www.bba-direct.com

Biochimica et Biophysica Acta 1691 (2004) 105–116

The role of p27Kip1 in maintaining the levels of D-type cyclins in vivo

Vıtezslav Bryjaa,b, Jirı Pachernıka,c, Ludmila Faldıkovad, Pavel Krejcıe, Robert Poguee,Iveta Nevrivac, Petr Dvoraka,b,c, Ales Hampla,b,c,*

aCenter for Cell Therapy and Tissue Repair, Charles University, V Uvalu 84, 150 06 Prague, Czech RepublicbDepartment of Molecular Embryology, Institute of Experimental Medicine, Academy of Sciences of the Czech Republic,

Vıdenska 1083, 142 20 Prague, Czech RepubliccLaboratory of Molecular Embryology, Mendel University Brno, Zemedelska 1, 613 00 Brno, Czech Republic

dVeterinary Research Institute, Hudcova 70, 621 32 Brno, Czech RepubliceDepartment of Medical Genetics, Cedars-Sinai Research Institute, Los Angeles, CA, USA

Received 2 April 2003; received in revised form 12 December 2003; accepted 9 January 2004

Abstract

This in vivo study employs p27-deficient mice to investigate the significance of p27 for the metabolism of D-type cyclins in differentiated

cells. The absence of p27 results in decreased levels of cyclins D2 and/or D3 in some organs. As demonstrated on Leydig cells of testis, such

dependency is only restricted to certain cell types including terminally differentiated ones, and the absence of p27 in these cells can interfere

with their differentiation. The decrease of cyclin D caused by the absence of p27 equals the amount of cyclin D physically associated with

p27 in non-mutant animals. The data indicate that it is the proportion of p27-associated cyclin D that determines the response to p27

deficiency. Cells in which the level of D-type cyclin is dependent on p27 do not up-regulate the activity of their CDK2 and CDK4 upon loss

of p27, and these cells have a negligible amount of p27 bound to CDK2 and/or cyclin A/E under normal conditions. Together, the findings

suggest the existence of a dual role for p27, one being a classical regulation of cell cycle via inhibition of cyclin-dependent kinases (CDK),

and the other being participation in the establishment and/or maintenance of differentiated status that is realized in conjunction with D-type

cyclins.

D 2004 Elsevier B.V. All rights reserved.

Keywords: D-type cyclin; p27; Differentiation; p27-deficient mouse; Leydig cell; Thymus; Lung

1. Introduction

Cyclin-dependent kinases (CDK), their cyclins, and

inhibitors of CDKs (CKI) were identified as the crucial

components of cell cycle-regulating machinery. In mam-

malian cells, the complexes of cyclins D1, D2, D3, A,

and E with relevant CDKs are well understood as motors

that drive cells to enter and to pass through S phase (for

review, see Sherr and Roberts [1] and references therein).

However, surprisingly high levels of D-type cyclins were

also found in various terminally differentiated non-prolif-

erating cells [2–5]. Moreover, differentiation-associated

0167-4889/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.bbamcr.2004.01.001

* Corresponding author. Laboratory of Molecular Embryology, Mendel

University Brno, Zemedelska 1, 613 00 Brno, Czech Republic. Tel./fax:

+420-54-513-3298.

E-mail address: hampl@mendelu.cz (A. Hampl).

up-regulation of D-type cyclin(s) is often accompanied

by an increase in levels of CKIs of Cip/Kip family [6–

10], which suggests the importance of such behavior for

establishment and/or maintenance of differentiated pheno-

type. Increased stability of ternary cyclin D–CDK–p27

complexes may then represent the mechanism that allows

for high levels of D-type cyclins to be maintained in

differentiated cells [11]. If this is the case, the levels of

D-type cyclins should be lower in the absence of Cip/Kip

CKIs due to their inability to stably associate with CDK4/

6. Although data that both support and contradict this

prediction exist in literature, no study directly addressing

this phenomenon has been published. For example, mouse

embryonic fibroblasts deficient for p27 and p21 contain

less cyclin D1 and D2 [11], as well as cyclin D3 [12]

than their normal counterparts. Similarly, mammary

glands of p27-deficient mice have a decreased level of

cyclin D1 [13]. In contrast, no alterations in levels of

V. Bryja et al. / Biochimica et Biophysica Acta 1691 (2004) 105–116106

cyclins D1 and D2 were found in brains of p27-deficient

mice [14]. While p27-deficient retinas show no changes

in levels of cyclins D1 and D3 [15,16], retinas of cyclin

D1/p27 double null mice have much less cyclin D3 then

retinas of cyclin D1-deficient mice [15]. Also, cyclin D3

is expressed in large amounts in normal Muller glia, but it

is not detectable in Muller glia of p27-deficient animals

[17]. Whereas the levels of D-type cyclins in liver and

thymus as well as the assembly of cyclin D–CDK

complexes in heart and kidney in p21/p27 double null

mutant mice are only slightly reduced, the assembly of

cyclin D–CDK complexes in liver and thymus of such

animals is heavily impaired [11].

In this in vivo study, we have employed p27-deficient

mice to further investigate the biological significance of

p27/cyclin D co-regulation by determining the conditions

under which p27 is required for proper levels of individual

D-type cyclins.

2. Materials and methods

2.1. Animals and organ lysates

Mice that were deficient in p27 gene [18] were

provided by Dr. James Roberts (Fred Hutchinson Cancer

Research Center, Seattle, WA). They were maintained,

bred, and PCR-genotyped in the animal facility of the

Department of Molecular Embryology, Mendel University,

Brno. Samples of 12 organs of adult (2-month-old) mice

(femoral muscle, heart muscle – right-hand side ventric-

ular region, stomach – fundus wall, kidney cortex, liver,

lung, spleen, testis, ovary, thymus, telencephalon, and

skin from apical part of tail) were lysed on ice in 10

mM Tris–HCl pH 8.0, 25 mM EDTA, 1% SDS, soni-

cated, and cleared by centrifugation at 15000� g for 10

min at 4 jC. After determining the concentrations of total

protein using DC Protein Assay Kit (Bio-Rad, Hercules,

CA), lysates were mixed with double-strength Laemmli

sample buffer, boiled for 5 min and stored at � 80 jCuntil use.

2.2. Antibodies and reagents

Mouse monoclonal antibody to mouse cyclin D1 (sc-

450); rabbit polyclonal antibodies to mouse cyclin D2 (sc-

593), cyclin A (sc-751), cyclin E (sc-481), p27 (sc-528)

and CDK2 (sc-163), and goat polyclonal antibody to

CDK4 (sc-601-G) were purchased from Santa Cruz Bio-

technology (Santa Cruz, CA). Mouse monoclonal anti-

body to mouse p27 (K25020) was purchased from

Transduction Laboratories (Lexington, KY). Mouse mono-

clonal antibodies against cyclin A (Ab-1, E23) and cyclin

D2 (Ab-4, DCS-3.1 + DCS-5.2) were purchased from

Neomarkers (Fremont, CA). Mouse monoclonal antibody

against a C-terminal part of human cyclin D3, which

cross-reacts with the mouse homologue (DCS-22), was

generously provided by Dr. Jiri Lukas (Danish Cancer

Society, Copenhagen, Denmark). Rabbit polyclonal anti-

body against rat p42/44 MAP kinase, which cross-reacts

with the mouse homologue (#9102), was purchased from

New England Biolabs (Beverly, MA). Horseradish perox-

idase-conjugated anti-immunoglobulins were from Sigma;

polyvinylidene difluoride (PVDF) membrane Hybond-P,

and chemiluminescence detection reagents (ECL + Plus)

were purchased from Amersham (Amersham, Aylesbury,

UK). Equine chorionic gonadotropin (eCG) was from

Bioveta (Ivanovice, Czech Republic). All other chemicals

were purchased from Sigma (St. Louis, MO) or Fluka

(Buchs, Switzerland).

2.3. Western blot analysis

For Western blot analysis, equal amounts of total

protein were subjected to 10% SDS-PAGE, electrotrans-

ferred onto Hybond-P, immunodetected using appropriate

primary and secondary antibodies, and visualized by

ECL + Plus reagent according to the manufacturer’s

instructions. When required, membranes were stripped in

62.5 mM Tris–HCl pH 6.8, 2% SDS, and 100 mM 2-

mercaptoethanol, washed, and reblotted with another an-

tibody from this selection. After immunodetection, each

membrane was stained by amido black to confirm equal

protein loading.

2.4. Immunodepletion, immunoprecipitation, and kinase

assay

For these assays, organ samples (lung, thymus) were

mechanically disintegrated, extracted for 30 min in ice-

cold lysis buffer (50 mM Tris–HCl (pH 7.4), 150 mM

sodium chloride, 0.5% Nonidet P-40, 1 mM EDTA, 0.1

mM dithiothreitol, 50 mM sodium fluoride, 8 mM h-glycerolphosphate, 100 mM PMSF, 1 Ag/ml leupeptin, 1

Ag/ml aprotinin, 10 Ag/ml soybean trypsin inhibitor, 10

Ag/ml tosylphenylalanine chloromethane). Then extracts

were cleared by centrifugation at 15000� g for 15 min at

4 jC and stored at � 80 jC until use. Concentrations of

total protein were determined using a DC Protein Assay

Kit (Bio-Rad). The extracts were first subjected to initial

absorption with Protein G agarose beads and then incu-

bated with appropriate antibodies for 1 h in an ice bath.

Immunoprecipitates were collected on Protein G agarose

beads by overnight rotation, washed five times with lysis

buffer, resuspended in 2� Laemmli sample buffer and

subjected to SDS-PAGE followed by Western blot anal-

ysis. For immunodepletions, aliquots of extracts were

taken before and after the immunoprecipitation procedure

and subjected to Western blot analysis for appropriate

regulator. Each membrane was reblotted with antibody

against p42/44 MAP-kinase in order to confirm an equal

protein content. For kinase assays, immunoprecipitates

V. Bryja et al. / Biochimica et Biophysica Acta 1691 (2004) 105–116 107

were prepared as above, except that the last two washes

were done using kinase assay buffer (50 mM HEPES, pH

7.5, 10 mM MgCl2, 10 mM MnCl2, 8 mM h-glycerol-phosphate, 1 mM dithiothreitol). For CDK2, kinase reac-

tions were carried out for 30 min at 37 jC in a total

volume of 25 Al in kinase assay buffer supplemented with

100 Ag/ml histone H1 (type III-S) and 40 ACi/ml [32P]

ATP. For CDK4, kinase reactions were carried out for 30

min at 30 jC in a total volume of 25 Al in kinase assay

buffer supplemented with 80 Ag/ml of GST-pRb and 40

ACi/ml [32P] ATP. Reactions were terminated by mixing

with 2� Laemmli sample buffer, and each total reaction

mix was subjected to SDS-PAGE and autoradiography.

When required, the intensities of signals were assessed by

densitometry using Intelligent Quantifier software (Bio-

Image, Ann Arbor, MI).

2.5. Isolation of Leydig cells and analysis of testosterone

production

Testicular cells of p27-deficient and normal males

were separated into Leydig and non-Leydig populations

in Percoll density gradients essentially as described by

Schumacher et al. [19]. Discontinuous gradient (20%,

37% and 53%) of Percoll was used. Testes from three

to six males were pooled in each sample. The cells in

Leydig cell fraction were counted and the production

of testosterone was assayed using Testosterone kit

(DIAMETRA, Italy) according to manufacturer’s

instructions both with and without hormonal stimulation

Fig. 1. Expression of D-type cyclins and p27 in selected organs of adult mouse.

stomach, kidney, liver, lung, spleen, ovary, testis, thymus, brain, and skin) of adult

analyzed and representative blots are presented.

by eCG. The purity of each Leydig cell fraction was

determined by cytochemical visualization of beta-

hydroxysterol-dehydrogenase as described by Levy et

al. [20]. The data are expressed as the production of

testosterone per 105 Leydig cells by taking into account

the number of all cells in sample and the proportion of

Leydig cells in Leydig cell fraction (this proportion

varied from 65% to 95%). Both Leydig and non-

Leydig cells were also extracted for Western blot

analyses as described above.

2.6. Real-time RT-PC

Total RNA was isolated from tissue using RNeasy Mini

Kit (Qiagen), treated with DNase I, and then re-purified

using RNeasy Mini Kit again (Qiagen). Random primed

cDNAwas synthesized from 1 Ag of total RNA. PCR primer

pairs were as follows: cyclin D1 (5V-AATGCCAGAGGCG-GATGAGAAC-3V, 5V-AAAGTGCGTTGTGCGGTAGC-3V); cyclin D2 (5V-TTACCTGGACCGTTTCTTGGC-3V,5V-CAATGAAGTCGTGAGGGGTGAC-3V); cyclin D3

(5V-CCTACTTCCAGTGCGTGCAAAAG-3V, 5V-GACAGGTAGCGATCCAGGTAGTTC-3V); GAPDH (5V-ACCACAGTCCATGCCATCAC-3V, 5V-TCCAC-

CACCCTGTTGCTGTA-3V). Real-time PCR reactions were

carried out by the SYBR-green method (Finnzymes, Fin-

land), using optimized PCR conditions. Each sample was

assayed in triplicate. Thermal cycling was carried out on a

MJ Opticon (MJ Research, Waltham, MA). The results

were analyzed using the Opticon Monitor software (MJ

The levels of D-type cyclins and p27 in 12 organs (skeletal muscle, heart,

mice were analyzed by Western blotting. Two females and two males were

V. Bryja et al. / Biochimica et Biophysica Acta 1691 (2004) 105–116108

Research). Gene expression for each tissue was expressed

in terms of the threshold cycle (Ct), normalized to GAPDH

(DCt). DCt values were then compared between normal and

Fig. 2. Levels of D-type cyclins in organs of normal and p27-deficient mice. The

females) of each genotype were analyzed by Western blotting. Expression of p27

proteins to confirm equal protein loading. Since no differences were found betwe

animals for one genotype).

p27�/� samples to calculate DDCt, (DCt [normal] �DCt

[p27�/� ]). The final comparison of transcript ratios

between samples was given as 2DDCt.

levels of D-type cyclins in organs of four individuals (two males and two

is shown to document the genotype. The membranes were stained for total

en the sexes, only one representative blot is shown for each genotype (two

Fig. 3. Expression of D-type cyclins in Leydig and non-Leydig cells of

testes. Testicular cells of p27-deficient and normal males were separated

into Leydig and non-Leydig populations. Both Leydig and non-Leydig cells

were extracted and Western blotted for all D-type cyclins. Expression of

p27 is shown to document the genotype. Data are representative of at least

three independent replicates.

V. Bryja et al. / Biochimica et Biophysica Acta 1691 (2004) 105–116 109

3. Results

3.1. Both D-type cyclins and p27 are broadly expressed in

differentiated tissues

The samples of 12 organs of adult (2-month-old) non-

mutant mice of both sexes were first analyzed by Western

blot for the levels of D-type cyclins and p27 (Fig. 1). Of all

three cyclins, cyclin D1 shows the most restricted expres-

sion, being detectable in five organs, with higher levels in

ovary and thymus. Cyclins D2 and D3 were found in all

organs analyzed except for the absence of cyclins D2 and

D3 in skeletal muscle and cyclin D3 in brain. Similar

widespread expression was seen also for p27 with strikingly

high amounts in lung, spleen, ovary, and thymus.

3.2. Lack of p27 is accompanied by decreased levels of

cyclin D2 and cyclin D3 in some organs

In order to determine to what extent p27 is required for the

maintenance of proper levels of D-type cyclins in vivo, the

same organs as above were obtained from both normal and

p27-deficient mice [18] and Western analyzed for all three

cyclins (Fig. 2). While the levels of cyclin D1 were not

affected by the absence of p27 in any analyzed organ, the

quantities of cyclins D2 and D3 were lowered in some organs

of p27-deficient animals. The extent of this reduction varied

from only slight, such as for cyclin D2 in brain and skin, to a

very large, as exampled by cyclins D2 and D3 in lung.

Notably, in all organs, the occurrence of such cyclin D2/D3

reduction is thoroughly unlinked with the normal level of

p27. For example, (1) cyclin D2 levels decrease in both heart

and ovary despite of the major difference in their normal p27

content, and (2) cyclin D2 remains unchanged in ‘‘high-p27’’

thymus as well as in ‘‘low-p27’’ liver. Moreover, although in

some organs the absence of p27 results in down-regulation of

only cyclin D2 (heart, stomach, brain) or cyclin D3 (spleen),

in others (lung, ovary), the levels of both these D-type cyclins

are lowered.

3.3. Terminally differentiated Leydig cells require p27 to

maintain their D-type cyclin levels at normal

Every organ is composed of various cell types, from

which only some contain p27 and D-type cyclin proteins

as evidenced by immunohistochemical studies [3–5,13,

17,21,22]. Realizing such heterogeneity, it was highly

probable that the overall D-cyclin picture in the partic-

ular p27-deficient organ is determined primarily by the

proportion of cell type(s) with pronounced p27-dependent

regulation of cyclin(s) D. Leydig cells of adult testis

were chosen to verify such a scenario mainly because:

(1) they are fully differentiated and mitotically inactive

[23], (2) they express high levels of both p27 and cyclin

D3 [21], (3) they can be easily isolated and identified,

(4) they represent less then 1% of all testicular cells, and

(5) the levels of cyclins D2 and D3 are the same in both

normal and p27-deficient testis (shown in Fig. 2). As

shown in Fig. 3, we found a major difference between

Leydig cells and the remaining cells of testis. While in

non-Leydig cells the absence of p27 had no effect on

the levels of D-type cyclins, in Leydig cells the absence

of p27 brought about the decrease in levels of cyclins

D2 and D3. Thus, although in Leydig cells the levels of

cyclins D2 and D3 are obviously p27-dependent, such

dependency is not observed in testis as a whole due to

the minute number of Leydig cells in this organ.

3.4. Loss of p27 results in deregulated production of

testosterone by Leydig cells

Quantitative data shown in Fig. 3 suggested that co-

regulation between p27 and D-type cyclins may have

some importance for the establishment and/or mainte-

nance of differentiated status. Therefore, we were inter-

ested in whether the differentiation of Leydig cells could

be altered in the absence of p27. Production of testoster-

one, assayed in vitro, was used as a measure of differ-

Fig. 5. Level of mRNA coding for D-type cyclins in lung and thymus. The

total RNA was isolated from lung and thymus and the amounts of mRNA

coding for D-type cyclins were determined by real-time RT-PCR. The results

are presented as a relative change of the amount of mRNA in p27-deficient

mouse compared to normal control. Two sets of independent analyses are

shown (#1 and #2), each with one normal and one p27-deficient animal.

Fig. 4. Production of testosterone by normal and p27-deficient Leydig cells.

Leydig cells were isolated on density gradient and their in vitro testosterone

production was measured by ELISA assay. Testes of 11 normal and of 6

p27-deficient males were analyzed. The data are presented as the

meansF standard deviations. Average age and weight of males, weight of

testis, and the number of testicular cells per male are also given.

V. Bryja et al. / Biochimica et Biophysica Acta 1691 (2004) 105–116110

entiation status of Leydig cells isolated from testes of

normal and p27-deficient males. As demonstrated in Fig.

4, the levels of testosterone produced by p27-deficient

Leydig cells are almost twice as high as the levels

produced by normal Leydig cells. This difference is

observable in non-stimulated Leydig cells as well as in

Leydig cells stimulated by eCG at concentrations of 7.8

and 31 IU/l. Thus, although p27 is not vital for the

differentiated phenotype of Leydig cells to develop, the

absence of p27 still causes significant abnormalities in

Leydig cell function.

3.5. Down-regulation of D-type cyclin proteins upon the

loss of p27 is not due to the alterations in the levels of their

mRNAs

Lung and thymus, the organs showing ‘‘large’’ versus

‘‘no’’ decrease of cyclins D2/D3 under p27-null condi-

tions (see Fig. 2) were selected for detailed analysis

Fig. 6. Physical association of D-type cyclins with p27 and CDK4 in lung and thy

extracts were subjected to biochemical analyses. (A) The extracts were and were n

p42/44 MAP-kinase was determined to confirm an equal protein content. To contr

were also included. Densitometry was used to assess the quantity of proteins. For ea

p27+/+ animals was defined as 1.0, and from this value all other values were calcu

bars. (B) The extracts were immunoprecipitated with p27- and CDK4-specific anti

blotting. (C) The extracts were immunoprecipitated with the antibodies specific for

The total amount of D-type cyclins in extracts was analyzed by Western blotting.

four independent replicates.

addressing the mechanism(s) by which p27 influences

the levels of D-type cyclins. First, the expression of

mRNAs coding for D-type cyclins was determined using

real-time RT-PCR in normal and p27-deficient organs.

Two sets of independent analyses were carried out (#1

and #2), each using one normal and one p27-deficient

animal. As shown in Fig. 5, the levels of mRNA for all

D-type cyclins were dramatically decreased in thymus of

p27-null animals, whereas only subtle variations were

observed in lung. Obviously, the expression of mRNAs

for D-type cyclins can be influenced by the absence of

p27 protein. Still, neither in lung nor in thymus, the

mus. Lung and thymus from normal mice were extracted and the resulting

ot immunodepleted of p27 and then analyzed by Western blotting. Level of

ol for a non-specific loss of D-type cyclins, organs from p27� /� animals

ch D-type cyclin, the average density of non-immunodepleted samples from

lated. Data represent the means with standard deviations indicated by error

bodies, respectively, and the immunoprecipitates were analyzed by Western

D-type cyclins and then Western blotted with the antibody against p27. (D)

Both immunodepletion and immunoprecipitation data are representative of

V. Bryja et al. / Biochimica et Biophy

changes in the levels of cyclin D proteins correspond to

the changes in the levels of their mRNAs. Therefore,

post-transcriptional rather than transcriptional mecha-

nism(s) are likely to underlie the p27/cyclin D co-regula-

tion observed here in lung.

3.6. The amount of D-type cyclin missing under p27�/�conditions equals its amount normally associated with p27

As it is well-documented on cyclin D3, the cyclin D

pool generally consists of (1) largely inactive cyclin D–

sica Acta 1691 (2004) 105–116 111

V. Bryja et al. / Biochimica et Biophysica Acta 1691 (2004) 105–116112

CDK4/6–CKI complexes, (2) active cyclin D–CDK4/6

complexes, and of (3) cyclin D that is free of both CDK

and CKI [12]. Therefore, it was possible that the reduc-

tion of levels of cyclin D2 and/or cyclin D3 under p27-

deficient conditions reflected a loss of p27-associated part

Fig. 7. Kinase activities of CDK2 and CDK4, and physical associations among CD

normal mice were extracted and the resulting extracts were immunoprecipitated us

assayed in vitro for their kinase activities using histone H1 and GST-pRb as substr

cyclin E antibodies, respectively, were probed by Western blotting for associating

CDK4 in extracts was analyzed by Western blotting. The data are representative

of cyclin D pool. If this is the case, an equation that

relates the amount of cyclin D normally associated with

p27 in certain an organ/cell, and the sensitivity of cyclin

D level in such an organ/cell to the absence of p27

should exist. To approach this hypothesis, the extracts

K2, cyclin A, cyclin E, and p27 in lung and thymus. Lung and thymus from

ing appropriate antibodies. (A) Immunoprecipitated CDK2 and CDK4 were

ates. (B) Immunoprecipitates made using anti-p27, -CDK2, -cyclin A, and -

p27 and cyclin A. (C) The total amount of cyclin A, cyclin E, CDK2, and

of three independent replicates.

V. Bryja et al. / Biochimica et Biophysica Acta 1691 (2004) 105–116 113

made of lung and thymus were immunodepleted of p27

and then Western blotted for D-type cyclins. As shown in

Fig. 6A, while in lung the quantities of both cyclins D2

and D3 decreased by about 80–90% upon depletion of

p27, this was not the case in thymus. Notably, the

quantity of cyclin D that is refractory to the depletion

by anti-p27 antibody equals the quantity of cyclin D in

p27-deficient animals (P>0.05, Student’s t-test). As dem-

onstrated by immunoprecipitation, cyclins D2 and D3 are

complexed with p27 in both lung and thymus (Fig. 6B,

C). However, while in lung the majority of cyclins D2

and D3 are complexed with p27, only a small portion of

these cyclins associates with p27 in thymus (Fig. 6A). As

documented here on cyclin D3, CDK4 also associates

with p27 and D-type cyclin in both lung and thymus (Fig.

6B), thus potentially connecting p27 to D-type cyclin in

these organs. To ensure the validity of the depletion/

precipitation data, the total amounts of D-type cyclins

were determined by Western blot in all samples used for

such analyses (Fig. 6D).

3.7. Lung and thymus differ from each other in their CDK2-

and CDK4-associated kinase activities and in the amounts

of cyclin A/E–CDK2–p27 complexes

To gain more detailed insight into the differences

between lung and thymus, the activity and composition

of CDK2- and CDK4-containing complexes were also

analyzed. First, CDK2- and CDK4-associated kinase ac-

tivities were determined in lung and thymus from normal

and p27-deficient animals. As we demonstrate in Fig. 7A,

CDK2- and CDK4-associated kinase activities are at

about the same low levels in lung and thymus in the

presence of p27. More importantly, although in thymus

the loss of p27 results in dramatic activation of CDKs, it

has no such effect in lung. Along with this difference in

p27-mediated regulation of the activities of CDKs,

corresponding differences exist in CDK2, cyclin A, and

cyclin E proteins, both in their total amounts and the

amounts physically associated with p27 (Fig. 7B, C). In

general, this data documents that the amounts of total and

p27-bound regulators are at (cyclin E) or below (CDK2,

cyclin A) a limit of detectability in lung but they are

quite high in thymus. We conclude that in cells of thymus

(but not of lung) both cyclin A/E–CDK2 and cyclin D–

CDK4 complexes are active in the absence of p27, and

normally p27 dynamically binds to these complexes in

order to regulate their activity.

4. Discussion

Although several previous studies already suggested the

involvement of p27–cyclin D interaction(s) in development

and/or maintenance of the differentiated phenotype [2,4,17,

21,24–26], this study may open a rather new dimension in

looking at such function. Specifically, it documents that: (1)

p27 is necessary for maintaining the proper levels of cyclins

D2 and D3 in vivo; (2) rather then being limited to only few

cell types, this p27-dependency is common to a wide variety

of cells/tissues in a body; (3) defect(s) in differentiation may

develop in the absence of p27 in cells typical by such p27-

dependency; and (4) the reduction of cyclin D amounts

under p27-deficient conditions equals the portion of cyclin

D pool normally associated with p27.

Well-pronounced dependency of D-type cyclins on p27

was found here in many organs. However, there were also

some organs, such as testis, that seemed to lack this type of

co-regulation. Therefore, the hypothesis that even these

organs may still contain cells with p27-dependency was

tested on terminally differentiated Leydig cells of testis. We

demonstrate that in Leydig cells, the levels of cyclins D2

and D3 are in fact p27-dependent. We also show that the

absence of p27 causes Leydig cells to produce about twice

as much testosterone when compared to their normal coun-

terparts. Although unexpected, such functional abnormality

can be explained by p27–cyclin D complexes being in-

volved in the feedback mechanism that negatively regulates

the production of testosterone. Importantly, regardless of the

exact mechanism, p27 (possibly in association with cyclins

D) in Leydig cells seems to serve primarily some purpose

other than regulating G1/S transition, similarly to what has

been summarized by Coqueret [27] for various other cell

types. Combining the data obtained on Leydig cells with

generally high incidence of p27-dependency in the body

leads to the notion that the interaction between p27 and D-

type cyclin described here is employed by differentiating

and/or differentiated cells.

We show here that with the lack p27 CKI, in some

organs the levels of cyclin D proteins are abnormally low,

but this is not due to the down-regulated transcription of

cyclin D genes (see Fig. 5). Instead, the reduction of the

amount of D-type cyclin under p27-deficient conditions

seems to reflect the absence of its pool normally associated

with p27 (see Fig. 6). Concerning the molecular nature of

such interaction between p27 and D-type cyclin, CDK4 is

an obvious candidate for a mediating molecule (both p27

and cyclin D3 physically interact with CDK4 in lung and

thymus; see Fig. 5B). However, although the level of total

cyclin D3 under p27-deficient conditions is appreciably

lowered only in lung and not in thymus, the level of

CDK4-associated cyclin D3 is decreased in both organs. In

other words, cyclin D–CDK4–p27 trimers undergo decay

in both organs and thus may hardly explain the dramatic

difference between lung and thymus in the dependency of

their total cyclins D2 and D3 on the presence of p27.

Therefore, it is highly probable that the p27–cyclin D

interaction observed here in lung (and also in other organs)

employs some molecule(s) other then CDK4.

Irrespective of whether it is CDK4 or some other mole-

cule that mediates the p27–cyclin D interaction in lung, the

ablation of p27 disrupts identical molecular complexes as

V. Bryja et al. / Biochimica et Biophysica Acta 1691 (2004) 105–116114

does the ablation of D-type cyclin. Consequently, if the

dependence of D-type cyclin(s) on p27 is of some biological

importance, major similarities in phenotypes should exist

between p27�/� and cyclin D� /� animals. In fact, the

findings of Muraoka et al. [13], Fantl et al. [28], and Sicinski

et al. [29], documenting that impaired lobuloalveolar differ-

entiation in the mammary gland and the inability to lactate

develop in both p27-deficient and cyclin D1-deficient ani-

mals, very much fulfill such a prediction. Notably, absence

of p27 in mammary glands is accompanied by a decreased

level of cyclin D1 [13]. Thus, deficiencies of D-type cyclins

and of p27 in mammary glands produce the same phenotype,

most likely due to the requirement of p27 for cyclin D to be

properly metabolized. In Fig. 8 we schematize such a

scenario and offer the concept of a dual role of complexes

involving p27 and D-type cyclins in cell proliferation and

differentiation. In this concept, D-type cyclins associated

with p27 not only contribute to the regulation of proliferation

via their well known link to cyclin A/E–CDK2 activities

(here represented by thymic cells), but they also play some,

although yet elusive, role in establishing and/or maintaining

differentiated status (here represented by Leydig cells and

cells of lung). Importantly, while D-type cyclins and p27 act

primarily in an opposite fashion as regulators of cell prolif-

eration (activation versus inhibition of CDKs), functioning

Fig. 8. Dual role of D-type cyclins and p27 in proliferation and differentiation—a

from the complexes containing CDK2, and thereby they control their proliferation

yet unidentified, function in differentiation-related processes. Thus, loss of p27 not

cyclin E/A but it may also elicit a decrease in levels of D-type cyclins that results in

titrate p27 from CDK2–cyclin A/E complexes and proliferation becomes suppress

cyclin are reminiscent to the defects produced by the absence of p27.

in the same direction seems to be typical for these two

molecules when driving differentiation. Obviously, the most

prominent phenotype in p27-deficient animals is multiorgan

hyperplasia [18,30,31], while cyclin D1-deficient animals

exhibit dwarfism of almost all organs [28,29], both charac-

terizing a loss of proliferation-associated functions of p27

and cyclin D1, respectively. In contrast, the main differen-

tiation-associated alteration in the cyclin D1-deficient

mouse, impaired lobuloalveolar development during preg-

nancy [28,29], is mimicked in p27-deficient mouse [13] as

discussed earlier. At the time of writing this paper, detailed

analysis of the phenotypes of cyclin D1-, cyclin D2-, and

cyclin D3-only mice was published by Ciemerych et al. [32].

Importantly, the concept given in our study allows under-

standing of two molecular abnormalities displayed by such

‘‘single-cyclin’’ mice: (a) level of the remaining, intact D-

type cyclin, is up-regulated; (b) down-regulation of p27

occurring under ‘‘single-cyclin’’ conditions does not produce

increased activity of CDK2. Our concept offers the following

scenario: since normally p27 is complexed with more then

one D-type cyclin, ‘‘single-cyclin D’’ conditions produce

relative overabundance of p27, which is resolved in part by

p27 degradation and shift of p27 to the remaining D-type

cyclin which in turn becomes stabilized and its amount

increases. No changes then occur to the activities of CDKs

model. Cells employ CDK4/6–cyclin D complexes to flexibly titrate p27

. However, mutual dependency between cyclin D and p27 serves also some,

only causes unrestricted growth due to the inefficient inhibition of CDK2–

differentiation defects. Upon ablation of cyclin D, cells loose their ability to

ed. On the other hand, defects in differentiation caused by the absence of D-

V. Bryja et al. / Biochimica et Biophysica Acta 1691 (2004) 105–116 115

simply because non-proliferating/differentiated cells employ

this particular type of interaction between p27 and D-type

cyclins.

Another line of evidence that may also support the

suggested role of D-type cyclins and p27 in cell differentia-

tion comes from the study of Fero et al. [33]. The organs

found in our study as being typical of p27-dependent regu-

lation of D-type cyclins (lung and ovary) are prone to

tumorigenesis in p27-deficient mice [33], whereas the organs

without such regulation only tend to be hyperplastic. Realiz-

ing that tumor formation involves not only the alterations in

proliferation rate but also the changes in differentiation status,

the tumorigenesis in p27-deficient mice can be viewed as a

result of the absence of p27 combined with the reduced levels

of D-type cyclins. In other words, an organ seems to be prone

to p27-mediated tumorigenesis when it employs complexes

containing p27 and D-type cyclin(s) for differentiation status

of its cells to be established and/or maintained.

In conclusion, the results of this study lead us to propose

that dependency of the levels of D-type cyclins on p27 in

fact reflects the role of these cyclins in differentiation-

related processes. Although our data provide clues only on

cyclins D2 and D3, it is still possible that cyclin D1 also

operates in a similar way, and the failure to detect its p27-

dependency is just due to the extremely small proportion of

a particular cell type. Importantly, when the changes in

levels of p27 and/or D-type cyclins occur, they may intro-

duce an alteration in equilibrium between proliferation and

differentiation processes that may finally result in tumor

formation.

Acknowledgements

We are very grateful to Dr. James M. Roberts for

providing us with p27-deficient mice, to Dr. J. Lukas for

providing us with antibodies and for critically reading the

manuscript, to Mrs. Eva Janska for excellent technical

assistance, and to Lucie Krenek for correction of the

English. This research was supported in part by Grant GA

204/01/0905 from the Grant Agency of the Czech Republic,

Grant AV 0Z5039906 from the Academy of Sciences of the

Czech Republic, and Grants MSM 432100001 and LN

00A065 from the Ministry of Education, Youth, and Sports

of the Czech Republic.

References

[1] C.J. Sherr, J.M. Roberts, CDK inhibitors: positive and negative reg-

ulators of G1-phase progression, Genes Dev. 13 (1999) 1501–1512.

[2] C.Y. Gao, P. Zelenka, Cyclins, cyclin-dependent kinases and differ-

entiation, BioEssays 19 (1997) 307–315.

[3] J. Bartkova, J. Lukas, M. Strauss, J. Bartek, Cyclin D3: requirement

for G1/S transition and high abundance in quiescent tissues suggest a

dual role in proliferation and differentiation, Oncogene 17 (1998)

1027–1037.

[4] J. Bartkova, E. Rajpert-de Meyts, N.E. Skakkebak, J. Bartek, D-type

cyclins in adult human testis and testicular cancer: relation to cell

type, proliferation, differentiation, and malignancy, J. Pathol. 187

(1999) 573–581.

[5] C. Doglioni, C. Chiarelli, E. Macri, A.P. Dei Tos, E. Meggiolaro, P.

Dalla Palma, M. Barbareschi, Cyclin D3 expression in normal, reac-

tive and neoplastic tissues, J. Pathol. 185 (1998) 159–166.

[6] P. Savatier, H. Lapillonne, L.A. van Grunsven, B.B. Rudkin, J.

Samarut, Withdrawal of differentiation inhibitory activity/leukemia

inhibitory factor up-regulates D-type cyclins and cyclin-dependent

kinase inhibitors in mouse embryonic stem cells, Oncogene 12

(1995) 309–322.

[7] O. Kranenburg, R.P. de Groot, A. Van der Eb, A. Zantema, Differen-

tiation of P19 EC cells leads to differential modulation of cyclin-

dependent kinase activities and to changes in the cell cycle profile,

Oncogene 10 (1995) 87–95.

[8] P.J. Hauser, D. Agrawal, M. Flanagan, W.J. Pledger, The role of

p27kip1 in the in vitro differentiation of murine keratinocytes, Cell

Growth Differ. 8 (1997) 203–211.

[9] M. Reichert, D. Eick, Analysis of cell cycle arrest in adipocyte dif-

ferentiation, Oncogene 18 (1999) 459–466.

[10] A. Hampl, J. Pachernık, P. Dvorak, Levels and interactions of p27,

cyclin D3 and CDK4 during the formation and maintenance of the

corpus luteum in mice, Biol. Reprod. 62 (2000) 1393–1401.

[11] M. Cheng, P. Olivier, J.A. Diehl, M. Fero, M.F. Roussel, J.M. Rob-

erts, C.J. Sherr, The p21Cip1 and p27kip1 ‘inhibitors’ are essential

activators of cyclin D-dependent kinases in murine fibroblasts,

EMBO J. 18 (1999) 1571–1583.

[12] T.K. Bagui, R.J. Jackson, D. Agrawal, W.J. Pledger, Analysis of

cyclin D3–cdk4 complexes in fibroblasts expressing and lacking

p27kip1 and p21cip1, Mol. Cell. Biol. 20 (2000) 8748–8757.

[13] R.S. Muraoka, A.E.G. Lenferink, J. Simpson, D.M. Brantley, L.R.

Roebuck, F.M. Yakes, C.L. Arteaga, Cyclin-dependent kinase in-

hibitor p27kip1 is required for mouse mammary gland morphogen-

esis and function, J. Cell Biol. 153 (2001) 917–931.

[14] P. Casaccia-Bonnefil, R.J. Hardy, K.K. Teng, J.M. Levine, A. Koff,

M.V. Chao, Loss of p27Kip1 function results in increased proliferative

capacity of oligodendrocyte progenitors but unaltered timing of dif-

ferentiation, Development 126 (1999) 4027–4037.

[15] W. Tong, J.F. Pollard, Genetic evidence for the interactions of cyclin

D1 and p27Kip1 in mice, Mol. Cell. Biol. 21 (2001) 1319–1328.

[16] Y. Geng, Q. Yu, E. Sicinska, M. Das, R.T. Bronson, P. Sicinski,

Deletion of the p27Kip1 gene restores normal development in

cyclin D1-deficient mice, Proc. Natl. Acad. Sci. U. S. A. 98

(2001) 194–199.

[17] M.A. Dyer, C.L. Cepko, Control of Muller glial cell proliferation and

activation following retinal injury, Nat. Neurosci. 3 (2000) 873–880.

[18] M.L. Fero, M. Rivkin, M. Tasch, P. Porter, C.E. Carow, E. Firpo, K.

Polyak, L.H. Tsai, V. Broudy, R.M. Perlmutter, K. Kaushansky, J.M.

Roberts, A syndrome of multiorgan hyperplasia with features of gi-

gantism, tumorigenesis, and female sterility in p27Kip1-deficient mice,

Cell 85 (1996) 733–744.

[19] M. Schumacher, G. Schaefer, A.F. Holzstein, H. Hilz, Rapid isolation

of mouse Leydig cells by centrifugation in Percoll density gradient

with complete retention of morphological and biochemical integrity,

FEBS Lett. 91 (1978) 333–338.

[20] H. Levy, H.W. Deane, B.L. Rubin, Visualization of 3beta-ol-dehydro-

genase activity in tissues of intact and hypophysectomized rats, En-

docrinology 65 (1959) 933–943.

[21] Q. Zhang, X. Wang, D.J. Wolgemuth, Developmentally regulated

expression of cyclin D3 and its potential in vivo interacting pro-

teins during murine gametogenesis, Endocrinology 140 (1999)

2790–2800.

[22] H. Nagahama, S. Hatakeyama, K. Nakayama, M. Nagata, K. Tomita,

K. Nakayama, Spatial and temporal expression patterns of the cyclin-

dependent kinase (CDK) inhibitors p27Kip1 and p57Kip2 during mouse

development, Anat. Embryol. 203 (2001) 77–87.

V. Bryja et al. / Biochimica et Biophysica Acta 1691 (2004) 105–116116

[23] L. Benton, L.-X. Shan, M.P. Hardy, Differentiation of adult Leydig

cells, J. Steroid Biochem. Mol. Biol. 53 (1995) 61–68.

[24] C. Cenciarelli, F. De Santa, P.L. Puri, E. Mattei, L. Ricci, F. Bucci, A.

Felsani, M. Caruso, Critical role played by cyclin D3 in the MyoD-

mediated arrest of cell cycle during myoblast differentiation, Mol.

Cell. Biol. 19 (1999) 5203–5217.

[25] J.M.T. Huard, C.C. Forster, M.L. Carter, P. Sicinski, M.E. Ross, Cer-

ebellar histogenesis is disturbed in mice lacking cyclin D2, Develop-

ment 126 (1999) 1927–1935.

[26] H. Preclikova, V. Bryja, J. Pachernik, P. Krejci, P. Dvorak, A. Hampl,

Early cycling-independent changes to p27, cyclin D2 and cyclin D3 in

differentiating mouse embryonal carcinoma cells, Cell Growth Differ.

13 (2002) 421–430.

[27] O. Coqueret, New roles for p21 and p27 cell-cycle inhibitors: a

function for each cell compartment? Trends Cell Biol. 13 (2003)

65–70.

[28] 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) 2364–2372.

[29] 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) 621–630.

[30] H. Kiyokawa, R.D. Kineman, K.O. Manova-Todorova, V.C. Soares,

E.S. Hoffman, M. Ono, D. Khanam, A.C. Hayday, L.A. Frohman, A.

Koff, Enhanced growth of mice lacking the cyclin-dependent kinase

inhibitor function of p27Kip1, Cell 85 (1996) 721–732.

[31] K. Nakayama, N. Ishida, M. Shirane, A. Inomata, T. Inoue, N. Shish-

ido, I. Horii, D.Y. Loh, K. Nakayama, Mice lacking p27Kip1 display

increased body size, multiple organ hyperplasia, retinal dysplasia and

pituitary tumors, Cell 85 (1996) 707–720.

[32] M.A. Ciemerych, A.M. Kenney, E. Sicinska, I. Kalaszczynska,

R.T. Bronson, D.H. Rowitch, H. Gardner, P. Sicinski, Develop-

ment of mice expressing a single D-type cyclin, Genes Dev. 16

(2002) 3277–3289.

[33] M.L. Fero, E. Randel, K.E. Gurley, J.M. Roberts, C.J. Kemp, The

murine gene p27Kip1 is haplo-insufficient for tumour suppression,

Nature 396 (1998) 177–180.