Reversal of preexisting hyperglycemia in diabetic mice by acute deletion of the Men1

Yuqing Yanga, Buddha Gurunga, Ting Wub, Haoren Wanga, Doris A. Stoffersc,d, and Xianxin Huaa,d,1

aAbramson Family Cancer Research Institute, Department of Cancer Biology, Abramson Cancer Center, cDepartment of Medicine, and dInstitute for Diabetes, Obesity, and Metabolism, University of Pennsylvania, Philadelphia, PA 19104; and bDepartment of Basic Medical Sciences, Medical College, Xiamen University, Xiamen 361005, China

Edited by Arnold J. Levine, Institute for Advanced Study, Princeton, NJ, and approved September 28, 2010 (received for review August 18, 2010) A hallmark of diabetes is an absolute or relative reduction in the Men1 excision differently from normal β cells. Thus, it is im- number of functional β cells. Therapies that could increase the num- portant to determine whether Men1 excision actually ameliorates ber of endogenous β cells under diabetic conditions would be desir- or reverses hyperglycemia in diabetic mice. able. Prevalent gene targeting mouse models for assessing β-cell How menin regulates β-cell proliferation is not well understood. proliferation and diabetes pathogenesis only address whether de- Although menin has been shown to be crucial for expression of letion of a gene prevents the development of diabetes. Models cyclin-dependent kinase inhibitor p18ink4c (p18 hereafter) and p27 testing whether acute excision of a single gene can ameliorate or in islets (10, 11) and liver cells (12), Men1 excision does not affect reverse preexisting hyperglycemia in established diabetes remain to liver cell proliferation (7). These findings raise the possibility that be explored, which could directly validate the effect of gene exci- menin may also regulate β cell proliferation through effectors sion on treating diabetes. Here, we report that acute and temporally other than p18 and p27 (12). Menin up-regulates gene transcrip- controlled excision of the Men1 gene, which encodes menin, ame- tion through histone H3 modifications, such as H3K4 methylation liorated preexisting hyperglycemia in streptozotocin-treated mice. (13, 14). However, it is unclear whether menin represses tran- Moreover, Men1 excision also improved the preexisting hypergly- scription of endogenous , especially proproliferative cell cy- cemia and glucose intolerance in genetic db/db diabetic mice. Fur- cle genes in β cells. thermore, acute Men1 excision reversed preexisting glucose in- In this study, we found that acute Men1 excision ameliorated

tolerance in high-fat diet-fed mice. Men1 excision improved glucose preexisting hyperglycemia in streptozotocin (STZ)-treated mice. CELL BIOLOGY metabolism at least partly through increasing proliferation of en- Moreover, acute Men1 excision also corrected preexisting glu- dogenous β cells and islet size. Acute Men1 excision up-regulated cose intolerance or hyperglycemia in genetic db/db or high-fat a group of proproliferative genes in pancreatic islets. Together, diet-fed diabetic mice. Acute Men1 ablation promoted β-cell these findings demonstrate that established hyperglycemia can be proliferation and increased β-cell number partly by coordinately reversed through repression of a single gene, Men1, in diabetic up-regulating multiple proproliferative cell cycle genes. Our conditions, and suggest that menin is a vital regulator in pathogen- findings suggest that menin actively regulates the process of di- esis of diabetes. abetes and could be manipulated to treat diabetes.

cell proliferation | db/db | high-fat diet | type 2 diabetes Results Insulin Secretion by Islets and Peripheral Insulin Sensitivity Are Not oth type 1 and type 2 diabetes ultimately result from an in- Affected by Men1 Ablation. Conventional mouse knockout models Bsufficient number of functional β cells in islets (1). Therefore, have been used to determine whether gene ablation can prevent approaches that promote β-cell regeneration or proliferation and development of diabetes (2). To further determine the effect of gene increase the number of endogenous β cells under diabetic con- ablation on reversing established abnormal glucose homeostasis, ditions would be desirable. Thus far, many factors including we used a conditional and inducible Men1 knockout model and de- multiple cell cycle regulators have been tested in mouse models, termined whether acute Men1 excision ameliorated preexisting hy- Men1l/l and their roles in β-cell proliferation and diabetes development perglycemia in diabetic mice. ;Cre-ER mice were generated by crossing Men1l/l mice to mice expressing the Ubc9 promoter-driven have been determined (2). For instance, cyclin D1, cyclin D2, Cre-ERT2 transgene (6, 15). Men1l/l;Cre-ER and control Men1l/l mice and cyclin-dependent kinase 4 (Cdk4) are crucial for β cell were fed tamoxifen (TAM), and Men1 excision in pancreatic islets proliferation and preventing the development of hyperglycemia cip1/kip1 was determined by quantitative real-time PCR (qRT-PCR) and (2). Deletion of p27 (p27 hereafter), a cyclin-dependent immunostaining, 30 d after TAM treatment. Menin expression was kinase inhibitor, prevents development of diabetes in db/db mice, l/l β markedly reduced in Men1 ;Cre-ER islets as compared with control partly by increasing the number of cells (3). Deletion of Lkb1, islets (Fig. 1 A–C), indicating effective Men1 excision. a tumor suppressor involved in AMP kinase activation, promotes To determine whether Men1 excision affects insulin secretion by β -cell proliferation and ameliorates glucose intolerance (4). How- pancreatic β cells in response to glucose, islet perifusion studies ever, no report has shown that acute deletion of a single gene reverses were performed by using islets isolated from mice, 30 d after TAM preexisting glucose intolerance or hyperglycemia in mouse models. feeding. Size-matched islets isolated from Men1-excised and con- Such a study would be desirable and is closely related to treating di- trol mice were perifused with glucose or potassium chloride (4). abetes, because it directly evaluates the impact of manipulating a sin- The first-phase, second-phase, or total insulin secretion was similar gle gene on treating preexisting diabetes. between the Men1-excised and control islets (Fig. 1D), suggesting Menin is a nuclear encoded by the Men1 gene that is mutated in patients with familial multiple endocrine neoplasia type 1 (MEN1) syndrome (5). Menin preferentially represses Author contributions: Y.Y., D.A.S., and X.H. designed research; Y.Y., B.G., T.W., and H.W. proliferation of endocrine cells including β cells (6, 7). Although performed research; Y.Y. and X.H. analyzed data; and Y.Y. and X.H. wrote the paper. Men1 excision after a long period promotes β-cell proliferation The authors declare no conflict of interest. and increases blood insulin levels under normal conditions (6, 8, 9), This article is a PNAS Direct Submission. little is known as to whether acute Men1 excision can correct pre- 1To whom correspondence should be addressed. E-mail: [email protected]. existing abnormal glucose homeostasis in diabetic mice. Stressed This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. endogenous β cells under diabetic conditions may respond to 1073/pnas.1012257107/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1012257107 PNAS Early Edition | 1of6 Downloaded by guest on September 23, 2021 Menin/Insulin excision on preexisting hyperglycemia in mice that were treated A *** Men1l/l Men1∆/∆ 1.0 Men1l/l with STZ, a β-cell cytotoxic agent that preferentially damages ∆/∆ l/l )l Men1 β

ANRm .8 B C cells (16). Both Men1 ;Cre-ER and control mice developed evel hyperglycemia 3 wk after STZ treatment (SI Appendix, Fig. 1A).

e .6

vitaler Men1

1neM We then examined whether excision could affect the .4 preexisting hyperglycemia in the diabetic mice (Fig. 2A). No ( .2 improvement in hyperglycemia was observed in the Men1l/l 0 control mice after TAM treatment over 6 wk of observation (Fig. 2B). In contrast, blood glucose levels started to decrease in 55% of D E Men1l/l;Cre-ER mice as early as 3 wk after TAM feeding (Fig. 2C, l/l enil 100 rescued). However, 45% of the Men1 ;Cre-ER mice remained

4 l l/l l/l )n

Men1 do esoculg evel Men1 esa e

i ∆/∆ ∆/∆ Men1 egatnecreP fo sab 90 Men1 hyperglycemic after TAM treatment (Fig. 2C, unrescued). Further m

/ 3

stelsi 001/ stelsi 80 investigation showed that Men1 was effectively excised in islets in e l

e Glucose ramp r

ni 2 (0-25mM) 70 the rescued mice after TAM treatment (Fig. 2D, Center), whereas

lu 60 Men1 excision was inefficient in the unrescued mice (Fig. 2D, Glucose (0mM) KCl (30mM) s 1 n fi g I

o 50 Right). The inef cient Men1 excision might be partly attributable to lb lb n ( β 0 40 STZ-induced alkylation of genomic DNA in cells (16). These 30 60 90 120 150 (min) 03060(min) Duration of perifusion Time after insulin injection results strongly indicate that Men1 excision ameliorates hypergly- cemia in mice with preexisting diabetes. Fig. 1. Men1 excision does not affect insulin secretion by islets and pe- Moreover, BrdU incorporation by β cells and the number of l/l l/l ripheral insulin sensitivity. Men1 or Men1 ;Cre-ER mice at the age of 12 wk insulin-positive cells were higher in the rescued mice than those in were fed tamoxifen (TAM) at 200 mg/kg of body weight per day. (A) Men1 l/l l/l the unrescued and control mice (Fig. 2 E and F and SI Appendix, mRNA levels in islets isolated from Men1 and Men1 ;Cre-ER mice 30 d after fi TAM treatment (n = 10 mice). (B and C) Immunostaining for menin and in- Fig. 1 B and C). Random serum insulin levels were not signi cantly sulin in islets from Men1l/l (B) and Men1l/l;Cre-ER mice (C) 30 d after TAM different among the three groups (Fig. 2G); however, Men1 ab- treatment (n = 6 mice). (Scale bar: 25 μm.) (D) Insulin release from islets lation increased the ratio of serum insulin to blood glucose levels isolated from Men1l/l and Men1l/l;Cre-ER mice 30 d after TAM treatment. in the rescued mice, consistent with improved islet function and/or One hundred size-matched islets of each genotype were hand-picked and total insulin secretion from islets in response to blood glucose cultured for 3 d. Insulin secretion stimulated by glucose (0–25 mM) and stimulation (Fig. 2H). As the Cre-ER transgene did not affect islet potassium chloride (KCl, 30 mM) were measured as described in SI Appendix. size (SI Appendix, Fig. 2 A and B), β cell proliferation (SI Appendix, (E) Insulin tolerance test (ITT; insulin at 0.75 U/kg of body weight, i.p.) was Fig. 2C), and number of glucagon-producing α cells (SI Appendix, performed in Men1l/l and Men1l/l;Cre-ER (n =4–5 mice) 4 mo after TAM treatment. Test results were expressed as the percentage of baseline non- Fig. 2D) in normal and STZ-treated mice (SI Appendix, Fig. 2), we fasting blood glucose levels. ***P < 0.001. conclude that Men1 excision rescues preexisting STZ-induced hyperglycemia at least partly through increasing β cell prolifer- ation and the number of β cells. that Men1 excision does not affect insulin secretion by an indi- vidual β cell in response to glucose. To further determine whether Acute Men1 Ablation Ameliorates Preexisting Hyperglycemia and Men1 excision affects overall insulin sensitivity in peripheral tissues Glucose Intolerance in Genetic db/db Diabetic Mice. To further other than pancreatic β cells, insulin tolerance tests (ITTs) were confirm our observation that acute Men1 excision reverses hy- performed in Men1l/l;Cre-ER and control mice 4 mo after TAM perglycemia in established diabetes, we chose db/db mice, a ge- feeding. There was no difference in insulin sensitivity between the netically defined type 2 diabetes mouse model with mutations in control and Men1-excised mice (Fig. 1E). the leptin receptor gene (17, 18) and evaluated the impact of acute Men1 excision on glucose homeostasis, as described in a schema Acute Men1 Ablation Ameliorates Preexisting STZ-Induced Hypergly- in Fig. 3A. Both db/db;Men1l/l;Cre-ER and control db/db;Men1l/l cemia. Next, we sought to determine the impact of acute Men1 mice developed glucose intolerance at 6 wk of age, and they

A B C l/l l/l Fig. 2. Acute Men1 excision ameliorates preexisting hyper- Men1l/l 600 Men1 600 Men1 ;Cre-ER Men1l/l // // 500 500 glycemia in STZ-induced diabetes. (A) A schematic of experi- Men1∆/∆ l/l l/l Men1l/l;Cre-ER // // 400 400 Unrescued mental design. Control Men1 or Men1 ;Cre-ER mice (n =16 // // 300 300 mice) at the age of 12 wk were i.p. injected with STZ at 40 mg/kg Age 12 16 19 20 21 22 (wk) 200 200 Rescued of body weight per day. Diabetic Men1l/l (n = 8 mice) or Men1l/l; rd th th th 100 100 STZ TAM3 4 5 6 034612 5 (wk) 0461235(wk) Cre-ER (n = 11 mice) mice whose blood glucose levels were Blood glucose (mg/dL) Blood glucose level Blood glucose (mg/dL) Time after TAM Time after TAM >250 mg/dL for two consecutive weeks after STZ injections were fed TAM. Blood glucose levels were monitored before Men1l/l; CreER D Men1l/l Rescued Unrescued and until 6 wk after the last dose of TAM treatment. (B) Av- erage blood glucose levels in the control mice. (C) Average blood glucose levels in the rescued and unrescued Men1 l/l;Cre- Insulin Men1l/l ER mice. Blood glucose levels <250 mg/dL for two consecutive Rescued Men1l/l;Cre-ER Unrescued Men1l/l;Cre-ER weeks after TAM treatment were defined as the rescued Menin/ phenotype. (D) Immunostaining for menin and insulin in islets from the control Men1l/l, rescued Men1l/l;Cre-ER, and unres- l/l μ E * * F + G 0.4 H 20 cued Men1 ;Cre-ER mice. (Scale bar: 25 m.) (E) Quantitation 3 40 * ** of BrdU incorporation by pancreatic β cells from the control, 0.3 15 ** ** 30 rescued, and unrescued Men1l/l; Cre-ER mice. (F) Quantitation 2 0.2 10 20 of the number of insulin-positive cells in islets. (G) Nonfasting 1 5 serum insulin levels 6 wk after TAM treatment. (H) Ratio of 10 0.1 glucose ratio cells per islet Serum insulin to serum insulin (ng/mL) to blood glucose (mg/dL) levels, multi- 0 Number of insulin 0 0 0 Serum insulin (ng/mL) < < BrdU incorporation (%) plied by 10,000. *P 0.05; **P 0.01.

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1012257107 Yang et al. Downloaded by guest on September 23, 2021 A D E Menin/Insulin db/db;Men1l/l db/db;Men1l/l ∆/∆

l/l )Ld/ Men1l/l Men1∆/∆ ) db/db;Men1 ;Cre-ER db/db;Men1 l/l L

Men1 d db/db;Men1l/l /g

g * ** ** B C 600 600 ∆/∆ * m( m( Men1 m

l/l ( db/db;Men1 ;Cre-ER dool esoculg 500 500 e 400 so 400 Age 6 14 (wk) c 300 ul 300 g TAM 8th 200 doo 200 GTT 100 100 lB l/l ∆/∆ Men1 Men1 B 04080120(min) 04080120(min)

IPAD F G J K L / l/l l/l l/l UdrB )%( noi )%( db/db;Men1 db/db;Men1 db/db;Men1 M;bd/bde 1n∆/∆ mu db/db;Men1nilusni ∆/∆ )Lm/gn( db/db;Men1∆/∆ aera aera 5 8 **

6 )ti **

ta 4 gnin

nu nu 6 r nilusnI op 4 3 yratibra H I iats

rocni Ud rocni 4 2

/UdrB 2 ni 2 l 1 ( usn r

rB 0 0 eS 0 I 0 5 10 15 (min)

Fig. 3. Acute Men1 excision ameliorates hyperglycemia in db/db mice. (A) A schematic of experimental design. Control db/db;Men1l/l and db/db;Men1l/l;Cre- ER mice (n =5–7 mice) were fed TAM at the age of 6 wk. Mice were killed 8 wk after TAM feeding. (B and C) Immunostaining for menin and insulin in islets from control mice (B) and db/db;Men1l/l;Cre-ER mice (C) 8 wk after TAM treatment. (Scale bar, 25 μm.) (D and E) Glucose tolerance test (GTT, glucose at 2 g/kg of body weight, i.p.) before (D) and 8 wk after (E) TAM treatment. (F–I) BrdU incorporation by β cells. (F and G) Immunostaining for BrdU and DAPI. (H and I) Immunostaining for BrdU and insulin. (Scale bar: 25 μm.) (J) Quantitation of BrdU incorporation by β cells. (K) Quantitation of insulin-staining area. (L) Glucose- stimulated insulin secretion (GSIS, glucose at 3 g/kg of body weight, i.p.), measured 8 wk after TAM treatment. *P < 0.05; **P < 0.01. CELL BIOLOGY were fed TAM at this age (Fig. 3 A and D). Two months after endogenous β cells in high-fat diet-fed mice because menin is not TAM feeding, Men1 was completely excised in the pancreas in the repressed in the islets of these mice. db/db;Men1l/l;Cre-ER mice, but not the control db/db;Men1l/l mice To detect the effect of acute Men1 excision in islets on high-fat (Fig. 3 B and C). Acute Men1 excision in db/db;Men1l/l;Cre-ER diet-induced glucose intolerance, we chose to generate mice in mice significantly improved glucose intolerance (Fig. 3 D and E), which Men1 could only be excised in pancreatic islets. To this end, and lowered fasting blood glucose levels in mice, 8 wk after TAM we generated Men1l/l;Pdx1-Cre-ER and control Men1l/l mice (23). treatment (0 min; Fig. 3E). Consistently, Men1 excision increased Pdx1 expression is restricted to pancreatic β cells in adult mice, BrdU incorporation by β cells (Fig. 3 F–J) and islet size (Fig. 3K). and Men1 was excised only in β cells, but not in the adjacent We also found that, at age of 14 wk or 8 wk after TAM acinar cells in the Men1l/l;Pdx1-Cre-ER mice after TAM feeding treatment, the fasting insulin level in the control db/db mice was (SI Appendix,Fig.5B and C). Men1l/l;Pdx1-Cre-ER and control 1.4 ng/mL (Fig. 3L, 0 min), consistent with the fact that com- Men1l/l mice were fed with either chow diet or high-fat diet for 16 pensatory high insulin levels (hyperinsulinemia) in db/db mice wk (Fig. 4A). Compared with mice on chow diet, the control and dropped because of deterioration of β cells at ≈14 wk of age (19). Men1l/l;Pdx1-Cre-ER mice on high-fat diet for 16 wk gained sub- Notably, the fasting insulin level in the Men1-excised db/db mice stantially more body weight and developed glucose intolerance was much higher than the control Men1-expressing mice (5.3 ng/ (SI Appendix, Fig. 5A and Fig. 4F). Eight weeks after TAM mL; Fig. 3L, 0 min), indicating that Men1 excision is crucial for feeding, Men1 was efficiently excised in islets of Men1l/l;Pdx1-Cre- maintaining a sufficient number of functional β cells (Fig. 3 J and ER mice but not in the control mice (Fig. 4 B–E). Notably, high- K) and compensatory hyperinsulinemia in db/db mice. Because fat diet-induced glucose intolerance was normalized 8 wk after Men1 excision did not change insulin secretion by size-match Men1 excision (Fig. 4 F and G), demonstrating that acute Men1 islets (Fig. 1D), body weight (SI Appendix, Fig. 3A), or peripheral excision corrected high-fat diet-induced glucose intolerance. insulin sensitivity (SI Appendix, Fig. 3B), our results strongly Men1 excision increased BrdU incorporation by β cells and suggest that acute Men1 excision improves preexisting glucose insulin staining areas in mice on either a high-fat diet or chow intolerance and hyperglycemia in db/db diabetic mice through diet (SI Appendix, Fig. 5 D–G and Fig. 4 H and I). Men1 excision promoting β-cell proliferation and increasing the number of also elevated fasting circulating insulin levels and glucose-stim- functional β cells. ulated insulin secretion in mice on either chow diet or high fat diet (Fig. 4 J and K). These results demonstrate that Men1 ex- Acute Men1 Ablation Reverses High-Fat Diet-Induced Preexisting cision increases β-cell proliferation, β-cell number, and circulat- Glucose Intolerance. To determine whether acute Men1 excision ing insulin concentrations and, subsequently, reverses glucose reverses diet-related abnormal glucose homeostasis, we extended intolerance in high-fat diet-fed mice. Although Pdx1-Cre-ER in- our study to a high-fat diet-induced type 2 diabetes mouse duced Men1 excision in β cells, but not in the adjacent acinar model. High-fat diet increases body weight and induces meta- cells in mice (SI Appendix, Fig. 5 B and C), we cannot completely bolic syndrome and diabetes in mice (20, 21). As menin ex- rule out the possibility that Pdx1-Cre-ER-induced Men1 excision pression is down-regulated in islets in obese Ay mice (22), we in tissues beyond β cells might also contribute to amelioration of examined the level of menin expression in C57BL6J mice on glucose intolerance in mice. a high-fat diet for 13 wk. Although the high-fat diet significantly increased body weight (SI Appendix, Fig. 4A) and induced glu- Menin Represses Multiple Proproliferative Genes in Pancreatic β Cells. cose intolerance in C57BL6J mice (SI Appendix, Fig. 4B), it did To understand how Men1 excision promotes β-cell proliferation, not repress menin expression levels in islets, as compared with and to identify early changes in profiles after the mice on chow diet (SI Appendix, Fig. 4 C and D). Therefore, acute Men1 excision, cDNA microarray analysis was performed acute Men1 excision might be able to promote proliferation of on Men1-excised and control islets isolated from mice 14 d after

Yang et al. PNAS Early Edition | 3of6 Downloaded by guest on September 23, 2021 Menin/Insulin A Men1l/l Men1∆/∆ F G Men1l/l & chow Men1l/l & chow B C Men1l/l; Pdx1-Cre-ER & chow Men1∆/∆ & chow Men1l/l & high fat Men1l/l & high fat l/l l/l ∆/∆ l/l Men1 Men1 ; Pdx1-Cre-ER Men1 Men1 // & high fat & high fat ∆/∆ Men1l/l; Men1 600 // 600 ** * Pdx1-Cre-ER 500 500 // *** Age 622 30 (wk) 400 400 * D E 300 300 Diet TAM 8th 200 200 GTT 100 100 0 0 High fat Chow 04080120(min) 04080120(min) Blood glucose (mg/dL) Blood glucose (mg/dL)

Men1l/l & chow I J K ∆/∆ H Men1 & chow Men1l/l & high fat Men1l/l * * Men1l/l Men1l/l Men1∆/∆ & high fat 1.5 Men1∆/∆ 8 Men1∆/∆ 2.0 * Men1∆/∆ 12 10 * * 6 1.5 1.0 8 4 1.0 6 0.5 4 2 .5 2 0

0 (arbitary unit) 0 0 HFD - - + + HFD - - + + HFD - - + + 051015(min) Insulin staining area BrdU incorporation (%) Serum insulin (ng/mL) Serum insulin (ng/mL)

Fig. 4. Acute Men1 excision reverses preexisting high-fat diet (HFD)-induced glucose intolerance. (A) A schematic of experimental design. Control Men1l/l and Men1l/l;Pdx1-Cre-ER mice (n =10–16 mice) were fed high-fat diet or chow diet at the age of 6 wk for 16 wk and followed by TAM treatment. Mice were killed 8 wk after TAM feeding. (B–E) Immunostaining for menin and insulin in islets. (B and C) Mice fed chow diet. (D and E) Mice fed high-fat diet. (Scale bar: 25 μm.) (F and G)GTT(n =8–15 mice) before (F) and 8 wk after (G) TAM treatment. (H) Quantitation of BrdU incorporation by β cells. (I) Quantitation of insulin-staining area. (J) Fasting insulin levels, 8 wk after TAM treatment. (K)GSIS(n =8–15 mice), 8 wk after TAM treatment. *P < 0.05; **P < 0.01; ***P < 0.001.

TAM feeding. Sixty-seven genes were significantly up-regulated However, we did not observe a reduction in mRNA levels of (Fig. 5A and SI Appendix, Fig. 6 and Table 1). Twenty-two genes either p18 or p27 in Men1-excised islets (Fig. 5C). were moderately down-regulated in Men1-excised islets, as com- As acute Men1 excision in islets up-regulated genes that pro- pared with control islets (SI Appendix, Table 2). mote each phase of the cell cycle (Fig. 5 A and B and SI Ap- analysis (24) showed that most of the up-regulated genes were pendix, Fig. 6), collectively, these results suggest that menin is involved in controlling cell cycle, DNA synthesis, and mitosis (SI a crucial player in repressing multiple genes governing G0/G1 Appendix, Table 3). Gene Set Enrichment Analysis (GSEA) (25) to S transition, S-phase progression, and mitosis in islets, and further revealed that several groups of genes involved in G /G Men1 excision may coordinately derepress these genes to pro- 0 1 β to S phase transition and DNA replication were up-regulated in mote -cell proliferation. Men1-excised islets (Fig. 5B and SI Appendix, Table 4). In- Discussion terestingly, many of the genes repressed by menin, including Ccne1, Ccne2, and Ccnd3 (Fig. 5B), are also target genes of the We used a temporally controlled Men1 excision approach and de- cell cycle-regulating transcription factor E2F3 (SI Appendix, Fig. termined the direct impact of acute Men1 excision on preexisting 7) (25–27), and E2F3 itself was mildly up-regulated by Men1 glucose intolerance or hyperglycemia in three distinct diabetes mouse models. Conventional gene targeting mouse models, such as excision (Fig. 5 B and D). the knockout models for D-cyclins and p27, were used to evaluate Pbk, a PDZ-binding kinase regulating mitosis (28) and cyclin the impact of gene ablation on the prevention of diabetes (2, 3). Fur- A were cell cycle genes up-regulated in Men1-excised islets (Fig. thermore, in those models, gene ablation is executed during em- 5A). Cyclin A forms a complex with Cdk2 and promotes G to S 1 bryonic development or before development of diabetes, and the phase and S phase progression (29). The effect of Men1 excision observed effect of prevention could also be attributed to develop- on regulating expression of cyclin A and Pbk was confirmed with fl mental or compensational changes in multiple organs including qRT-PCR, Western blotting, or immnuno uorescent staining pancreatic islets. In contrast, we used a different experimental ap- (Fig. 5 C and D; Fig. 5 E–H for Pbk immunostaining). Most β proach and determined the effect of acute Men1 ablation in di- BrdU-positive cells in islets were also stained positive for Pbk abetic mice after they have developed glucose intolerance or hy- (Fig. 5 G and H), suggesting that the Pbk expression level was perglycemia. This method directly evaluates whether manipulation increased in proliferating cells. Consistent with a positive role for of a single gene can be used for treating preexisting diabetes. We β Pbk in regulating -cell proliferation, ectopic expression of Pbk found that Men1 excision in pancreatic β cells ameliorated preex- β in a -cell line led to an increased cell number (Fig. 5I) and also isting STZ-induced hyperglycemia in diabetic mice. Acute Men1 the percentage of cells in S phase (SI Appendix, Fig. 8). Con- excision corrected preexisting hyperglycemia in db/db diabetic mice versely, Pbk knockdown in β cells resulted in a reduction in the and reversed preexisting glucose intolerance in high-fat diet-fed number of β cells (Fig. 5J), indicating that the Pbk expression mice (21). To our knowledge, this is the first study to show that acute level is correlated with increased β-cell proliferation in Men1- and temporally controlled excision of a single gene reverses preex- excised β cells. Moreover, Hmmr, which encodes the receptor of isting hyperglycemia or glucose intolerance in diabetic mice. hyaluronic acid mediated motility (Rhamm), was up-regulated in Our results showed that Men1 excision ameliorated preexisting Men1-excised islets (Fig. 5K). hyperglycemia or glucose intolerance mainly through increasing the Consistent with a previous report by Scacheri et al. (12), ex- number of functional endogenous β cells. Men1 excision consistently pression of Cdc20 (SI Appendix, Table 1), another menin-regu- increased BrdU incorporation, a marker of cell proliferation, in three lated cell cycle regulator, was also increased by Men1 excision. diabetic models, suggesting that regeneration or replication is in-

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1012257107 Yang et al. Downloaded by guest on September 23, 2021 Men1l/l Men1∆/∆ Men1l/l Men1∆/∆ 14 d after TAM A Nusap1 B Mcm5 C Pbk* Mcm2 6 ** Men1l/l Hmmr Mcm6 Men1∆/∆ Top2a Pole2 Rbl1 4 Kif2c Pola2 Fig. 5. Proproliferative genes are up-regulated in Men1- Cenpf E2f3 BC094916 fi > Ccne1 2 excised islets. (A)Apro le of genes up-regulated by 5-fold in Bub1b Creb3l3 Ccnb2 Mcm7 Men1-excised islets. Pancreatic islets were isolated from con- Prc1 Cdc45l l/l l/l Mki67 Rpa2 0 trol Men1 (n = 8 mice) or Men1 ;Cre-ER mice (n = 6 mice) 14 d Cdca5 Cdk2 25 30 d after TAM Bub1 Orc6l after TAM feeding and processed for RNA isolation and cDNA Ccnb1 Ccnd3 20 Cep55 Mcm4 microarray analysis, as described in Materials and Methods and Rad51ap1 Prim2a Kif20b mRNA (relative level) 15 Aspm Ccnd1 SI Appendix. Genes labeled in bold and with an asterisk were Crebl1 10 ** Dtl Cdkn1c further investigated. (B) Genes involved in G /G to S phase Ccna2* Myc ** 0 1 Cdca8 Ccne2 5 transition were enriched in Men1-excised islets. Gene expres- Pcna Tp53 0 sion profiles of control Men1l/l and Men1-excised islets were n1 18 27 e p cna2 Pbkcnd1 p M C C analyzed against c2 of Molecular Signatures Database with Men1l/l Men1∆/∆ Gene Set Enrichment Analysis (GSEA). (C) Gene expression

D l/l ∆/∆ profiles were verified by qRT-PCR. mRNA levels in islets from E F control Men1l/l or Men1l/l;Cre-ER mice 14 or 30 d after TAM kDa Men1 Men1 70 Menin feeding (n =6–10 mice), using TaqMan primers and probes with Gapdh levels as internal controls. (D) Protein levels in 50 Ponceau S islets isolated from control Men1l/l or Men1 l/l;Cre-ER mice, 8

50 Cyclin A Pbk/ Insulin / DAPI wk after TAM treatment (n =4–5 mice). Cyclin A protein levels 50 Ponceau S G H in islets were quantitated by measuring signal density and area of cyclin A2 bands in Western blotting (1.65-fold, Men1-excised 50 E2F3 DAPI vs. control islets). (E–H) Immunostaining for Pbk and BrdU in 50 Ponceau S Men1l/l and Men1l/l;Pdx1-Cre-ER mice (n =10–11 mice) fed with TAM at the age of 12 wk and killed 8 wk after TAM treatment. Pbk/ BrdU/ (E and F) Immunostaining for Pbk and insulin. (G and H) ) ) 5

5 Immunostaining for Pbk and BrdU. Arrow denotes cells stained I Vector J Scramble K 5 μ ** Pbk ** shRNA 1 positively for Pbk and BrdU. (Scale bar: 25 m.) (I and J) Growth 4 30 shRNA 2 shRNA 3 curves for mouse Pancreatic Islet-derived Men1 Excisable cells 3 30 d after TAM 20 (PIME1 cells), as described in SI Appendix.(I) Growth curves of CELL BIOLOGY 2 *** *** Men1l/l 10 Men1∆/∆ control and Pbk overexpressing PIME1 cells. (Upper) Cell 1 *** *** 8 0 0 * numbers after seeding. (Lower) Protein levels of Pbk in control 0246(d) 0246(d) 6 Cell number (1X10 Cell number (1X10 and Pbk overexpressing cells. (J) Growth curves of control and or le 1 3 4 A A2 A Pbk knockdown PIME1 cells. (Upper) Cell numbers after seed- ect mb N N mRNA kDa V Pbk RN R 2 cra hR h ing. (Lower) Pbk protein level in control and Pbk knockdown kDa S s s sh 37 Pbk (relative level) 0 l/l 37 Pbk 1 cells. (K) Hmmr mRNA levels in islets from control Men1 or n l/l 37 Beta-actin Me Men1 ;Cre-ER mice 30 d after TAM feeding, as described in Beta-actin Hmmr 37 Fig. 5C (n = 6 mice). *P < 0.05; **P < 0.01; ***P < 0.001.

creased due to Men1 excision in stressed β cells. Moreover, the riod (10, 11). Gene profiling analysis on islets from mice with number of β cells, islet size, and circulating insulin levels were sig- floxed Men1;Rip-Cre shows that p18, but not p27, is down-regu- nificantly increased in Men1-excised diabetic mice. These findings lated in pancreatic islets 15 wk after Men1 excision (12). Although suggest that Men1 excision reversed preexisting hyperglycemia in both Scacheri et al. and our group found that cell cycle regulator diabetic mice, at least partly through increased β cell proliferation and Cdc20 was up-regulated in Men1-excised islets (12) (SI Appendix, the number of functional β cells. However, we cannot rule out that Table 1), we did not observe a reduction in p18 or p27 mRNA Men1 excision in β cells or in other tissues also improves β-cell levels in islets 14 d after Men1 excision. Different Cre systems function, insulin secretion, or peripheral insulin sensitivity under di- (Ubc9-Cre-ER vs. Rip-Cre) used for the studies and/or different abetic conditions, contributing to the improvement of abnormal times when Men1 was excised (14 d vs. 15 wk) could partly ac- glucose metabolism in diabetic mice. It is also possible that non-β cells count for the discrepancy (12). We previously found that mRNA in the pancreas can reprogram into insulin-secreting β cells in diabetic levels of p18 and p27 were reduced in total pancreatic tissues conditions, because Men1-excised glucagon-secreting α cells can from Men1l/l; Ubc9-Cre-ER mice that were fed TAM (31). Thus, transdifferentiate into β cells and insulinomas (30). we cannot rule out the possibility that the expression of p18 and Our molecular analysis using an acute Men1 excision system p27 is reduced in β cells a long period after Men1 excision and revealed that menin controls expression of multiple cell cycle that Men1 excision influences expression of p18 and p27 in non- genes, and Men1 excision may derepress these genes and promote endocrine cells in the pancreas. β-cell proliferation. For instance, expression levels of Ccna2 and We found that Men1 excision rapidly ameliorated hyperglycemia multiple Mcms, genes important for S phase progression or DNA within 3 wk after Men1 excision. Moreover, Men1 expression has replication, were increased in Men1-excised islets (Fig. 5 A and B). been reported to be physiologically suppressed in islets during and Pbk, Ccnb1, and Ccnb2, a group of genes involved in G2/M tran- after pregnancy in mice by the prolactin signaling pathway to pre- sition and mitosis, were also up-regulated by Men1 ablation (Fig. vent gestational diabetes (22). These findings suggest that menin 5A). Because ectopic Pbk expression and Pbk knockdown signif- inhibition might be a valuable means of promoting β-cell pro- icantly affected β-cell proliferation (Fig. 5 I and J; SI Appendix, liferation and treating diabetes. However, great caution must be Fig. 8), and Pbk is mainly expressed in β cells but not in acinar cells exercised because β-cell hyperplasia, insulinomas, and hyper- (Fig. 5 E–H), Pbk may act as one of tissue-specific effector of insulinemia have been detected in Men1 knockout mouse models menin in controlling β cell proliferation. However, further studies and in MEN1 patients (5, 6, 8, 31). This finding raises a substantial are necessary to determine whether Pbk is essential for Men1 ex- concern that permanent menin inhibition as a therapeutic approach cision-induced β-cell proliferation. may eventually lead to insulinomas (6, 32). If the aforementioned It has been reported that p18 and p27 are down-regulated in situation is the case, targeting menin itself as a means to treat di- islets in which the Men1 gene had been lost for a prolonged pe- abetes would not be viable.

Yang et al. PNAS Early Edition | 5of6 Downloaded by guest on September 23, 2021 However, because menin has multiple functions including re- Biomedicals) at 200 mg/kg of body weight per day for two consecutive days, gulation of β-cell proliferation and DNA repair (33), it might be followed by 1 d off and then for another two consecutive days (31). possible to modulate the function of menin in regulating β-cell proliferation, yet retain its other functions, such as DNA repair, STZ-Induced Hyperglycemia. Multiple low doses of STZ (Sigma) were injected fi thus reducing the possibility of tumor formation. In this regard, (i.p.) at 40 mg/kg of body weight per day for ve consecutive days (36). Mice were diagnosed as diabetic when their nonfasting blood glucose levels menin has been reported to repress gene expression through were >250 mg/dL for two consecutive weeks. interaction with repressive histone deacetylases (HDACs) (34) and inhibit cell proliferation (35). It is possible that menin re- High-Fat Diet Feeding. C57BL6J, control Men1l/l, and Men1l/l;Pdx1-Cre-ER mice presses expression of proproliferative genes, such as Ccna2 and at the age of 6 wk were fed either high-fat diet (51 kcal% fat; Research Pbk, through repressive histone modifications via HDACs or Diets) or chow diet (Harlan) for 13–16 wk. perhaps other hitherto unidentified histone modifiers. Therefore, targeting or repressing interaction between menin and repressive Physiological Measurements. Blood glucose levels were assayed from tail vein histone modifiers could lead to increased expression of propro- blood. Serum insulin levels were measured by ELISA using a mouse insulin kit liferative genes and β-cell proliferation. These approaches could (Crystal Chem). Glucose tolerance test (GTT; glucose at 2 g/kg of body weight, reduce the concern over directly targeting menin and thereby i.p.) and glucose-stimulated insulin secretion (GSIS; glucose at 3 g/kg of body shedding light on improving diabetes therapy. weight, i.p.) were performed on mice fasting overnight. Insulin tolerance test (ITT, insulin at0.75 U/kgof body weight, i.p.) was performed on nonfasting mice. Materials and Methods Methods on islets perifusion, immunohistochemistry, cDNA microarray analysis, qRT-PCR, Gene Ontology, GSEA, and Western blotting were de- Mice. Men1l/l;Cre-ER mice were generated by crossing floxed Men1 mice scribed in SI Appendix. (Men1l/l, FVB/129Sv, from Francis Collins) to Cre-ERT2 mice (129Sv/C57BL6J) (6, 15). db/db; Men1l/l;Cre-ER mice were generated by crossing Men1l/l;Cre-ER Statistical Analysis. Results are expressed as mean ± SEM. For two-group mice to db/+ mice (BKS.Cg-Dock7m+/+Leprdb/J; The Jackson Laboratory) (17). comparison, unpaired Student’s t test or rank sum test was used. For mul- Men1l/l;Pdx1-Cre-ER mice were generated by crossing floxed Men1 mice to tiple-group comparison, one-way ANOVA or ANOVA on ranks was used. P < Pdx1-Cre-ER mice (23). Only male mice were used for the following experi- 0.05 was considered statistically significant. ments. Genotyping of mice was performed by PCR on mouse tail DNA. All mouse studies were approved by the University Laboratory Animal Resource and the University of Pennsylvania committee on animal care; the animal ACKNOWLEDGMENTS. We thank Jizhou Yan and Wei Qin for technical assis- tance and Drs. Gary Koretzky, Steven Reiner, Mitchell A. Lazar, Morris Birnbaum, care was in accordance with institutional guidelines. and Klaus Kaestner for stimulating discussions. This work was supported in part by National Institutes of Health Grants R01-CA-113962, R56-DK08512, Excision of the Floxed Men1 Locus. Men1l/l;Cre-ER, db/db;Men1l/l;Cre-ER, and R01-DK085121 (to X.H.), American Diabetes Association Grant 7-07-RA-60 Men1l/l;Pdx1-Cre-ER mice and their littermate controls were fed TAM (MP (to X.H.), and National Institutes of Health Grant R01 DK068157 (to D.A.S.).

1. Butler AE, et al. (2003) β-cell deficit and increased β-cell apoptosis in humans with type 20. Biddinger SB, et al. (2005) Effects of diet and genetic background on sterol regulatory 2 diabetes. Diabetes 52:102–110. element-binding protein-1c, stearoyl-CoA desaturase 1, and the development of the 2. Heit JJ, Karnik SK, Kim SK (2006) Intrinsic regulators of pancreatic β-cell proliferation. metabolic syndrome. Diabetes 54:1314–1323. Annu Rev Cell Dev Biol 22:311–338. 21. Winzell MS, Ahrén B (2004) The high-fat diet-fed mouse: A model for studying 3. Uchida T, et al. (2005) Deletion of Cdkn1b ameliorates hyperglycemia by maintaining mechanisms and treatment of impaired glucose tolerance and type 2 diabetes. compensatory hyperinsulinemia in diabetic mice. Nat Med 11:175–182. Diabetes 53(Suppl 3):S215–S219. 4. Fu A, et al. (2009) Loss of Lkb1 in adult β cells increases β cell mass and enhances 22. Karnik SK, et al. (2007) Menin controls growth of pancreatic β-cells in pregnant mice glucose tolerance in mice. Cell Metab 10:285–295. and promotes gestational diabetes mellitus. Science 318:806–809. 5. Marx SJ (2005) Molecular genetics of multiple endocrine neoplasia types 1 and 2. Nat 23. Gu G, Dubauskaite J, Melton DA (2002) Direct evidence for the pancreatic lineage: Rev Cancer 5:367–375. NGN3+ cells are islet progenitors and are distinct from duct progenitors. 6. Crabtree JS, et al. (2003) Of mice and MEN1: Insulinomas in a conditional mouse Development 129:2447–2457. knockout. Mol Cell Biol 23:6075–6085. 24. Dennis G, Jr., et al. (2003) DAVID: Database for annotation, visualization, and 7. Scacheri PC, et al. (2004) Homozygous loss of menin is well tolerated in liver, a tissue integrated discovery. Genome Biol 4:P3. not affected in MEN1. Mamm Genome 15:872–877. 25. Subramanian A, et al. (2005) Gene set enrichment analysis: A knowledge-based 8. Bertolino P, et al. (2003) Pancreatic β-cell-specific ablation of the multiple endocrine approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA neoplasia type 1 (MEN1) gene causes full penetrance of insulinoma development in 102:15545–15550. mice. Cancer Res 63:4836–4841. 26. Bild AH, et al. (2006) Oncogenic pathway signatures in human cancers as a guide to 9. Biondi CA, et al. (2004) Conditional inactivation of the MEN1 gene leads to pancreatic targeted therapies. Nature 439:353–357. and pituitary tumorigenesis but does not affect normal development of these tissues. 27. Bracken AP, Ciro M, Cocito A, Helin K (2004) E2F target genes: Unraveling the biology. Mol Cell Biol 24:3125–3131. Trends Biochem Sci 29:409–417. 10. Karnik SK, et al. (2005) Menin regulates pancreatic islet growth by promoting histone 28. Gaudet S, Branton D, Lue RA (2000) Characterization of PDZ-binding kinase, a mitotic methylation and expression of genes encoding p27Kip1 and p18INK4c. Proc Natl Acad kinase. Proc Natl Acad Sci USA 97:5167–5172. Sci USA 102:14659–14664. 29. Rosenberg AR, et al. (1995) Overexpression of human cyclin A advances entry into S 11. Milne TA, et al. (2005) Menin and MLL cooperatively regulate expression of cyclin- phase. Oncogene 10:1501–1509. dependent kinase inhibitors. Proc Natl Acad Sci USA 102:749–754. 30. Lu J, et al. (2010) Alpha cell-specific Men1 ablation triggers the transdifferentiation of 12. Scacheri PC, et al. (2006) Genome-wide analysis of menin binding provides insights glucagon-expressing cells and insulinoma development. Gastroenterology 138: into MEN1 tumorigenesis. PLoS Genet 2:e51. 1954–1965. 13. Hughes CM, et al. (2004) Menin associates with a trithorax family histone 31. Schnepp RW, et al. (2006) Mutation of Men1 acutely enhances methyltransferase complex and with the hoxc8 locus. Mol Cell 13:587–597. proliferation of pancreatic islet cells. Cancer Res 66:5707–5715. 14. Yokoyama A, et al. (2005) The menin tumor suppressor protein is an essential 32. Harding B, et al. (2009) Multiple endocrine neoplasia type 1 knockout mice develop oncogenic cofactor for MLL-associated leukemogenesis. Cell 123:207–218. parathyroid, pancreatic, pituitary and adrenal tumours with hypercalcaemia, 15. Lois C, Hong EJ, Pease S, Brown EJ, Baltimore D (2002) Germline transmission and hypophosphataemia and hypercorticosteronaemia. Endocr Relat Cancer 16: tissue-specific expression of transgenes delivered by lentiviral vectors. Science 295: 1313–1327. 868–872. 33. Busygina V, Kottemann MC, Scott KL, Plon SE, Bale AE (2006) Multiple endocrine 16. Lenzen S (2008) The mechanisms of alloxan- and streptozotocin-induced diabetes. neoplasia type 1 interacts with forkhead transcription factor CHES1 in DNA damage Diabetologia 51:216–226. response. Cancer Res 66:8397–8403. 17. Lee GH, et al. (1996) Abnormal splicing of the leptin receptor in diabetic mice. Nature 34. Kim H, Lee JE, Cho EJ, Liu JO, Youn HD (2003) Menin, a tumor suppressor, represses 379:632–635. JunD-mediated transcriptional activity by association with an mSin3A-histone 18. Coleman DL, Hummel KP (1974) Hyperinsulinemia in pre-weaning diabetes (db) mice. deacetylase complex. Cancer Res 63:6135–6139. Diabetologia 10(Suppl):607–610. 35. Wu T, Zhang X, Huang X, Yang Y, Hua X (2010) Regulation of cyclin B2 expression and 19. Kobayashi K, et al. (2000) The db/db mouse, a model for diabetic dyslipidemia: cell cycle G2/m transition by menin. J Biol Chem 285:18291–18300. Molecular characterization and effects of Western diet feeding. Metabolism 49: 36. Liadis N, et al. (2005) Caspase-3-dependent β-cell apoptosis in the initiation of 22–31. autoimmune diabetes mellitus. Mol Cell Biol 25:3620–3629.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1012257107 Yang et al. Downloaded by guest on September 23, 2021