Targeted Elimination of Peroxisome Proliferator-Activated Receptor {gamma} in {beta} Cells Leads to Abnormalities in Islet Mass without Compromising Glucose Homeostasis

The nuclear hormone receptor peroxisome proliferator-activatedreceptor ${gamma}$ (PPAR ${gamma}$ ) is an important regulator of lipid and glucosehomeostasis and cellular differentiation. Studies of many celltypes in vitro and in vivo have demonstrated that activationof PPAR ${gamma}$ can reduce cellular proliferation. We show here thatactivation of PPAR ${gamma}$ is sufficient to reduce the proliferationof cultured insulinoma cell lines. We created a model with micein which the expression of the PPARG gene in ß cellswas eliminated (ß ${gamma}$ KO mice), and these mice were foundto have significant islet hyperplasia on a chow diet. Interestingly,the normal expansion of ß-cell mass that occurs incontrol mice in response to high-fat feeding is markedly bluntedin these animals. Despite this alteration in ß-cellmass, no effect on glucose homeostasis in ß ${gamma}$ KO micewas noted. Additionally, while thiazolidinediones enhanced insulinsecretion from cultured wild-type islets, administration ofrosiglitazone to insulin-resistant control and ß ${gamma}$ KOmice revealed that PPAR ${gamma}$ in ß cells is not requiredfor the antidiabetic actions of these compounds. These datademonstrate a critical physiological role for PPAR ${gamma}$ functionin ß-cell proliferation and also indicate that themechanisms controlling ß-cell hyperplasia in obesityare different from those that regulate baseline cell mass inthe islet.

INTRODUCTION

Top

Introduction

Peroxisome proliferator-activated receptor ${gamma}$ (PPAR ${gamma}$ [NR1C3], encodedby PPARG) is a member of the nuclear hormone receptor superfamilyof ligand-gated transcription factors. PPAR ${gamma}$ has been shown toplay a role in diverse biological processes, including adipogenesisand glucose and lipid homeostasis (33). Although the endogenousligand of PPAR ${gamma}$ is not known, there are several synthetic compoundsthat bind PPAR ${gamma}$ with high affinity and activate the receptor.These include the thiazolidinedione (TZD) class of drugs, whichare insulin sensitizers currently in use for the treatment oftype 2 diabetes.

In addition to insulin sensitization, other actions of PPAR ${gamma}$ have recently come to light, including the control of cellularproliferation. Activation of PPAR ${gamma}$ has been shown to reduce growthin a variety of cell lines cultured from cancers of adiposetissue (40), colon (35), breast (29), prostate (30), liver (34),lung (41), and pancreatic acinar tissue (13). These findingshave been extended to some human tumors, such as liposarcoma(11) and prostate cancer (30), and clinical trials with otherforms of cancer are in progress. Interestingly, loss-of-functionmutations in PPARG have been demonstrated for some human coloncancers (36) and loss of heterozygosity at 3p25, a broad regionthat includes PPARG, has been seen in many human prostate (30)and pancreatic endocrine (9, 10, 31, 38, 46) tumors.

The latter observation led us to wonder whether the activationof PPAR ${gamma}$ in pancreatic ß cells would also regulatecellular proliferation. PPAR ${gamma}$ is expressed in the ßcells of both rodents and humans (12, 47), and treatment ofdiabetic or prediabetic humans and rodents with PPAR ${gamma}$ agonistsleads to improvements in islet architecture, insulin content,and glucose-stimulated insulin secretion (GSIS) (6, 8, 17).Although these effects could be explained as the beneficialsequelae of reducing peripheral insulin resistance, troglitazoneadded directly to the culture medium of islets isolated fromZucker fatty rats also improves GSIS (37). Additionally, activationof PPAR ${gamma}$ directly induces the expression of the glucose transporterGlut2 and glucokinase, critical participants in GSIS, in culturedprimary rat islets and insulinoma cell lines (20, 21).

These studies suggested to us that PPAR ${gamma}$ may exert a significanteffect on ß-cell function and that at least part ofthe therapeutic effect of TZD administration may be mediatedthrough the pancreatic islet. We therefore sought to determinewhether PPAR ${gamma}$ could regulate the growth and function of ßcells by using both cultured insulinoma cell lines and targetedgene elimination in mice.

MATERIALS AND METHODS

Top

Materials and Methods

Animal protocols. Study populations of mice were generated by crossing animalswith exon 2 of PPARG flanked by loxP sites (PPAR ${gamma}$ ^fl/fl) to miceexpressing Cre recombinase driven by the rat insulin promoter(PPAR ${gamma}$ ^cre). F₁ PPAR ${gamma}$ ^fl/+ mice were then crossed with F₁ PPAR ${gamma}$ ^fl/+cre mice, thus generating the four study groups (PPAR ${gamma}$ ^fl/fl,PPAR ${gamma}$ ^+/+, PPAR ${gamma}$ ^{+/+ cre}, and PPAR ${gamma}$ ^{fl/fl cre} [heretofore and hereincalled ß ${gamma}$ KO]). Because mice were maintained on a mixedFVB-129-C57 background, littermates served as controls for allexperiments. For studies involving a high-fat diet, mice werefed Harlan Teklad TD93075 special diet (55% fat by caloric content).Mice were genotyped by PCR using the following primers (seeFig. 2): F, CTCCAATGTTCTCAAACTTAC; R1, GATGAGTCATGTAAGTTGACC;and R2, GTATTCTATGGCTTCCAGTGC. All animal work was performedwith the approval of the Animal Use Committees of Harvard andBrandeis Universities.

View larger version (43K):
[in this window]
[in a new window]
FIG. 2. Deletion of PPARG exon 2 in ß ${gamma}$ KO mice. (A) Scheme showing exon 2 of PPARG targeted with loxP sites (triangles) before and after Cre-mediated recombination. Heavy lines indicate the expected PCR products from primers F, R1, and R2. The 400-bp band is a specific recombination product seen only when exon 2 is appropriately excised, while the 275-bp band represents the nonrecombined floxed allele. (B) Results of PCR from whole islets isolated from PPAR ${gamma}$ ^fl/fl and ß ${gamma}$ KO mice. (C) Dispersed islet cells were loaded with a calcium-sensitive fluorophore and treated with glucose, followed by flow sorting. RT-PCR for insulin and glucagon shows that cells with high fluorescence are primarily ß cells but that those with low fluorescence are primarily non-ß cells. RT-PCR of RNA from these cells yields a band of 387 bp for the wild-type transcript and 217 bp for the recombined transcript missing exon 2.

Blood glucose, plasma insulin, glucose tolerance tests, and in vivo insulin secretion. Blood glucose values were determined from whole venous bloodby using an automated glucose monitor (Glucometer Elite; Bayer).Insulin levels in serum were measured by enzyme-linked immunosorbentassay using mouse insulin as a standard (Crystal Chem, Chicago,Ill.). Glucose tolerance tests and acute insulin secretion testswere performed on animals that had been fasted overnight for16 h. The peritoneal cavities of the animals were injected with2 g of glucose/kg of body weight. Glucose levels were measuredfrom blood collected from the animals' tails immediately beforeand 15, 30, 60, and 120 min after the injections (24).

Islet morphology and immunohistochemistry. Tissues were fixed in Bouin's solution and 10% buffered formalinand embedded in paraffin. Sections of pancreas were stainedfor non-ß-cell hormones with a cocktail of antibodiesto glucagon, somatostatin, and pancreatic polypeptide (24, 27).Immunofluorescent staining for insulin was achieved with anti-insulinantibodies and detected by fluorescein antibodies (Jackson Immunoresearch).ß-Cell growth was evaluated by bromodeoxyuridine (BrdU)labeling. Mice were injected with BrdU (Sigma) $~$ 6 h before sacrifice,followed by dissection of the pancreas for staining and sectioningas described above. BrdU-positive cells were identified withan anti-BrdU antibody (cell proliferation kit; Amersham, Piscataway,N.J.).

ß-cell mass determination. ß-Cell mass was evaluated by point counting morphometryon immunoperoxidase-stained sections of pancreas (27, 28). Multiplesections (separated by 80 µm each) were obtained fromeach pancreas and analyzed systematically by using a grid systemcovering at least 175 fields per mouse. Separate images wereacquired with a BX60 microscope (Olympus, New Hyde Park, N.Y.)as described previously (28). Relative volumes were calculatedfor ß cells, non-ß cells, and exocrine tissue.The extent of contaminating tissue (pancreatic ducts, lymphnodes, adipose, and intestine) was recorded to correct for thepancreatic weight. ß-Cell mass was calculated by thefollowing formula: islet ß-cell mass = relative ß-cellvolume x corrected pancreatic weight.

Fluorescence sorting of islet cells. We used the observation that glucose stimulation promotes greaterintracellular calcium concentration in ß cells thanin non-ß cells to sort islet cells (4, 18). Briefly,islets were dispersed after isolation to obtain single cellsas described previously (19). Dispersed cells were suspendedin Krebs-Ringer buffer (KRB) containing 0.1 mM glucose, 2 µMFluo-4, and 0.02% Pluronic F-127 (Molecular Probes, Eugene,Oreg.) and incubated for 30 min at 37°C. Cells were thenwashed with 0.1 mM glucose KRB for 30 min, followed by stimulationwith 20 mM glucose for 10 min. Samples were washed and keptat 37°C until analysis. Flow cytometry was used for sorting(excitation, 488 nm; emission, 502 to 542 nm), and islet cellswere separated based on low (non-ß cells) versus high(ß cells) fluorescence. We extracted RNA from cellpellets and analyzed gene expression by reverse transcription-PCR(RT-PCR). We used nested PCR for amplifying PPAR ${gamma}$ cDNA. Primerdetails are available on request.

Insulin release and content of islets. Islets were isolated by using the intraductal collagenase methodas described previously (27). The insulin released in vitrofrom isolated islets was measured by static incubation of isletscultured overnight (27). Insulin content in whole pancreas inacid-ethanol extracts was measured by enzyme-linked immunosorbentassay (Crystal Chem).

Cell culture and proliferation experiments. INS-1 and HIT-T15 cells were grown in RPMI 1640 with 10% fetalbovine serum, 0.5 M HEPES, 102 mM L-glutamine, 50 mM sodiumpyruvate, and 2.5 mM ß-mercaptoethanol. RINmF andßTC3 cells were grown in Dulbecco's modified Eagle'smedium with 10% fetal calf serum. Cells were plated on day 0and counted with a hemocytometer on the days indicated in Fig1. Troglitazone (10 µM) or vehicle (0.1% dimethyl sulfoxide)was added as appropriate. Retroviral supernatants expressingPPAR ${gamma}$ 1 or vector (pMSCVpuro) only were generated, and INS-1 cellswere infected and selected in puromycin as described previously(32).

View larger version (27K):
[in this window]
[in a new window]
FIG. 1. Effect of TZDs on cultured insulinoma cell growth. (A) INS-1 cells were treated with various doses of troglitazone for 15 days before being counted. Results are given as mean numbers of cells (10⁴) ± standard errors of the means from at least three independent experiments. *, P < 0.05; **, P < 0.005; ***, P < 0.001. (B) INS-1 cells were infected with a retrovirus expressing PPAR ${gamma}$ 1 or vector only and then treated with vehicle (solid lines) or 10 µM troglitazone (Tro) (dashed lines) before being counted as described for panel A. (C to E) ßTC3, RINmF, and HIT-T15 cells were cultured in the presence of 10 µM troglitazone or vehicle for the indicated numbers of days before being counted as described for panel A.

RESULTS

Top

Results

PPAR ${gamma}$ regulates cellular proliferation in cultured insulinoma cells. We first assessed the effect of PPAR ${gamma}$ activation on the proliferationof insulinoma cells in vitro. Rat INS-1 cells were treated withvarious doses of troglitazone or vehicle for 15 days, and cellnumbers were counted; a dose-dependent reduction in cellularproliferation was noted (Fig. 1A). Inhibition of cellular proliferationcould be seen with as little as 0.03 µM troglitazone,significantly below the K_d of the receptor for this ligand,making a nonspecific effect of troglitazone very unlikely. Tomodulate PPAR ${gamma}$ activity in an independent way, we also expressedthe receptor by retrovirally mediated gene transfer into INS-1cells. This step had an effect identical to that of ligand administration,and the combination of PPAR ${gamma}$ and ligand caused an additive repressionof cell growth (Fig. 1B). This phenomenon is not restrictedto INS-1 cells, as other ß-cell-derived lines (ßTC3,HIT-T15, and RINmF) responded similarly to the addition of 10µM troglitazone (Fig. 1C to E).

PPAR ${gamma}$ regulates cellular proliferation in islets in vivo. The data above show that PPAR ${gamma}$ exerts strong antiproliferativeeffects on ß cells when it is expressed via viralvectors or activated by ligand. If this effect of PPAR ${gamma}$ is biologicallyimportant, then the absence of the receptor in ß cellsmay lead to islet hyperplasia. Unfortunately, this hypothesiscould not be tested with a global PPAR ${gamma}$ knockout model, as thesemice die early in embryogenesis (5, 22). To test our hypothesis,we crossed mice bearing a version of PPARG with loxP sites flankingexon 2 (PPAR ${gamma}$ ^fl/fl) (Fig. 2A) with mice expressing Cre recombinasefrom a ß-cell-specific promoter (rat insulin promoter[RIP]-Cre). Both of these lines have been described previously(2, 15). F₁ PPAR ${gamma}$ ^fl/+ mice with and without Cre were mated toone another to generate PPAR ${gamma}$ ^fl/fl, PPAR ${gamma}$ ^+/+, PPAR ${gamma}$ ^cre, and ß ${gamma}$ KOanimals. ß ${gamma}$ KO mice were born in the expected Mendelianratio and were not found to have a survival advantage or disadvantageduring more than 2 years of study. Figure 2B shows the resultsof whole-islet PCR demonstrating that the recombined allele(represented by the 400-bp band) appears only in ß ${gamma}$ KOmice, as expected. The 275-bp band represents the nonrecombinedfloxed allele, present as expected in the PPAR ${gamma}$ ^fl/fl mice butvirtually absent from ß ${gamma}$ KO mice. As the number of isletsincreases, a small amount of nonrecombined product in the ß ${gamma}$ KOmice becomes apparent, most likely reflecting the presence ofnon-ß cells, which do not express the RIP-Cre transgene.In order to ascertain whether PPAR ${gamma}$ mRNA reflected the recombinationevent in ß cells, we took advantage of the observationthat intracellular calcium levels are higher in ßcells than in non-ß cells after a glucose challenge(4, 18). This difference allows one to separate ßcells from non-ß cells using flow sorting on the basisof fluorescence in the presence of a calcium-activated fluorophoreand glucose. Figure 2C shows that dispersed islet cells characterizedby high and low fluorescence were characterized by the presenceof insulin and glucagon mRNA, respectively. Using RT-PCR primersfor PPAR ${gamma}$ complementary to exons 1 and 3, we show that thereis virtually no recombination of PPARG in non-ß cellsof ß ${gamma}$ KO mice. Conversely, ß cells from ß ${gamma}$ KOmice yield no detectable nonrecombined PPAR ${gamma}$ transcripts.

Islets from ß ${gamma}$ KO mice were approximately twice as largeas those from PPAR ${gamma}$ ^fl/fl mice on a chow diet (Fig. 3A and B).Morphometric analysis and immunostaining for insulin revealedthat the differences in islet sizes were largely accounted forby ß-cell hyperplasia. Consistent with these results,BrdU incorporation was proportionately greater in ß ${gamma}$ KOislets than in PPAR ${gamma}$ ^fl/fl islets; this finding is also consistentwith the notion that an increase in cell number rather thancell size contributes to the discrepancy in islet size (Fig.3C). Insulin content was significantly greater in ß ${gamma}$ KOislets (3.45 ± 0.28 ng of insulin/ng of DNA) than inPPAR ${gamma}$ ^fl/fl islets (1.75 ± 0.13 ng of insulin/ng of DNA)(P = 0.001).

View larger version (43K):
[in this window]
[in a new window]
FIG. 3. Islet morphology and histomorphometry in control and ß ${gamma}$ KO mice. (A) Hematoxylin and eosin staining of pancreata from PPAR ${gamma}$ ^fl/fl and ß ${gamma}$ KO mice maintained on either a chow or high-fat diet. (B) Quantification of ß-cell mass from PPAR ${gamma}$ ^fl/fl and ß ${gamma}$ KO mice maintained on either a chow or high-fat diet. Four mice were tested for each determination. *, P < 0.05 for PPAR ${gamma}$ ^fl/fl versus ß ${gamma}$ KO mice on the chow diet; **, P < 0.001 for PPAR ${gamma}$ ^fl/fl mice on the chow versus the high-fat diet; $§$ , P < 0.001 for ß ${gamma}$ KO mice on the chow versus the high-fat diet; ${dagger}$ , P < 0.001 for PPAR ${gamma}$ ^fl/fl versus ß ${gamma}$ KO mice on the high-fat diet. (C) BrdU labeling of islets from chow-fed control and ß ${gamma}$ KO mice (*, P < 0.01; n = 4).

Surprisingly, when the mice were placed on a high-fat diet,quite a different effect was seen. Islets from PPAR ${gamma}$ ^fl/fl miceunderwent the expected hyperplastic response that typifies theobese, insulin-resistant state, increasing in mass 8.3-fold.Islets from obese ß ${gamma}$ KO mice, on the other hand, showedonly a 2.1-fold increase in ß-cell mass from baseline.Thus, PPAR ${gamma}$ is required for the normal expansion of islet massseen in obesity, demonstrating an unexpected and exciting newrole for this receptor in ß-cell biology.

Metabolic function in ß ${gamma}$ KO mice. The islet hyperplasia in chow-fed ß ${gamma}$ KO mice led usto examine whether there were corresponding metabolic alterationsin these animals. Male ß ${gamma}$ KO mice, both fasting andfed, had normal glucose and insulin levels in their sera comparedto levels in the three control genotypes, PPAR ${gamma}$ ^fl/fl, PPAR ${gamma}$ ^+/+,and PPAR ${gamma}$ ^cre (Table 1), as well as normal body weight (Fig. 4A).Switching mice to a high-fat diet for up to 55 weeks resultedin impaired glucose tolerance in all genotypes, with no differencesin glucose or insulin levels noted between ß ${gamma}$ KO miceand controls. We then performed glucose and insulin tolerancetesting to attempt to unmask a diabetic diathesis. In both chow-fedanimals and mice fed a high-fat diet for 55 weeks, no significantdifferences were seen in glucose excursions after intraperitonealglucose (Fig. 4B and C) or insulin administration or in glucose-stimulatedacute-phase insulin secretion (data not shown). For all of theseparameters, trends among female mice did not differ from thoseseen among males (data not shown).

View this table:
[in this window]
[in a new window]
TABLE 1. Glucose and insulin levels in the sera of ß ${gamma}$ KO and control mice^a

View larger version (17K):
[in this window]
[in a new window]
FIG. 4. Metabolic characteristics of male PPAR ${gamma}$ ^fl/fl and ß ${gamma}$ KO mice. (A) Body weights of PPAR ${gamma}$ ^fl/fl (triangles) and ß ${gamma}$ KO (circles) mice on the chow (open symbols) and high-fat (filled symbols) diets. (B) Glucose tolerance tests of male mice on chow (16 weeks of age). Phenotypes (numbers of mice) are as follows: PPAR ${gamma}$ ^+/+ (n = 14), PPAR ${gamma}$ ^cre (n = 13), PPAR ${gamma}$ ^fl/fl (n = 11), and ß ${gamma}$ KO (n = 16). (C) Glucose tolerance tests of male mice on the high-fat diet (44 weeks of age). Phenotypes (numbers of mice) are as follows: PPAR ${gamma}$ ^+/+ (n = 6), PPAR ${gamma}$ ^cre (n = 10), PPAR ${gamma}$ ^fl/fl (n = 15), and ß ${gamma}$ KO (n = 10).

We next assessed the effect of a PPAR ${gamma}$ agonist on insulin secretionfrom isolated islets. Troglitazone (10 µM) significantlyimproved insulin secretion from PPAR ${gamma}$ ^fl/fl islets in the presenceand absence of glucose, consistent with previous reports (45).In contrast, troglitazone had no effect on ß ${gamma}$ KO islets(Fig. 5A). This result demonstrates that PPAR ${gamma}$ is the sole targetof troglitazone in ß cells (at least with respectto effects on insulin secretion), and the knockout is likelyto be virtually complete in those cells that can synthesizeand secrete insulin. More importantly, these results suggestedthat TZDs might exert their antidiabetic actions, in part, throughPPAR ${gamma}$ in ß cells. In order to definitively addressthat possibility, we took male PPAR ${gamma}$ ^fl/fl and ß ${gamma}$ KO miceand placed them on a high-fat diet for 10 weeks. The mice werethen treated with either rosiglitazone (3 mg/kg of body weight/day)or vehicle for 3 weeks. For fasting mice, insulin levels insera obtained after the drug treatment period were higher thanthose found immediately preceding drug administration in vehicle-treatedmice, demonstrating that insulin resistance was still developing(Fig. 5B). Rosiglitazone, however, eliminated this rise in seruminsulin levels equally in PPAR ${gamma}$ ^fl/fl and ß ${gamma}$ KO mice,thus demonstrating that TZD action within ß cellsdoes not significantly contribute to the antidiabetic effectsof these drugs.

View larger version (26K):
[in this window]
[in a new window]
FIG. 5. Effect of TZD drugs on insulin secretion and glucose homeostasis in vivo and in vitro. (A) Isolated islets from PPAR ${gamma}$ ^fl/fl (Flox) and ß ${gamma}$ KO (KO) mice were cultured overnight in 10 µM troglitazone (a TZD drug) or vehicle prior to exposure to glucose at the indicated concentrations. Insulin release is expressed as a percentage of the total content. *, P < 0.05; number of mice for all conditions, 4. (B) Male PPAR ${gamma}$ ^fl/fl and ß ${gamma}$ KO mice were fed a high-fat diet for 10 weeks, followed by 3 weeks of rosiglitazone (3 mg/kg/day) or vehicle. Serum insulin levels of fasting mice were measured before and after drug administration, and the net changes ( ${Delta}$ ) over the treatment period are depicted. Results are means ± standard errors of the means. For each group, the number of mice was 8. *, P < 0.05; **, P < 0.01; RSG, rosiglitazone.

DISCUSSION

Top

Discussion

Despite intense investigation, we have a very incomplete understandingof the specific roles played by PPAR ${gamma}$ in many of the tissueswhere it is expressed. The identification of the pancreaticß cell as a site of PPAR ${gamma}$ expression, combined withdata demonstrating the effects of TZD administration on isletarchitecture and function, suggested to us that PPAR ${gamma}$ could beimportant in the biology of the ß-cell. We have criticallyanalyzed this issue by using Cre-lox technology to target PPAR ${gamma}$ in a ß-cell-specific manner. Counter to expectations,we could detect no significant metabolic abnormality in ß ${gamma}$ KOmice relative to the metabolism of PPAR ${gamma}$ ^fl/fl, PPAR ${gamma}$ ^+/+, or PPAR ${gamma}$ ^crecontrols. Our analysis included measurements of insulin andglucose in fasting and fed mice on both the lean (chow) andinsulin-resistant (high-fat) diets. Glucose and insulin tolerancetesting and levels of acute-phase insulin secretion similarlyrevealed no differences between ß ${gamma}$ KO and control mice.Although it is not possible to confirm that we have deletedPPAR ${gamma}$ expression in every ß cell, several lines ofevidence indicate that we have eliminated its expression inthe vast majority of cells. First, RT-PCR of ß cellsfrom the dispersed islets of ß ${gamma}$ KO mice reveals theabsence of wild-type PPAR ${gamma}$ mRNA. Second, PCR of whole islets(which contain numerous non-ß cells, such as ${alpha}$ and ${delta}$ cells, endothelial cells, etc.) reveals a significant reductionin the expression of the floxed allele, while the 400-bp PCRproduct indicating the presence of the deleted allele is presentonly in ß ${gamma}$ KO islets. Third, the effect of troglitazoneon insulin secretion is totally eliminated in the ß ${gamma}$ KOislets. Fourth, in other studies, the expression of the RIP-Cretransgene has been demonstrated to be virtually completely penetrant(24, 25).

The absence of metabolic consequences in ß ${gamma}$ KO micedoes not exclude the ß cell as an antidiabetic targetof TZD drugs. To assess this possibility, we measured insulinsecretion after TZD administration from isolated whole isletsof control or ß ${gamma}$ KO mice. In control islets, troglitazonecaused a roughly twofold induction of insulin secretion, evenin the absence of glucose. This effect was totally eliminatedin ß ${gamma}$ KO islets. When these studies were extended tothe intact mouse, however, we found that the improved glucosehomeostasis associated with TZD administration was preservedin ß ${gamma}$ KO animals. This finding demonstrates that PPAR ${gamma}$ in ß cells is unlikely to play a major role in thetherapeutic response to TZDs. Even if there is a slight improvementin insulin secretion after TZD administration in vivo, thiseffect is likely masked by the peripheral insulin sensitizationthat dominates the clinical response to TZDs.

The most striking difference between control mice and ß ${gamma}$ KOmice was islet mass. Loss of PPAR ${gamma}$ is associated with a hyperplasticresponse in the targeted ß cells. PPAR ${gamma}$ is known toregulate growth in a variety of cell types, including preadipocytes,myeloid leukemia cells, and epithelial tumor lines derived frombreast, colon, and prostate. The fact that ß ${gamma}$ KO isletsdo not proliferate indefinitely suggests that other growth-regulatingpathways eventually come into play, which may explain why tumorformation in the islet tissue of ß ${gamma}$ KO mice was notobserved.

ß-Cell hyperplasia is most commonly seen as a responseto obesity and insulin resistance; it may be caused by a circulatinggrowth factor or factors (14). Despite several theories, nospecific factor has been definitively identified. Recent evidencehas suggested that insulin itself might be important as a mediatorof ß-cell hyperplasia (1, 26). All the known componentsof the insulin signaling pathway are present in ßcells (16) (reviewed in reference 23). Loss of insulin receptorsubstrate 1 (IRS-1) results in hyperplastic, but dysfunctional,islets (3, 27, 39), while loss of IRS-2 results in diabeteswithout compensatory ß-cell hyperplasia (44). Theoverexpression of a constitutively active form of Akt increasesß-cell mass in transgenic mice (7, 42). Furthermore,loss of ß-cell insulin receptors (ßIRKO)is associated with reduced ß-cell hyperplasia duringnormal aging (24). Crossing ßIRKO mice to animalswith hepatic insulin resistance (due to the inactivation ofinsulin receptors in their livers) results in progeny that areinsulin resistant but which do not appropriately expand theirß-cell mass (R. N. Kulkarni, unpublished data). Giventhe obvious defect in ß-cell hyperplasia after high-fatfeeding in ß ${gamma}$ KO mice, it is tempting to speculate thatthe lack of PPAR ${gamma}$ may cause local insulin resistance within ßcells.

Other mechanisms may also contribute to our findings, of course.For example, excess lipid accumulation in the islet has beenassociated with the lipoapoptosis of ß cells (43).TZD treatment of islets from Zucker fatty rats causes reductionsin cellular triglyceride levels as well as reduced lipoapoptosis(17, 37). It is possible, therefore, that for ß ${gamma}$ KOanimals on the high-fat diet, the smaller islet mass than thatof controls results from enhanced apoptosis in the setting ofreduced PPAR ${gamma}$ activity.

One of the more intriguing aspects of our study is the discordancebetween islet mass and function, especially for mice in theobese, insulin-resistant state. Specifically, the fact thatglucose and insulin levels are identical in ß ${gamma}$ KO andPPAR ${gamma}$ ^fl/fl mice on a high-fat diet implies that there must bea compensatory gain in ß-cell function in the ß ${gamma}$ KOmice. This observation is supported by the elevated insulincontent on a per cell basis that was seen in ß ${gamma}$ KO islets(at least in the chow-fed mice). PPAR ${gamma}$ can thus be identifiedas a factor that couples peripheral insulin resistance to changesin ß-cell proliferation, perhaps by promoting lipiddeposition in the islet. In this scenario, removal of PPAR ${gamma}$ makesthe ß cell work more efficiently. However, obese ß ${gamma}$ KOmice were still glucose intolerant, indicating that the absenceof PPAR ${gamma}$ cannot compensate fully for the ß-cell dysfunctionseen in states of peripheral insulin resistance.

ACKNOWLEDGMENTS

We thank K. C. Hayes and the staff of the animal facility ofthe Foster Biomedical Laboratories at Brandeis University.

This work was supported by NIH grants KO8 DK02535 (to E.D.R.),RO3 DK58850 (to E.D.R.), R37DK31405 (to B.M.S.), RO1 DK33201(to C.R.K.), and KO8 DK02885 (to R.N.K.).

FOOTNOTES

* Corresponding author. Mailing address for Evan D. Rosen: Division of Endocrinology, Beth Israel Deaconess Medical Center, 99 Brookline Ave., Boston, MA 02215. Phone: (617) 667-3221. Fax: (617) 667-2927. E-mail: erosen{at}bidmc.harvard.edu. Mailing address for Bruce M. Spiegelman: Department of Cancer Biology, Dana-Farber Cancer Institute, 1 Jimmy Fund Way, Boston, MA 02115. Phone: (617) 632-3748. Fax: (617) 632-5363. E-mail: bruce_spiegelman{at}dfci.harvard.edu.

REFERENCES

Top