Introduction

In mammals, control of cell proliferation in response to growth-stimulatory signals is mediated by a complex molecular machinery that must regulate exit from quiescence as well as progression through G1 into the S phase (Pardee, 1989). Evidence, derived mostly from cell culture experiments, indicates that cyclin-dependent kinase 4 (Cdk4) plays a central role as a sensor of at least some of these proliferative signals in early G1, regulating the transition from quiescence to proliferation (Reed, 1997; Sherr and Roberts, 1999; Malumbres and Barbacid, 2001). Cdk4 activity requires D-type cyclins (D1, D2 and D3), whose synthesis is regulated by mitogenic signaling pathways. Cdk4–CyclinD complexes phosphorylate and partially inactivate the pocket proteins pRb, p107 and p130, thereby activating the E2F/DP transcription factors that modulate the expression of genes required for cell cycle progression (Weinberg, 1995; Dyson, 1998; Nevins, 1998; Sherr, 2000). Early E2F-responsive genes include E- and A-type cyclins that bind and activate the Cdk2 kinase, an enzyme thought to be essential for progression through the S phase (Lundberg and Weinberg, 1998; Sherr and Roberts, 1999).

The activity of Cdk4 can be negatively regulated by proteins of the INK4 family (p16INK4a, p15INK4b, p18INK4c and p19INK4d) that prevent its binding to the D-type cyclins by an allosteric mechanism (Morgan, 1995; Pavletich, 1999). The maintenance of proper Cdk4 activity during G1 to S phase is critical to avoid abnormal cell proliferation that may lead to neoplastic development (reviewed in Ruas and Peters, 1998; Malumbres and Barbacid, 2001). For instance, in certain human cancers such as hereditary melanoma, Cdk4 is mutated (substitution of arginine 24 by cysteine) in a way that prevents INK4 binding (Wolfel et al., 1995; Zuo et al., 1996). Mice in which this mutation has been engineered by gene targeting of the Cdk4 locus (Cdk4^R/R) develop multiple tumors and become highly susceptible to the development of invasive melanoma and angiosarcomas (Sotillo et al., 2001a, 2001b). Mouse embryonic fibroblasts (MEFs) derived from Cdk4^R/R mice do not undergo senescence and show certain characteristics of neoplastic cells (Sotillo et al., 2001a).

On the other hand, loss of Cdk4 expression has no significant consequences for a wide variety of cell types (Rane et al., 1999; Tsutsui et al., 1999). Cdk4 null (Cdk4^n/n) mice are viable, albeit considerably smaller than wild-type animals. Cdk4 null MEFs proliferate normally, although they enter from quiescence into S phase with a slight delay. Yet, Cdk4 is necessary for the proliferation of certain endocrine cell types. Adult Cdk4^n/n mice become diabetic because of reduced numbers of insulin-producing pancreatic cells. These mice also have decreased male fertility because of defective spermatogenesis and reduced numbers of Leydig cells. Furthermore, female Cdk4^n/n mice are sterile because of limited prolactin production, a consequence of reduced numbers of pituitary lactotrophs (Moons et al., 2002a, 2002b).

In the present study, we analyse the molecular mechanisms responsible for the proliferative defects observed in Cdk4 null (Cdk4^n/n) mice. Using genetic crosses to restore Cdk4 expression to the affected cell types, we were able to revert the diabetes and infertility phenotypes suffered by the Cdk4^n/n animals, thus indicating that the proliferative defects observed in cells and pituitary cells are cell autonomous. Yet, these mice retain their small size phenotype, suggesting that Cdk4 may play a role in controlling homeostatic cell numbers.

Results

Embryonic -cell development is not affected in Cdk4 mutant mice

In mice, endocrine cell differentiation in the developing pancreas is initiated at day 8.5 of gestation (E8.5). At E9.5, scattered cells expressing insulin or glucagon are already detectable but the typical architecture of the islets of Langerhan is not apparent until late in gestation (E18.5) (Herrera et al., 1991; Kim and Hebrok, 2001). A rapid increase in -cell mass takes place during the perinatal period (Montanya et al., 2000; Bonner-Weir, 2000a, 2000b). At E15.5, Cdk4^n/n and Cdk4^R/R embryos have similar numbers of insulin- and glucagon-producing cells as wild-type embryos (Figure 1A). A statistical analysis of the number of these cells per mm² in equivalent pancreatic sections of E15.5 embryos does not reveal significant differences among genotypes (P>0.05, n=8). Likewise, the -cell mass at day 5 postnatum (P5) of Cdk4^n/n and Cdk4^R/R is comparable to that of wild-type animals (Figure 2B). Thus, Cdk4 does not appear to be essential for -cell neogenesis in the embryo. Analysis of the subcellular localization of Cdk4 protein reveals that in E15.5 wild-type embryos there is no detectable nuclear Cdk4 expression in insulin-producing cells. Moreover, none of these insulin-producing cells stain positively for Ki67 (Figure 1B). At E17.5, Cdk4 is present in the nucleus and, coincident with this, the insulin-producing cells begin to proliferate, as determined by Ki67 staining (Figure 1B). These observations suggest that nuclear Cdk4 expression is developmentally regulated to coincide with, and presumably promote the initiation of the proliferative program of pancreatic cells.

Figure 1.

Embryonic -cell differentiation and proliferation in Cdk4 mutant mice. (A) Representative paraffin sections of pancreas from Cdk4^+/+ (a,d), Cdk4^n/n (b,e) and Cdk4^R/R (c,f) E15.5 embryos immunostained for insulin (a–c) and glucagon (d–f). Magnification 400. (B) Coimmunostaining for Cdk4 (a–c) or Ki67 (d–f) (brown) and insulin (blue) of pancreas sections from wild-type E15.5 embryos (a,d), E17.5 embryos (b,e) and P1 mice (c, f). Arrows indicate –cell nuclei positive for Cdk4 (b,c) or Ki67 (e,f). Magnification 1000

Full figure and legend (430K)

Figure 2.

Altered postnatal islet growth and proliferation in Cdk4 mutant mice. (a) Percentage of islets containing 1–40, 41–400 or more than 400 cells/section in Cdk4^+/+ (left), Cdk4^n/n (middle) and Cdk4^R/R (right) mice at P5 (open bars), P20 (light dotted bars), P60 (heavy dotted bars) and P120 (solid bars). (b) Islet cell mass in Cdk4^+/+ (open bars) Cdk4^n/n (solid bars), and Cdk4^R/R (hatched bars) in P5 (left) and P120 (right) mice. (c) Percentage of BrdU-positive cells in islets and acini in pancreas from Cdk4^+/+ (open bars), Cdk4^n/n (solid bars) and Cdk4^R/R (hatched bars) P10 mice

Full figure and legend (68K)

Altered postnatal growth of endocrine islets in Cdk4 mutant mice

Analysis of the number and size of pancreatic islets at different ages (P5, P20, P60 and P120) revealed that over time the islets of Cdk4^+/+ mice grow in size, evolving from mostly small islets (1–40 cells/islet-section) to medium- and large-sized islets (41–400 cells/islet-section and >400 cells/islet-section, respectively) (Figure 2a). This progression is accelerated in Cdk4^R/R mice (Figure 2a). In contrast, islets of Cdk4^n/n mice remain small in size and fail to grow to medium- and large-sized islets. As a consequence, the total -cell mass of Cdk4^n/n mice at P120 is only 10% of that of their wild-type littermates (Figure 2b).

At any given time, -cell mass is a balance between the rate of production of new -cells, either by neogenesis or proliferation, and the rate of apoptosis (Bonner-Weir, 2000b). Levels of -cell proliferation were determined by measuring BrdU incorporation. At P10, islets of Cdk4^n/n mice display a 3.5-fold reduction in the proportion of cells that incorporate BrdU compared with Cdk4^+/+ islets, while Cdk4^R/R islets display a 2.5-fold increase relative to Cdk4^+/+ islets (Figure 2c). In the exocrine pancreas the proportion of acinar cells labeled with BrdU is similar in Cdk4^n/n mice and Cdk4^+/+ and slightly higher (10%) in Cdk4^R/R animals compared to Cdk4^+/+. Immunohistochemistry experiments using double labeling with antibodies against Ki67 and either insulin, glucagon, somatostatin or pancreatic polypeptide confirmed the previous results for cells and showed that the level of proliferation of the other endocrine cells is not altered in the Cdk4 mutant mice (data not shown). Taken together, these results indicate that Cdk4 activity plays a key role in controlling the postnatal proliferation rate of cells, but not of other pancreatic endocrine cells. The apoptotic index in the pancreas of the Cdk4^n/n mice was determined by TUNEL assays. No significant difference with wild-type mice was detected in either the islets or the exocrine tissue (data not shown), indicating that loss of Cdk4 does not significantly affect the survival of pancreatic cells.

Expression of the G1 to S phase cyclin-dependent kinases in the pancreas

Although it is generally assumed that cell cycle Cdks are ubiquitous enzymes, little is known about their expression pattern in vivo or that of their direct regulators. To address this issue, the levels of Cdk4, Cdk6 and Cdk2 mRNAs were measured in the islets of P60–P90 wild-type mice using real-time quantitative RT–PCR. The low abundance of certain populations of endocrine cells (e.g. somatostatin or PP-producing cells) prevented us from isolating populations of specific cell types (see below). Cdk4 mRNA levels were the most abundant, being at least two orders of magnitude higher than those of Cdk6, which were undetectable under the conditions used (Figure 3). Cdk4 is also the most abundant G1 Cdk in acinar cells but these cells do express Cdk6 mRNA, albeit at low levels (data not shown).

Figure 3.

Expression of G1/S transition regulators in pancreatic islets. G1 Cdks (top left), D-type cyclins (top right), INK4 inhibitors (bottom left) and Cip/Kip inhibitors (bottom right) genes. Data were obtained by real-time quantitative RT–PCR and are expressed as the number of cDNA copies per g of total RNA. Error bars represent the standard deviation of data from two independent assays performed in duplicates

Full figure and legend (38K)

We also analysed the pattern of expression of Cdk4 regulators in islet cells. Cyclin D2 mRNA was the most abundant among the D-type cyclins. Cyclin D1 mRNA was also present, albeit at about six-fold lower levels. However, the levels of cyclin D3 mRNA were found to be below the limit of detection (Figure 3). Among the negative regulators, we detected mRNA for all members of the INK4 family, with p15INK4b and p19INK4d being the most abundant. Finally, the Cip/Kip inhibitors p21Cip1 and p27Kip1 were found to be expressed at levels similar to those of p15INK4b and p19INK4d (Figure 3).

Pituitary abnormalities in Cdk4 mutant mice

The pituitary glands of Cdk4^n/n mice are about half the size of those of Cdk4^+/+ mice of the same age and gender. Histological and morphometric analyses of the pituitaries of P120 Cdk4^n/n mice reveal that the cell density in the neurohypophysis and the intermediate lobe (pars intermedia) is normal (P=0.19 and P=0.57, respectively, n=18). In contrast, the cellularity of the adenohypophysis of Cdk4^n/n mice is reduced by at least 30% when compared with wild-type littermates (P<0.0003, n=18) (Figure 4a). Cdk4^R/R mice show increased cellularity in the adenohypophysis and to a lesser extent in the pars intermedia (P<0.0001 and P<0.001 respectively, n=18) (Figure 4a). Analysis of specific cellular subtypes within the adenohypophysis of the Cdk4^n/n mice reveals a drastic reduction (70%) in PRL-producing cells and a slight reduction in those producing GH (Figure 4b), but not in cells producing either FSH or LH (data not shown). These observations correlate well with a report that attributes infertility in Cdk4^n/n females to a defect in pituitary lactotrophs rather than in gonadotrophs (Moons et al., 2002a, 2002b).

Figure 4.

Altered cell numbers in the pituitary of Cdk4 mutant mice. (a) Number of cells in the three functional regions in pituitaries from P120 mice. (b) Number of adenohypophysis cells positive for immunostaining with specific antibodies against prolactin (PRL) and growth hormone (GH) in pituitary sections from P120 mice. GH-producing cells are representative of the rest of the hormone-producing cells of the pituitary in relative numbers. Cdk4^+/+ (open bars), Cdk4^n/n (solid bars) and Cdk4^R/R (hatched bars) mice

Full figure and legend (83K)

Restoration of Cdk4 expression in Cdk4 null mice

Loss of Cdk4 expression in Cdk4^n/n mice is because of the insertion of a floxed PGK-neo-cassette in intron 1 of the Cdk4 locus (Figure 5A) (Rane et al., 1999). Thus, Cdk4 expression can be specifically restored by crossing these mice with transgenic animals that express the Cre recombinase under the control of tissue-specific promoters. It should be noted however that the Cre-activated Cdk4 allele will express the mutant Cdk4R24C protein present in Cdk4^R/R mice (Rane et al., 1999). However, since these mice do not present relevant physiological alterations (mainly tumor development) until almost a year of age, this fact should not interfere with our experiments that involve mice up to 4 months of age (see below).

Figure 5.

Activation of Cdk4R24C expression in Cdk4^n/n;+/Cre mice. (A) Generation of the transcriptionally active Cdk4R24C allele by excision of the PGK-neo-cassette of the silent Cdk4neo-allele in tissues of Cdk4^n/n;+/Cre mice derived from crossing Cdk4^n/n mice with a transgenic line expressing the Cre recombinase under the control of the rat insulin promoter. Solid boxes represent Cdk4-coding sequences; hatched boxes, Cdk4-noncoding sequences and triangles, loxP sequences. (B) Amplification by PCR of specific bands for Cdk4neo and Cdk4R24C alleles in a series of tissues: finger (Fg), pancreatic acinar cells (Ac), pancreatic islets (Is), pituitary (Pt), ovary (Ov), testis (Ts), seminal vesicle (Sv), tail (Tl), brain (Br), heart (Ht), intestine (In), kidney (Kd), liver (Lv), lung (Ln), muscle (Ms), skin (Sk), spleen (Sp) and uterus (Ut). (C) Nuclear Cdk4 immunostaining (brown) in representative pancreatic islet (a–c) and adenohypophysis (d–f) sections from Cdk4^+/+;+/+ (a,d), Cdk4^n/n;+/+ (b,e) and Cdk4^n/n;+/Cre (c,f) P120 littermates. Magnification 1000

Full figure and legend (278K)

In the present studies, we were primarily interested in restoring Cdk4 expression in pancreatic cells. To this end, we crossed Cdk4^n/n mice with a transgenic strain that expresses the Cre recombinase under the control of the rat insulin promoter (Rip-Cre mice) (Postic et al., 1999). The Cre recombinase is expressed in the cells of the Rip-Cre mice, but not in other pancreatic endocrine or acinar cells (Kulkarni et al., 1999). Fortuitously, this recombinase is also expressed in all three regions of the pituitary gland: adenohypophysis, neurohypophysis and the pars intermedia (Gannon et al., 2000). Excision of the neo-cassette in the pancreatic islets and the pituitary, but not in a variety of other tissues, was confirmed by PCR of genomic DNA of the resulting Cdk4^n/n;+/Cre mice (Figure 5B). The restoration of expression of Cdk4R24C in cells and in the adenohypophysis was confirmed by immunohistostaining with anti-Cdk4 antibodies (Figure 5C).

Phenotypic characterization of Cdk4^n/n;+/Cre mice

Expression of Cdk4R24C in pancreatic cells and pituitary cells rescued the diabetic and infertility phenotypes characteristic of Cdk4^n/n mice. Elevated serum glucose levels in Cdk4^n/n mice start to be detectable at approximately 1.5 months of age (data not shown) and more than 80% of the Cdk4^n/n mice develop hyperglycemia and diabetes and die within the first 6 months of life (Rane et al., 1999). The average glucose concentration in blood of these mutant mice is 426150 versus 12812 mg/dl in wild-type littermates. The wide standard deviation in Cdk4^n/n animals is because of the variable time at which hyperglycemia develops. In contrast, Cdk4^n/n;+/Cre mice are normoglycemic and survive for at least 18 months of age (data not shown). These results suggest that the early lethality of Cdk4^n/n mice is caused by their diabetic phenotype. An analysis of the pancreatic sections of P120 Cdk4^n/n;+/Cre animals indicates that the cells of these animals have regained their proliferative capacity (Figure 6a) and the architecture of the islets in Cdk4^n/n;+/Cre mice is normal (data not shown). Similar results were observed in the pituitary gland (Figure 6b). Interestingly, the number of cells in Cdk4^n/n;+/Cre mice is lower than that observed in Cdk4^R/R animals (Figure 6a). These differences could be because of the differences in the genetic background of these strains. However, we cannot rule out that a percentage of cells in Cdk4^n/n;+/Cre mice are either null (Cdk4^n/n) or heterozygous (Cdk4^n/R) for Cdk4 R24C expression because of incomplete activity of the Cre recombinase.

Figure 6.

Phenotype of Cdk4^n/n;+/Cre mice. (a) Percentage of islet area in pancreas sections. (b) Cell density in the adenohypo-physis of Cdk4^+/+;+/+, Cdk4^+/+;+/Cre, Cdk4^n/n;+/+, Cdk4^n/n;+/Cre P120 littermates. Data for Cdk4^R/R mice of a similar age are shown in A and B (diagonally hatched bars). (c) Postnatal growth curves of male (left) and female (right) Cdk4^+/+;+/+ (open squares), Cdk4^+/+;+/Cre (open circles), Cdk4^n/n;+/+ (solid squares) and Cdk4^n/n;+/Cre (solid circles) mice. 10 males and 10 females of each genotype were weighed

Full figure and legend (68K)

All female Cdk4^n/n;+/Cre mice mated so far (n=8) are fertile and have yielded several litters. The ovaries of Cdk4^n/n;+/Cre females have a normal morphology (data not shown). The rescue of female fertility in Cdk4^n/n;+/Cre is most likely because of the fortuitous expression of the Cre recombinase in the pituitary of these animals, since there is no Cre recombinase activity in the ovaries of Rip-cre transgenic mice and no Cdk4 expression was detected in this tissue in Cdk4^n/n;+/Cre females (data not shown).

Cdk4^n/n;+/Cre mice retain small size and body weight

In spite of rescuing the diabetic and infertility phenotypes of Cdk4^n/n mice, Cdk4^n/n;+/Cre animals continue to be small in size. The average length from tip of nose to base of tail of adult (4–6 months) Cdk4^n/n or Cdk4^n/n;+/Cre male mice is 8.600.36 cm. This value represents a 17% reduction in body length compared to that of wild-type littermates (average length 10.400.42 cm; P<0.001, n=12). A similar reduction is observed in female Cdk4^n/n or Cdk4^n/n;+/Cre animals (8.50.4 versus 10.170.29 cm, P<0.001, n=12). In addition, these mice display as much as 40% reduction in weight as illustrated in Figure 6c. On the other hand, adult Cdk4^R/R mice are not significantly longer than their wild-type littermates, in spite of having a 10% increase in overall body weight (Rane et al., 1999).

Morphometric examination of a variety of tissues reveals that this phenotype is not because of reduced cell size, indicating that these animals must have reduced cell numbers. Indeed, the size of Cdk4 null cells in most tissues (including spleen, liver and kidney) appears to be slightly larger (by approximately 10%) than normal cells (Figure 7). Interestingly, the cells of Cdk4^n/n mice are almost twice as large as those of their wild-type counterparts. This observation suggests that Cdk4 does not play a role in regulating the increase in cell size that many cells undergo before entering S phase. Instead, the increased size of the Cdk4^n/n cells may reflect an uncoupling of cell size regulation and progression through the cell cycle and cell division. Furthermore, these results establish that the smaller body weight and organ size of Cdk4^n/n and Cdk4^n/n;+/Cre mice cannot be accounted for by their hormonal defects, at least those derived from pancreatic or pituitary hormones. Therefore, it is likely that the reduced size of these mice is a direct consequence of the lack of Cdk4 expression.

Figure 7.

Cell size in Cdk4^n/n mice. Cell size in the indicated tissues of Cdk4^n/n mice (solid bars) relative to the cell size in the same tissues of Cdk4^+/+ animals (open bars)

Full figure and legend (28K)

Discussion

Our current view of how Cdks regulate commitment of quiescent cells to a new round of cell division has been primarily derived from cell culture studies. Gene ablation in mice has revealed a more complex scenario in which different members of the Cdk/D-type cyclin families play cell-type-specific roles (Fantl et al., 1995; Sicinski et al., 1995; Rane et al., 1999; Tsutsui et al., 1999). For instance, whereas Cdk4 is not essential for mouse development, adult mice lacking this kinase develop diabetes because of a severe decrease in pancreatic -cell mass (Rane et al., 1999; Tsutsui et al., 1999). Here, we show that this defect is because of a cell autonomous requirement of Cdk4 for postnatal -cell proliferation. Reconstitution experiments in which Cdk4^n/n mice have been crossed with Rip-Cre transgenic animals leading to re-expression of Cdk4 in pancreatic cells, results in restoration of -cell proliferation with the concomitant disappearance of the diabetic phenotype. The long survival of Cdk4^n/n;+/Cre mice also suggests that the limited lifespan of Cdk4 null mice is because of their early onset of diabetes.

Cdk4 null embryos have normal levels of insulin-producing cells. Since at this time the major source of cells is neogenesis from precursor ductal epithelial cells (Finegood et al., 1995), our results suggest that Cdk4 may not be required for differentiation of cells. In midgestation embryos, E15.5, cells do not display nuclear Cdk4 immunostaining, suggesting that Cdk4 is not active during this developmental period. However, Cdk4 appears in the nucleus of cells as they begin to form islet structures in late gestation embryos (E17.5/E18.5), coinciding with the appearance of insulin- and Ki67- double-positive cells. These results suggest that Cdk4 activity is developmentally regulated at a time when expansion of the -cell population switches from neogenesis from ductal cell precursors to proliferation of pre-existing cells.

The requirement of Cdk4 activity for -cell proliferation may be because of the fact that the other cyclin D-dependent kinase, Cdk6, which potentially could compensate for the loss of Cdk4, is not expressed in this cell type. Indeed, pancreatic acinar cells, which proliferate normally in Cdk4^n/n mice, do express Cdk6. The lack of suitable anti-Cdk6 antibodies has precluded us from studying Cdk6 expression in other endocrine cell types that are not affected by lack of Cdk4 expression. Analysis of the expression of Cdk4 regulators in pancreatic islets has also revealed that these cells express cyclin D1 and D2, but not cyclin D3. As expected, neither cyclin D1 or D2 null mice appear to have defects in cells, since both Cdk regulatory subunits are likely to play compensatory roles (Fantl et al., 1995; Sicinski et al., 1995). Whether double mutant mice lacking cyclin D1 and D2 display proliferative defects in pancreatic cells remains to be examined. Finally, mutations in negative regulators of Cdk4 activity also lead to abnormal -cell proliferation, albeit with limited penetrance, again most likely because of their respective functional redundancy (Latres et al., 2000; Franklin et al., 2000).

In addition to impaired -cell proliferation, mice lacking Cdk4 are infertile (partial fertility in males) and about 40% smaller than their wild-type littermates (Rane et al., 1999; Tsutsui et al., 1999). Early studies involving ovarian transplantations or progesterone injection suggested that female infertility was because of deficits in the hypothalamus–pituitary axis that resulted in low levels of circulating progesterone (Rane et al., 1999; Moons et al., 2002a). More recently, Moons et al. (2002b) have shown that this defect is a consequence of reduced numbers of pituitary prolactin-producing lactotrophs, a result in agreement with those reported in this study. Restoration of normal fertility levels in Cdk4^n/n;+/Cre mice provides the definitive proof that infertility in Cdk4-defective mice is because of proliferative defects in the adenohypophysis.

The small size of Cdk4^n/n mice is due to a reduced number of cells rather than to a decrease in cell size. Indeed, cells of Cdk4^n/n mice appear to be about 10% larger than those of wild-type littermates, at least in several major organs such as spleen, liver and kidneys. However, the reason for the decreased homeostatic cell numbers of Cdk4 mutant mice remains to be resolved. Despite having increased numbers of pituitary cells, including somatotrophs, adult Cdk4^n/n;+/Cre mice are as small as Cdk4^n/n animals, indicating that this phenotype is not because of a defect in growth hormone production. We would therefore like to argue that the small size of these mutant mice is a cell autonomous defect. Ablation of the DmCdk4 locus in Drosophila results in flies that develop and eclose normally, but present a phenotype remarkably similar to that of Cdk4 null mice (Meyer et al., 2000). Mutant flies are smaller because of reduced cell numbers and females are mostly infertile. In addition, the differentiated adult wing cells of these mutant flies are slightly larger than their normal counterparts. As in mice, these observations do not rule out an indirect effect of loss of DmCdk4 on certain specific cell types whose dysfunction may affect the size of these mutant flies. However, the close similarity between the phenotypes of flies and mice defective for Cdk4 strongly support the concept that this kinase may play a role in controlling homeostatic cell numbers. Additional studies will be required to unveil the molecular mechanisms by which Cdk4 may control cell numbers in multicellular organisms such as flies and mice.

Materials and methods

Mouse strains and genotyping

Cdk4^n/n and Cdk4^R/R mice were maintained in a mixed 129/Sv CD1 background and genotyped by PCR as previously described (Rane et al., 1999). Transgenic Rip-Cre animals (Postic et al., 1999) were maintained in a C57BL/6 background. Genotyping of the Rip-Cre transgene was carried out by PCR using forward oligomer Rip-1 (5'-GGCCT CACCCTCTCTGAGACAAT-3') and reverse oligomer Cre 1054 (5'-ACCGTCAGTACGTGAGATATCTT-3').

Histology and immunohistochemistry

At least six sections from six animals (or three embryos) for each genotype and age were analysed. Error bars represent the standard deviation. Immunohistochemistry was carried out with the following antibodies and dilutions: anti-Cdk4 (C22, rabbit polyclonal, Santa Cruz Biotechnology) 1 : 600; anti-Ki67 (NLC-Ki67-MM1, mouse monoclonal, Novo Castra Laboratories) 1 : 50; anti-insulin (A564, guinea-pig polyclonal, Dako) 1 : 5000; anti-glucagon (A565, rabbit polyclonal, Dako or mouse monoclonal, Sigma), 1 : 2500 and 1 : 1000, respectively; antisomatostatin (A566, rabbit polyclonal, Dako) 1 : 500; antipancreatic polypeptide (rabbit polyclonal, Monosan) 1 : 100; rabbit polyclonal anti-ACTH, anti-FSH, anti-GH, anti-PRL, anti-LH and anti-TSH (National Hormone and Peptide Program, Harbor-UCLA Medical Center) 1 : 20.000.

Islet and cell size determination

Islet size (cells/islet section) and -cell size were measured in at least three randomly cut anti-insulin-stained sections from same pancreatic regions of four to six animals per genotype at each stage of development. Sections were more than 500 m apart from each other to avoid overestimating larger islets in these analysis. Islet and total pancreas areas were measured by morphometry using Q-win software (Leica). Islet cell mass was calculated by multiplying the proportion of pancreas area occupied by islets (area of islet cells/area of pancreas) by the average total pancreas weight measured for each genotype at each stage of development. Cell size was calculated by counting the number of nuclei per field of microscope view. Cellularity of the pituitary was likewise determined in at least three sections from six animals of each genotype at each stage of development.

BrdU incorporation

P10 mice were injected intraperitoneally with 3 one hourly doses of 10 mM BrdU in PBS (20 l per gram of body weight) starting 4 h before being sacrificed. BrdU incorporation was detected in fixed tissue sections with anti-BrdU monoclonal antibody (Amersham).

Quantitative real time RT–PCR

Animals (2–3 months old) were killed and the pancreas perfused with 2 ml of 1 g/ml collagenase P (Roche) in Hank's buffered saline solution (GibcoBRL) by injection through the bile duct. The pancreas was digested in 1 ml of 1 g/ml collagenase P solution at 37^oC for 5 min with constant agitation. Following digestion, islets and acinar cells were separated under a binocular microscope. Islets and acinar cells from three animals were resuspended in 100 l TRIzol reagent (GibcoBRL Life Technologies) and total cellular RNA extracted according to the recommended protocol. Reverse transcription was carried out in 20 l reactions using 2 g total RNA and 1 l MuLV reverse transcriptase (Roche). Quantitative PCR was carried out using the LightCycler (Roche) and the LightCycler-FastStart DNA Master SYBR Green I kit (Roche). Standard curves were constructed with a series of known concentrations of plasmid carrying the cDNA of each gene. Primer sequences were as follows: Cdk2 forward 5'-CGC CTC ACT AGC GCT C-3', reverse 5'-GGT ACA CCT TCA GTC TCA-3'; Cdk4 forward 5'-AGC CGA GCG TAA GAT CCC CT-3', reverse 5'-CAG CTG CTC CTC CAT TAG GA-3'; Cdk6 forward 5'-TAT TGA CGA ACT AGG CAA AG-3', reverse 5'-GGG GAC TCG CAG CCG C-3'; Cyclin D1 forward 5'-GGC AGC CCC AAC AAC TTC-3', reverse 5'-TCC CGC CTG CCC GGT GG-3'; Cyclin D2 forward 5'-AGG ATG ATG AAG TGA ACA CA-3', reverse 5'-AGA AGG GGC TAG CAG ATG A-3'; Cyclin D3 forward 5'-CGC CTG CTC TAT GTC TGC-3', reverse 5'-AGA TAT AGC ATG GAT TGT TCT-3'; p15INK4b forward 5'-TCC TGG AAG CCG GCG CAG AT-3', reverse 5'-TCA ATC TCC AGT GGC AGC GT-3'; p16INK4a forward 5'-CGA GGA AAG CGA ACT CGA-3', reverse 5'-AAT CTG CAC CGT AGT TGC G-3'; p18INK4c forward 5'-AGC CTG CAA TGT GGG G-3', reverse 5'-CTG CAG GCT TGT GGC T-3'; p19INK4d forward 5'-CTT AAC TGG GCT TGG GG-3', reverse 5'-TTG CTT CAG GAG CTC CAA A-3'; p21Cip1 forward 5'-GAG GCC CAG TAC TTC CTC TG-3', reverse 5'-AAG GCC GAA GAT GGG GAA GA-3'; p27Kip1 forward 5'-AGG CGG TGC CTT TAA TTG GG-3', reverse 5'-TTA CGT CTG GCG TCG AAG GC-3'.

References

1. Bonner-Weir S. (2000a). J. Mol. Endocrinol., 24, 297–302.
2. Bonner-Weir S. (2000b). Trend. Endocrinol. Metab., 11, 375–378.
3. Dyson N. (1998). Genes Dev., 12, 2245–2262. | PubMed | ISI | ChemPort |
4. Fantl V, Stamp G, Andrews A, Rosewell I and Dickson C. (1995). Genes Dev., 9, 2364–2372. | Article | PubMed | ISI | ChemPort |
5. Finegood DT, Scaglia L and Bonner-Weir S. (1995). Diabetes, 44, 249–256. | Article | PubMed | ISI | ChemPort |
6. Franklin DS, Godfrey VL, O'Brien DA, Deng C and Xiong Y. (2000). Mol. Cell. Biol., 20, 6147–6158. | Article | PubMed | ISI | ChemPort |
7. Gannon M, Shiota C, Postic C, Wright CVE and Magnuson M. (2000). Genesis, 26, 139–142. | Article | PubMed | ISI | ChemPort |
8. Herrera PL, Huarte J, Sanvito F, Meda P, Orci L and Vassalli J-D. (1991). Development, 113, 1257–1265. | PubMed | ISI | ChemPort |
9. Kim SK and Hebrok M. (2001). Genes Dev., 15, 111–127. | Article | PubMed | ISI | ChemPort |
10. Kulkarni RH, Brüning JC, Winnay JN, Postic C, Magnuson MA and Kahn CR. (1999). Cell, 96, 329–339. | Article | PubMed | ISI | ChemPort |
11. Latres E, Malumbres M, Sotillo R, Martín J, Ortega S, Martín-Caballero J, Flores JM, Cordón-Cardo C and Barbacid M. (2000). EMBO J., 19, 3496–3506. | Article | PubMed | ISI | ChemPort |
12. Lundberg AS and Weinberg RA. (1998). Mol. Cell. Biol., 18, 753–761. | PubMed | ISI | ChemPort |
13. Malumbres M and Barbacid M. (2001). Nat. Rev. Cancer, 1, 222–231. | Article | PubMed | ChemPort |
14. Meyer CA, Jacobs HW, Datar SA, Du W, Edgar BA and Lehner CF. (2000). EMBO J., 19, 4533–4542. | Article | PubMed | ISI | ChemPort |
15. Montanya E, Nacher V, Biarnes M and Soler J. (2000). Diabetes, 49, 1341–1346. | PubMed | ISI | ChemPort |
16. Moons DS, Jirawatnotai S, Tsutsui T, Franks R, Parlow AF, Hales DB, Gibori G, Fazleabas AT and Kiyokawa H. (2002a). Endocrinology, 143, 647–654.
17. Moons DS, Jirawatnotai S, Parlow AF, Gibori G, Kineman RD and Kiyokawa H. (2002b). Endocrinology, 143, 3001–3008. | Article | PubMed | ChemPort |
18. Morgan DO. (1995). Nature, 374, 131–134. | Article | PubMed | ISI | ChemPort |
19. Nevins JR. (1998). Cell Growth Differentiation, 9, 585–593.
20. Pardee AB. (1989). Science, 246, 603–608. | Article | PubMed | ISI | ChemPort |
21. Pavletich NP. (1999). J. Mol. Biol., 287, 821–828. | Article | PubMed | ISI | ChemPort |
22. Postic C, Shiota M, Niswender KD, Jetton TL, Chen Y, Moates JM, Shelton KD, Lindner J, Cherrington AD and Magnuson MA. (1999). J. Biol. Chem., 274, 305–315. | Article | PubMed | ISI | ChemPort |
23. Rane SG, Dubus P, Mettus RV, Galbreath EJ, Boden G, Reddy EP and Barbacid M. (1999). Nat. Genet., 22, 44–52. | Article | PubMed | ISI | ChemPort |
24. Reed SI. (1997). Cancer Survey, 29, 7–23.
25. Ruas M and Peters G. (1998). Biochem. Biophys. Acta, 1378, F115–F177. | PubMed | ChemPort |
26. Sherr CJ. (2000). Cancer Res., 60, 3689–3695. | PubMed | ISI | ChemPort |
27. Sherr CJ and Roberts JM. (1999). Genes Dev., 13, 1501–1512. | PubMed | ISI | ChemPort |
28. Sicinski P, Donaher JL, Parker SB, Li T, Fazeli A, Gardner H, Haslam SZ, Bronson RT, Elledge SJ and Weinberg RA. (1995). Cell, 82, 621–630. | Article | PubMed | ISI | ChemPort |
29. Sotillo R, Dubus P, Martin J, De la Cueva E, Ortega S, Malumbres M and Barbacid M. (2001a). EMBO J., 20, 6637–6647. | Article | PubMed | ISI | ChemPort |
30. Sotillo R, García JF, Ortega S, Martín J, Dubus P, Barbacid M and Malumbres M. (2001b). Proc. Natl. Acad. Sci. USA, 98, 13312–13317. | Article | PubMed | ChemPort |
31. Tsutsui T, Hesabi B, Moons DS, Pandolfi PP, Hansel KS, Koff A and Kiyokawa H. (1999). Mol. Cell. Biol., 19, 7011–7019. | PubMed | ISI | ChemPort |
32. Weinberg RA. (1995). Cell, 81, 323–330. | Article | PubMed | ISI | ChemPort |
33. Wolfel T, Hauer M, Schneider J, Serrano M, Wolfel C, Klehmann-Hieb E, De Plaen E, Hankeln T, Buschenfelde KH and Beach D. (1995). Science, 269, 1281–1284. | Article | PubMed | ISI | ChemPort |
34. Zuo L, Wedger J, Yang Q, Goldstein AM, Tucker MA, Walker GJ, Hayward N and Dracopoli NC. (1996). Nat. Genet., 12, 97–99. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We thank Ernesto de la Cueva, Maribel Muñoz, Marta San Román, Ignacio Segovia, Michelle Turmo and Raquel Villar for their excellent technical assistance. This work was supported by Grants from the Spanish Ministerio of Educación y Cultura (to SO and MB), Ligue contre le Cancer (Comité de Dordogne) and INSERM (to PD) and Fundación Pfizer (to MB). JM and RS were supported in part by fellowships from the Fondo de Investigación Sanitaria, SLH by a postdoctoral fellowship from the Comunidad Autónoma de Madrid and FN-P by a fellowship from the Foundation pour la Recherche Médicale.