PII: S0165-3806(98)00032-7

Developmental Brain Research 108 1998 7787.Research reportRegulation of cyclin dependent kinase inhibitor proteins during neonatal cerebella developmentGenichi Watanabe a, Pilar Pena a, George E. ShambaughIII b, George K. HainesIII c,Richard G. Pestell a,)aThe Albert Einstein Cancer Center, Department of Medicine and Department of Deelopmental and Molecular Biology, Albert Einstein College ofMedicine, Chanin 302, 1300 Morris Park Ae., Bronx, NY 10461, USAbNorthwestern Uniersity Medical School, Veterans Affairs Lakeside Medical Center, 303 E. Chicago Ae., Chicago, IL 60611, USAcDepartment of Pathology, Northwestern Uniersity Medical School, 303 E. Chicago Ae., Chicago, IL 60611, USAAccepted 3 February 1998

AbstractThe cyclin dependent kinase holoenzymes CDKs , composed of catalytic cdk and regulatory cyclin subunits, promote cellular . . . proliferation and are inhibited by cyclin dependent kinase inhibitor proteins CDKIs . The CDKIs include the Ink4 family. p15Ink4b, p16Ink4a, p18Ink4c, p19Ink4d. and the KIP family p21Cip1 and p27Kip1.. The sustained induction of p21 and p18 during myogenesis implicates these CDKI in maintaining cellular differentiation. Herein we examined the CDK cyclin D1, cdk5 and CDKI expression . profiles during the first 24 days of postnatal rat cerebella development. Cdk5 abundance increased and cyclin D1 decreased from day 9 through to adulthood. The CDKIs increased transiently during differentiation. p27 increased 20-fold between days 4 and 24, whereas p21 rose twofold between 6 to 11 days. p19, p18 and p16 increased approximately two- to threefold, falling to low levels in the adult. Immunostaining of cyclin D1 was localized in the external granular cells, whereas p27, was found primarily in the Purkinje cells. The period of maximal differentiation between days 9 to 13 was associated with a change in p21 and p16 staining from the external granular and Purkinje cells to a primarily Purkinje cell distribution. Protein-calorie malnutrition, which was previously shown to arrest rat cerebella development, reduced cyclin D1 kinase activity and p27 levels. However, p16 and p21 levels were unchanged. We conclude that the CDKIs are induced with distinct kinetics in specific cell types and respond differentially to growth factors during cerebella development, suggesting discrete roles for these proteins in normal cerebella development. q1998 Elsevier Science B.V. All rights reserved.

84G. Watanabe et al.rDeelopmental Brain Research 108 1998 7787()G. Watanabe et al.rDeelopmental Brain Research 108 1998 7787()85

Keywords: Cell-cycle; Cyclin dependent kinase inhibitor protein; Cerebella

IntroductionThe family of cyclin dependent kinases CDKs and the . associated cyclin dependent kinase inhibitor CDKI pro- . teins play a critical role in cellular proliferation and differentiation 61 . The CDK holoenzymes are composed of aw x catalytic and regulatory subunit which bind and phosphorylate target substrates including the retinoblastoma protein pRB . The cyclin D1 gene encodes a labile mitogen-in-. ducible regulatory subunit of the cdk holoenzyme required for normal progression through the early G1 phase of the cell-cycle 44,46,61 . Cyclin D1 is expressed primarily inw x

) Correspondingauthor.Fax:q1-718-430-8674;E-mail:pestell@aecom.yu.edu0165-3806r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved.the developing brain, the sympathetic nervous system and in the external granular layer of the developing cerebellum which gives rise to neurons within the molecular layer w41,47,53 . Transgenic mice with homozygous deletion ofx the cyclin D1 gene develop symptoms of neurological impairment 47 and fail to develop normal retina 15,47 .w xw x The retina is reduced in thickness with hypoplasia of the ganglion cells, nuclear and plexiform layers 15 , suggest-w x ing cyclin D1 is required for normal cell number and neural development 47 . Cdk5 was previously shown tow x associate with cyclin D1 in WI 38 cells w63x and cdk5 prepared from brain exhibits pRB kinase activity w55 .x Cdk5 abundance and activity are increased during neuronal differentiation and in postmitotic neurons 25,26,56 . cdk5w x activity is required for neurite outgrowth in cultured cortical neurons 32 , and heterozygous deletion of the cdk5w x gene w34x severely disrupted the cytoarchitecture of the cerebellum and cerebrum. Together these findings indicate that cdk5 and cyclin D1 can play an important role in obtaining the differentiated neural phenotype.The activity of the cyclin dependent kinases cdks. is modulated in vitro by two families of inhibitor proteins. The Ink4 family, which contains a highly conserved motif of ankyrin rings, includes p16Ink4a, p15Ink4b, p18Ink4c, and p19Ink4d. The Ink4 proteins are specific inhibitors of Cdk4 and Cdk6 activity w20,40 . The second family, the Cipx proteins, includes p27Kip1 and p21Cip1. p27 and p21 are inhibitors of Cdk2 and Cdk4 in vitro and bind with high affinity to cyclin D1 19,35,49,54 . The Cip proteins alsow x function as assembly factors and regulate subcellular compartmentalization of the cdks suggesting complex functions of these proteins 24,39 . The abundance of both thew x Ink4 and Cip proteins are induced during paradigms of cellular differentiation. During oligodendrocyte differentiation p27 levels increase and recent studies of mice homozygously deleted of the p27 gene have suggested an important role for p27 as an inhibitor of neural cellular proliferation 9 . p21 was induced during differentiation ofw x myocytes w19,35,49 ,x hepatic and hematopoietic cells w23,51 and neural cell lines 38,64 . In myelomonocyticx w x cells, differentiation induced by 1,25-dihydroxyvitamin D3 was associated with an induction of p21 and overexpression of p21 induced a component of the differentiated phenotype 28 . During myocyte differentiation p18 levelsw x were increased and p19 levels were decreased 16 . Cyclinw x D1 levels were decreased during myocyte differentiation, and overexpression of cyclin D1 antagonized components of the myocyte differentiation process 49 . These studiesw x suggest that differentiation occurring in cultured cells is the result of a dynamic interplay between the abundance of the cdk holoenzymes and the CDKIs.The neonatal cerebellum represents an ideal in vivo model to study the role of the CDKs and CDKIs in normal neural development. Two main characteristics of this in vivo system contribute to the utility of this model in understanding the mechanisms of neural development. Firstly, the cerebellum undergoes highly reproducible spatial and temporal profiles of neural differentiation. Secondly, development of the cerebellum can be arrested in a reproducible manner using maternal starvation; thus, the mechanism of starvation induced developmental arrest can be analyzed in detail. The developing cerebellum contains both proliferating cells and differentiating cells which occupy specific topographic positions 36,17 . During nor-w x mal postnatal cerebellar development the cerebellar cortex is composed of the external granular layer, molecular, Purkinje cell and internal granular layers w36,17 . Thex external granular layer is composed of proliferative and pre-migratory zones. Proliferation in the external granular layer continues up to day 9 and then gradually decreases up to day 21. Upon the completion of mitosis, the external granular cell descends to the pre-migratory zone. The cell soma then undergoes dramatic elongation perpendicular to the previous cellular axis to pass through the molecular layer and thereby reach the internal granular layer, whereupon it extends axonal processes back into the molecular layer. Maternal starvation is capable of delaying andror arresting cerebellar development resulting in distinctive changes in the cerebellar cellular topography w8,30,31,37,52 . Maternal protein restriction in the rat durxing late gestation and the first 2 weeks of postnatal life resulted in a 40% decrease in cerebellar DNA and protein levels and a reduction in molecular layer thickness 42 .w xPrevious studies have documented the induction of specific CDKIs during differentiation in cultured cells. The aim of these studies was firstly to systematically assess the expression profile of the known members of the CDKI family in vivo. It was hypothesized that the sustained induction of the CDKIs would be required for the maintenance of the differentiated neural phenotype. These studies also aimed to utilize the distinctive topographic changes that occurred during postnatal cerebellar development to identify the cell types expressing the CDKIs. It was hypothesized that the induction of the CDKIs in the proliferating external granular cells, expressing the CDKs, may be responsible for the cessation of cellular proliferation and subsequent differentiation. Finally, these studies aimed to determine whether starvation, that induced developmental arrest, induced the CDKI, thereby directly blocking proliferation in the affected cells.Materials and methods2.1. Animals and tissue preparationAll dietary manipulations and procedures described were approved by the Animal Care Committee of the Veterans Administration Lakeside Medical Center. The details of nutritional manipulation of the Female SpragueDawley rats weighing 150250 g Holtzman, Madison, WI. used in these studies were previously described 41 . Virus-freew x female SpragueDawley rats weighing 150250 g Holtz- man, Madison, WI were housed at 21. "18C with alternating lightdark cycles of 12 h each. Animals were received on gestational day 17 delivery day 21 r22.. The standard diet PROLAB-3000 Agway Country Foods, Syracuse, NY , contained 22% protein. The control diet,.a960260, ICN Biochemicals, Cleveland, OH , was a 20% protein. pellet diet and the test experimental diet, a960254., contained 4% protein supplemented with additional carbohydrates to make equal weights of the pellets isocaloric with the test control.Pregnant rats were randomly assigned on gestational day 21 to either the experimental group, assigned a 4% protein diet, or the controls which were given a 20% protein diet. For refeeding, 8-day-old animals from malnourished mothers were foster fed to a control mother.Non-pregnant females at 60 days of age and nursing mothers were anesthetized with carbon dioxide prior to sacrifice. Pups were weighed, then sacrificed on postnatal days 4 through 24 days as indicated in the figures. Animals were anesthetized with carbon dioxide prior to sacrifice and following decapitation, the brain including the cerebellum was removed and tissues placed on a block of dry ice. Following decapitation, the brain including the cerebellum was removed and tissues placed on a block of dry ice. Frozen tissues were stored at y708C until ready for sectioning. Sectioning of the lateral cerebellar hemispheres was performed in the coronal plane beginning at the caudal surface of the cerebellum at the vermis. Contiguous sections were stained using specific antibodies as described below. Tissues were sectioned in 8 mm thickness at y178C, and mounted on silan coated slides. Slides were stored at y808C until ready for immunohistochemical staining.2.2. Tissue cultureC2C12 myoblasts CRL 1772; American Type Culture Collection, Rockville, MD. were maintained in growth media Dulbeccos modified Eagles medium with 4500 mgrl glucose wDMEM-H , 15% fetal bovine serum andx antibiotics. Cells were grown on 150-mm tissue culture dishes at low density and induced to differentiate by switching medium to DMEM-H with 2% horse serum and 10 mgrml bovine insulin Life Technologies, Gaithersburg, MD ..2.3. Western blots and immune-complex assaysWestern blotting was performed as previously described w59x using antibodies to p21 C-19 , p21 187 , p27 M-197, . . C-19 , p16 C-20 -G, p16 M-156 and cyclin D1 HD-11 ,. . . . Santa Cruz, CA , HD-1 a gift from Dr. J. Koh and Dr. E.. Harlow. and horse radish peroxidase conjugated second antibody. In addition, the p35 antibody 5H8, a gift from Li-Huei Tsai and Dr. M. Nicolic , the p21 antibodies. Cp36 and CP68. a gift from Dr. B. Dynlacht and tubulin. monoclonal antibody 5H1 . w60x were used in Western blot analysis or immunoprecipitation kinase assays as previously described. The p18 antibody 11256 and p19 12077 . . antibodies 16 were a generous gift from Dr. Y. Xiong.w x Reactive proteins were visualized by the enhanced chemiluminescence system Amersham, Arlington Heights, IL. and quantitation was performed by densitometry using a Bio-Rad Molecular Analysis.Immunoprecipitation kinase assays were performed essentially as previously described w29,58,59x using either Histone H1 Sigma, St. Louis, MO. or pRB substrate prepared from the vector pGEX-Rb 14w x a gift from Dr. E. Harlow . Cells were suspended at 1. =106 to 5=106rml in immunoprecipitation RIPA buffer 150 mM NaCl, 1% . NP-40, 0.5% Deoxycholate, 0.1% SDS, 50 mM Tris pH7.5. with 0.1 mM phenylmethylsulfonylfluoride, 1 mg of leupeptin per ml, and 1 mM sodium orthovanadate Sigma , . at 48C. Lysates were centrifuged at 10,000=g for 5 min and the supernatants were precipitated for 4 h at 48C with protein A- sepharose beads precoated with saturating amounts of the cyclin D1 antibody, DCS-11 NeoMarker, Fremont, CA . Phosphorylated proteins were visualized. after exposure to KODAK XAR film and quantitation performed by densitometry Image Quant version 1.11, Molecular Dynamics Computing Densitometer, Sunnyvale, CA ..2.4. Immunohistochemical stainingImmunohistochemical analysis was performed with the cyclin D1 and the CDKI p27, p21 and p16. antibodies essentially as previously described w41x. The cyclin D1 antibody clone DSC-6 was purchased from Vector laboratories Novacastra Laboratories, Newcastle upon Tyne, UK and this antibody was used at a 1:10 dilution. The p16. M-156. and the p21 C-19. antibodies were purchased from Santa Cruz Biotechnology Santa Cruz, CA . The. CDKI antibodies were used at a 1:50 dilution. Immunohistochemical analysis was carried out using a biotinylated secondary antibody, and an avidinrbiotin-linked horseradish peroxidase Vectastain ABC system, Vector laboratories, Burlingame, CA . The complex was stained. with diaminobenzamidine tetrahydrochloride DAB. substrate obtained from Kirkegaard and Perry laboratories Gaithersburg, MD . The sections were wiped dry then. fixed in either acetone for 20 min at 48C, followed by rehydration in phosphate buffered saline PBS , or fixation . with formalinracetonermethanol 2r19r19 vrv. for 5 min, at 48C, followed by a PBS rinse, then in methanol for 10 min followed by rehydration in PBS 41 . Rehydratedw x samples were blocked with serum directed against the species producing the secondary antibody. Control sera included either normal rabbit serum or isotype matched monoclonal antibodies . Following a 15 min incubation at. 378C, the primary antibody was added and sections incubated for 24 h at 48C. Following incubation with the primary antibodies, slides were rinsed in PBS then treated with a biotinylated secondary antibody and held at 378C for 15 min, washed again in PBS, and incubated with horseradish peroxidase in an avidinbiotin matrix. Following a 15-min incubation with a biotinylated secondary antibody and horse radish peroxidase, the slides were immersed in DAB 100 mg in 200 ml 0.1 M phosphate buffer pH 7.3 to which was added 5 ml 1% cobalt chloride then 4.0 ml of 1% nickel ammonium sulfate 1 . Optimalw x incubation time was found to be 15 min. Then 0.66 ml 3% hydrogen peroxidase was added and slides were incubated at room temperature for 5 min. The slides were then placed in water for 2 min and stained for 3 min with neutral red or 1 min with light green. Slides were then placed in alcohol and dehydrated with isopropanol.Results3.1. Cdk5 and cyclin D1 leels are inersely related during cerebellar deelopmentPrevious studies had examined the expression of cdk5 mRNA in the developing embryonic brain and in the adult cerebellum w22,55 . Western blotting was performed tox examine the cdk5 protein abundance in the developing cerebellum. Cdk5 protein was readily detectable at day 4 and cdk5 abundance increased between postnatal days 9 and 13 Fig. 1a,b . We examined the abundance of cyclin . D1 using a specific antibody HD-1 . Cyclin D1 protein. abundance by Western blotting increased transiently between days 4 and 9 postnatally, then decreased between days 9 and 11 and subsequently into adulthood Fig. 1a,c , . consistent with previous studies in which cyclin D1 mRNA levels began to decrease between days 9 and 11 41 .w x3.2. p27 and p21 are induced transiently during cerebellar deelopmentTo examine the expression profile of p27 and p21 in the developing cerebellum, Western blotting was performed. p27 levels increased 10-fold from day 4 to day 9 Fig. 2a . . The induction of p27 was sustained through day 24. In the

Fig. 1. Expression of cdk5 and cyclin D1 proteins during postnatal cerebellar development. The mean values for Western blot analysis of either b . cdk5 and .c cyclin D1 from cerebella or from differentiated C2C12 cells a . The adjacent figures show the raw data for the animals . at each time point. Fold induction is shown normalized to the abundance of cyclin D1 in the adult as 1.

Fig. 2. Expression of p27 and p21 during postnatal cerebellar development. Western blot analysis of a p27 and b p21 in the developing rat . . cerebellum. The data are also shown as normalized mean data for the separate animals at each time point.adult, p27 levels were reduced to levels approaching that seen at day 4 Fig. 2a . p21 abundance increased from. days 6 through 9 and then decreased to very low levels in the adult Fig. 2b . .3.3. p18 leels are relatiely constant during cerebellar deelopmentp19 and p18 were readily detected in the day 4 cerebellum. Between days 4 and 9, the abundance of p19 and p18 changed relatively little compared with the changes in p27 Fig. 2a observed during this time Fig. 3a,b . p19 levels. . increased twofold between days 4 and 9 and remained elevated until day 24, decreasing by 50% in the adult Fig. 3a . Experiments were also conducted in C2C12 myocytes. induced to differentiate by switching to differentiation medium Fig. 3a, inset . The abundance of p19 in cerebel- . lar samples was significantly less than that observed in equal amounts of C2C12 cell extracts Fig. 3a inset . p19 . levels decreased with the addition of differentiation medium in C2C12 cells Fig. 3a inset, and data not shown. con-Fig. 3. Expression of Ink4 proteins during postnatal cerebellar development. Western blot analysis of either a p19 b p18 from cerebella or c p16 levels . . . from rat cerebella. The adjacent figures show data for the mean of two to three separate animals. The inset of a shows a comparison of the relative . abundance of p19 from differentiated C2C12 cells and cerebella in which equal amounts of total protein were loaded. The C2C12 cells were induced to

differentiate as described in Section 2.firming recently published observations w16 . p18x levels were relatively constant from days 4 through 24 then decreased by 50% from days 24 to the adult levels Fig. 3b . These results contrast with the dramatic increases in. p18 levels that occurred during C2C12 myocyte differentiation 16 . p16 levels increased twofold between days 6w x and 9 and remained elevated until day 16, after which time levels decreased to barely detectable in the adult Fig. 3c . . p15 was detected at low levels and increased between days 13 and 16 data not shown . The peak induction of p15 at . days 13 to 16 was much later than found with the other Ink4 family members. Together these results indicate that the Ink4 proteins are induced at different times in the developing cerebellum.Fig. 4. Immunohistochemical localization of p27 in the developing cerebellum. Cyclin D1 a and p27 b immunostaining is shown in the rat cerebellum at . . 11 days. The external granular cells EG , the molecular layer M , and the Purkinje cells P are shown. The cyclin D1 immunopositive cells, indicated by. . . the arrow, are shown primarily in the external granular cell layer and the p27 immunopositive cells, which are yellowrbrown, are shown predominantly in

the Purkinje cells. The magnification is 400=.3.4. Immunolocalization of p27, p21 and p16 in the deeloping rat cerebellumImmunohistochemical studies were performed to identify the cell type in which the CDKIs were expressed. The layers of the cerebellum, external granular EG , molecular . .M , and the Purkinje P are shown Fig. 4 . In cerebella . . at 11 days, cyclin D1 staining was seen only in the external granular layer Fig. 4a . p27 staining at this time . of development was seen predominantly in the Purkinje cells with no staining apparent in the molecular or external granular layers Fig. 4b . The predominant distribution of . cyclin D1 and p27 are therefore in different cell types at the same time during cerebellar development. p21 and p16 staining were next examined in the developing cerebellum Fig. 5 . Comparison was made between days 9 and 13, as. considerable differentiation during this time results in increasing thickness of the molecular layer Fig. 5a vs. b . . p21 staining at 9 days was seen in scattered cells within the external granular layer, in occasional cells migrating across the molecular layer and in Purkinje cells Fig. 5a . . By 13 days, the predominant staining for p21 was local-

Fig. 5. Immunohistochemical analysis of p21 and p16 in the developing cerebellum. The cell types expressing p21 a, b and p16 c, d in the developing . . rat cerebellum are shown. Immunostaining is shown for days 9 a, c and 13 b, d . Positively staining cells are brown cells indicated by the arrow. p21 . . staining at 9 days a is found predominantly in the Purkinje cells P with some immunopositive cells in the external granular layer EG . At 13 days, p21 . . . staining is found in the Purkinje cells b . The p16 positive cells at day 9 are located in several different cell types, including the external granular cells . EG , the molecular layer M , the Purkinje cells P and the internal granular cells IG . At day 13, p16 staining is observed primarily in the Purkinje cells. . . . with infrequent staining in the internal granular layer. The magnification is 400=.

ized to the Purkinje cell layer Fig. 5b . The external. granular layer was essentially free of p21 staining, with occasional cells immunopositive for p21 found migrating across the molecular layer Fig. 5b . In the sections of. 9-day cerebella Fig. 5c , p16 immunostaining was found . in both the external granular and the Purkinje cell layer Fig. 5c . At 13 days, p16 staining was found only in the. Purkinje cells, with no remaining staining in the granular layer Fig. 5d . Together these studies indicate that in the . developing cerebellum, cyclin D1 levels decrease as expression of the CDKI p16 and p21 is induced and subsequently maintained in the differentiated cells.3.5. Reduced cyclin D1 kinase actiity in protein-calorie malnutrition associated with deelopment delayMalnutrition during critical periods of brain development has adverse effects on the developing brain and can

Fig. 6. The effect of protein calorie malnutrition on cyclin D1 kinase activity and cdki abundance in the developing cerebellum. a Cyclin D1 . immune kinase assays were performed on cerebella from fed 20% and . protein calorie malnourished 4% animals. The cyclin D1 immunoprecip- . itation was performed using the DCS-11 antibody and GST-pRB substrate as described in Section 2. The phosphorylated pRB band is shown for animals either fed 20% or starved 4% animals. In b d Western . . . . blotting was performed on cell extracts derived from the cerebella of fed or starved animals. The abundance of p27, p21, and p16 is shown graphically as mean data.result in lasting deficits 12 . In the neonatal rat, cerebellarw x growth increases more rapidly during the first two postnatal weeks than the remainder of the brain. Changes in cerebellar weight during malnutrition are relatively greater than in extracerebella regions of the brain w36,42x and malnutrition results in defective maturation of Purkinje cell circuitry 43 . In previous studies, nutritional deprivationw x markedly reduced cerebella mass, reduced cellular proliferation and resulted in a delay in neural maturation 7,41,48 .w x The effects of protein-calorie deprivation were manifested by a marked decrease in the width of the molecular layer, and thinning of the external granular layer 41 . To deter-w x mine whether nutritional deprivation-induced growth arrest was associated with changes in cyclin D1 kinase activity, immune-precipitation kinase assays were performed on the cerebella of either wild-type or nutritionally-deprived animals. Cyclin D1-immune kinase activity CD1K was de- . termined in the developing rat cerebellum from fed controls 20% . A direct comparison between the malnour.ished and control animals at day 11 demonstrated a 50% decrease in CD1K in the malnourished animals Fig. 6a . .CD1K activity is determined in part by the relative abundance of cyclin D1 and the CDKIs. Because CD1K activity was reduced with malnutrition we examined the abundance of the CDKIs to determine whether an induction of the CDKIs contributed to the reduction in CD1K activity during starvation. p27 levels were reduced to 30%, and p16 and p21 levels were unchanged Fig. 6bd . As . previous studies had shown a reduction in cyclin D1 abundance in starved animals at 11 days w41x together these studies suggest the reduction in CD1K activity during starvation may be due to the reduction in cyclin D1 abundance.DiscussionThe current studies were performed in order to characterize the temporal and spatial profile of the CDKIs in vivo during cerebellar development. It was hypothesized that the sustained induction of the CDKIs may be required for the maintenance of the differentiated phenotype. These studies demonstrated that the induction of differentiation but not the maintenance of the differentiated phenotype was associated with the expression of the CDKIs. Secondly, we had hypothesized that the induction of the CDKIs within the proliferating external granular cells may attenuate their proliferation resulting in cell-cycle arrest and subsequent differentiation. Although the p27 levels increased by Western blotting, the results of the immunostaining suggested that the induction of the CDKI, in particular p27, did not occur in the external granular cells, either within the proliferating external layer or the inner postmitotic layer. These results suggest that p27, although previously described as capable of inducing cellular arrest in other cell types, is not directly involved in this process of cellular proliferative arrest in the external granular cells of the cerebellum. The induction of p16 and p21 may contribute to the cellular arrest and the subsequent induction of differentiation. Finally, we hypothesized that maternal starvation may induce the CDKIs, and thereby contribute to the growth arrest observed in vivo. Although the phosphorylation of pRB, assessed through immune-precipitation kinase assays, decreased with starvation, the CDKIs were not induced. In conjunction with previous studies demonstrating a reduction in cyclin D1 protein levels upon maternal starvation 41 , these results suggest the growthw x arrest may involve a reduction in proliferation secondary to a fall in the abundance of cyclin D1.The relative abundance of cyclin D1 compared with the CDKIs, decreased during differentiation, is consistent with several other models of differentiation including cultured myocytes. Cyclin D1 protein abundance gradually decreased in the developing cerebellum consistent with previous studies showing a reduction in cyclin D1 mRNA levels during postnatal cerebellar development. In previous studies cyclin D1 was expressed in the proliferating external granular cells, but not in the molecular layer derived from the external granular cells 41 . Recent studies performedw x with cultured rat neural cells demonstrated a decrease in cyclin D1 levels as the oligodendrocyte precursor cells differentiated into non-dividing oligodendrocytes w13 .x Similar types of observation have been made during myocyte differentiation, wherein cyclin D1 mRNA levels decreased as myotubules form 19,35,49 . cdk5 levels in-w x creased gradually in the developing cerebellum, with increased levels maintained in the adult. cdk5 associates with cyclin D1 and p35 63 . When bound to cyclin D, cdk5w x associates with p27, p21 and p57 63 .w xIn our studies, p27 levels increased in postmitotic neurons, commensurate with increased cdk5 activity and reduced cyclin D1 levels. p27 was found in the Purkinje cells and some cells of the internal granular layer. In our studies, total cerebella p27 protein levels were induced 20-fold, concurrently with differentiation of the external granular cells. p27 and cyclin D1 did not co-localize in the proliferating external granular cells, however, suggesting p27 is not involved in proliferative arrest of these cells. Furthermore, p27 levels were reduced in the adult cerebellum suggesting that p27 function is not required for the maintenance of the terminally differentiated neural phenotype. These findings contrast with cultured myocytes in which the induction of p27 abundance was induced in a sustained manner suggesting a role for p27 in both the induction and maintenance of the differentiated phenotype w16,19 . We had hypothesized that p27 level may bex increased in the postmitotic layer of the external granular cells implicating p27 in the proliferative arrest of these cells because p27 can inhibit cyclin D1rcdk4 activity and cellular proliferation in vitro and in cultured fibroblasts w54x. Recent observations, however, suggest that p27 is present in active kinase complexes in proliferating cells w50x and can bind cyclin D1rcdk4 without inhibiting its kinase activity, functioning as an assembly factor 24,39 .w x Alternatively, the accumulation of p27 may occur as part of an intrinsic counting system that determines when a precursor cell stops proliferating and terminally differentiates 13 . p27 levels accumulated in proliferating oligoden-w x drocyte precursor cells and p27 was also present in differentiated oligodendrocytes w13x, thus the role of p27 in proliferative arrest may be cell-type specific.The induction of p21 in the developing cerebellum peaked at 9 days and then decreased in abundance. During myocyte differentiation the abundance of p21 was increased in a sustained manner, implicating this protein in maintenance of the differentiated phenotype w19,35,49 .x Homozygous disruption of p21 did not block normal myocyte differentiation, however, and p21 expression occurred after myoblasts began differentiating 10 . Althoughw x functional redundancy may contribute to the lack of phenotype in these animals, these results suggest p21 may not be important in the maintenance of the differentiated phenotype. In the present studies, p21 was found in the external granular cells and was also found in the molecular and internal granular cells. p21 was therefore found in cell types known to be proliferating and differentiating. The expression of p21 is known to occur in proliferating cells with maximal synthesis during G -phase and the nadir1 during S-phase 27,33 . In addition to binding and inhibit-w x ing G cyclin1 rcdk complexes, p21 also binds the processivity factor proliferating cell nuclear antigen PCNA. w19,35,49,57 . In previous studies, PCNA was also localxized to the proliferating external granular cells 41 and itw x will be of interest to determine whether p21 is bound to PCNA in these cells. The presence of p21 in the differentiated cerebellar cells may be the result of its function to induce the differentiated phenotype, as suggested in myocytes 19,35,49 . Alternatively, recent studies in the neu-w x roblastoma cell line, SH-SY5Y, demonstrated a neuronal survival function of p21, inhibiting programmed cell death w38 . p21x may convey a similar function in the developing cerebellum.p18 levels were relatively unchanged between days 4 to 24 and then decreased 50% in the adult. p19 increased between days 4 and 9 postnatally, and then decreased to low levels in the adult. These results contrast with differentiating cultured myocytes in which we observed increased p18 levels and decreased p19 levels continually upon differentiation as previously described 16 . As the magni-w x tude of induction during differentiation is greater than 50-fold, p18 is thought to function as the major CDK inhibitor involved in myogenic differentiation. In contrast, the abundance of the other CDKIs including p27 was only modestly altered w16x. Both p18 and p19 inhibit cyclin D-dependent kinases, and overexpression of p18 or p19 induces G -phase arrest 18,21 . Because cyclin D1 levels1 w x decreased from day 9, the relative abundance of p18 and p19 compared with cyclin D1 increased during the time of differentiation after day 9, consistent with a model in which the relative abundance of the CDKI compared with the CDK holoenzyme, increase in paradigms of differentiation. The role of additional cyclins such as p39, which plays a role in models of hippocampal cell differentiation w62x, remains to be investigated in cerebellar development.The molecular mechanisms by which protein calorie malnutrition arrests cerebellar development is complex. In these studies, starvation was associated with a 60% reduction in CD1K activity. The reduction in CD1K activity associated with malnutrition is consistent with several studies in which growth factors induced CD1K activity reviewed in Ref. 45 . A reduction in CD1K activity mayw x. be due to an increase in the CDKI abundance or a decrease in cyclin D1rcdk abundance. In the current studies, in the cerebellum of starved animals, p16 and p21 levels were unchanged, p27 levels decreased, and in previous studies, cyclin D1 immunopositivity was decreased w2,41 . Tox gether these results suggest the reduction in CD1K activity found in the starved animals cerebella may be due to reduced cyclin D1 protein levels rather than an induction in CDKI abundance. Insulin-like growth factor I IGF-I. plays an important role in the normal growth of the developing cerebellum, and malnutrition is associated with reduced bioavailable IGF-I. Cyclin D1 is induced by IGF-I w2,41 and has a short half life upon withdrawal of growthx factors 11 . The reduction in cyclin D1 protein levels inw x the cerebella of starved animals is consistent with a model in which cyclin D1 levels are induced in the developing cerebella by IGF-I.In summary, these studies demonstrate that postnatal cerebellar development is associated with an increase in cdk5 abundance and transiently enhanced expression of each of the CDKI examined. The temporal profile of expression of each of the CDKIs was distinct, consistent with different roles for each CDKI in promoting but not in maintaining the differentiated phenotype. These studies provide useful information about the temporal and spatial profile of expression of the CDKI proteins during cerebellar development and provide supportive evidence that the CDKI may play a role in the induction of differentiation of specific cell types.AcknowledgementsWe thank Dr. Li-Huei Tsai, Dr. M. Nicolic, Dr. B. Dynlacht and Dr. L. Binder for gifts of antibodies and cell lines, and Dr. A. Koch and Dr. E. Mugnaini for their expert advice. This work was supported in part by the NIH FIRST award 1R29CA70897 and grant R55CA75503 to R.G.P. and Merit Reviews from the Department of Veter-. ans Affairs G.E.S. . Work at the Albert Einstein College . of Medicine was also supported by Cancer Center Core National Institute of Health grant 5-P30-CA13330-26.References w x1 J.C. Adams, Heavy metal intensification of DAB-based HRP reaction product, J. Histochem. Cytochem. 29 1981 775. . w x2 C. Albanese, J. Johnson, G. Watanabe, N. Eklund, D. Vu, A. Arnold, R.G. Pestell, Transforming p21ras mutants and c-Ets-2 activate the cyclin D1 promoter through distinguishable regions, J. Biol.Chem. 270 1995 2358923597..w x3 J. Altman, Autoradiographic and histological studies of postnatal neurogenesis. Dating the time of production and onset of differentiation of cerebellar microneurons in rats, J. Comp. Neurol. 136 1969 . 269294. w x4 J. Altman, Postnatal development of the cerebellar cortex in the rat: I. The external granule layer and the transitional molecular layer, J.Comp. Neurol. 145 1972 353..w x5 J. Altman, Postnatal development of the cerebellar cortex in the rat: II. Phases in the maturation of Purkinje cells and of the molecular layer, J. Comp. Neurol. 145 1972 399. . w x6 J. Altman, Postnatal development of the cerebellar cortex in the rat: III. Maturation of the components of the granular layer, J. Comp.Neurol. 145 1972 465. . w x7 D. Barnes, J. Altman, Effects of different schedules of early undernutrition on the preweaning growth of the rat cerebellum, Exp.Neurol. 38 1973 406419..w x8 D. Barnes, J. Altman, Effects of two levels of gestational-lactational undernutrition on the postweaning growth of the rat cerebellum, Exp. Neurol. 38 1973 420428. .w x9 P. Casaccia-Bonnefil, R. Tikoo, H. Kiyokawa, V. Friedrich, M.V. Chao, A. Koff, Oligodendrocyte precursor differentiation is perturbed in the absence of the cyclin-dependent kinase inhibitor p27Kip1, Genes Dev. 11 1997 23352346. . w10x C. Deng, P. Zhang, J.W. Harper, S.J. Elledge, P. Leder, Mice lacking p21CIP1rWAF1 undergo normal development, but are defective in G checkpoint control, Cell 82 1995 675685.1 . w11x J.A. Diehl, F. Zindy, C.J. Sherr, Inhibition of cyclin D1 phosphorylation on threonine-286 prevents its rapid degradation via the ubiquitin-proteasome pathway, Genes Dev. 11 1997 957972. . w12x J. Dobbing, The later development of the brain and its vulnerability, in: J.A. Davis, J. Dobbing Eds. , Scientific Foundation of Paedi-. atrics, Univ. Park Press, Baltimore, 1981, pp. 744759.w13x B. Durand, F.-B. Gao, M. Raff, Accumulation of the cyclin-dependent kinase inhibitor p27rKIP1 and the timing of oligodendrocyte differentiation, EMBO J. 16 1997 306317. . w14x M.E. Ewen, H.K. Sluss, C.J. Sherr, D.M. Livingston, H. Matsushime, Functional interactions of the retinoblastoma protein with mammalian D-type cyclins, Cell 73 1993 487497. . w15x V. Fantl, G. Stamp, A. Andrews, I. Rosewell, C. Dickson, Mice lacking cyclin D1 are small and show defects in eye and mammary gland development, Genes Dev. 9 1995 23642372. . w16x D.S. Franklin, Y. Xiong, Induction of the p18INK4c and its predominant association with CDK4 and CDK6 during myogenic differentiation, Mol. Biol. Cell 7 1996 15871599. . w17x S. Fujita, Quantitative analysis of cell proliferation in the cortex of the postnatal mouse cerebellum, J. Cell. Biol., 1967 277288. .w18x K.L. Guan, C.W. Jenkins, Y. Li, M.A. Nichols, X. Wu, C.L. OKeefe, A.G. Matera, Y. Xiong, Growth suppression by p18, a p16INK4rMTS1- and p14INK4BrMTS2-related CDK inhibitor, correlates with wild-type pRb function, Genes Dev. 8 1994. 2939 2952. w19x O. Halevy, B.G. Novitch, D.B. Spicer, S.X. Skapek, J. Rhee, G.J. Hannon, Correlation of terminal cell cycle arrest of skeletal muscle with induction of p21 by MyoD, Science 267 1995 10181021. .w20x G.J. Hannon, D. Beach, p15INK4B is a potential effector of TGFbeta-induced cell cycle arrest, Nature 371 1994 257261. .w21x H. Hirai, M.F. Roussel, J.Y. Kato, R.A. Ashmun, C.J. Sherr, Novel INK4 proteins, p19 and p18, are specific inhibitors of the cyclin D-dependent kinases CDK4 and CDK6, Mol. Cell. Biol. 15 1995 . 26722681.w22x H. Ino, T. Ishizuka, T. Chiba, T. Chiba, M. Tatibana, Expression of CDK5 PSSALRE kinase , a neural cdc2-related protein kinase in. the mature and developing mouse central and peripheral nervous systems, Brain Res. 661 1994 196206. . w23x H. Jiang, J. Lin, Z.-Z. Su, F.R. Collart, E. Huberman, P.B. Fisher, Induction of differentiation in human promyleocytic HL-60 leukemia cells activates, p21, WAF1rCIP1, expression in the absence of p53, Oncogene 9 1994 33973406. .w24x J. LaBaer, M.D. Garret, L.F. Stevenson, J.M. Slingerland, C. Sandhu, H.S. Chou, A. Fattaey, E. Harlow, New functional activities for the p21 family of CDK inhibitors, Genes Dev. 11 1997 847862. .w25x M.-H. Lee, M. Nikolic, C.A. Baptista, E. Lai, L.-H. Tsai, J. Massague, The brain-specific activator p35 allows Cdk5 to escape inhibition by p27Kip1 in neurons, Proc. Natl. Acad. Sci. U.S.A. 93 1996 32593263..w26x J. Lew, Q.-Q. Huang, Z. Qi, R.J. Winkfein, R. Aebersold, T. Hunt, J.H. Wang, A brain-specific activator of cyclin-dependent kinase 5, Nature 371 1994 423426. . w27x Y. Li, C.W. Jenkins, M.A. Nichols, Y. Xing, Cell cycle expression and p53 regulation of the cyclin-dependent kinase inhibitor p21, Oncogene 9 1994 22612268. . w28x M. Liu, M.-H. Lee, M. Cohen, M. Bommakanti, L.P. Freedman, Transcriptional activation of the Cdk inhibitor p21 by vitamin D3 leads to the induced differentiation of the myelomonocytic cell line U937, Genes Dev. 10 1996 142153. .w29x H. Matsushime, D.E. Quelle, S.A. Shurtleff, M. Shibuya, C.J. Sherr, J.-K. Kato, D-Type cyclin-dependent kinase activity in mammalian cells, Mol. Cell. Biol. 14 1994 20662076. . w30x P. McConnell, M. Berry, The effect of refeeding after neonatal starvation on purkinje cell dendrtitic growth in the rat, J. Comp.Neurol. 178 1978 759772. . w31x P. McConnell, M. Berry, The effects of undernutrition on purkinje cell dendritic growth in the rat, J. Comp. Neurol. 177 1978. 159172.w32x M. Nikolic, H. Dudek, Y.T. Kwon, Y.F.M. Ramos, L.-H. Tsai, The cdk5rp35 kinase is essential for neurite outgrowth during neuronal differentiation, Genes Dev. 10 1996 816825. . w33x A. Noda, Y. Ning, S.F. Venable, O.M. Pereira-Smith, J.R. Smith, Cloning of senescent cell-derived inhibitor of DNA synthesis using an expression screen, Exp. Cell. Res. 211 1994 9098. . w34x T. Ohshima, J.M. Ward, C.-G. Huh, G. Longenecker, Veerana, H.C. Pant, R.O. Brady, L.J. Martin, A.B. Kulkarini, Targeted disruption of the cyclin-dependent kinase 5 gene results in abnormal corticogenesis, neuronal pathology and perinatal death, Proc. Natl. Acad. Sci. U.S.A. 93 1996 11731188. . w35x S.B. Parker, G. Eichele, P. Zhang, A. Rawls, A.T. Sands, A. Bradley, E. Olson, J. Harper, S. Elledge, p53-independent expression of p21cip1 in muscle and other terminally differentiating cells, Science 267 1995 10241027. . w36x A.N. Peeling, J.L. Smart, Vulnerability of developing brain to undernutrition: claims of selective sensitivity of the cerebellum found to be misleading on theoretical and empirical grounds, Growth Dev. Aging 55 1991 105116. . w37x L. Persson, A. Sima, The effect of pre- and postnatal undernutrition on the development of the cerebellar cortex in the rat, Neurobiology 5 1975 151166. .w38x W. Poluha, D.K. Poluha, B. Chang, N.E. Crosbie, C.M. Schonhoff, D.L. Kilpatrick, A.H. Ross, The cyclin-dependent kinase inhibitor p21WAF1 is required for survival of differentiating neuroblastoma cells, Mol. Cell. Biol. 16 1996 13351341. .w39x I. Reynisdottir, J. Massague, The subcellular locations of p15Ink4b and p27Kip1 coordinate their inhibitory interactions with cdk4 and cdk2, Genes Dev. 11 1997 492503. . w40x M. Serrano, G.J. Hannon, D. Beach, A new regulatory motif in cell-cycle control causing specific inhibition of cyclin DrCDK4, Nature 366 1993 704707. .w41x G.E.I. Shambaugh, R.J. Lee, G. Watanabe, F. Erfurth, A.N. Karnezis, A.E. Koch, G.K. Haines III, M. Halloran, B.A. Brodie, R.G. Pestell, Reduced cyclin D1 expression in the cerebella of nutritionally deprived rats correlates with developmental delay and decreased cellular DNA synthesis, J. Neuropathol. Exp. Neurol. 55 1996. 10101021. w42x G.E.I. Shambaugh, N. Natarajan, M.L. Davenport, D. Oehler, T. Unterman, Nutritional insult and recovery in the neonatal rat cerebellum: Insulin-like growth factors IGFs. and their binding proteins IGFBPs , Neurochem. Res. 20 1995 475489.. . w43x S.K. Sharma, U. Nayar, B. Singh, Effect of undernutrition on developing rat cerebellum. Some electrophysiological and neuromorphological correlates, J. Neurophys. Sci. 78 1987 261272. . w44x C.J. Sherr, G1 phase progression: cycling on cue, Cell 79 1994.551555. w45x C.J. Sherr, Mammalian G cyclins, Cell 73 1993 10591065.1. w46x C.J. Sherr, J.M. Roberts, Inhibitors of mammalian G cyclin-depen-1 dent kinases, Genes Dev. 9 1995 11491163..w47x P. Sicinski, J.L. Donaher, S.B. Parker, T. Li, A. Fazeli, H. Gardner, S.Z. Haslam, R.T. Bronson, S.J. Elledge, R.A. Weinberg, Cyclin D1 provides a link between development and oncogenesis in the retina and breast, Cell 82 1995 621630. . w48x A. Sima, L. Persson, The effect of pre- and postnatal undernutrition on the development of the rat cerebellar cortex: I. Morphologic observations, Neurobiology 5 1975 2334. . w49x S. Skapek, J. Rhee, D. Spicer, A. Lassar, Inhibition of myogenic differentiation in proliferating myoblasts by cyclin D1-dependent kinase, Science 267 1995 10221024. . w50x T.J. Soos, H. Kiyokawa, J.S. Yan, M.S. Rubin, A. Giordano, A. DeBlasio, S. Bottega, B. Wong, J. Mendelsohn, A. Koff, Formation of p27-cdk complexes during the human mitotic cell cycle, Cell Growth Differ. 7 1996 135146. . w51x R. Steinman, A.B. Hoffman, A. Iro, C. Guillouf, D.A. Liebermann, M.E. El-Houseini, Induction of p21 WAF-1 rCIP1 during differen-. tiation, Oncogene 9 1994 33893396. . w52x W. Sue, T Griffin, D.J. Woodward, R. Chanda, Malnutrition-induced alterations of developing purkinje cells, Exp. Neurol. 56 1977 298311.. w53x T. Tamaru, S.K. Trigun, H. Nakagawa, Differential expression of D type cyclins during neuronal maturation, Neurosci. Lett. 168 1994 . 229232.w54x H. Toyoshima, T. Hunter, p27, a novel inhibitor of G1 cyclin-Cdk protein kinase activity, is related to p21, Cell 78 1994 6774. .w55x L.-H. Tsai, T. Takahashi, V.S. Caviness, E. Harlow, Activity and expression pattern of cyclin-dependent kinase 5 in the embryonic mouse nervous system, Development 119 1993 10291040. . w56x L.H. Tsai, I. DFelalle, V.S. Caviness, T. Chae, E. Harlow, p35 is a neural specific regulatory subunit of cyclin-dependent kinase 5, Nature 371 1994 419423. . w57x S. Waga, G.J. Hannon, D. Beach, B. Stillman, The p21 inhibitor of cyclin-dependent kinases controls DNA replication by interaction with PCNA, Nature 369 1994 574578. . w58x G. Watanabe, A. Howe, R.J. Lee, C. Albanese, I.-W. Shu, A. Karnezis, L. Zon, J. Kyriakis, K. Rundell, R.G. Pestell, Induction of cyclin D1 by SV40 small t antigen, Proc. Natl. Acad. Sci. U.S.A. 93 1996 1286112866.. w59x G. Watanabe, R.J. Lee, C. Albanese, W.E. Rainey, D. Batlle, R.G. Pestell, Angiotensin II AII. activation of cyclin D1-dependent kinase activity, J. Biol. Chem. 271 1996 2257022577. . w60x G. Watanabe, P.P. Pena, C. Albanese, L.D. Wilsbacher, J.B. Young,R.G. Pestell, Adrenocorticotropin induction of stress activated protein kinase in the adrenal cortex in vivo, J. Biol. Chem. 272 1997 . 2006320069.w61x R.A. Weinberg, The retinoblastoma protein and cell cycle control, Cell 81 1995 323330. . w62x W. Xiong, R.G. Pestell, M.R. Rosner, Role of cyclins in neuronal differentiation of immortalized hippocampal cells, Mol. Cell. Biol.17 1997 65856597..G. Watanabe et al.rDeelopmental Brain Research 108 1998 7787()7990G. Watanabe et al.rDeelopmental Brain Research 108 1998 7787()G. Watanabe et al.rDeelopmental Brain Research 108 1998 7787()91w63x Y. Xiong, H. Zhang, D. Beach, D-Type cyclins associate with multiple protein kinases and the DNA replication and repair factor PCNA, Cell 71 1992 505514. . w64x G.-z. Yan, E.B. Ziff, NGF regulates the

PC12 cell cycle machinery through specific inhibition of the Cdk kinases and induction of cyclin D1, J. Neurosci. 15 1995 62006212. .