanisms of

-cell decompensation and

glucolipotoxicity

P. obesusgerbils Yes Genetic model induced by

relative nutritional excess

Particular genetic background

adapted to desert life (thrifty

genotype); limited availability

and high cost of animals

Studies on the mechanisms of

metabolic decompensation induced

by changes in nutritional

environment

GK rats No Genetic model without obesity

derived by selective breeding of

Wistar rats; stable moderate

hyperglycaemia

Intrinsic-cell dysfunction on

unknown genetic basis; poor

insulin responsiveness to

glucose in vivo and in vitro

Studies on diabetic complications

and other effects of chronic

hyperglycaemia; comparison with

non-genetic models (e.g., effects ofHFD or aging)

LP IUGR rats Yes Easy to be induced; possibility to

introduce selected nutritional

supplementations

Difficult choice of protocol

variants regarding severity and

duration of LP diet; variability

of effects; hyperglycaemia only

at old age

Studies on the late metabolic

effects of nutritional restriction and

underlying mechanisms

IUGR rats Yes Reproduction of utero-placental

insufficiency of humans with

low-birth weight; progressive

hyperglycaemia with insulin

resistance

No major disadvantage Studies on the late metabolic effects

of altered intrauterine environment

and underlying mechanisms

Px rats No Previously healthy animals,

without genetic background and

-cell dysfunction; induction inyoung adults; intact residual

-cells

Invasive technique with

removal of exocrine pancreas

and endocrine non--cells;need for animal welfare

measures; unsuitable for HFD

experiments; lack of insulin

responsiveness to glucose and

tolbutamide in perfused

pancreas

Studies of the mechanisms of

compensation/failure of a primarily

reduced healthy-cell mass;suitable for treatment with drugs

stimulating cell growth

n-STZ rats No Stable moderate hyperglycaemia Residual-cells deriving from

regeneration/neogenesis

apparently not well

differentiated; lack of insulin


tolbutamide in perfused

pancreas

Studies on diabetic complications

and other effects of chronic

hyperglycaemia

STZ-NA rats No Previously healthy animals,

without genetic background and

-cell dysfunction; easy

induction in young adults; mild

stable hyperglycaemia; residual

-cells well differentiated; in

vivo and in vitro insulin


tolbutamide

Residual-cells potentially

damaged by STZ; lack of

insulin resistance, however

inducible with HFD;

hyperglycaemia only in the

non-fasting state

Studies on long-term effects of

mild/moderate hyperglycaemia on

-cell function and gene

expression; studies on

compensation/failure upon HFD;

studies on new insulin

secretagogues or drugs stimulating

cell growth; studies on the effects

of aging in animals with reduced

-cell mass

ZDF: Zucker diabetic fatty; GK: Goto-Kakizaki; LP IUGR: low-protein intrauterine growth retardation; IUGR: intrauterine growth retardation by

uterine artery ligation in pregnancy; Px: partially pancreatectomized; n-STZ: neonatal streptozotocin; STZ-NA: streptozotocin-nicotinamide; HFD:

high-fat diet.


15/21


All the above-mentioned studies using exendin-4

have been performed in neonate or very young animals

that are likely more responsive to growth-stimulating

factors. However, 9-week-old Zucker obese rats treated

with exendin-4 (Gedulin et al., 2005) and aging glucose-

intolerant rats treated with GLP-1 (Perfetti, Zhou, Doyle,

& Egan, 2000)also showed beneficial effects. Further-more, GLP-1 and exendin-4, administered for 4 weeks to

adult STZ-NA rats, were recently reported to normalize

plasma glucose levels (Ozyazgan, Kutluata, Af, Ozda, &

Akkan, 2005).

1.2.5. Advantages and limitations of the most

commonly used animal models of type 2 diabetes

with-cell mass reduction

A scheme summarizing the main advantages, disad-

vantages and limitations of the most common animal

models described above is shown inTable 2.

2. Conclusions

From the analysis of the animal models described

above, it appears that among the factors involved in

pancreas development, Pdx-1 emerges for its role not

only in the maintenance of fully differentiated and func-

tional -cells but also in their physiological postna-

tal expansion and their survival. It has to be under-

lined that Pdx-1 appears as a critical regulator of -

cell plasticity, given its essential contribution to ensureappropriate -cell proliferation and/or neogenesis in

the adaptive response to insulin resistance or -cell

injury.

IRS-2 and its complex downstream signalling path-

way are also relevant for -cell plasticity and survival,

likely in an integrated fashion with Pdx-1 activity. Also

cAMP-dependent CREB signalling appears to play a sig-

nificant role in-cell survival and is particularly attrac-

tive as a mediator of the stimulating effect of GLP-1 and

its long-acting analog exendin-4 on -cell proliferation

and neogenesis.

Defects in specific factors involved in cell cycle pro-gression such as cyclin D2 might have pathogenic sig-

nificance for-cell mass reduction only within a small

temporal window corresponding to the period of maxi-

mal postnatal-cell growth.

The animal models of reduced -cell mass associ-

ated with amyloid formation or ER stress support the

possible intervention of these factors in the pathogene-

sis of type 2 diabetes. The role of ER stress as a possible

general mechanism of induction of-cell death is prob-

ably underestimated and should be further investigated

in animal models of reduced -cell mass, especially in

situations of increased insulin demand and consequent-cell overstimulation.

The animal models with reduction of -cell mass

induced by nutritional, surgical or chemical means show

that compensatory adaptation takes place in the resid-

ual -cell mass. Partial recovery of the lost mass may

result from post-injury induction of-cell regenerationand/or neogenesis and eventually from treatment with

growth-stimulating pharmacological agents. Hypertro-

phy and increased insulin responsiveness to glucose and

free fatty acids may also occur in residual -cells. Nev-

ertheless, the adaptive response might be insufficient

or temporary because of incomplete differentiation of

newly formed -cells and/or acquired dysfunctions of

the residual -cells chronically exposed to a metabol-

ically altered environment. An increased frequency of

apoptosis due to prolonged overstimulation of residual

-cells, chronic hyperglycaemia/hyperlipidaemia and/oramyloid formation might accelerate decompensation.

Interestingly, recent data obtained in partially pancreate-

ctomized rats and STZ-NA rats indicate that mild hyper-

glycaemic conditions are compatible with prolonged

successful -cell compensation from a reduced -cell

mass.

An important consideration that stems from the anal-

ysis of a number of experimental models of -cell

mass reduction is that a genetically or environmentally

induced permanent alteration of the developmental pat-

tern of cellular proliferation and differentiation in theendocrine pancreas may result in early-onset reduction

of-cell mass and/or inability to undergo cellular and

metabolic compensatory adaptations to increased insulin

demand. It is possible that the pathogenesis of type 2 dia-

betes developing in one-third of obese subjects relies on

this basis.

Further investigations are required to fully elucidate

the mechanisms of-cell compensation and subsequent

failure in animal models of reduced -cell mass. Given

our current inability to accurately determine functional

human -cell mass in a non-invasive manner, these mod-

els are also of paramount importance for longitudinalstudies aimed at correlating -cell mass and metabolic

parameters in a number of pathophysiologicalconditions

or upon treatment with pharmacological agents.

Thus, it is expected that animal models of type 2 dia-

betes with reduced -cell mass will continue to have a

relevant place in diabetes research.

Acknowledgement

The helpful assistance of Dr. Michela Novelli

and Ms. Valentina DAleo in the proper arrange-


16/21


ment of the references is gratefully acknow-

ledged.

References

Adler, J. H., Kalman, R., Lazarovici, G., Bar-On, H., & Ziv, E. (1991).

Achieving predictablemodelof type 2 diabetes in thesand rat. In E.

Shafrir (Ed.),Frontiers in diabetic research: Lessons from animal

diabetes III(pp. 212214). London: Smith-Gordon.

Ahlgren, U., Jonsson, J., Jonsson, L., Simu, K., & Edlund, H. (1998).

Beta-cell-specific inactivation of the mouse Ipf1/Pdx1 gene results

in loss of the beta-cell phenotype and maturity onset diabetes.

Genes Dev.,12, 17631768.

Araki, E., Oyadomari, S., & Mori, M. (2003). Impact of endoplas-

mic reticulum stress pathway on pancreatic -cells and diabetes

mellitus.Exp. Biol. Med.,228, 12131217.

Arantes, V. C., Teixeira, V. P. A., Reis, M. A. B., Latorraca, M. Q.,

Leite, A. R., Carneiro, E. M., et al. (2002). Expression of PDX-1 is

reduced in pancreatic isletsfrom pups of ratdams fed a low protein

diet during gestation and lactation. J. Nutr.,132, 30303035.Armitage,J. A.,Khan,I. Y., Taylor, P. D.,Nathanielsz, P. W., & Poston,

L. (2004).Developmentalprogramming of the metabolicsyndrome

by maternal nutritional imbalance: How strong is the evidence

from experimental models in mammals? J. Physiol., 561, 355

377.

Barker, D. J., Hales, C. N., Fall, C. H., Osmond, C., Phipps, K., &

Clark, P. M. (1993). Type 2 (non-insulin-dependent) diabetes mel-

litus, hypertension and hyperlipidemia (syndrome X): Relation to

reduced fetal growth.Diabetologia,36, 6267.

Becker, E. B., & Bonni, A. (2004). Cell cycle regulation of neuronal

apoptosis in development and disease.Prog. Neurobiol.,72, 125.

Bernard-Kargar, C., & Ktorza, A. (2001). Endocrine pancreas plas-

ticity under physiological and pathological conditions. Diabetes,

50(Suppl. 1), S30S35.Bertin, E., Gangnerau, M. N., Bellon, G., Bailbe, D., Arbelot De Vac-

quer, A., & Portha, B. (2002). Development of beta-cell mass in

fetuses of rats deprived of protein and/or energy in last trimester

of pregnancy.Am. J. Physiol. Regul. Integr. Comp. Physiol.,283,

R623R630.

Bonner-Weir, S. (2000a). Islet growth and development in the adult. J.

Mol. Endocrinol.,24, 297302.

Bonner-Weir, S. (2000b). Life and death of the pancreatic beta cells.

Trends Endocrinol. Metab.,11, 375378.

Bonner-Weir, S. (2000c). Perspective: Postnatal pancreatic beta cell

growth.Endocrinology,141, 19261929.

Bonner-Weir, S. (2001).-Cell turnover. Its assessment and implica-

tions.Diabetes,50, S20S24.

Bonner-Weir, S., Baxter, L. A., Schuppin, G. T., & Smith, F. E. (1993).

A secondpathway for regeneration of adult exocrineand endocrine

pancreas. A possible recapitulation of embryonic development.

Diabetes,42, 17151720.

Bonner-Weir, S.,Trent,D. F., Zmachinski, C. J., Clore, E. T., & Weir, G.

C. (1981). Limited -cell regeneration in a -cell deficient model:

Studies with dexamethasone.Metabolism,30, 914918.

Boujendar, S., Reusens, B., Merezak, S., Ahn, M.-T., Arany, A., Hill,

D., et al. (2002). Taurine supplementation to a low protein diet

during foetal and early postnatallife restores a normal proliferation

and apoptosis of rat pancreatic islets.Diabetologia,45, 856866.

Brazil, D. P., & Hemmings, B. A. (2001). Ten years of protein kinase

B signalling: A hard Akt to follow. Trends Biochem. Sci., 26,

657664.

Briaud, I., Kelpe, C. L., Johnson, L. M., Tran, P. O. T., & Poitout, V.

(2002). Differential effects of hyperlipidemia on insulin secretion

in islet of Langerhans from hyperglycemic versus normoglycemic

rats.Diabetes,51, 662668.

Broca, C., Gross, R., Petit, P., Sauvaire, Y., Manteghetti, M., Tournier,

M., et al. (1999). 4-Hydroxyisoleucine: Experimental evidence of

its insulinotropic and antidiabetic properties.Am. J. Physiol.,277,

E617E623.Burke, J. P., Williams, K., Gaskill, S. P., Hazuda, H. P., Haffner, S.

M., & Stern, M. P. (1999). Rapid rise in the incidence of type 2

diabetes from 1987 to 1996: Results from the San Antonio

Animal Models of DM2 With Reduced Pancreatic Β-cell Mass - 2006

Documents