Cdk4 Is Indispensable for Postnatal Proliferation of the Anterior Pituitary*

For proper development and tissue homeostasis, cell cycle progression is controlled by multilayered mechanisms. Recent studies using knock-out mice have shown that animals can develop relatively normally with deficiency for each of the G1/S-regulatory proteins, D-type and E-type cyclins, cyclin-dependent kinase 4 (Cdk4), and Cdk2. Although Cdk4-null mice show no embryonic lethality, they exhibit specific endocrine phenotypes, i.e. dwarfism, infertility, and diabetes. Here we have demonstrated that Cdk4 plays an essential non-redundant role in postnatal proliferation of the anterior pituitary. Pituitaries from wild-type and Cdk4-null embryos at embryonic day 17.5 are morphologically indistinguishable with similar numbers of cells expressing a proliferating marker, Ki67, and cells expressing a differentiation marker, growth hormone. In contrast, anterior pituitaries of Cdk4-null mice at postnatal 8 weeks are extremely hypoplastic with markedly decreased numbers of Ki67+ cells, suggesting impaired cell proliferation. Pituitary hyperplasia induced by transgenic expression of human growth hormone-releasing hormone (GHRH) is significantly diminished in the Cdk4+/– genetic background and completely abrogated in the Cdk4–/– background. Small interfering RNA (siRNA)-mediated knockdown of Cdk4 inhibits GHRH-induced proliferation of GH3 somato/lactotroph cells with restored expression of GHRH receptors. Cdk4 siRNA also inhibits estrogen-dependent cell proliferation in GH3 cells and closely related GH4 cells. In contrast, Cdk6 siRNA does not diminish proliferation of these cells. Furthermore, Cdk4 siRNA does not affect GHRH-induced proliferation of mouse embryonic fibroblasts or estrogen-dependent proliferation of mammary carcinoma MCF-7 cells. Taken together, Cdk4 is dispensable for prenatal development of the pituitary or proliferation of other non-endocrine tissues but indispensable specifically for postnatal proliferation of somato/lactotrophs.

Proliferation-stimulatory and -inhibitory signals control the G 1 phase of the cell cycle, which is critical for development and homeostasis of essentially all animal tissues (1). Cyclin-de-pendent kinases (Cdks) 1 activated by association with the regulatory cyclin subunits play central roles in promoting cell cycle progression, and Cdk4 is a critical regulator of the G 1 phase (2). In complex with D-type cyclins, Cdk4 phosphorylates G 1 -specific substrates, including the retinoblastoma protein (Rb). Unphosphorylated or hypophosphorylated forms of Rb participate in transcriptional repressor complexes with E2F-1-3. Rb phosphorylation in collaboration with cyclin D/Cdk4 (or closely related Cdk6) and cyclin E/Cdk2 results in release of Rb from the E2F complex, leading to transactivation of the E2F target genes important for the S-phase, e.g. cyclins E and A, dihydrofolate reductase and DNA polymerase-␣ (3). Furthermore, several Cdk inhibitor proteins, including p16 INK4a , p27 Kip1 , and p21 Cip1/Waf1 , play roles in coordinating temporal activation of Cdk4 and Cdk2 by D-type and E-type cyclins, respectively (4). The multilayered G 1 -Cdk regulation ensures proper control of proliferation, as well as timely transition between proliferation and quiescence, and should be crucial for development and function of every tissue.
Two groups, including ours, recently demonstrated that Cdk4-deficient mice survive embryogenesis, suggesting that Cdk4 is dispensable for embryonic development (5,6). Consistently, mouse embryonic fibroblasts (MEFs) prepared from Cdk4-null E12.5 embryos proliferate normally, with the exception of a modest delay in cell cycle entry from serum deprivation-induced quiescence. These data suggest that some other genes, possibly Cdk6 and Cdk2, may compensate for Cdk4 deficiency. However, studies on postnatal Cdk4-null mice (5, 6) unraveled unique phenotypes: 1) spontaneous degeneration of the pancreatic ␤ cells, first detectable at several weeks of age, resulting in insulin-dependent diabetes mellitus; 2) postnatal growth retardation; and 3) infertility in both females and males. Cdk4Ϫ/Ϫ male mice display severe testicular atrophy with a meiotic defect, whereas females exhibit normal ovulation and fertilization but fail implantation with complete sterility (7). Our data indicated that prolactin deficiency in Cdk4Ϫ/Ϫ female mice, associated with pituitary hypoplasia, plays a major role in the sterility phenotype (7,8). Thus, Cdk4 is required for homeostasis of particular endocrine tissues such as the pancreatic islet and pituitary even though Cdk4 is expressed ubiquitously.
Studies on other knock-out mice deficient in the G 1 control pathways also suggest that the pituitary gland is extremely sensitive to perturbation of G 1 regulation. Mice heterozygous for Rb deletion develop pituitary tumors originating from the melanotrophs in the intermediate lobe (9). Moreover, mice deficient for p27 Kip1 or p18 INK4c also develop melanotroph adenomas (reviewed in Ref. 10). Interestingly, double deficiency for p27 Kip1 and p18 INK4c or Rb results in more aggressive pituitary tumors, suggesting genetic cooperation. Thus, proper control of the Cdk/Rb-dependent G 1 regulatory pathway seems to be critical for homeostasis and tumor suppression in the pituitary gland, although the mechanism of this unique tissue specificity remains to be defined.
Here we have reported that Cdk4 is required specifically for hormonally regulated proliferation of somatotrophs and lactotrophs in postnatal pituitary glands, whereas Cdk4 is dispensable in embryonic development of the pituitary. Together with the requirement of Cdk4 for maintenance of postnatal pancreatic ␤ cells (11,12), the data suggest a unique nature of cell cycle control in particular endocrine tissues.
Cell Culture-GH3 and GH4 cell lines were a gift from Drs. Eun Jig Lee and J. Larry Jameson, Northwestern Medical School. MCF-7 cells were a gift from Dr. William T. Beck, University of Illinois at Chicago. GH3 and GH4 cell lines were cultured in 1:1 DMEM/Ham's F12 (Cambrex, Walkersville, MD) supplemented with 10% FBS (Sigma) and penicillin/streptomycin. MCF-7 cells and MEFs were cultured in DMEM, supplemented with 10% FBS, and penicillin/streptomycin.
Immunohistochemistry-Pituitary tissues were processed and sectioned with 5 m thickness according to a standard protocol. The sections were dewaxed in Safeclear II (Fisher Scientific, Pittsburgh, PA) and rehydrated in a series of decreasing concentrations of ethanol. Endogenous peroxidase activity was quenched by immersing slides in 0.3% H 2 O 2 in deionized H 2 O for 30 min. Slides were rinsed in phosphate-buffered saline, pH7.4, containing 0.01% Triton X-100 or boiled 10 min in 10 mM pH 6.0 sodium citrate for Ki67 staining. The sections were then treated for 30 min with 1.5% normal goat serum and incubated for 2 h with monkey anti-GH antibody (14) at a dilution of 1:100,000 or with mouse anti-Ki67 antibody (Vector Laboratories, Burlingame, CA) at 1:1,000. Slides were then rinsed in phosphate-buffered saline/Triton and incubated for 30 min with biotinylated secondary antibody (Vector Laboratories) at a dilution of 1:200, rinsed in phosphate-buffered saline, and incubated in avidin-peroxidase conjugate (Elite Vectastain; Vector Laboratories). The color reaction was developed using 3,3Ј-diaminobenzidine for 2 min. The sections were then subjected to counterstaining with Nuclear Fast Red (Vector Laboratories) or hematoxylin-eosin.
Small Interfering RNA (siRNA) and Cell Proliferation Assay-si-RNAs that specifically target mouse/rat Cdk4, human Cdk4, and rat Cdk6 mRNAs were designed according to the manufacturer's protocol (Dharmacon Research, Lafayette, CO). The sense sequences correspond to residues 801-819 of the coding region of the rat Cdk4 mRNA, which are identical to the murine counterpart, residues 632-650 of the human Cdk4 mRNA, and residues 96 -116 of the rat Cdk6 mRNA. At 24 h prior to siRNA transfection, cells were plated in 6-cm cell culture dishes that contain cover glasses (Fisher Scientific, Pittsburgh, PA) at the bottom. Cell lines were transfected twice, with a 24-h interval, with each of the specific siRNA or a pool of 4 nonspecific 21-mer double strand RNA control (Dharmacon) using Oligofectamine (Invitrogen). At 24 h posttransfection, bromodeoxyuridine (BrdUrd) was added to the culture medium at a concentration of 5 g/ml, followed by a 60-min incubation. BrdUrd incorporation was visualized by immunofluorescence using mouse anti-BrdUrd antibody (Roche Applied Science) and fluorescein isothiocyanate-conjugated anti-mouse antibody (Vector Laboratories). Percentages of BrdUrd ϩ cells were quantified in random fields (Ͼ300 total cells) of at least three different slides/treatment and subjected to statistical analyses.
Adenovirus-mediated Expression of the Human GHRH Receptor-Generation and characterization of adenovirus encoding human GHRH receptor (Ad GHRH-R) were described previously (15). GH3 cells were plated in a 10-cm culture dish and starved in minimal medium (DMEM/ F12 medium containing 10% charcoal/dextran-treated FBS (HyClone, Logan, UT)). 96 h later, cells were infected with the adenovirus at a dosage of 10 plaque-forming units/cell. Cells were incubated for 48 h for the expression of hGHRH-R before being replated in 6-cm culture dishes in the medium supplemented with or without GHRH (Sigma). Cdk4 siRNA knockdown and cell proliferation assay were performed 24 h later. MEF cells were cultured in DMEM ϩ 10% FBS and infected with Ad GHRH-R. Cell numbers were determined 4 days after the treatments of 50 nM GHRH.
Apoptosis Assay-Apoptotic cells were detected in paraffin-embedded sections by the terminal deoxynucleotidyltransferase-mediated dUTP nick-end labeling (TUNEL) assay using the Apoptosis Detection System, Fluorescein (Promega, Madison, WI) according to the manufacturer's protocol.
Statistics-Data are described as means ϩ S.E. The differences between groups were evaluated with Student's t test and analysis of variance with a significance level of p Ͻ 0.05.

Cdk4 Is Dispensable for Prenatal Pituitary Development and
Somatotroph Differentiation-Somatotrophs and lactotrophs of the pituitary in adult Cdk4Ϫ/Ϫ mice are severely hypoplastic, whereas gonadotrophs remain undisturbed (8). However, small numbers of growth hormone (GH)-and prolactin-immunoreactive cells are present in Cdk4Ϫ/Ϫ pituitaries, suggesting that cell fate determination to the somato/lactotroph lineages is still functional in the absence of Cdk4. To examine whether Cdk4 deficiency affects embryonic development of the pituitary, we examined the morphology of pituitaries from Cdk4Ϫ/Ϫ and Cdk4ϩ/ϩ embryos at day 17.5 post coitus (E17.5). We chose the developmental stage in which the lineage commitment of all anterior pituicytes is completed (16). However, at E17.5 the somatotroph population has yet to be fully expanded because the expression of the somatotroph-specific mitogen, growth hormone-releasing hormone (GHRH), is not increased until a later stage of development (17,18). Interestingly, transverse sections of heads demonstrated that pituitaries of Cdk4Ϫ/Ϫ embryos were morphologically indistinguishable from those of wild-type embryos (Fig. 1, A-F). Immunohistochemistry showed that GH expression was observed in similar numbers of cells in E17.5 Cdk4ϩ/ϩ and Cdk4Ϫ/Ϫ anterior pituitaries, indicating that differentiation of the somatotroph lineage takes place normally at this developmental stage (Fig. 1, A-F). In addition, proliferation in E17.5 Cdk4Ϫ/Ϫ anterior pituitaries appeared intact. Immunohistochemistry for a proliferation marker, Ki67 (19), showed similar degrees of proliferation in both E17.5 Cdk4ϩ/ϩ and Cdk4Ϫ/Ϫ anterior pituitary (Fig. 1, G  and H). Especially at the area of residual Rathke's pouch, similar numbers of cells displayed Ki67 immunoreactivity in wild-type and mutant pituitaries. These data indicate that Cdk4 is dispensable for pituitary morphogenesis up to E17.5 and for the commitment to the somatotroph lineage.
Pituitary Cells in Cdk4Ϫ/Ϫ Mice Are Insensitive to Mitogenic Signals by a GHRH Transgene-The complex endocrine network involving the hypothalamus and target organs regulates the homeostasis of the postnatal pituitary (20). Hypoplastic changes of somatotrophs and lactotrophs observed in Cdk4Ϫ/Ϫ pituitaries could result from perturbation of mitogenic stimuli from the endocrine network, insensitivity of pituicytes to mitogenic signals, or both. To assess the mechanism, we examined effects of ectopic expression of GHRH upon pituitaries of Cdk4Ϫ/Ϫ mice. Cdk4-null mice were cross-bred with human GHRH transgenic mice under the control of the metallothi-onein-I promoter (MT-hGHRH) (13); MT-hGHRH mice in Cdk4ϩ/ϩ, Cdk4 ϩ/Ϫ, and Cdk4Ϫ/Ϫ genetic backgrounds were generated. A previous immunohistochemical study (21) demonstrated that the MT promoter expresses hGHRH in most tissues, including the pituitary, pancreas, kidney, duodenum, lung, gonads, adrenal, heart, and brain. The pituitaries from 8-week-old Cdk4Ϫ/Ϫ females were extremely small, weighing about 30 -40% of Cdk4ϩ/ϩ pituitary weights (Fig. 2, A and B) as demonstrated previously (8). The MT-hGHRH transgene induced a significant increase in pituitary sizes and weights in Cdk4ϩ/ϩ mice. Pituitaries of Cdk4ϩ/Ϫ mice expressing the transgene displayed diminished hyperplastic responses, whereas Cdk4Ϫ/Ϫ;MT-hGHRH mice failed to show any significant response (Fig. 2, A and B). Histological analyses of Cdk4ϩ/ϩ;MT-GHRH mice revealed that major hyperplastic changes were within the anterior lobe, displaying increased numbers of GH-immunoreactive cells (Fig. 2, C and D). Counting monodispersed pituitary cells (Fig. 2E) showed that cell numbers of the somatotroph (GH ϩ ), lactotroph (PRL ϩ ), and thyrotroph (TSH ϩ ) in Cdk4ϩ/ϩ;MT-hGHRH mice were increased by more than 2-fold compared with wild-type mice without the transgene (Fig. 2E, Cdk4ϩ/ϩ). The MT-hGHRH transgene did not significantly increase numbers of lactotrophs, somatotrophs, or thyrotrophs in Cdk4Ϫ/Ϫ mice. To further address the mechanism of negative impact of Cdk4 deficiency on GHRH-induced hyperplasia, we examined proliferation and apoptosis in transgenic pituitaries by Ki67 immunohistochemistry and TUNEL assay, respectively. Pituitaries of MT-hGHRH;Cdk4ϩ/ϩ mice at 8 weeks of age showed more proliferating (Ki67 ϩ ) cells than those of Cdk4ϩ/ϩ mice without the transgene (Fig. 3, A and B). In contrast, pituitaries of Cdk4Ϫ/Ϫ mice exhibited minimum numbers of Ki67 ϩ cells regardless of the absence or presence of the transgene. Pituitaries of Cdk4ϩ/ϩ and Cdk4Ϫ/Ϫ mice, with or without the MT-GHRH transgene, did not show any difference in apoptosis according to the TUNEL assay (data not shown). These data suggest that Cdk4 is essential and rate limiting for GHRHinduced proliferation in the postnatal pituitary.
GHRH-induced Proliferation of Cultured Somato/Lactotroph Cells Depends on Cdk4 -The MT-GHRH transgenic study suggests that impaired somatotroph proliferation in Cdk4Ϫ/Ϫ mice is related to insensitivity of somatotrophs to the specific mitogen, GHRH. To further examine the role of Cdk4 in proliferation of the somatotroph lineage in an environment without a complex endocrine network, we used rat somato/lactotroph GH3 cells. This immortalized cell line exhibits characteristics of both somatotroph and lactotroph and expresses detectable levels of GH and PRL, whereas it shows minimum expression of GHRH receptors (22,23). Thus, to study GHRH-mediated proliferation in vitro, we restored the expression of GHRH receptors in GH3 cells by Ad GHRH-R (15). Infection with Ad GHRH-R sensitized GH3 cells to the GHRH treatment and increased numbers of S-phase cells by 300% (Fig. 4C). To assess whether GHRH-induced proliferation requires Cdk4, Cdk4 was knocked down using siRNA. Cdk4 siRNA reduced levels of endogenous Cdk4 protein by about 80% (Fig. 4A) and decreased the numbers of GHRH-treated cells in the S-phase by 60% (Fig. 4, B and C). It should be noted that Cdk4 siRNA also decreased proliferation of cells cultured in the minimum medium without GHRH to a lesser extent than that of GHRH-treated GH3 cells (Fig. 4C). These data suggest GHRH-induced proliferation of somato/lactotroph cells requires Cdk4. Because we previously reported that Cdk4Ϫ/Ϫ MEFs proliferate normally under optimal growth-promoting conditions (5), we attempted to clarify whether Cdk4 is essential for the GHRH-mediated proliferation of MEFs. Upon ectopic expression of GHRH receptors, fibroblasts are capable of proliferating in response to GHRH (24). Using Ad GHRH-R, we expressed GHRH receptors in MEFs prepared from Cdk4Ϫ/Ϫ and Cdk4ϩ/ϩ E12.5 embryos and then cultured cells without or with GHRH (50 nM) for 4 days. GHRH treatment significantly increased cell numbers in Ad GHRH-R-infected MEF cultures regardless of the Cdk4 status (Fig. 4D). These observations indicate that, unlike GH3 cells, MEFs do not depend on Cdk4 for GHRH-mediated proliferation.
Cdk4 Is Required for Estrogen-mediated Proliferation of Pituitary Cell Lines-Estrogen (E2) is a potent mitogen that stimulates lactotroph proliferation, mediating physiological lactotroph hyperplasia during pregnancy. We previously demonstrated that E2 administrations into mature Cdk4Ϫ/Ϫ female mice failed to increase numbers of lactotrophs, whereas the same treatment effectively induced lactotroph hyperplasia in Cdk4ϩ/ϩ females (8). Thus, similar to GHRH-induced somatotroph hyperplasia, E2-induced lactotroph hyperplasia may require Cdk4. To elucidate whether estrogen signaling in lactotrophs requires intrinsic Cdk4 for proliferative response, we performed siRNA-mediated knockdown of Cdk4 in E2treated GH3 cells. We also examined closely related GH4 somato/lactotroph cells (25) and human mammary carcinoma MCF-7 cells as a non-endocrine cell type responsive to estrogen. All of these cell lines depend largely on E2 for proliferation, and culture under E2-deprived conditions using charcoaltreated serum markedly inhibits proliferation (Fig. 5A). We also examined expression of G 1 -Cdks and D-type cyclins in E2-stimulated GH3 and MCF-7 cells (Fig. 5B). GH3 cells exhibited high levels of Cdk4, Cdk6, and Cdk2, whereas Cdk6 was undetectable in MCF-7 cells. Cyclin D1 was up-regulated by E2 in GH3 cells. Cyclin D3 levels were not altered by E2 in GH3 and MCF-7 cells, whereas cyclin D2 was barely detectable. GH4 cells showed similar expression patterns of these G 1 -regulatory proteins as GH3 cells (data not shown). Transfection with Cdk4 siRNA effectively decreased Cdk4 levels and brought down numbers of GH3 and GH4 cells in the S-phase (BrdUrd ϩ ) by 40 and 75%, respectively (Fig. 6A). We also found that the Cdk4 siRNA did not affect continuous proliferation of MEFs (Fig. 6A) despite Ͼ80% reduction in Cdk4 protein levels (data not shown). Although human Cdk4-specific siRNA decreased Cdk4 protein levels by 90% in MCF-7 cells, it did not alter BrdUrd incorporation under E2-induced or -deprived conditions (Fig. 6B). These data suggest that Cdk4 is dispensable for E2-responsive proliferation of MCF-7 cells. Consistently, stable silencing of Cdk4 using retrovirus-based Cdk4 siRNA did not impair proliferation of MCF-7 (data not shown). Moreover, we observed normal in vivo proliferation of mammary and uterine epithelial cells of Cdk4-null mice in response to estro- gen administration. 2 Therefore, Cdk4 is required specifically for postnatal proliferation of the somato/lactotroph in response to hormonal signals in a manner unique to these cell types. As GH3 and GH4 cells express high levels of Cdk6, we examined effects of siRNA-mediated knockdown of Cdk6 in these cells. Although Cdk6 siRNA specifically suppressed Cdk6 expression by 70 -80% (Fig. 6C), E2-dependent proliferation of GH3 and GH4 cells was not affected (Fig. 6D). These results suggest that, unlike Cdk4, Cdk6 does not play a rate limiting role for proliferation of somato/lactotroph cells. DISCUSSION The central machinery of the cell cycle is highly conserved in all eukaryotic cells. Indeed, mammalian Cdk1 (Cdc2) and Cdk2 can complement the fission yeast cdc2 mutant and budding yeast Cdc28 mutant (26,27). In contrast to the single Cdk protein functioning in yeast cells, multiple Cdk proteins appear to collaborate in the mediation of cell cycle progression in mammalian cells. The observations that Cdk4-null mice are viable and develop almost normally except for particular endocrine organs (5,6,8) imply that presumably other Cdks compensate for the absence of Cdk4. Cdk2-null mice and Cdk6-null mice are also viable (28 -30), as are mice deficient in cyclin D1, D2, D3, E1, or E2 (31)(32)(33)(34)(35). These data further support the notion that the mammalian Cdks and cyclins play overlapping roles in G 1 cell cycle progression. However, expression patterns of multiple Cdk proteins do not simply explain the unique dependence of pituitary cells on Cdk4. GH3 somato/lactotroph cells express relatively high levels of Cdk4, Cdk6, and Cdk2 (Fig. 5B), whereas the cells depend on Cdk4 for GHRH-or estrogen-induced proliferation. In contrast, Cdk6 is undetectable in MCF-7 breast carcinoma cells, which could proliferate normally with minimum levels of Cdk4. Moreover, Cdk2-null mice exhibit neither pituitary hypoplasia nor diabetes. 3 Although details of collaborations between different Cdk proteins remain to be clarified, characterizations of Cdk4-null mice revealed that Cdk4 plays unique indispensable roles in postnatal homeostasis of particular endocrine tissues, i.e. pituitary somato/lactotrophs and pancreatic ␤ cells.
During embryogenesis, precursor cells for the anterior pituitary lineages undergo dynamic regulation of proliferation and differentiation (36). Precursor cells committed to the anterior pituitary actively proliferate in the Rathke's pouch, which is controlled by the transcription factor Pitx2 with influences of growth factors, including BMP2 and FGF8. Subsequent regulation by lineage-specific transcription factors, such as Prop-1, Pit-1, and GATA-2, govern the cellular commitment process to each pituitary lineage. Apparently normal development of the embryonic pituitary in Cdk4-null mice indicates that Cdk4 is dispensable for proliferation associated with this series of developmental events. After terminal differentiation of each pituitary lineage, a variety of extracellular signals regulate homeostasis of the tissue. The pituitary gland responds to both hypothalamic and peripheral signals by undergoing reversible and adaptive changes in proliferation. Multiple lines of evidence suggest that differentiated pituitary cells with expression of specific hormone markers can continue mitosis. For example, adult pituitary glands slowly undergo hyperplastic changes during pregnancy, associated with estrogen-mediated proliferation of lactotrophs. It has been described that about 30% of rat pituitary cells arise from "adaptive" proliferation of terminally differentiated cells, whereas others are originated from undifferentiated precursors or stem cells (20). The proliferation defects in somatotrophs and lactotrophs of postnatal Cdk4-null mice suggest that proliferation of adult pituitary cells, possibly by mitosis of differentiated cells, depends on Cdk4 in a non-redundant manner. Similarly, the proliferation defect of Cdk4-null pancreatic ␤ cells is restricted to postnatal animals (11,12). Proliferation of terminally differentiated ␤ cells has recently been clearly demonstrated (37,38). This unique requirement of the endocrine tissues for Cdk4 may imply that slow adaptive cell cycle progression is controlled by a mechanism distinct from the rapid cell cycle of regenerative tissues, e.g. the skin and digestive tracts. Possible fundamental differences in cell cycle control between differentiated endocrine cells and non-endocrine epithelial cells might be consistent with clinical data that tumors that arise from the pituitary or pancreatic islet rarely exhibit malignant phenotypes, whereas carcinomas can frequently develop from the regenerative tissues. Although the Rb/E2F pathway had been long thought to be the only target of Cdk4, a recent report described that Cdk4 phosphorylates and regulates Smad3, a transcription factor mediating the transforming growth factor ␤/activin signaling (39). Thus, it will be important to determine whether Cdk4 regulates proliferation and function of postnatal endocrine tissues in a manner dependent on Rb or on other undefined substrates specific to the cell lineages. Studies on the genetically engineered mice with genomics and proteomics should provide us with further insight into tissue-specific control of the mammalian cell cycle.