Role of Insulin-Like Growth Factor-Binding Protein 5 (IGFBP5) in Organismal and Pancreatic β-Cell Growth

Abstract

A family of IGF-binding proteins (IGFBP) exerts biological actions both dependent on and independent of IGF-I. A major effector of the insulin/IGF-I signaling pathway, the serine/threonine protein kinase Akt, mediates cellular processes such as glucose uptake, protein synthesis, cell survival, and growth. IGF-I is required for normal organismal growth, and in the pancreatic β-cell, the insulin/IGF-I signaling pathway is critical for normal and adaptive maintenance of β-cell mass. Expression of myrAkt1, an activated form of Akt, in the endocrine pancreas drives β-cell expansion through dramatic increases in both islet and β-cell size and number. Herein we present a comparative expression profiling of myrAkt1 transgenic islets that demonstrates the increased abundance of transcripts encoding proteins associated with growth, suppression of apoptosis, RNA processing, and metabolism. Although IGFBP5 is identified as a gene induced by Akt1 activation in the β-cell, Igfbp5 expression is not necessary for myrAkt1 to augment β-cell size or mass in vivo. However, in the absence of Igfbp5, mice demonstrate an increase in size and mild glucose intolerance. This is accentuated during diet-induced obesity, when Igfbp5-deficient mice have increased adiposity compared with wild-type mice on the same diet. These studies reveal a novel role for Igfbp5 in the control of growth and metabolism.

IGFBP5 is not required for enhanced beta cell growth in response to active Akt, but is important to normal organismal growth and glucose metabolism.

The insulin signaling pathway is responsible for regulation of glucose homeostasis via stimulation of glucose uptake and storage in peripheral tissues and by inhibition of glucose production in the liver, whereas IGF-I is a major regulator of organ and organismal growth. In addition to the classical insulin target tissues, insulin and IGF-I also act on the β-cell and are critical for the regulation of insulin secretory function and maintenance of normal and adaptive β-cell mass (1). For example, conditional deletion of either the insulin receptor (IR) (2) or the IGF-I receptor (3) in the β-cell impairs β-cell glucose-stimulated insulin secretion with only minor or no effect, respectively, on maintenance of β-cell mass. However, combined deletion of both receptors leads to severe hyperglycemia due to both β-cell functional defects and an inability to maintain normal β-cell mass (4). In wild-type animals and humans, when insulin demand is increased during physiological or pathological states such as pregnancy, obesity, or insulin resistance, insulin output must increase to maintain glucose homeostasis. This compensatory adaptation involves β-cell hyperplasia, hypertrophy, suppression of apoptosis, and enhanced insulin secretion. It is now generally accepted that inadequate β-cell compensation is the critical factor in the progression from insulin resistance to type 2 diabetes (5).

The canonical insulin/IGF-I signaling pathway involves activation of receptor tyrosine kinase activity and recruitment of the scaffolding proteins, IR substrates 1 (IRS1) and 2 (IRS2) (6). Receptor phosphorylation of the IRS proteins then leads to activation of phosphatidylinositol-3 kinase (PI3K) and increased levels of phosphoinositide-3,4,5-phosphate (PIP3) at the plasma membrane. A major target of PI3K activation is the serine/threonine kinase Akt (6). Previously, mice with β-cell-specific expression of an activated form of Akt1, myrAkt1, were generated to investigate the function of Akt1 in the β-cell (7,8). These mice display an approximately 8-fold increase in β-cell mass, which is due to an increase in the number and size of islets and a doubling in β-cell size. In addition, the proliferation rate is increased, suggesting enhanced cell division contributes to the augmentation of β-cell mass (7). Consistent with the increase in β-cell mass, the myrAkt1 transgenic (TG) mice are hyperinsulinemic and hypoglycemic and exhibit improved glucose tolerance. Thus, Akt1 positively affects β-cell expansion via stimulation of β-cell growth (size), proliferation, and potentially neogenesis.

The ability of Akt to mediate the effects of growth factors and nutrients on cell growth, proliferation, and survival is well documented in various cell types. These studies have also led to the identification of several targets of Akt, both direct and indirect, that are candidates for transmitting the growth and survival signals (9). For instance, in addition to its role in regulation of glycogen breakdown, the Akt substrate glycogen synthase kinase 3 (GSK3) is also involved in targeting various substrates for ubiquitin-mediated degradation. Many of the proteins that GSK3 phosphorylates are regulators of cell growth, such as c-Myc, β-catenin, and cyclin D1 and D2 (9). Akt also regulates the cell cycle inhibitors p21^CIP and p27^KIP1 by phosphorylation and sequestration in the cytosol (10,11). In addition, Akt influences transcription of cell cycle regulators such as p27^KIP1 via phosphorylation and inhibition of the Foxo family of forkhead transcription factors (12).

Specific to pancreatic β-cells, the G1 cell cycle proteins cyclin-dependent kinase 4 (CDK4) and the cyclins D1 and D2 are involved in postnatal β-cell proliferation and maintenance of β-cell mass (13,14). Recently, Fatrai et al. (15) demonstrated that CDK4 activity is required for the enhanced β-cell proliferation observed in mice with β-cell-specific overexpression of myrAkt1. In these mice, activation of Akt in the β-cell is associated with increased phosphorylation of GSK3β, increased protein expression of cyclins D1 and D2, and CDK4 activation. Although this study suggests one potential mechanism for the effect of Akt activation on β-cell proliferation, enhanced proliferation is only one component contributing to the overall increase in β-cell mass. Thus, the precise mechanism by which myrAkt1 elicits its TG phenotype, particularly the increase in cell size, remains uncertain (16). For example, haploinsufficiency of FOXO1, a transcription factor negatively regulated by Akt, restores PDX-1 expression and β-cell mass in IRS2 knockout mice; this suggests a potential link between Akt activation and PDX-1 expression (17). PDX-1 is an important transcription factor for both pancreas and islet development and normal β-cell function.

In this study, we have used comparative expression profiling of myrAkt1 TG islet mRNA to identify potential targets and gene networks that might mediate the effect of Akt1 activation on pancreatic β-cell mass expansion. To narrow the list of candidates, we have compared the transcriptional profile secondary to expression of active Akt1 with a physiological condition that causes enhanced growth, i.e. pregnancy. Analysis of differentially regulated genes provides insight into specific cellular functions such as growth, suppression of apoptosis, RNA processing, and metabolism that are influenced by Akt1 activation in the β-cell. Furthermore, we have identified the IGF-binding protein 5 (IGFBP5) as a gene highly up-regulated by myrAkt1 in the β-cell and have investigated its role in mediating the effect of myrAkt1 on augmentation of β-cell mass. In the course of these studies, we have serendipitously identified an unexpected role for Igfbp5 in the regulation of growth and metabolism.

Results

Effect of Akt1 activation on gene expression in pancreatic β-cells

To identify genes that are differentially regulated by Akt1 activation in the β-cell, we performed large-scale expression profiling on pancreatic islets isolated from 4- to 6-wk-old RIP-myrAkt1 TG mice. This microarray expression analysis identified 131 genes of 6299 distinct genes spotted on the PancChip 4.0 as differentially regulated (2.1%) (supplemental Fig. 1, published as supplemental data on The Endocrine Society’s Journals Online web site at http://mend.endojournals.org). Of the 131 genes, 68 were significantly (>1.4-fold) up-regulated and 63 significantly down-regulated in the RIP-myrAkt1 TG islets; genes encoding multiple classes of proteins were represented (supplemental Fig. 2).

To elucidate the physiological processes in the β-cell that are influenced by Akt1 activation, functional annotation analysis was performed to determine gene ontology functions and to identify functional groups that are overrepresented within the set of differentially expressed genes (Table 1). As expected, genes affecting both cell growth and cell death were overrepresented among the genes differentially regulated by myrAkt1 in TG islets. Importantly, the thymoma viral protooncogene protein 1 gene, encoding Akt1, was up-regulated 1.53-fold. The increased Akt1 expression is a good positive control and lends confidence to the validity of results from this analysis. Energy availability and the capacity to enhance protein synthesis are requirements for cellular growth. Therefore, it is interesting that several genes involved in the regulation of carbohydrate metabolism and RNA processing were differentially regulated in the myrAkt1 TG islets (Table 1).

Table 1.

Classified list of known genes identified by PancChip 4.0 analysis as significantly differentially expressed (>1.4 and 20% false discovery rate) in myrAkt1 TG islets compared with NTG islets

Entrez ID	Gene name	Fold change
Cell adhesion/extracellular matrix
20750	Secreted phosphoprotein 1 (Spp1)	3.13
13602	SPARC-like 1 (Sparcl1)	2.10
20733	Serine protease inhibitor, Kunitz type 2 (Spint2)	1.66
22329	Vascular cell adhesion molecule 1 (Vcam1)	1.58
53378	Syndecan binding protein (Sdcbp)	1.43
15894	Intercellular adhesion molecule (Icam1)	−1.45
Membrane- and organelle-associated proteins
12653	Chromogranin B (Chgb)	2.94
12505	CD44 antigen (CD44)	2.27
17748	Metallothionein 1 (Mt1)	2.14
16783	Lysosomal membrane glycoprotein 1 (Lamp1)	1.52
12010	β2-Microglobulin (B2m)	1.42
67182	Membrane-associated protein 17 (Pdzk1ip1)	−1.47
68585	Reticulon 4 (Rtn4; Nogo)	−1.59
80876	Interferon-induced transmembrane protein 2 (Ifitm2)	−1.60
216350	Transmembrane 4 superfamily member 3 (Tspan8)	−2.20
Receptors
67168	Purinergic receptor (family A group 5) (Plac8)	1.67
16004	IGF-II receptor (Igfr2)	−1.50
23920	Insulin receptor-related receptor (Insrr)	−1.89
77773	IGF-I receptor 1 (Igf1r)	−1.93
	Glucose-dependent insulinotropic polypeptide receptor (Gipr)	−2.30
Transporters
74413	Membrane-targeting (tandem) C2 domain containing 1 (Mtac2d1)	1.93
11927	ATX1 (antioxidant protein 1) homolog 1 (yeast) (Atox1)	1.89
69150	Sorting nexin 4 (Snx4)	1.60
66868	Major facilitator superfamily domain containing 1 (Mfsd1)	1.51
20515	Solute carrier family 20, member 1 (Slc20a1)	1.41
69354	Solute carrier family 38, member 4 (Slc38a4)	−1.59
18111	Neuronatin (Nnat)	−1.63
22139	Transthyretin (Ttr)	−1.89
Signal transduction
83965	Ectonucleotide pyrophosphatase/phosphodiesterase 5 (Enpp5)	1.57
106572	RIKEN cDNA 1700093E07 gene (Rab31)	1.52
19344	RAB5B, member RAS oncogene family	1.41
57436	γ-Aminobutyric acid (GABA(A)) receptor-associated protein-like 1 (Gabarapl1)	−1.40
66120	FK506 binding protein 11 (Fkbp11)	−1.98
Cytokines and chemokines
17319	Macrophage migration inhibitory factor (Mif)	1.57
67213	Chemokine-like factor super family 6 (Cmtm6)	1.42
Metabolism
Glycolysis
18746	Pyruvate kinase, muscle (Pkm2)	1.51
13806	Enolase 1, α nonneuron (Eno1)	1.45
14121	Fructose bisphosphatase 1 (Fbp1)	−1.56
TCA cycle
17449	Malate dehydrogenase 1, NAD (soluble) (Mdh1)	1.41
Oxidative phosphorylation
11951	ATP synthase, H+ transporting, mitochondrial F0 complex, subunit c (subunit 9), isoform 1 (Atp5g1)	1.64
228033	ATP synthase, H+ transporting, mitochondrial F0 complex, subunit c (subunit 9), isoform 3 (Atp5g3)	1.43
11973	ATPase, H+ transporting, V1 subunit isoform 1 (Atp6v1e1)E	1.41
Lipid
26971	Phospholipase A2, group IIF (Pla2g2f)	1.50
16956	Lipoprotein lipase (Lpl)	−3.61
(Continued)

Table 1A.

Continued

Entrez ID	Gene name	Fold change
Amino acid
67092	l-Arginine:glycine amidinotransferase (Gatm)	2.86
14719	Glutamate oxaloacetate transaminase 2, mitochondrial (Got2)	1.79
12411	Cystathionine β-synthase (Cbs)	−1.50
Vitamin
27973	Vitamin K epoxide reductase complex, subunit 1 (Vkorc1)	−1.43
Coenzyme
14871	Glutathione S-transferase, θ1 (Gstt1)	1.41
Other
68133	Glycine cleavage system protein H (aminomethyl carrier) (Gcsh)	1.51
432720	Similar to 3(20)α-hydroxysteroid/dihydrodiol/indanol dehydrogenase (Akr1c19)	−1.76
RNA-protein complex
76936	Heterogeneous nuclear ribonucleoprotein M (Hnrpm)	3.39
20813	Signal recognition particle 14 (Srp14)	1.75
54208	ADP-ribosylation factor-like 6 interacting protein 1 (Arl6ip1)	1.55
12972	Crystallin, ζ(Cryz)	1.45
11837	Acidic ribosomal phosphoprotein P0 (Arbp)	−1.40
67891	Ribosomal protein L4 (Rpl14)	−1.41
17938	Nascent polypeptide-associated complex-α polypeptide (Naca)	−1.41
12696	Cold inducible RNA-binding protein (Cirbp)	−1.57
19981	Ribosomal protein L37a (Rpl37a)	−1.62
13627	Eukaryotic translation elongation factor 1 α 1 (Eef1a1)	−1.66
20102	Ribosomal protein S4, X-linked (Rps4x)	−1.67
319195	Ribosomal protein L17 (Rpl17)	−1.85
Nuclear and DNA-binding proteins
21888	Transducin-like enhancer of split 4, homolog of Drosophila E(spl) (Tle4)	1.64
68024	Histone 1, H2bc (Hist1h2bc)	1.50
17536	Myeloid ecotropic viral integration site-related gene 1 (Mrg1)	−1.42
18124	Nuclear receptor subfamily 4, group A, member 3 (Nr4a3)	−1.42
18129	Notch gene homolog 2 (Drosophila) (Notch2)	−1.45
50496	E2F transcription factor 6 (E2f6)	−1.46
15331	High-mobility group nucleosomal binding domain 2 (Hmgn2)	−1.48
18033	Nuclear factor of κ−light chain gene enhancer in B cells 1, p105 (Nfkb1)	−1.65
15081	H3 histone, family 3B (H3f3b)	−1.76
13170	D site albumin promoter binding protein (Dbp)	−1.79
14009	Ets variant gene 1 (Etv1)	−1.98
Cell growth and death
16011	IGFBP5 (Igfbp5)	4.30
77579	Myosin, heavy polypeptide 10, nonmuscle (Myh10)	2.26
21987	Tumor protein D52-like 1 (Tpd52l1)	2.16
22350	Villin 2 (Vil2)	1.85
12700	Cytokine inducible SH2-containing protein (Cish)	1.75
12759	Clusterin (Clu)	1.58
11651	Thymoma viral protooncogene 1 (Akt1)	1.53
22195	Ubiquitin-conjugating enzyme E2 liter 3 (Ube213)	1.50
13643	Ephrin B3 (Efnb3)	−1.41
75723	Angiomotin-like 1 (Amotl1)	−1.54
21761	Mortality factor 4 like 1 (Morfl1)	−1.55
11746	Annexin A4 (Anxa4)	−1.75
18616	Paternally expressed 3 (Peg3)	−2.63
Cytoarchitecture
18040	Neurofilament 3, medium (Nefm)	2.72
11867	Actin-related protein 2/3 complex, subunit 1B (Arpc1b)	1.63
22145	Tubulin, α 4 (Tuba4)	1.53
19241	Thymosin, β 4, X chromosome (Tmsb4x)	−1.49
Hormones
16334	Insulin II (Ins2)	−1.73
14526	Glucagon (Gcg)	−2.08
20604	Somatostatin (Sst)	−2.33
56373	Carboxypeptidase B2 (plasma) (Cpb2)	4.64
(Continued)

Table 1B.

Continued

Entrez ID	Gene name	Fold change
94284	UDP glycosyltransferase 1 family polypeptide A6 (Ugt1a6a),	1.99
26879	UDP-Gal:βGlcNAc β1,3-galactosyltransferase, polypeptide 3 (B3galnt1)	1.87
13039	Cathepsin L (Cts1)	1.79
11883	Arylsulfatase A (Arsa)	−1.45
12876	Carboxypeptidase E (Cpe)	−1.56
76485	Glycosyltransferase 8 domain containing 1 (Glt8d1)	−1.62
Miscellaneous
103537	Mbt domain containing 1 (Mbtd1)	2.03
66442	Spindle pole body component 25 homolog (S. cerevisiae) (Spbc25)	2.00
100604	RIKEN cDNA E430036I04 gene (Lrrc8c)	1.98
55951	Brain protein 44-like (Brp441)	1.87
225131	WW domain containing adaptor with coiled-coil (Wac)	1.83
20363	Selenoprotein P, plasma, 1 (Sepp1)	1.79
16061	Ig heavy chain (J558 family) (Igf-VJ558)	1.53
67702	Ring finger protein 149 (Rng149)	1.44
81018	Zinc finger protein 313 (Zfp313)	−1.43
27886	Expressed sequence 2 embryonic lethal	−1.43
406217	RIKEN cDNA 2410004M13 gene (Bex4)	−1.48
373070	miRNA-containing gene (Mirg)	−1.98

Pregnancy is a condition of increased body mass and metabolic stress that stimulates increases in β-cell mass (18). Because overexpression of activated Akt is a very artificial means to increase growth, we compared the myrAkt1 microarray data set with an array comparing differential gene expression in islets experiencing a physiological form of enhanced growth, i.e. pregnancy. Interestingly, we found that mRNA encoding Akt1 was increased in the islets from pregnant mice and that many genes are differentially regulated and change in a similar direction in both paradigms. This analysis is most informative when the Ingenuity metabolic/cell-to-cell signaling network for each array is compared (supplemental Fig. 3). In addition to Akt1 induction, IGFBP5, cytokine-inducible Src homology-2-containing protein (CISH), leucine-rich repeat containing 8C (LRRC8C), and villin (Vil2) are up-regulated in both experimental systems. The genes encoding lipoprotein lipase (Lpl), insulin 2 (Ins), preproglucagon (GCG), and somatostatin (SST) are similarly decreased in both expression profiles. This analysis demonstrates that Akt1 expression is induced in pregnancy, a condition during which β-cell mass expands and islet function is optimized. Comparison of both gene expression profiles points to novel Akt targets, such as IGFBP5 and lipoprotein lipase, that might contribute to the effect of Akt1 on β-cell growth and function. Furthermore, the fact that these genes are similarly regulated in two situations in which β-cell mass is increased provides additional support for an important role downstream of Akt1 activation.

As described above, Igfbp5 was identified both as a gene induced in the islet during pregnancy and as one of the most highly up-regulated genes in the myrAkt1 TG islet array. IGFBP5 is a member of the IGF-I-binding protein family that bind with high affinity to both IGF-I and IGF-II (19). Igfbp5 mRNA expression is induced by treatment with IGF-I in both aortic and vascular smooth muscle cells in a wortmannin- and rapamycin-sensitive manner, suggesting that induction of Igfbp5 is downstream of the PI3K/mTOR signaling pathway (20,21). Considering the numerous studies implicating IGFBP5 as a regulator of cellular growth and differentiation (22,23), we selected Igfbp5 as a plausible candidate to mediate the effects of myrAkt1 in the β-cell.

In vivo and in vitro confirmation of myrAkt1-induced increase in Igfbp5 expression

Igfbp5 mRNA and protein expression was assessed in islets isolated from non-TG (NTG) or myrAkt1 TG mice (Fig. 1). Igfbp5 mRNA expression was increased 7-fold in myrAkt1 TG islets compared with NTG islets, in agreement with the microarray analysis (Fig. 1A). Western blot analysis also showed a dramatic increase in IGFBP5 protein expression in the myrAkt1 TG islets (Fig. 1B). Additionally, Igfbp5 mRNA expression was examined in the MIN6 β-cell line, in which the myrAkt1 transgene was stably expressed. Figure 2A shows confirmation of myrAkt1 protein expression by Western blot analysis. The myrAkt1 transgene migrates faster than the endogenous protein due to deletion of the pleckstrin homology domain. Increased Akt1 phosphorylation and phosphorylation of the downstream target GSK indicates activation of the Akt1 signaling pathway compared with cells expressing the empty vector (pLNCX). Analysis of Igfbp5 expression by quantitative PCR demonstrated an increase of approximately 2-fold in the myrAkt1-expressing cells compared with controls (Fig. 2B). The relatively lesser induction of Igfbp5 mRNA in the myrAkt1 stable MIN6 cells compared with the myrAkt1 TG islets might be due to differences in the level of expression of the transgene or a higher basal Igfbp5 expression in the immortalized β-cell line. To ask whether regulation of Igfbp5 by Akt is a generalized phenomenon, we examined the effect of acute Akt1 activation on induction of Igfbp5 expression in primary mouse embryonic fibroblasts (MEFs). Wild-type MEFs were infected with either adenovirus encoding green fluorescent protein (ad-GFP) or myrAkt1 (ad-myrAkt1). The myrAkt1 cDNA expressed by this virus contains the full-length Akt1 sequence. As shown in Fig. 3A, ad-myrAkt1 infection increased total and phosphorylated Akt1 expression compared with ad-GFP-infected MEFs. Adenoviral overexpression of myrAkt1 induced a dramatic 50-fold increase in Igfbp5 mRNA expression compared with ad-GFP-infected cells (Fig. 3B). Together, these data confirm that Akt1 activation increases expression of Igfbp5 in the β-cell.

Figure 1.

Igfbp5 mRNA and protein is up-regulated in the myrAkt1 TG islets, and expression is absent in Igfbp5^−/− islets. A, Islets from 4- to 6-wk-old mice were isolated and total RNA was extracted and assessed by quantitative PCR for Igfbp5. Igfbp5 mRNA was corrected for Gapdh expression (n = 3 per genotype). Error bars represent sd. *, P < 0.01. B, Islets were isolated from 12- to 14-wk-old mice, and protein lysates were prepared from a total of 30 islets per genotype and processed for immunoblotting as described in Materials and Methods. WT, Wild-type; P-Akt, phospho-Akt.

Figure 2.

Igfbp5 expression is increased in MIN6 cells by stable expression of myrAkt1. MIN6-myrAkt1 cells were plated into six-well plates and harvested at approximately 80% confluency after an overnight culture in DMEM containing 3 mm glucose. A, Cell extracts were processed for immunoblotting as described in Materials and Methods. B, Total RNA was extracted, and Igfbp5 expression was assessed by quantitative PCR. Igfbp5 mRNA level was corrected for TBP expression. The graph reflects a representative experiment from a total of three separate experiments performed in triplicate. Error bars represent sd. *, P < 0.01. P-GSK, Phospho-GSK.

Figure 3.

MyrAkt1 induces Igfbp5 mRNA expression in primary MEFs. Primary MEFs were infected with either ad-GFP or myrAkt1 and then subcultured for 48 h before harvesting. A, Cells were lysed and processed for immunoblotting as described in Materials and Methods. B, Igfbp5 expression was assessed by quantitative PCR. Igfbp5 mRNA levels were normalized to TBP expression. The graph represents the average of four separate experiments. The error bars indicate sd. *, P < 0.001. P-Akt, Phospho-Akt.

Derivation and analysis of myrAkt1 TG mice deficient for Igfbp5

We next sought to determine whether the induction of Igfbp5 is required for increased β-cell size or islet number stimulated by myrAkt1. The most convincing strategy to determine whether IGFBP5 is required for changes in β-cell mass in the myrAkt1 TG islets is to examine, in vivo, the consequences of myrAkt1 expression in Igfbp5-deficient mice. To this end, we intercrossed the myrAkt1 TG mice with mice deficient for Igfbp5 (Igfbp5^−/−;myrAkt1 TG). The Igfbp5-null mice have been characterized in regard to the effect of Igfbp5 on whole-body and organ growth as well as mammary gland development and involution after lactation (24). Development and growth is normal in the absence of Igfbp5, but there is a small delay in mammary gland involution. The effect of Igfbp5 deletion on regulation of glucose homeostasis has not been examined previously (24).

Deletion of Igfbp5 was confirmed in the Igfbp5^−/− mice by examining Igfbp5 protein expression in both kidney and islet tissue. Igfbp5 expression is abundant in the kidney and was detected by Western blot in Igfbp5 wild-type samples. No Igfbp5 was detected in tissues taken from the Igfbp5^−/− mice (data not shown). Although Igfbp5 levels were below the detection limit in wild-type islets, myrAkt1 greatly increased IGFBP5 protein expression. However, in the Igfbp5^−/−;myrAkt1 TG islets, IGFBP5 protein expression was absent, confirming deletion of Igfbp5 in these mice (Fig. 1B).

Analysis of β-cell mass and size in Igfbp5^−/−;myrAkt1 TG mice

To determine whether β-cell mass was augmented in Igfbp5^−/−;myrAkt1 TG mice, pancreas sections were stained with insulin and glucagon (Fig. 4). Igfbp5^−/−;myrAkt1 TG mice displayed a similar expansion in islet size and number as the Igfbp5^+/+;myrAkt1 TG. Of note, although the Igfbp5^−/− mice exhibited impaired glucose tolerance (see below), islet size and number did not appear different from the Igfbp5^+/+ mice.

Figure 4.

Islet size and β-cell mass are similarly augmented in Igfbp5^−/−;myrAkt1 TG mice. Indirect immunofluorescence of pancreas sections from 9- to 10-wk-old male mice of the indicated genotypes. Insulin is green, glucagon is red, and Hoechst 33358 (nuclei) is blue.

To determine whether β-cell size was altered in either the Igfbp5^−/−;NTG mice or the Igfbp5^−/−;myrAkt1 TG compared with Igfbp5^+/+ mice, we stained pancreas sections with β-catenin to highlight the cell periphery and hemagglutinin (HA) to confirm transgene expression (Fig. 5). β-Cell size appeared similar in Igfbp5^+/+;NTG and Igfbp5^−/−;NTG pancreata. As expected, expression of the myrAkt1 transgene led to a dramatic increase in β-cell size, but there was no difference between cells expressing the myrAkt1 transgene in the presence or absence of Igfbp5. Together, these results demonstrate that deletion of Igfbp5 does not impair the ability of myrAkt1 to augment both β-cell size and mass in vivo.

Figure 5.

β-Cell size is similarly increased in Igfbp5^−/−;myrAkt1 TG mice as in Igfbp5^+/+;myrAkt1 TG mice. Indirect immunofluorescence of pancreatic sections from male 9- to 10-wk-old mice. A and C, Igfbp5^+/+;NTG; B and D, Igfbp5^−/−;NTG; E and G, Igfbp5^+/+;myrAkt1 TG; F and H, Igfbp5^−/−;myrAkt1 TG. Cells were stained for HA (green) to detect expression of the myrAkt1 transgene and for β-catenin (red) and with Hoechst 33358 (nuclei, blue). A, B, E, and F show β-catenin staining only.

Because the β-cell expansion in myrAkt1 TG mice leads to metabolic abnormalities such as hyperinsulinemia, hypoglycemia, and improved glucose tolerance, we assessed glucose homeostasis in myrAkt1 mice deficient for Igfbp5 (7,8). As demonstrated previously, mice wild-type for Igfbp5 and expressing the myrAkt1 transgene (Igfbp5^+/+;myrAkt1 TG) exhibit fasting hypoglycemia. Igfbp5^−/− mice expressing the myrAkt1 transgene (Igfbp5^−/−;myrAkt1 TG) did not exhibit a significant decrease in fasting blood glucose level, although there was a trend in this direction (Fig. 6A). Random fed blood glucose levels were not significantly different regardless of myrAkt1 transgene expression and/or the absence of Igfbp5 (Fig. 6B). As expected, glucose tolerance in the Igfbp5^+/+;myrAkt1 TG mice was significantly improved compared with the Igfbp5^−/−;NTG mice. Likewise, the Igfbp5^−/−;myrAkt1 TG mice displayed improved glucose tolerance compared with either the Igfbp5^+/+;NTG mice or Igfbp5^−/−;NTG mice, such that the glucose tolerance curves in myrAkt1 TG mice were indistinguishable whether or not they expressed Igfbp5. Surprisingly, Igfbp5^−/−;NTG mice were significantly glucose intolerant (Fig. 6C). As measured using the euglycemic-hyperinsulinemic clamp technique, Igfbp5^−/− mice showed a trend toward reduced glucose disposal and enhanced hepatic glucose output in the presence of insulin that did not achieve statistical significance (data not shown). Serum insulin levels after an overnight fast or during the glucose tolerance test were increased in the Igfbp5^+/+;myrAkt1 TG mice compared with Igfbp5^+/+;NTG mice; however, this increase did not reach significance, possibly due to the small number of mice. In contrast, the serum insulin levels in the Igfbp5^−/−;myrAkt1 TG mice were significantly elevated at both fasting and 15 min after glucose injection compared with Igfbp5^−/−;NTG mice (Fig. 6D). The improved insulin secretion observed in the latter mice suggests that Igfbp5 is not required for the effects of myrAkt1 on the β-cell.

Figure 6.

Regulation of glucose metabolism in male Igfbp5^−/−;myrAkt1 TG mice. Glucose values were obtained through tail bleeds from either 14- to16-h fasted (n = 9–24) (A) or random fed (n = 5–12) (B) mice. *, P < 0.001 for NTG vs. myrAkt1 TG. C, Intraperitoneal glucose tolerance tests were performed on 14- to 16-h fasted male mice (8–10 wk old) of the indicated genotypes (n = 9–24). Error bars indicate se. *, P < 0.05; **, P < 0.01 vs. NTG values. D, Serum insulin levels in blood obtained by tail bleeds were ascertained on animals immediately before and 15 min after glucose injection (n = 8–19). Error bars represent se. *, P < 0.01 for 0 vs. 15 min; **, P < 0.01 for Igfbp5^−/−;NTG vs. Igfbp5^−/−;myrAkt1 TG mice.

Both the Igfbp5^−/− and Igfbp5^−/−;myrAkt1 TG mice exhibited a small but significant increase in body weight (8.5%) at 8–10 wk of age compared with wild-type mice (Fig. 7A). To determine whether the increased body weight is present at birth or develops postnatally, we weighed pups daily from d 1–7, once per week for 12 wk, and then every 2 wk. The weights of the pups were not significantly different during the first 7 d after birth (data not shown). Male Igfbp5^−/− mice did not show a significant difference in body weight until 6 wk of age with the increase reaching 11% by 12 wk. In contrast, female Igfbp5^−/− mice exhibited significantly enhanced weight gain by the end of the first week (Fig. 7C). At 5 wk of age, there was an approximately 12% increase in body weight in the Igfbp5^−/− mice compared with wild type; this difference grew to about 19% by 7 wk of age. At 20 wk of age, male Igfbp5^−/− mice exhibit increased length (from rump to snout) and a significant increase in lean mass as determined by dual-energy x-ray absorptiometry (DEXA) analysis (Figs. 7D and 10E). The increased weight in the knockout mice can be specifically attributed to the changes in lean mass, because fat mass and bone mineral composition and density were the same in both groups (data not shown). We also monitored energy expenditure in these mice to examine whether changes in thermogenesis, respiration, or activity attributed to the enhanced growth phenotype. As shown in Table 2, there were no differences between genotypes in respiration, heat generation, or food intake. However, the Igfbp5^−/− mice exhibited decreased total activity compared with wild type during the 24 h they were monitored.

Figure 7.

Igfbp5^−/− mice exhibit increased postnatal body weight. A, Body weights were measured on 14- to 16-h fasted male mice between 8 and 10 wk of age (n = 9–24). Error bars indicate se. *, P < 0.01 for Igfbp5^−/− and Igfbp5^−/−;myrAkt1 TG compared with NTG or myrAkt1 TG mice. Growth curves for male (B) and female (C) mice. Weights were monitored on random fed mice beginning 1 wk after birth and once per week thereafter for 12 wk and then every 2 wk until 14 wk (males, n = 10–12; females, n = 7–9). Error bars indicate se. *, P < 0.01 for Igfbp5^+/+ vs. Igfbp5^−/− mice. D, DEXA analysis; E, body length.

Figure 10.

Igfbp5^−/− mice have significantly elevated fasting blood glucose and are more insulin resistant than Igfbp5^+/+ mice after 9 wk on a high-fat diet. A, Glucose levels were determined from blood obtained through tail bleeds on 5- to 6-h fasted mice (Igfbp5^+/+ n = 5; Igfbp5^−/− n = 4). *, P < 0.05. B, Plasma insulin levels in blood obtained from tail bleeds were ascertained on 5- to 6-h fasted mice (Igfbp5^+/+ n = 5; Igfbp5^−/− n = 4). *, P < 0.04. C, Blood glucose levels after a single injection of insulin (Igfbp5^+/+ n = 5; Igfbp5^−/− n = 4). Error bars indicate se. *, P < 0.04.

Table 2.

Energetics in Igfbp5-deficient mice

Genotype	Igfbp5+/+ perbody mass	Igfbp5−/− perbody mass	Igfbp5+/+ perlean mass	Igfbp5−/− perlean mass
VO2 (ml/kg · h)	125,336 ± 6,066.5	117,029 ± 4,046.3	146,931 ± 6,441.1	131,370 ± 3,909.7
VCO2 (ml/kg · h)	103,830 ± 7,434.3	101,193 ± 3,251.8	114,251 ± 5,008.4	113,507 ± 7,009.6
RER	46.1 ± 1.18	48.3 ± 1.24	NA	NA
Heat (kcal/h · kg)	659.8 ± 25.08	596.2 ± 23.01	802.8 ± 27.86	731.0 ± 22.14
Food intake (g)	2.40 ± 0.375	3.00 ± 0.368	NA	NA
Normalized food intake (g/kg)	89.0 ± 16.06	92.5 ± 14.26	106.7 ± 17.68	112.5 ± 15.06
Total activity	71,405 ± 13,914.6	31,886 ± 5,758.0a	NA	NA

Wild-type (n = 4) and Igfbp5^−/− (n = 4) male, approximately 20-wk-old mice were used to assess food intake and respiratory exchange ratio (RER) under conditions of free access to food and water using the Comprehensive Laboratory Animal Monitoring System (CLAMS; Columbus Instruments, Columbus, OH) for 24 h. Oxygen consumption (VO₂) and carbon dioxide production (VCO₂) were measured in CLAMS and used to calculate respiratory quotient/RER. NA, Not appropriate; RER, respiratory exchange ratio. 

^aP < 0.05, wild type vs. Igfbp5^−/−. 

To elicit a more dramatic metabolic phenotype, wild-type and Igfbp5^−/− mice were placed on a high-fat diet. Igfbp5^−/− gained significantly more weight than wild-type mice and demonstrated greater linear growth (Fig. 8). During diet-induced obesity, Igfbp5^−/− had a different body composition compared with wild-type mice, with a significant increase in adiposity (Fig. 9, A and B). This was evident by DEXA scanning or by assessing the weight of the fat pad relative to body weight (Fig. 9, B and C). As expected, associated with the enhanced obesity, Igfbp5^−/− mice had elevated serum glucose and insulin concentrations compared with wild-type mice on the same diet (Fig. 10). Moreover, they were insulin resistant as assessed by an insulin tolerance test. Thus, Igfbp5 is required for maintenance of normal body size and insulin sensitivity, particularly when challenged by a high-fat diet.

Figure 8.

Igfbp5^−/− (BP5−/−) mice gain more weight on a high-fat diet compared with Igfbp5^+/+ (BP5+/+) mice. A, Mice were fasted overnight, and weight gain was calculated by subtracting the weight at the indicated time from weight at the initiation of high-fat diet feeding. *, P < 0.01. B, Body weight expressed as percentage of weight at initiation of diet (n = 5 per genotype, except at 4 wk, Igfbp5^+/+ n = 2 and Igfbp5^−/− n = 3). *, P < 0.04. C, Igfbp5^+/+ and Igfbp5^−/− mice at 16 wk on high-fat diet. D, Body length of mice at 16 wk on HFD (n = 5 per genotype). E, Body weight at 16 wk on high-fat diet (random fed value) (n =5 per genotype). Error bars indicate se. *, P < 0.05.

Figure 9.

Body composition analysis of Igfbp5^−/− mice after 16 wk on a high-fat diet. A, Total fat and lean mass (n =5 per genotype); B, fat and lean mass corrected for total mass; C, isolated liver and epididymal fat pad weights; D, liver and epididymal fat pad weights corrected for body weight (n = 5 per genotype). Error bars represent se. *, P < 0.01.

Discussion

Regulation of glucose homeostasis throughout life requires dynamic fluctuations in pancreatic β-cell mass. The endocrine organ achieves this plasticity through a careful balance of β-cell renewal and β-cell loss. The serine/threonine kinase Akt has been implicated as a critical factor linking diverse nutrient, growth factor, and incretin stimuli with processes related to β-cell mass expansion such as proliferation, hypertrophy, and cell survival. Importantly, the factors downstream of Akt1 that mediate these growth-related effects have not been defined completely. In this study, by assessing the gene expression profile of RIP-myrAkt1 TG islets, we identify potential targets and/or gene networks that might contribute to the effect of Akt1 activation on enhancing β-cell mass and/or function. We demonstrate that Akt1 activation induces gene networks with functions related to cell growth, inhibition of apoptosis, RNA processing, and metabolism. From this analysis, we also identify the IGFBP5 as a gene induced by Akt1 in the β-cell. We find that Igfbp5 deficiency does not prevent the effect of myrAkt1 on expansion of β-cell mass in vivo. However, IGFBP5 is important for regulation of postnatal growth and possibly glucose homeostasis.

Gene expression analysis identified genes with interesting functions that might contribute to the myrAkt1 TG phenotype. We decided to focus our efforts on IGFBP5, based on its induction in the array analyses and its purported effects on growth. IGFBP5 is a member of a family of six IGFBPs that bind to and modulate the activity of IGF-I and IGF-II (19). Both IGF-I and IGF-II are potent stimuli for pancreatic β-cell growth (25). In β-cell tissue culture lines, IGF-I potentiates the mitogenic effect of glucose (26). Overexpression of IGF-I in the β-cell protects them from streptozotocin-induced β-cell death (27). Furthermore, IGF-I mRNA increases rapidly in the area of the regenerating pancreas after 90% pancreatectomy (28). Lastly, Devedjian et al. (29) demonstrated that overexpression of IGF-II in the β-cell leads to a 3-fold increase in β-cell mass. Together, these data indicate that IGF-I can positively influence β-cell growth and development under various conditions. The IGFBPs have been postulated to modulate IGF activity by various mechanisms, including protecting IGFs from proteolytic degradation, targeting IGFs in serum to specific tissues, and regulating local IGF availability to receptors by sequestration in extracellular storage pools (19). There has been limited investigation into the effects of IGFBP on β-cell growth and/or function.

Many studies have demonstrated a role for IGFBP5 in promoting aspects of growth such as differentiation and proliferation in non-β-cell lines and tissues. For example, Yin et al. (30) report that endogenous IGFBP5 is important for maintaining bone cell survival and differentiation in human osteoblastic cells. Igfbp5 has also been identified by serial analysis of gene expression (SAGE) as an abundant transcript during cardiac development (31). The ability of IGFBP5 to regulate growth has been divided into both IGF-dependent and -independent effects. IGFBP5 stimulates proliferation of murine calvarial cells, a primary osteoblast cell line, independently of IGF-I (32). IGFBP5 expression is also up-regulated in various tumor sections, and cell lines such as pancreatic islet adenocarcinomas, thyroid carcinoma, and prostate and breast cancers (33,34,35,36).

IGF-I induces IGFBP5 expression in aortic and vascular smooth muscle cells. This induction is blocked by both LY294002 and rapamycin, suggesting that the PI3K/mTOR signaling pathway is involved in stimulation of IGFBP5 expression (20,21,37). IGFBP5 has also been identified as a gene induced by TG overexpression of myrAkt1 in the heart, a model of cardiac hypertrophy (38,39). Furthermore, acute overexpression of myrAkt1 by adenovirus infection induces IGFBP5 expression in neonatal rat ventricular cardiomyocytes (38). However, the role of IGFBP5 in regulating cardiac hypertrophy in these models has not been assessed.

Igfbp5 mRNA and protein are expressed in islets and ductal cells in the developing and adult mouse pancreas (40). Interestingly, Igfbp5 is recognized as a transcript enriched in the developing pancreatic endoderm (41). However, the role of IGFBP5 in β-cell growth and/or function has not been examined previously. In this study, we present data that IGFBP5 is induced by overexpression of myrAkt1 in the β-cell and during pregnancy, both conditions of augmented β-cell mass. However, even in mice deficient for IGFBP5, myrAkt1 overexpression still augmented β-cell mass. The lack of IGFBP5 does not compromise β-cell function because the Igfbp5^−/−;myrAkt TG mice exhibit fasting hyperinsulinemia and improved glucose tolerance. These experiments demonstrate that Igfbp5 is not necessary for the effect of myrAkt1 on expansion of β-cell mass in vivo. However, it is possible that other members of the Igfbp family could have compensated for the absence of Igfbp5, because IGFBPs 1–5 are expressed in the β-cell and surrounding islet cell types (40).

Ubiquitous overexpression of IGFBP5 inhibits growth and development in vivo, suggesting that IGFBP5 might function primarily as an inhibitor of IGF action (42). A requirement for IGFBP5 in normal growth was not detected in several previous studies (24,43). The most likely explanation for the different findings reported herein are the additional crosses of Igfbp5^−/− alleles onto a C57BL/6 genetic background. In this study, we have demonstrated that Igfbp5 is required for normal growth only during the postnatal period. Of note, overexpression of IGFBP5 results in decreased prenatal growth and maximal growth retardation that occurred by about 2–3 wk after birth, suggesting that IGFBP5 regulates growth before the onset of GH-stimulated IGF synthesis (42). From this analysis, the authors surmised that the effect of IGFBP5 on growth was either through regulation of IGF-II action or through IGF-independent means. However, our results show that in the absence of IGFBP5, enhanced growth does not develop until after GH-stimulated IGF synthesis has begun. In contrast to the IGFBP5 overexpression study, the Igfbp5^−/− phenotype presents the antithesis of Igf1-null mice, which exhibit progressively increased growth retardation after about 3 wk postnatal development (44). Thus, although artificially high levels of IGFBP5 lead to IGF-I-independent inhibition of the early stages of growth, we suggest that IGFBP5 influences growth during postweaning development most likely through IGF-I.

Interestingly, during this study, we observed that IGFBP5-deficient mice exhibit impaired glucose tolerance. Given this, it was surprising that the euglycemic hyperinsulinemic clamp showed no abnormalities. Perhaps this simply reflects the subtlety of the defect; alternatively, the resistance might have been related to a change in a circulating factor, for example glucose, that is held constant in the clamp but not the glucose tolerance test. In any case, the insulin-resistance phenotype was greatly enhanced when the mice were placed on a high-fat diet. In addition to its well-known anabolic effects, IGF-I also influences carbohydrate metabolism. For example, mice with either liver- or skeletal muscle-specific deletion of IGF-I exhibited impaired glucose intolerance and insulin action (45,46). Intraperitoneal injection of IGF-I induces a pronounced decrease in blood glucose level in IR-deficient mice (47). IGFBP proteins influence glucose homeostasis as well. For instance, hyperglycemia and glucose intolerance occur as a result of global overexpression of either IGFBP1 or IGFBP3 (48,49). Our findings demonstrate that in addition to effects on growth, IGFBP5 is also important for normal glucose homeostasis.

Materials and Methods

Materials

DMEM, fetal bovine serum (FBS), penicillin/streptomycin solution, and sodium pyruvate solution were obtained from Invitrogen (Carlsbad, CA). Fugene transfection reagents were from Roche (Mannheim, Germany). The Igfbp5 antibodies were from R&D Systems (Minneapolis, MN) and Santa Cruz Biotechnology (Santa Cruz, CA). The GFP antibody was from BD Biosciences (Palo Alto, CA), and the actin antibody was from Abcam (Cambridge, MA). Horseradish peroxidase-conjugated secondary antibodies were from Santa Cruz Biotechnology. Secondary antibodies for use with the Odyssey infrared imaging system (LICOR Biosciences, Lincoln, NE) were purchased from Rockland Inc. (Gilbertsville, PA). All other antibodies were from Cell Signaling Technology, Inc. (Beverly, MA).

Animals and genotype analysis

The derivation of both the RIP-myrAkt1 TG mice (8) and the Igfbp5 knockout mice (24) have been reported previously. To generate Igfbp5^−/−;myrAkt TG mice, Igfbp5^−/− mice were mated to wild-type C57BL/6 females (The Jackson Laboratory, Bar Harbor, ME). Igfbp5^+/− progeny were then mated inter se to generate Igfbp5^−/− mice. The Igfbp5^+/− or Igfbp5^−/− mice were mated among each other to propagate the Igfbp5 line. Next, either Igfbp5^+/− or Igfbp5^−/− mice were intercrossed to RIP-myrAkt1 TG mice (>10 generations on C57bL/6 background). The resulting Igfbp5^+/−;myrAkt1 TG mice were mated to either Igfbp5^+/− or Igfbp5^−/− mice to generate mice of the appropriate genotypes. Heterozygous Igfbp5 mice were bred together to generate cohorts for growth-curve analysis. Weights for the growth-curve analysis were obtained on normally fed mice from 1400–1600 h. Genotyping was performed by PCR analysis using genomic DNA isolated from tail snips of newborn mice. All procedures involving mice were conducted in accordance with approved Institutional Animal Care Use Committee protocols.

Islet isolation and Western blot

Pancreatic islets of Langerhans were isolated from 4- to 6-wk-old NTG or RIP-myrAkt1 TG mice by collagenase (Crescent Chemicals, Hauppauge, NY) digestion followed by purification through a Ficoll gradient as described previously (50). Purified islets were handpicked using a dissecting microscope. For Western blot analysis, islets were washed once in ice-cold 1× PBS and then lysed in ice-cold buffer containing 140 mm NaCl, 10 mm Tris (pH 7.4), 200 mm NaF, 10% glycerol, 1% Nonidet P-40, 1× Complete protease inhibitor cocktail (Roche) and 1× phosphatase inhibitor cocktail 1 and 2 (Sigma Chemical Co., St. Louis, MO). The islet lysates were homogenized by freezing-thawing three times in liquid nitrogen followed by a 37 C water bath. Insoluble material was pelleted by centrifugation at 10,000 rpm at 4 C for 10 min. The supernatant was collected and protein was measured by bicinchoninic acid assay (Pierce, Rockland, IL).

RNA isolation and real-time PCR analysis

Total RNA from islets was isolated in Trizol (Invitrogen) according to the manufacturer’s instructions. Total RNA from MIN6 cells or MEFs was isolated using the RNeasy Plus kit (Ambion, Austin, TX). RNA was reverse transcribed using random decamer primers, Retroscript II reverse transcriptase, and the accompanying reagents (Ambion) following the manufacturer’s directions. PCR mixes were assembled using the Brilliant SYBR Green QPCR Master Mix (Stratagene, La Jolla, CA). Reactions were performed using the SYBR Green (with dissociation curve) program on the Mx3000 quantitative PCR system (Stratagene). All reactions were performed in triplicate with the reference dye normalization, and median cycle threshold values were used for analysis. The primer sequences were as follows: Igfbp5 forward, 5′-GCAAAGCGTTGGAATAGAGG-3′; Igfbp5 reverse, 5′-AGATTCAGCCTTGGTTGTGG-3′; Gapdh forward, 5′-GGCAAAGTGGAGATTGTTGC-3′; Gapdh reverse, 5′-TGACAAGCTTCCCATTCTCG-3′; TATA-binding protein (TBP) forward, 5′-CCCCTTGTACCCTTCACCAAT-3′; TBP reverse, 5′-GAAGCTCGGTACAATTCCAG-3′.

Microarray expression profiling, data analysis, and gene ontology analysis

Total RNA (100 ng) for each sample (n = 3 per genotype) was amplified using the MessageAmp aRNA kit (Ambion). The concentration and integrity of each sample was determined both before and after amplification using an Agilent Bioanalyzer Lab-ON-A-Chip Nano 6000 chip. Microarray expression profiling, data, and gene ontology analysis were described previously (51). Fluorescently labeled (CyDye; Amersham Pharmacia Biotech Ltd., Piscataway, NJ) cDNA was hybridized to the PancChip version 4.0. This chip contains a set of 13,824 cDNAs including 3450 IMAGE clones chosen for their expression in pancreas and liver, 300 samples from the University of Pennsylvania’s Functional Genomics Core in-house collection, and 7226 nonredundant clones collected from pancreas libraries and control genes. The data were normalized using the print-tip lowess method using the SMA (Statistical Microarray Analysis) package in R (52). For statistical analysis, genes were called differentially expressed using the latest version of Significance Analysis of Microarrays (SAM) with a false discovery rate of 20% (53).

Cell culture and Western blot analysis

MIN6 cells (kindly provided by Prof. Jun-Ichi Miyazaki, Osaka University, Osaka, Japan) were used between passages 28 and 40 at approximately 80% confluence. MIN6 cells were grown in DMEM containing 25 mm glucose supplemented with 15% FBS, 1 mm sodium pyruvate, 50 μm β-mercaptoethanol, 100 μg/ml streptomycin, and 100 U/ml penicillin, equilibrated with 5% CO₂, 95% air at 37 C. MIN6 myrAkt1 stable cells were generated by retroviral infection followed by antibiotic selection with neomycin. The myrAkt1 construct was generated by inserting cDNA for human Akt1 (containing a deletion of the pleckstrin homology domain (amino acids 4–129) and addition of a 14-amino-acid myristoylation sequence (8) into the retroviral vector pLNCX1. The generation of retrovirus was as described previously with slight modifications (54). Briefly, ecotropic BOSC cells (a gift from Warren S. Pear, University of Pennsylvania, Philadelphia, PA) were transiently transfected with pVSV G and pCgP pantropic retroviral packaging constructs and retroviral vector (pLNCX1-HAmyrAkt1). Cell-free viral supernatants were harvested at both 24 and 48 h and used to infect MIN6 cells (passage 29). Primary wild-type MEFs were established from embryonic d-13.5 embryos as described previously (55). MEFs were cultured in DMEM containing 25 mm glucose supplemented with 10% FBS and 100 μg/ml streptomycin, 100 U/ml penicillin. For western blot analysis, cells were washed twice in ice-cold 1× PBS and then lysed in ice-cold buffer (see islet isolations above for recipe). 5 μl of lysate was removed for analysis of total protein by BCA assay (Pierce, Rockland, IL).

Adenovirus infection

The myrAkt1 adenoviral constructs have been described previously (56). The CMV-eGFP adenovirus was obtained from the University of Pennsylvania Viral Vector Core. For adenovirus infection of MEFs, the infection was optimized by preincubation of the virus in DMEM containing 0.5% BSA (Sigma) and 0.5 μg/ml poly-l-lysine (Sigma) for 1.5 h at room temperature. Wild-type primary MEFs were washed once with unsupplemented DMEM before addition of the adenovirus mixture. MEFs were incubated with adenovirus overnight and then washed twice in DMEM plus 10% FBS and subcultured an additional 48 h before harvesting for preparation of protein and total RNA.

Analytical techniques

Blood glucose levels were measured with a Bayer Glucometer Elite. Insulin levels on serum obtained from tail bleeds were determined by ultrasensitive ELISA (Crystal Chem, Downers Grove, IL). Intraperitoneal glucose tolerance tests were performed on 14- to 16-h fasted mice at a glucose dose of 1 g/kg body weight. Insulin tolerance tests were performed on 5- to 6-h fasted mice. Animals were injected with 1 U/kg body weight with Humulin-R (Eli Lilly, Indianapolis, IN). For the high-fat diet study, male mice were placed on a diet containing 58% of energy in kcal as fat from coconut oil and high sucrose (catalog number D12331; Research Diets, Inc., New Brunswick, NJ) at 5 wk of age. The hyperinsulinemic euglycemic clamp studies, DEXA, and energy expenditure analysis were performed by the Institute for Diabetes, Obesity, and Metabolism Mouse Phenotyping and Physiology Core facility.

Histological analyses

For all histological studies, pancreata were dissected and fixed in 4% paraformaldehyde for 24 h and then laid flat for paraffin embedding and sectioning by the University of Pennsylvania Morphology Core. For immunofluorescence microscopy, after deparaffinization, all samples were exposed to boiling 10 mm citric acid buffer (pH 6.0) for 6 min. Slides were then blocked in protein blocking solution reagent 1481 for 10 min at room temperature. Primary antibodies were diluted in PBS containing 0.1% BSA and 0.2% Triton X-100 and incubated overnight at 4 C. The appropriate secondary antisera were also diluted in PBS containing 0.1% BSA and 0.2% Triton X-100 and added to slides for 2 h at room temperature. The following antisera were used: insulin (Linco, St. Charles, MO), glucagon (Biodesign, Saco, ME), HA (Santa Cruz Biotechnology), β-catenin (BD Biosciences). The secondary antibodies Cy2-conjugated donkey anti-guinea pig IgG, Cy3-conjugated donkey antirabbit IgG, and Cy3-conjugated donkey antimouse IgG were from Jackson ImmunoResearch, West Grove, PA.

Statistical analysis

Data are expressed as means ± sd or ± sem as indicated in the figure legends. Statistically significant differences between groups were analyzed using ANOVA (JMP Start Statistics). P < 0.05 was considered statistically significant.