skp2 is required for incretin hormone-mediated β-cell ......exendin-4 is known to increases camp...

Skp2 Is Required for Incretin Hormone-Mediated�-Cell Proliferation

Shuen-Ing Tschen, Senta Georgia, Sangeeta Dhawan, and Anil Bhushan

Departments of Medicine (S.-I.T., S.G., S.D., A.B.) and Molecular, Cell and Developmental Biology (A.B.),University of California, Los Angeles; Los Angeles California 90095-7073

The glucoincretin hormone glucagon-like peptide-1 (GLP-1) and its analog exendin-4 (Ex-4)promote �-cell growth and expansion. Here we report an essential role for Skp2, a substraterecognition component of SCF (Skp, Cullin, F-box) ubiquitin ligase, in promoting glucoincre-tin-induced �-cell proliferation by regulating the cellular abundance of p27. In vitro, GLP-1treatment increases Skp2 levels, which accelerates p27 degradation, whereas in vivo, loss ofSkp2 prevents glucoincretin-induced �-cell proliferation. Using inhibitors of phosphatidylino-sitol 3-kinase and Irs2 silencing RNA, we also show that the effects of GLP-1 in facilitatingSkp2-dependent p27 degradation are mediated via the Irs2-phosphatidylinositol-3 kinasepathway. Finally, we show that down-regulation of p27 occurs in islets from aged mice andhumans, although in these islets, age-dependent accumulation of p16Ink4a prevent glucoin-cretin-induced �-cell proliferation; however, ductal cell proliferation is maintained. Takentogether, these data highlight a critical role for Skp2 in glucoincretin-induced �-cellproliferation. (Molecular Endocrinology 25: 2134 –2143, 2011)

Pancreatic �-cell mass is dynamic and responds to varia-tions in insulin demand due to physiological and patho-

logical challenges such as aging, pregnancy, and obesity.Failure of the endocrine pancreas to adapt the �-cell mass tochanging insulin demand can lead to hyperglycemia anddiabetes. Hence, understanding the mechanisms that regu-late �-cell plasticity is important in developing effectivetherapeutic approaches for the treatment of diabetes. Theglucoincretin hormone glucagon-like peptide-1 (GLP-1)and its long-acting peptide analog exendin-4, both wellknown therapeutic candidates, have pleiotropic effectsthat include potentiation of glucose-dependent insulin re-lease as well as �-cell proliferation and survival (1–3).Activation of GLP-1 receptor by glucoincretins results inthe induction of cAMP second messenger pathway. In-crease in cAMP stimulates the expression of various genesthat play a role in glucose sensing [GLUT2 (glucose trans-porter 2) and glucokinase] and insulin secretion (Pdx1and insulin itself). GLP-1-mediated increase in cAMP alsopromotes �-cell survival by inducing Irs2 expression (4)and enhanced activation of AKT (5). However, despite

the significant progress in our understanding of the sig-naling cascade initiated by GLP-1, the mechanisms bywhich glucoincretins induce �-cell proliferation to resultin expansion of �-cell mass are not clear.

A number of studies have indicated that expansion of�-cell mass in adults is due to the replication of existing�-cells (6, 7). The cellular abundance of cell cycle inhibi-tor, p27, is a critical determinant of the transition fromquiescence to a proliferative state (8). The level of p27 inthe �-cells is controlled by a ubiquitin ligase complex thatregulates p27 degradation (9–11). Skp2, an F box pro-tein, functions as a receptor component of an SCF ubiq-uitin ligase complex, resulting in p27 ubiquitination anddegradation (12, 13). In Skp2�/� mice, accumulation ofp27 prevents �-cell proliferation, resulting in decreased�-cell mass, hypoinsulinemia, and a corresponding de-crease in the ability to dispose glucose from blood (14).The insulin receptor substrate 2 (Irs2) branch of insulinsignaling has been linked to cell cycle regulation, becauseIrs2-deficient mice show progressive accumulation of p27in the nuclei of �-cells (15). Deletion of p27 allele rescues

ISSN Print 0888-8809 ISSN Online 1944-9917Printed in U.S.A.Copyright © 2011 by The Endocrine Societydoi: 10.1210/me.2011-1119 Received June 1, 2011. Accepted September 12, 2011.First Published Online October 6, 2011

Abbreviations: cdk, Cyclin-dependent kinase; CKI, cyclin kinase inhibitors; DAPI, 4�,6-diamidino-2-phenylindole; GLP-1, glucagon-like peptide-1; PI3-kinase, phosphatidylino-sitol 3-kinase; siRNA, silencing RNA; TUNEL, terminal deoxynucleotide transferase-medi-ated dUTP nick end labeling.

O R I G I N A L R E S E A R C H

2134 mend.endojournals.org Mol Endocrinol, December 2011, 25(12):2134–2143

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 30 April 2014. at 18:18 For personal use only. No other uses without permission. . All rights reserved.

the diabetic phenotype and �-cell mass of Irs2-deficientmice, indicating that p27 functions downstream of Irs2-mediated signaling to regulate �-cell proliferation (11).Glucoincretins also are known to induce Irs2 expressionvia the cAMP pathway, correlating with the induction ofproliferation (5).

In the light of these studies, we hypothesized that GLP-1signaling via Irs2 could regulate the stability of p27 andthereby control glucoincretin-induced �-cell proliferationand expansion of �-cell mass. In this study, we have ana-lyzed whether glucoincretins regulate the degradation ofp27 via Skp2 to mediate the proliferative effects of glucoin-cretins on pancreatic �-cells. The results presented hereshow that glucoincretins induce the degradation of p27 me-diated by Skp2 via the Irs2 phosphatidylinositol 3-kinase(PI3-kinase) pathway. Using Skp2�/� animals and silencingRNA targeting Irs2 in Min6 cells, we show that Skp2 isdownstream of the Irs2-PI3-kinase pathway and that Skp2mediated degradation of p27 is required for the proliferativeeffects of GLP-1/exendin-4 on �-cell mass expansion. Fur-thermore, we show that glucoincretins can also mediate thedegradation of endogenous p27, by using prediabetic db/dbmice (where p27 levels are normally very high), leading to�-cell proliferation in these animals. Finally, we present dataon how the levels of another cell cycle inhibitor, p16Ink4a,may override glucoincretin-dependent p27 degradation in�-cells in old animals. We also show that degradation of p27is also the mechanism leading to ductal and exocrine prolif-eration in response of exendin-4 treatment. These resultssuggest that the proliferative effect of glucoincretins inyoung animals on pancreatic �-cells is mediated by Skp2-dependent degradation of p27. In summary, our study de-scribes the molecular connection between the Irs2-PI3-ki-nase signaling cascade and Skp2-mediated p27 degradationas a means of induction of cellular proliferation in responseto glucoincretins.

Results

GLP-1 regulates Skp2-mediated p27 degradationthrough the Irs2-PI3-kinase pathway

Our previous work has shown a critical role for Skp2-mediated p27 degradation in regulating �-cell proliferation(14). Glucoincretins such as GLP-1 induce proliferation of�-cells (16), and long-term treatment of isolated mouse isletswith long-acting GLP-1 analog exendin-4 has been shownto cause elevated Skp2 levels (17). These data suggest thatglucoincretins induce �-cell proliferation by modulatingSkp2 levels. To address whether a short-term GLP-1 treat-ment also induces Skp2-mediated p27 degradation in cul-tured islets, isolated human islets were incubated with

GLP-1 for 24 and 48 h, and the protein levels of Skp2 andp27 were measured by immunoblotting. GLP-1 treatmentincreased Skp2 levels 2.6-fold after 24 h of treatment, and to4-fold after 48 h (Fig. 1, A and B). Levels of p27 protein weredecreased after 24 h of GLP-1 treatment and reduced 3.5-fold after 48 h (Fig. 1, A and C). Similar effects of short-termGLP-1 treatment were observed in cultured mouse islets(data not shown). Glucoincretins utilize the PI3-kinase cas-cade to mediate diverse effects, including �-cell proliferation(1, 4, 5, 18). To assess whether the effect of GLP-1 on Skp2is also mediated through the PI3-kinase pathway, we in-cluded the PI3-kinase inhibitor along with GLP-1 treatment(16). The effects of GLP-1 on Skp2 and p27 levels wereblocked by PI3-kinase inhibitor (LY294002, 50 �mol/liter)(Fig. 1, A–C), indicating the involvement of the PI3-kinase-signaling pathway in GLP-1-induced Skp2-mediated degra-dation of p27. We further examined whether reduction ofp27 levels in human islets after GLP-1 treatment leads to�-cell proliferation. We did not observe any Ki67 and insulindouble-positive cells in human islets after 48 h of GLP-1treatment (Fig. 1, D and E). Therefore, we reasoned thatanother cell cycle inhibitor, p16Ink4a, may constrain isletproliferation. Previous studies have shown that p16Ink4a pre-vented adult mouse and human islets proliferation (20–22).To test this idea, we measured levels of p16 in control andGLP-1-treated human islets by immunoblotting. Indeed,high levels of p16 Ink4a were detected in these islets, and thelevels of p16 Ink4a did not change after 48 h of GLP-1 treat-ment (Fig. 1F).

Exendin-4 is known to increases cAMP levels, whichpromotes Irs2 expression and stimulates Akt phosphory-lation via PI3-kinase, resulting in �-cell proliferation (4,5). We, therefore, reasoned that glucoincretins promoteSkp2-mediated degradation of p27 and �-cell prolifera-tion by inducing Irs2 expression. We used loss-of-func-tion of Irs2 by transfecting specific silencing RNA(siRNA) in Min6 cells to determine whether the effects ofexendin-4 on proliferation, mediated by Skp2 and p27,are directly regulated by Irs2-signaling pathway in �-cells.Min6 cells were transfected with Irs2 siRNA or scram-bled siRNA 24, and cell samples were collected 24 h or48 h after transfection and measured by real-time PCR formRNA transcription and Western blotting for proteinlevels. Both mRNA transcription and protein levels weredecreased (Fig. 2A). We then measured protein levels ofp27 and Skp2 in Irs2 siRNA or scrambled siRNA-trans-fected Min6 cells, treated with or without exendin-4.Min6 cells were transfected with Irs2 siRNA or scram-bled siRNA 24 h before exendin-4 (2.5 nM) treatment.Exendin-4 treatment was associated with 6-fold increasein Skp2 levels and 3.5-fold decrease in p27 levels in con-trol Min6 cells transfected with scrambled siRNA (*P �

Mol Endocrinol, December 2011, 25(12):2134–2143 mend.endojournals.org 2135


0.05). By contrast, transfection of Irs2 siRNA in Min6cells reduced the Skp2 levels in sham treatment group andblocked the effect of exendin-4 to increase Skp2 and re-duce p27 levels (Fig. 2B). Taken together, these data showthat glucoincretins accelerated p27 degradation via up-regulation of Skp2 levels, which is downstream of PI3-kinase cascade mediated via Irs2.

Skp2-mediated p27 degradation is required forglucoincretin-mediated �-cell mass expansionin mice

Several reports have suggested that continuous short-term glucoincretin treatments in mice leads to �-cell mass

expansion (3, 16). Because Skp2 levels quicklyincrease in isolated islets in response to GLP-1treatment, we further examined whether Skp2mediates glucoincretins-induced �-cell mass ex-pansion in vivo by regulating cellular abundanceof p27. The long-acting form of GLP-1, exen-din-4 (10 nmol/kg body weight) or PBS shamcontrol were administered ip to young (7–8 wkof age) Skp2�/� mice and control littermatesdaily for 7 d. Skp2�/� mice were used as controllittermates because these heterozygous mice donot exhibit any phenotypes (14). No significantweight differences were observed between PBS-or exendin-4-treated groups (Fig. 3A). Pancreatafrom both groups were collected, and pancreaticsections were stained with antiinsulin antibodyto measure �-cell mass in each group. Exendin-4-treated control littermates displayed a 2.5-foldincrease in �-cell mass when compared withPBS-treated controls. In contrast, exendin-4treatment of Skp2�/� mice did not show any dif-ferences in �-cell mass compared with PBS treat-ment in these mice (Fig. 3B). To assess whetherexendin-4-mediated expansion of �-cell mass isdependent on �-cell proliferation, the number ofproliferating �-cells was further determined byKi67 staining. Exendin-4-treated control Skp2�/�

mice showed a 7.2-fold increase in Ki67-positive�-cells compared with PBS-treated controls (Fig.3C). In contrast, Skp2�/� mice showed little orno differences between exendin-4 and PBS treat-ments in the number of Ki67-positive �-cells.These observations show a requirement for Skp2in glucoincretin-induced �-cell mass expansion.

To assess the levels of p27 protein in exendin-4-dependent �-cell mass expansion, pancreaticsections from exendin-4 or PBS-treated 7-wk-old wild-type mice were analyzed for the expres-sion of p27 and insulin by immunohistochemis-try. A number of �-cells in the PBS-treated

control mice showed normal accumulation of p27 in thenucleus (Fig. 3D, left). In contrast, �-cells in the exendin-4-treated mice showed loss of p27 in the nucleus (Fig. 3D,right), which correlates with the increment in �-cell prolif-eration (Fig. 3C). Levels of another cyclin kinase inhibitor,p21, were also analyzed by immunohistochemistry, and nodifferences were observed between exendin-4 or PBS-treatedmice (data not shown), thus indicating specificity for p27.To rule out the possibility of antiapoptotic effects of exen-din-4 in �-cell mass expansion, terminal deoxynucleotidetransferase-mediated dUTP nick end labeling (TUNEL) as-say was performed on pancreatic sections from exendin-4-

FIG. 1. Skp2-mediated p27 degradation is activated in response to GLP-1 treatmentand is regulated posttranscriptionally via a PI3-kinase-dependent pathway. Culturedhuman islets were treated with GLP-1 (10 nmol/liter) with or without Ly294002 (50�mol/liter) for indicated time periods. Protein levels of Skp2 and p27 were measuredby Western blotting (A). Ly294002 was added 30 min before GLP-1 stimulation.Data show representative from three individual experiments. Bar graphs and theerror bars represent the mean � SE of Skp2 (B) and p27 (C) protein levels normalizedwith �-tubulin from two identical experiments. Control (D) and GLP-1-treated (panelE, 10 nmol/liter) human islets were immunostained with insulin (green) and ki67(red). No insulin-ki67 double positive cells were observed after 48 h of GLP-1treatment. Protein levels of p16 from control and GLP-1-treated human islets (48 h)were measured by Western blotting (F). Data show representatives from threeindividual experiments. Bar graphs and the error bars represent the mean � SE ofp16 protein levels normalized with �-tubulin from three experiments. *, P � 0.12.�-Tub, �-Tubulin; Ins, insulin.

2136 Tschen et al. GLP-1 Induced �-Cell Proliferation via Skp2 Mol Endocrinol, December 2011, 25(12):2134–2143


or PBS-treated Skp2�/� mice and control littermates, and nodifferences in apoptosis were observed (data not shown).These data, therefore, demonstrate that Skp2 is required forexendin-4-dependent expansion of �-cell mass via �-cellproliferation.

Exendin-4 promotes �-cell replication in db/dbmice by promoting degradation of p27

The nuclei of �-cells of Irs2�/� mice and mice lacking thelong form of the leptin receptor (Lepr�/� or db/db) show

progressive accumulation of p27 and decreased�-cell proliferation, resulting in the inability ofthe pancreas to respond to insulin resistance(11). It is likely that endogenous accumulation ofp27 in these animals occurs due to Skp2 defi-ciency. To confirm whether a deficit in Skp2 as-sociates with the accumulation of p27 in db/dbmice, we measured Skp2 and p27 levels by West-ern blotting using isolated islets from youngdb/db and wild-type mice. These experimentsshowed that indeed, the levels of Skp2 in theislets isolated from db/db mice were negligible,accounting for the high levels of p27 in thismodel. This was in contrast to islets from wild-type mice, which showed normal levels of Skp2protein and correspondingly low levels of p27(Fig. 4A). Elevated p27 levels in islets of db/dbmice were further confirmed by immunohisto-chemistry (Fig. 4B). These data suggest thataccumulation of p27 in db/db mice resultsfrom low levels of Skp2 leading to reduce�-cell proliferation.

We next investigated whether treatment ofdb/db mice with exendin-4 could induce Skp2and promote degradation of endogenous p27.We used 4-wk-old db/db mice that displayednormal blood glucose levels to avoid any com-plications from hyperglycemia. These 4-wk-old prediabetic mice were treated with exen-din-4 ip for 7 d, and pancreatic sections fromcontrol and treated mice were analyzed for theexpression of p27 and insulin. As expected,immunostaining showed that p27 protein ac-cumulated in the nuclei of �-cells in controldb/db mice (Fig. 5A), as previously reported(11). Interestingly, most of the �-cells in theexendin-4-treated db/db mice showed a signif-icant reduction of p27 levels (Fig. 5B), suggest-ing that exendin-4 can reduce endogenous p27in �-cells. No significant differences were ob-served in TUNEL staining between two groups,suggesting that apoptosis rates were not af-fected by short-term exendin-4 treatment (Fig.

5, C and D). To further examine whether the exendin-4treatment decreased accumulation of p27 protein by in-creasing Skp2 levels, islets isolated from control and ex-endin-4-treated db/db mice were subjected to immuno-blot analysis. Islets from exendin-4-treated db/db miceshowed dramatic increase in Skp2 levels and decreasedp27 levels compared with islets from PBS-treated litter-mates (Fig. 5E). To further examine whether Skp2-induced p27 degradation resulted in increased �-cell pro-

FIG. 2. Irs2-signaling pathway mediates the effects of glucoincretins-induced Skp2-mediated p27 degradation. Min6 cells were transfected with Irs2 siRNA (500 nM) for24 h or 48 h. RNA and protein were isolated. Quantitative RT-PCR analysis showinglevels of Irs2 transcript in Min6 cells treated with Irs2 siRNA (A, left). Westernblotting was also performed to evaluate IRS2 protein levels (A, middle). The errorbars represent SEM (n � 3) for each experiment; P � 0.05 (panel A). Min6 cells weretransfected with scramble siRNA or Irs2 siRNA 24 h before exendin-4 (2.5 nM)treatment. Cell lysates were collected 24 h after exendin-4 treatment, and levels ofSkp2 and p27 were measured by Western blotting. Data show representatives fromfour individual experiments. Bar graphs and the error bars represent the mean � SE

of Skp2 and p27 protein levels normalized with �-tubulin from two identicalexperiments (panel B). ctrl, Control; Ex-4, exendin-4; Ins, insulin; �-Tub, �-tubulin.



liferation in db/db mice, �-cell proliferation wasevaluated in both groups by Ki67 staining. Exendin-4 treat-ment in prediabetic young db/db mice resulted in signifi-cantly more Ki67-positive �-cells compared with controllittermate mice treated with PBS (Fig. 5F, 2.6-fold in-crease, P � 0.01), suggesting that prediabetic db/db micehave the capacity to respond to exendin-4 to increaseSkp2 levels, reduce p27 levels, and increase their �-cellproliferation.

Exendin-4 fails to induce �-cell proliferation inaged mice despite p27 degradation

Previously published work from our group and othershas shown that young mice are able to respond to glucoin-

cretin treatment by �-cell proliferation,which leads to an increase in �-cell mass.However, aged animals do not expand�-cell mass in response to glucoincretintreatment, primarily due to accumula-tion of p16Ink4a, an inhibitor of cyclinD/cyclin-dependent kinase (Cdk)4 (3,23). We first checked whether glucoin-cretins mediate any reduction of p27 lev-els in the islets, in aged mice. Wild-typemice (8 months of age) were treated withPBS or exendin-4 ip, and pancreas sam-ples were collected. Exendin-4 treatmentresulted in the reduction of p27 proteinlevels in the islets (Fig. 6, A and B); how-ever, p16Ink4a levels in the islets were un-changed (Fig. 6, C and D). These datasuggest that despite reductions in p27levels, high p16Ink4a levels prevent cellcycle reentry of islet cells. Thus, glucoin-cretin treatment is unable to expand isletmass in aged animals.

To test whether this phenomenon wasrestricted to islets in the pancreas, we ex-amined whether exendin-4 treatment inaged mice resulted in changes in p27 andp16 levels in other compartments of thepancreas. Several studies have shownthat GLP-1 receptors are expressedthroughout the pancreas (24–27), sug-gesting that GLP-1 treatment could af-fect the exocrine and ductal tissue. Ex-pression analysis of p16Ink4a revealedgenerally low levels of expression in exo-crine and ductal tissue and in contrast toislet cells, no increase in p16Ink4a levelswere observed in these tissues with aging(data not shown). Pancreatic sectionsfrom control and exendin-4-treated 8-

month-old animals revealed that exendin-4 treatment re-sulted in reduction of p27 levels in the exocrine and ductaltissue, in a manner similar to islets (Fig. 6, A, B, E, and F).Furthermore, the low levels of p16Ink4a in the ductal andexocrine tissue remained unchanged in response to exen-din-4 (Fig. 6, G and H). Costaining of Ki67 and Mucin-1(a ductal marker) revealed a 3-fold increase in ductal pro-liferation in exendin-4-treated animals (Fig. 6, I–K) thatcorrelated with the decreased levels of p27. These datasuggest that glucoincretin treatment in aged mice can notinduce proliferation of �-cells due to high p16Ink4a levelsin islets, although other cell types in the pancreas respondto glucoincretin to increase proliferation.

FIG. 3. Exendin-4-induced �-cell mass expansion is blocked in Skp2�/� mice. Young (7–8wk of age) wild-type control (Skp2�/�) and Skp2�/� mice were ip injected with PBS orexendin-4 (10 nmol/kg body weight) daily for 7 d. Skp2�/� mice were used as wt controlbecause no phenotype was observed. Body weights were measured after 7 d of exendin-4injection. No significant weight differences were observed between PBS or exendin-4-treatedgroups. Black bar represents wt control; gray bar, Skp2 �/� (A). Mice were killed and �-cellmass was analyzed. A 2.5-fold increase in �-cell mass was observed in exendin-4-treated wtcontrol littermates when compared with PBS-treated group. No differences in �-cell masswere found in Skp2�/� mice treated with exendin-4 or PBS. Data are shown as mean � SE

from five to seven sections per mouse, three to four mice per group. *, P � 0.05. Black bar,PBS treated; gray bar (B). Exendin-4 treated �-cell replication was evaluated byimmunohistochemistry with antiinsulin and anti-Ki67 antibodies. Numbers of Ki67-positive�-cells were quantified in pancreata from Skp2�/� and Skp2�/� mice after daily injectionof PBS or exendin-4 for 7 d. Wild-type control mice show a 7.2-fold increase in �-cellproliferation in response to exendin-4 treatment whereas no differences were observed inSkp2�/� mice. Black bar, PBS; gray bar, exendin-4. Data are shown as mean � SE from fivesections per mouse, three mice per group; *, P � 0.05 (C). Exendin-4 treatment in vivo leadsto p27 degradation. Pancreatic sections from wt control mice (�7–8 wk of age) treated withPBS (left) or exendin-4 (right) were immunostained with antimouse insulin (green) and anti-mouse p27 (red) antibodies. Data show representatives from three individual experiments (D).Ex4, Exendin-4.



Discussion

As GLP-1 analogs have been approved by the US Foodand Drug Administration (FDA) and are currently beingprescribed to patients with type 2 diabetes to improveglycemic control, it is important to understand their mo-lecular mechanism of action. In this study, we analyze themitogenic pathways activated by glucoincretins and theireffect on �-cell proliferation. Several previous studieshave shown that GLP-1 functions via the Irs2-PI3-kinasepathway to regulate the growth and survival of pancreatic�-cells (4, 5, 16). What was not clear was how incretinhormone signaling to the PI3-kinase pathway impingeson the cell cycle machinery to regulate �-cell prolifera-tion. In this study, we show that glucoincretins actthrough Irs2-PI3-kinase pathway to promote degradationof cell cycle inhibitor, p27, via the SCF ubiquitin ligasecomplex containing Skp2. Interestingly, the Irs2-PI3-ki-nase pathway is also used by insulin and IGF family tomediate cell growth. This suggests a possibility that var-ious mitogenic signals utilize a common mechanism ofIrs-PI3-kinase signaling, to cause Skp2-mediated p27degradation.

We also demonstrate that reduction of endogenouslevels of p27, mediated by an up-regulation of Skp2 inresponse to glucoincretin signaling, is a requirement for�-cell mass expansion, as the Skp2�/� mice are unable toexpand their �-cell mass in response to glucoincretins.Skp2 also acts as a binding partner of Cyclin A in cancercells and stabilizes it from inhibition by p27 via compet-itive binding (28). Although it is possible that Cyclin A2could also be affected by the absence of Skp2, several linesof evidence suggest that Cyclin A2 and Skp2 are likely indistinct pathways. First, double knockouts of Skp2 andp27 completely rescued the phenotype of skp2-null miceand �-cell replication as we demonstrated in our previousstudy (14). This indicates that in �-cells, p27 was theprimary target of Skp2 for protein degradation to regulate�-cell replication. Second, unlike p27, which acts at theG1/S transition (29), cyclin A2 is most abundant in S/G2

phase and acts primarily at the G2/M transition, and thusthese two factors are separated in cell cycle phases (30).Third, direct protein interactions of cyclin A2 with Skp2complex were only observed in cancer cells and may notnecessarily occur during a normal cell cycle. Finally, theobservation of Hussain and co-workers (17) that GLP1functions through cyclinA2 by transcriptional regulationis on a different time scale than the rapid protein degra-dation during progression of the cell cycle. Thus, al-though Cyclin A2 can also play a role in GLP1-inducedproliferation in the pancreas (17), it is likely to be a dis-tinct pathway that is separated in time from the Skp2-mediated p27 degradation. We also show that the db/dbmice have very low endogenous levels of Skp2, resultingin accumulation of p27 and low proliferation in the db/dbislets, leading to a failure in adaptive expansion. Surpris-ingly, these mice are able to up-regulate Skp2 levels inresponse to a glucoincretin treatment, leading to reducedp27 levels and induction of �-cell proliferation. This is aninteresting observation in terms of glucoincretins actingdifferently in diabetic animal models.

Our work also highlights that although different pan-creatic compartments respond similarly to glucoincretintreatment by reducing the p27 levels, the endocrine andductal compartments respond to the reduction of p27differently. �-Cell replication is dependent on the balancebetween the cyclin D2-Cdk4 complex that forms in re-sponse to mitotic signals, and cyclin kinase inhibitors(CKI) that block the activity of the cyclin E-Cdk2 com-plex. Two groups of CKI have been described (31). Theseinclude the Ink4 family, members of which specificallyinhibit cyclin D-Cdk4/6 activity, and the CIP/KIP (CDKinteracting protein/kinase inhibitory protein) family, thatincludes p21, p27, and p57, which exhibit promiscuousCDK-inhibitory activity. A member of the Ink4a family,

FIG. 4. Accumulation of p27 levels in �-cells of db/db mice. Isletsfrom young (5 wk of age) prediabetic db/db and wild-type mice wereisolated, and protein levels of Skp2 and p27 were measured byWestern blotting. Data show a representative from three individualexperiments. Bar graphs and the error bars represent the quantificationand average deviation from three identical experiments. Black bar, db/db mice; white bar, wild-type mice. **, P � 0.01 (panel A). Pancreaticsections from db/db (panel B) and wild type (panel C) control micewere immunostained with antimouse insulin (green) and antimousep27 (red) antibodies. �-Tub, �-tubulin; WT, wild type.



p16Ink4a, accumulates in the �-cells with age, indicatingthat high levels of p16Ink4a in older �-cells shift the bal-ance, and mitotic stimuli that lead to changes in levels ofother CKI would have minimal effect (22). This is indeedthe case for pancreatic �-cells, as �-cells from old miceexhibit high levels of p16Ink4a, and are resistant to reenterthe cell cycle (3, 23). We show that GLP-1 induced deg-radation of p27 is insufficient to promote �-cell replica-tion in older mice that display high levels of p16Ink4a intheir islets. Data from our work and others (20) suggestthat the high endogenous p16 Ink4a levels in adult humanislets may also account for the observations that human

islets do not proliferate in response toglucoincretins, even though they areable to induce Skp2 and reduce p27 inresponse to such treatments.

GLP-1 receptors are expressed inmany cell types within the pancreas,including ductal, acini, and � cells (24,25). Unlike pancreatic � cells, ductalcells do not display the age-dependentincrease in p16Ink4a and thus are capa-ble of responding to GLP-1 even in agedmice to increase proliferation. Recentstudies in rats treated with exendin-4 orsitagliptin, which stabilizes GLP-1, havedemonstrated increased ductal prolifera-tion, ductal metaplasia, and chronic pan-creatitis (25, 32). Moreover, increasedductal proliferation is also strongly asso-ciated with degradation of p27 levels(33). Incretins have long been consideredto enhance glucose-mediated insulin se-cretion; however, they also have the po-tential to activate cell cycle in nonendo-crine pancreatic cells types. It remains tobe tested whether PI3-kinase inhibitorcould be used to block the effects ofglucoincretins on cell cycle prolifera-tion while maintaining enhancement ofglucose-mediated insulin secretion.

Materials and Methods

Mice. Skp2�/� mice between 7 and 8 wk ofage were used in this study and were main-tained on a C57BL/6J/CD1 mixed back-ground and genotyped by PCR as previousdescribed, and Skp2�/� littermates wereused as control because no phenotypeswere observed (14). Young (4 wk of age)Lepr�/� (db/db) mice were purchased fromThe Jackson Laboratory (Bar Harbor,

ME). All mice were maintained ad libitum on a standard dietand kept under a 12-h light, 12-h dark cycle. All animal proto-cols were approved by Chancellor’s Animal Research Commit-tee at UCLA.

Cell culture and human islets culture. Isolated human islets wereobtained from ABCC Islet Cell Resource (three donors, between26 and 67 yr of age; purity, �80–90% upon arrival). Humanislets were hand picked and cultured in CMRL medium with10% fetal calf serum and 100 U/ml Pen-Strep (Invitrogen, Carls-bad, CA) at 37 C for 1–2 d before treatment with GLP-1 (10nmol/liter; Sigma-Aldrich, St. Louis, MO) for 24 or 48 h asindicated. Min6 cells were maintained in DMEM (Invitrogen)

FIG. 5. Endogenous p27 degradation is required for glucoincretins action on �-cellproliferation. Pancreatic sections from 4-wk-old db/db mice daily treated with PBS (A) orexendin-4 (B) for 7 d were immunostained with antimouse p27 (red) and antimouse insulin(green) antibodies. �-Cell apoptosis was examined by TUNEL staining (red) costained withantimouse insulin (green) antibodies in 4-wk-old db/db mice treated with PBS (C) or exendin-4(D). Islets were isolated from 4-wk-old db/db mice treated with PBS or exendin-4 for 7 d,and protein levels of Skp2 and p27 were measured by Western blotting. Data showrepresentatives from three individual experiments. Bar graphs and the error bars represent thequantification and average deviation from three identical experiments (E). �-Cell replicationwas evaluated by immunohistochemistry with antiinsulin and anti-Ki67 antibodies. Number ofKi67-positive �-cells was quantified in pancreata from db/db mice (4 wk of age) after dailyinjection of PBS or exendin-4 for 7 d. A 2.5-fold increment was observed in exendin-4-treatedwild-type animals. Data are shown as mean � SE from four sections per mouse, four mice pergroup; **, P � 0.01 (F). Nuclear staining with DAPI is shown in blue. Ex-4. Exendin-4.



with 10% fetal calf serum, 50 �M �-mercaptoethanol, and 100U/ml Pen-Strep at 37 C under 5% CO2.

Drug treatment. Skp2�/� and control mice (7–8 wk of age) oryoung (4 wk of age) Lepr�/� (db/db) mice were injected ip withexendin-4 (10 nmol/kg body weight; Sigma-Aldrich) or vehiclecontrol (PBS) daily for 7 d. Mice were killed and pancreata werecollected for histological analysis.

Immunohistochemistry. Pancreatic tissue was processed as pre-viously described (14). In brief, the pancreas was dissected andfixed in 4% formaldehyde before being embedded in paraffin.Sections (5 �m) were deparaffinized and rehydrated, followedby antigen retrieval by using Antigen Unmasking Buffer (VectorLaboratories, Inc., Burlingame, CA), permeabilized in 0.4%Triton X-100/TBS. Slides were blocked with 3% IgG-free BSA(Jackson ImmunoResearch Laboratories, West Grove, PA), fol-lowed by antimouse insulin (DAKO Corp., Carpinteria, CA),Ki67 (BD Pharmingen, Franklin Lakes, NJ), Mucin-1, p27, p21,p16 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) antibody.Fluorescein isothiocyanate- or Cy3-conjugated antimouse anti-bodies (Jackson ImmunoResearch Laboratories) were used as asecondary antibody. In situ cell death detection assay (TUNEL)was performed followed by manufacturer’s instruction (Roche,Indianapolis, IN). Slides were mounted with Vectashield with4�,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories),and images were obtained using a Leica DM 6000 microscope(Leica Corp., Deerfield, IL) with Openlab software.

�-Cell mass. �-Cell mass was measured as previously described(14). In brief, four to six sections from each pancreas werestained with antimouse insulin antibody and scanned by a LeicaDM6000 microscope. Montage images were made by ImageJsoftware. The cross-sectional areas of pancreas and �-cells weredetermined by ImagePro software. �-Cell mass per pancreas wasestimated as the product of the relative cross-sectional area of�-cells per total tissue and the weight of the pancreas and cal-culated by examining pancreata from at least three animals foreach genotype.

Immunoblotting. Lysates from human islets or Min6 cells wereextracted and quantified by Qubit assay (Invitrogen), and equalamount of proteins were resolved by SDS-PAGE, followed bytransfer to polyvinylidene difluoride membrane for immuno-blotting. The indicated proteins were detected by incubatingspecific antibodies against Skp2, p16, p27 (Santa Cruz), Irs2(Cell Signaling Technology, Inc., Danvers, MA) and �-Tubulin(Sigma-Aldrich). Protein levels were normalized to the proteinlevels of housekeeping gene �-Tubulin. Quantifications weredone by ImageJ and ImagePro software.

siRNA and transfection. Knockdown of Irs2 expression wasachieved by gene-specific siRNA. Irs2 siRNA (Dharmacon,Lafayette, CO) was designed corresponding to the target se-quences 5�-AATAGCTGCAAGAGCGATGAC-3�, and scram-bled siRNA targeting firefly luciferase (5�-CTGACGCG-GAATACTTCGA-3�) was used as a control. Min6 cells andhuman islets were transfected by using Lipofectamine-2000(Invitrogen).

FIG. 6. Exendin-4 fails to induce �-cell proliferation in aged micedespite p27 degradation. Pancreatic sections from 7-month-oldwild-type mice daily treated with PBS (control) or exendin-4 for 7 dwere immunostained with various antibodies. Immunostaining ofp27 (A and B) and p16 (C and D) within islets in PBS (A and C) orexendin-4 (B and D)-treated animals were examined. Levels of p27in islets are reduced in exendin-4-treated wild-type mice (B)compared with PBS-treated wild-type mice (A). No differences wereobserved in p16 levels (C and D). Data were representative fromfour sections per mouse and three mice per group. Antimouse p27(E and F) and antimouse p16 (G and H) were also costained withantimouse Mucin-1 in PBS (E and G)- or exendin-4 (F and H)-treatedpancreatic sections. Arrows indicate ductal structures in control (Eand G) and exendin-4-treated (F and H) wild-type mice. Levels ofp27 in ductal cells is reduced in exendin-4-treated wt mice (F)compared with PBS-treated wild-type mice (E). No differences wereobserved in p16 levels (G and H). Data were representative fromfour sections per mouse and three mice per group. Numbers ofKi67-positive ductal cells were quantified by using antimouse Ki67and antimouse Mucin-1. A 3-fold increase in ductal cellproliferation was observed in exendin-4-treated wild-type controllittermates when compared with PBS-treated group. Data areshown as mean � SE from three sections per mouse and three miceper group. *, P � 0.05 (I). Arrows indicate ductal structures incontrol (J) and exendin-4-treated (K) wild-type mice. DAPI stainingmarking nuclei is shown in blue. Muc-1, Mucin-1.



RNA isolation, RT-PCR, and real-time PCR. RNA was isolatedfrom cultured Min6 cells or islets using TRI Reagent (Molecular Re-search Center, Cincinnati, OH). cDNA was synthesized by Super-script IIIReverse transcriptase (Invitrogen)witholigodTpriming.TaqPCR Master Mix Kit (QIAGEN, Chatsworth, CA) was used for PCRwith the following primers: Irs2 forward, AGTAAACGGAGGTG-GCTACA; Irs reverse, 5�-AAGCTGCTGAGAAGTCAGGT; Tubu-lin forward, 5�-GTTGGCCAGGCTGGTGTCCAG-3�; Tubulinreverse, 5�-CTGTGATGAGCTGCTCAGGGTGG-3�. Real-timeRT-PCR were performed using the 7900HT Fast Real-Time PCRSystem(AppliedBiosystems,FosterCity,CA).Theexpressionlevelsofeach transcript were normalized to the housekeeping gene Cyclophi-lin. Data shown are the averages from three experiments, each per-formed in triplicate.

Acknowledgments

We thank Lendy Le for technical help and Dr. Beter C. Butler(both from Larry Hillblom Islet Research Center, UCLA, LosAngeles, CA) for helpful discussions and advice.

Address all correspondence and requests for reprints to: AnilBhushan, Department of Medicine, University of California,Los Angeles, Los Angeles California 90095-7073. E-mail:[email protected].

This work was supported by grants from the National Insti-tutes of Health (R01 DK-068763), Helmsley Trust, and JuvenileDiabetes Research Foundations (to A.B.). S.G. is supported byan National Institute of Diabetes and Digestive and Kidney Dis-eases Career Development Award.

Disclosure Summary: The authors disclosed no conflict ofinterest.

References

1. Brubaker PL, Drucker DJ 2004 Minireview: glucagon-like peptidesregulate cell proliferation and apoptosis in the pancreas, gut, andcentral nervous system. Endocrinology 145:2653–2659

2. Holz GG, Chepurny OG 2005 Diabetes outfoxed by GLP-1? SciSTKE 2005:pe2

3. Tschen SI, Dhawan S, Gurlo T, Bhushan A 2009 Age-dependentdecline in � cell proliferation restricts the capacity of � cell regen-eration in mice. Diabetes 58:1312–1320

4. Jhala US, Canettieri G, Screaton RA, Kulkarni RN, Krajewski S,Reed J, Walker J, Lin X, White M, Montminy M 2003 cAMPpromotes pancreatic �-cell survival via CREB-mediated inductionof IRS2. Genes Dev 17:1575–1580

5. Park S, Dong X, Fisher TL, Dunn S, Omer AK, Weir G, White MF2006 Exendin-4 uses Irs2 signaling to mediate pancreatic � cellgrowth and function. J Biol Chem 281:1159–1168

6. Dor Y, Brown J, Martinez OI, Melton DA 2004 Adult pancreatic�-cells are formed by self-duplication rather than stem-cell differ-entiation. Nature 429:41–46

7. Georgia S, Bhushan A 2004 � cell replication is the primary mech-anism for maintaining postnatal � cell mass. J Clin Invest 114:963–968

8. Georgia S, Bhushan A 2006 p27 Regulates the transition of �-cellsfrom quiescence to proliferation. Diabetes 55:2950–2956

9. Montagnoli A, Fiore F, Eytan E, Carrano AC, Draetta GF, HershkoA, Pagano M 1999 Ubiquitination of p27 is regulated by Cdk-

dependent phosphorylation and trimeric complex formation. GenesDev 13:1181–1189

10. Sheaff RJ, Groudine M, Gordon M, Roberts JM, Clurman BE 1997Cyclin E-CDK2 is a regulator of p27Kip1. Genes Dev 11:1464–1478

11. Uchida T, Nakamura T, Hashimoto N, Matsuda T, Kotani K,Sakaue H, Kido Y, Hayashi Y, Nakayama KI, White MF, Kasuga M2005 Deletion of Cdkn1b ameliorates hyperglycemia by maintain-ing compensatory hyperinsulinemia in diabetic mice. Nat Med 11:175–182

12. Carrano AC, Eytan E, Hershko A, Pagano M 1999 SKP2 is requiredfor ubiquitin-mediated degradation of the CDK inhibitor p27. NatCell Biol 1:193–199

13. Sutterluty H, Chatelain E, Marti A, Wirbelauer C, Senften M, Mul-ler U, Krek W 1999 p45SKP2 promotes p27Kip1 degradation andinduces S phase in quiescent cells. Nat Cell Biol 1:207–214

14. Zhong L, Georgia S, Tschen SI, Nakayama K, Nakayama K,Bhushan A 2007 Essential role of Skp2-mediated p27 degradationin growth and adaptive expansion of pancreatic � cells. J Clin Invest117:2869–2876

15. Kulkarni RN, Holzenberger M, Shih DQ, Ozcan U, Stoffel M,Magnuson MA, Kahn CR 2002 �-cell-specific deletion of the Igf1receptor leads to hyperinsulinemia and glucose intolerance but doesnot alter �-cell mass. Nat Genet 31:111–115

16. Buteau J, Spatz ML, Accili D 2006 Transcription factor FoxO1mediates glucagon-like peptide-1 effects on pancreatic �-cell mass.Diabetes 55:1190–1196

17. Song WJ, Schreiber WE, Zhong E, Liu FF, Kornfeld BD, Wondis-ford FE, Hussain MA 2008 Exendin-4 stimulation of cyclin A2 in�-cell proliferation. Diabetes 57:2371–2381

18. Hui H, Nourparvar A, Zhao X, Perfetti R 2003 Glucagon-likepeptide-1 inhibits apoptosis of insulin-secreting cells via a cyclic5�-adenosine monophosphate-dependent protein kinase A- and aphosphatidylinositol 3-kinase-dependent pathway. Endocrinology144:1444–1455

19. Tuttle RL, Gill NS, Pugh W, Lee JP, Koeberlein B, Furth EE, Po-lonsky KS, Naji A, Birnbaum MJ 2001 Regulation of pancreatic�-cell growth and survival by the serine/threonine protein kinaseAkt1/PKB�. Nat Med 7:1133–1137

20. Chen H, Gu X, Su IH, Bottino R, Contreras JL, Tarakhovsky A,Kim SK 2009 Polycomb protein Ezh2 regulates pancreatic �-cellInk4a/Arf expression and regeneration in diabetes mellitus. GenesDev 23:975–985

21. Dhawan S, Tschen SI, Bhushan A 2009 Bmi-1 regulates the Ink4a/Arf locus to control pancreatic �-cell proliferation. Genes Dev 23:906–911

22. Krishnamurthy J, Ramsey MR, Ligon KL, Torrice C, Koh A,Bonner-Weir S, Sharpless NE 2006 p16INK4a induces anage-dependent decline in islet regenerative potential. Nature 443:453–457

23. Rankin MM, Kushner JA 2009 Adaptive �-cell proliferation is se-verely restricted with advanced age. Diabetes 58:1365–1372

24. Xu G, Kaneto H, Lopez-Avalos MD, Weir GC, Bonner-Weir S2006 GLP-1/exendin-4 facilitates �-cell neogenesis in rat and hu-man pancreatic ducts. Diabetes Res Clin Pract 73:107–110

25. Matveyenko AV, Dry S, Cox HI, Moshtaghian A, Gurlo T, GalassoR, Butler AE, Butler PC 2009 Beneficial endocrine but adverseexocrine effects of sitagliptin in the human islet amyloid polypep-tide transgenic rat model of type 2 diabetes: interactions with met-formin. Diabetes 58:1604–1615

26. Korner M, Stockli M, Waser B, Reubi JC 2007 GLP-1 receptorexpression in human tumors and human normal tissues: potentialfor in vivo targeting. J Nucl Med 48:736–743

27. Singh G, Eng J, Raufman JP 1994 Use of 125I-[Y39]exendin-4 tocharacterize exendin receptors on dispersed pancreatic acini andgastric chief cells from guinea pig. Regul Pept 53:47–59

28. Ji P, Goldin L, Ren H, Sun D, Guardavaccaro D, Pagano M, Zhu L



2006 Skp2 contains a novel cyclin A binding domain that directlyprotects cyclin A from inhibition by p27Kip1. J Biol Chem 281:24058–24069

29. Kossatz U, Dietrich N, Zender L, Buer J, Manns MP, Malek NP2004 Skp2-dependent degradation of p27kip1 is essential for cellcycle progression. Genes Dev 18:2602–2607

30. Frouin I, Montecucco A, Biamonti G, Hubscher U, Spadari S, MagaG 2002 Cell cycle-dependent dynamic association of cyclin/Cdkcomplexes with human DNA replication proteins. EMBO J 21:2485–2495

31. Sherr CJ, Roberts JM 1999 CDK inhibitors: positive and negativeregulators of G1-phase progression. Genes Dev 13:1501–1512

32. Nachnani JS, Bulchandani DG, Nookala A, Herndon B, Molteni A,Pandya P, Taylor R, Quinn T, Weide L, Alba LM 2010 Biochemicaland histological effects of exendin-4 (exenatide) on the rat pan-creas. Diabetologia 53:153–159

33. Karamitopoulou E, Zlobec I, Tornillo L, Carafa V, Schaffner T,Brunner T, Borner M, Diamantis I, Zimmermann A, Terracciano L2010 Differential cell cycle and proliferation marker expression inductal pancreatic adenocarcinoma and pancreatic intraepithelialneoplasia (PanIN). Pathology 42:229–234

Members can search for endocrinology conferences, meetings and webinars on the

Worldwide Events Calendar.www.endo-society.org/calendar



skp2 is required for incretin hormone-mediated β-cell ......exendin-4 is known to increases camp...

Documents