Testing the importance of p27 degradationby the SCFskp2 pathway in murine modelsof lung and colon cancerInke Timmerbeula,b,c, Carrie M. Garrett-Engelec,d, Uta Kossatzb, Xueyan Chend,e, Eduardo Firpod,e, Viktor Grunwaldf,Kenji Kaminog, Ludwig Wilkensg, Ulrich Lehmannh, Jan Bueri, Robert Geffersi, Stefan Kubickaa, Michael P. Mannsa,Peggy L. Porterj, James M. Robertsd,e,k, and Nisar P. Maleka,b,l

aDepartment of Gastroenterology, Hepatology, and Endocrinology, bInstitute for Molecular Biology, fDepartment of Hematology and Oncology, hInstitutefor Pathology, and gInstitute of Cell and Molecular Pathology, Hannover Medical School, D-30625 Hannover, Germany; iDepartment of Cell Biology andImmunology, German Research Centre for Biotechnology, D-38124 Braunschweig, Germany; dDivision of Basic Sciences and jDivision of Human Biology,Fred Hutchinson Cancer Research Center, Seattle, WA 98109; eDepartment of Biochemistry, University of Washington, Seattle, WA 98195; andkHoward Hughes Medical Institute, Seattle, WA 98195

Communicated by George J. Todaro, Targeted Growth, Inc., Seattle, WA, July 25, 2006 (received for review February 22, 2006)

Decreased expression of the CDK inhibitor p27kip1 in human tumorsdirectly correlates with increased resistance to chemotherapies,increased rates of metastasis, and an overall increased rate ofpatient mortality. It is thought that decreased p27 expression intumors is caused by increased proteasomal turnover, in particularactivation of the pathway governed by the SCFskp2 E3 ubiquitinprotein ligase. We have directly tested the importance of theSCFskp-mediated degradation of p27 in tumorigenesis by analyzingthe tumor susceptibility of mice that express a form of p27 thatcannot be ubiquitinated and degraded by this pathway(p27T187A). In mouse models of both lung and colon cancerdown-regulation of p27 promotes tumorigenesis. However, wefound that preventing p27 degradation by the SCFskp2 pathwayhad no impact on tumor incidence or overall survival in eithertumor model. Our study unveiled a previously unrecognized rolefor the control of p27 mRNA abundance in the development ofnon-small cell lung cancers. In the colon cancer model, the fre-quency of intestinal adenomas was similarly unaffected by thep27T187A mutation, but, unexpectedly, we found that it inhibitedprogression of intestinal adenomas to carcinomas. These studiesmay guide the choice of clinical settings in which pharmacologicinhibitors of the Skp2 pathway might be of therapeutic value.

cell cycle � lung cancer � ubiquitylation � F-box protein

The CDK inhibitor p27kip1 controls the progression of cellsthrough the G1 phase of the cell cycle by regulating the

activities of cyclin E and cyclin A�Cdk2 complexes (1). Loss ofp27 in the mouse leads to a significant increase in the prolifer-ation of all tissues (2). It also predisposes the mouse to thespontaneous development of pituitary carcinomas and the oc-currence of tumors in multiple tissues after irradiation ortreatment with the carcinogen N-ethyl-N-nitrosourea (ENU).Importantly, reduction of p27 dosage by 50% in p27 heterozy-gous mice also leads to accelerated carcinogenesis making p27 ahaploinsufficient tumor suppressor protein (3). By crossing p27knockout mice to mouse strains with defects in other tumorsuppressor proteins, cooperative functions of p27 with proteinslike PTEN, CBP, and Rb in preventing tumor formation wereuncovered (4–6).

The results obtained in the p27 knockout mouse correlate closelywith the observation that low expression levels of p27 alone or incombination with other defects predict an overall reduced survivalof the human cancer patients. An important difference between p27and other tumor suppressor proteins, like p53 or p16, is that thedecrease in p27 expression in tumor tissue is not caused bymutations of the p27 gene or epigenetic silencing. It is generallybelieved that down-regulation of p27 in cancers occurs throughaccelerated degradation of the p27 protein (7). Nevertheless, direct

evidence in support of this hypothesis is limited to in vitro experi-ments in which extracts from increasingly aggressive primary hu-man tumors had a correspondingly increased ability to degraderecombinant p27 (8).

These observations have suggested that pharmacologic stabi-lization of p27 in tumor tissues may be a useful therapeuticstrategy (9). However, such strategies require a detailed under-standing of the molecular mechanisms involved in p27 turnover.It is now clear that the degradation of p27 during late G1 and Sphases depends on the phosphorylation of p27 at threonine-187(T187). This phosphorylation leads to binding of p27 to theF-box protein Skp2 and subsequent polyubiquitylation by theSCF (Skp1–Cullin–F-box) E3 ligase (10–12). Previously, wegenerated a mouse knockin system in which T187 was replacedby alanine (T187A) and characterized how loss of this SCFskp2-mediated pathway for p27 degradation affected normal devel-opment (13). Since that time other mechanisms for regulatingp27 levels have been described. In addition to proteolyticdegradation by the SCFskp2 E3 ligase, translational, transcrip-tional, mRNA stability, and alternative proteolytic mechanismshave been shown to regulate the abundance of the p27 protein(14–17). Furthermore, the subcellular localization of the p27protein can be regulated through phosphorylation at differentsites (18). The relative contributions of these many regulatorypathways to controlling p27 levels in both normal and tumor cellshas not been investigated.

ResultsLoss of p27 Promotes Tumorigenesis in the Lung Independent ofSCFskp2-Mediated p27 Degradation. In this study we have focused onthe importance of the SCFskp2-mediated pathway for down-regulation of p27 during tumorigenesis. Our strategy was todetermine whether expression of p27T187A would affect tumorincidence, progression, and overall survival in two differenttumor models. The first model system was one in which spon-taneous activation of a latent copy of the activated K-rasoncogene causes 100% of all transgenic animals (K-ras*) todevelop lung tumors (19). These tumors are reminiscent ofhuman non-small cell lung carcinomas (NSCLC), with theearliest lesions developing after 1 week of age. We crossed thep27-null (p27���) and p27T187A alleles into the K-ras* back-

Conflict of interest statement: No conflicts declared.

Abbreviations: ENU, N-ethyl-N-nitrosourea; NSCLC, non-small cell lung carcinoma; PCNA,proliferating cell nuclear antigen.

cI.T. and C.M.G.-E. contributed equally to this work.

lTo whom correspondence should be addressed. E-mail: malek.nisar@mh-hannover.de.

© 2006 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0606316103 PNAS � September 19, 2006 � vol. 103 � no. 38 � 14009–14014

CELL

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

May

23,

202

1

ground to generate homozygous p27�/��K-ras* or homozygousp27T187A�K-ras* (p27T187A�K-ras*) mice. To detect differ-ences in survival we followed the entire colony of transgenic mice(K-ras*, p27�/��K-ras*, and p27T187A�K-ras*), and when ananimal appeared moribund it was killed and subjected to afocused necropsy. The tumors from all three genotypes werestage-1, moderately differentiated, noninvasive adenomas, andno difference in proliferation as measured by Ki67 staining wasobserved among the tumors (data not shown). As shown in Fig.1A, loss of p27 accelerated the course of the disease significantly,leading to the death of 50% of all mice after 6 months (P �0.0001). In contrast, most K-ras* and p27T187A�K-ras* micewere still alive at 6 months. This result indicated that loss of p27cooperated with the activation of K-ras in the formation of lungtumors. Importantly, the p27T187A�K-ras* strain, expressing apotentially stabilized form of p27 protein, did not show asignificant advantage in overall survival compared with wild-type p27 mice expressing the activated K-ras gene (P � 0.3).

p27 Levels in NSCLC Are Regulated at the mRNA Level. Next wecompared the expression levels of p27 and p27T187A in lungcancer tissue and surrounding normal noncancerous tissue by

immunohistochemistry using an antibody specific to p27. Thistechnique has been shown to accurately measure the amounts ofp27 protein expressed in lung carcinomas (20). Specifically, itdistinguishes expression of p27 in tumor cells from tumor-infiltrating cells like lymphocytes or blood vessels. As shown inFig. 1B, no differences in the expression of p27 were detectablebetween K-ras* and p27T187A�K-ras* mice. Both strains showedclear p27 staining in surrounding normal lung tissue, whereaslittle or no p27 expression was detectable in the cancerouslesions.

Cks1 has previously been shown to be rate-limiting for turn-over of p27 by the SCFskp2 pathway (21, 22) in vitro and in vivo.Also, recent studies in human NSCLC tissue suggested aninverse relation between cks1 and p27 expression (23). Wetherefore measured the protein expression levels of this criticalcomponent of the T187- and SCFskp2-dependent degradationmechanism in lung tumor and surrounding normal tissue ofK-ras* and p27T187A�K-ras* mice. As shown in Fig. 2A, wefound cks1 expression unchanged or even slightly decreased inlung tumor tissue in both wild-type and p27T187A mice. Theseexperiments indicated that although p27 acted as a tumorsuppressor protein in the development of NSCLC, its down-regulation in tumor tissue was independent of the Skp2-mediated T187-dependent degradation pathway.

We therefore wondered whether regulation at the RNA levelrather than increased protein turnover might account for theobserved changes in p27 expression. To test this hypothesis wemeasured p27 mRNA levels by quantitative real-time PCR inlung tumor tissue from K-ras* and p27T187A�K-ras* mice andcompared it to mRNA levels in K-ras* normal lung tissue. Wefound a strong down-regulation of p27 mRNA levels in K-ras*(76–92%) and T187A�K-ras* (65–93%) lung cancer tissue com-pared with normal lung (Fig. 2B). Identical results were obtainedwhen p27 mRNA expression in lung cancer tissue microdissectedfrom the ENU-treated mice was compared with surroundingwild-type tissue (see Stabilization of p27 Interferes with ColonCancer Development; data not shown), indicating that, indepen-dent of the inducing agent, p27 down-regulation occurs bymodulating the p27 transcript levels in these tumors.

Regulation of the p27 Transcript in Human Breast Cancers. Our studyof K-ras*-induced tumorigenesis pointed toward a role of p27mRNA regulation as a mechanism involved in the control of p27expression in cancer cells. To ascertain whether p27 mRNAregulation occurs in human cancers, the p27 protein and mRNAlevels were assayed in breast cancers, a tumor entity in which p27protein expression levels predict the clinical outcome of theaffected patients (24). To determine whether low p27 proteinlevels correlate with decreased p27 mRNA, the quantity of p27transcript was analyzed in tumors with known p27 proteinamounts.

Twenty-seven breast tumors were assayed for p27 proteinexpression. From this analysis, 13 tumors were identified thatclearly exhibited either high nuclear or low nuclear p27 proteinstaining (Fig. 2C). The other 14 specimens had intermediatelevels of p27 protein or cytoplasmic staining. Cytoplasmic mis-localization of p27 in response to oncogenic signals contributesto functional inactivation of the p27 protein in breast cancertissues (25, 26). In such tumors overall p27 levels are of ques-tionable significance in predicting patients’ survival. We there-fore focused our analysis on breast cancer samples with pre-dominant nuclear p27 staining. The relative amount of p27transcript in the 13 tumors with nuclear staining was determinedby quantitative RT-PCR (Fig. 2D). Although the transcript levelsvary among the samples, all of the tumors in the lowest quartileof p27 mRNA amounts were also low for p27 protein, suggestingthat the decreases in transcript levels in these specimens causedthe decrease in p27 protein. Additionally, a previous study

Fig. 1. Expression of p27T187A does not prevent p27 down-regulation inlung cancer. (A) Survival curve of homozygous p27KO, p27T187A, and wild-type mice bred to K-ras*-expressing mice. Mice were killed when signs ofmorbidity were obvious. p27�/��K-ras* mice died significantly faster thanK-ras* and p27T187A�K-ras* mice, which died with comparable kinetics. (B)Immunohistochemical staining of lung cancer and surrounding normal tissueof the indicated strains with an antibody specific for p27kip1. No expression ofp27 or p27T187A was detectable in the tumor tissue, whereas normal sur-rounding tissues stained positive with a p27-specific antibody.

14010 � www.pnas.org�cgi�doi�10.1073�pnas.0606316103 Timmerbeul et al.

Dow

nloa

ded

by g

uest

on

May

23,

202

1

analyzing only p27 mRNA levels in breast cancers indicated thatoverall survival was significantly decreased for patients in the lowestquartile of p27 mRNA compared with patients with higher p27transcript levels (25). Overall, this pilot analysis indicated that inmore than half (62%) of human breast cancers down-regulation ofthe p27 protein may be caused by decreased abundance of the p27mRNA. This finding was consistent with our mouse studies, whichshowed that inactivation of the major pathway for p27 proteolysisdid not affect its down-regulation during lung tumorigenesis.

Stabilization of p27 Interferes with Colon Cancer Development. Toextend our analyses of the p27T187A mutation to a broader

spectrum of tumors, we injected 15-day-old mice i.p. with thecarcinogen ENU. We chose this agent because its carcinogeniceffect had previously been shown to be increased by loss of p27(3). These mice predominantly develop colon tumors and ulti-mately succumb to this disease by 1 year of age. The effect of theT187A mutation on colon tumorigenesis is considered in detailin Discussion. However, these mice also developed liver carci-nomas, albeit at a relatively lower frequency. Nevertheless, wefound that the p27 protein was down-regulated in these livercarcinomas equally in the wild-type and p27T87A mice (data notshown), which was consistent with what we had observed duringlung tumorigenesis.

As in our initial study in the K-ras* strain, our primaryquestion was whether p27 stabilization would lead to an im-proved overall survival of the mutant mouse as compared withwild-type controls. Therefore, we observed the entire colony ofENU-treated animals for a total of 58 weeks after injection of thecarcinogen and killed moribund mice only when clear signs offatal disease were obvious. Fig. 3A shows that wild-type andp27T187A animals died with comparable frequencies between18 and 58 weeks; thus, expression of p27T187A did not lead toan improved overall survival. Necropsies were performed on allmice followed by a complete histological analysis of all internalorgans. As shown in Fig. 3B most animals showed multipletumors at the time of necropsy with no obvious differences in the

Fig. 2. Lung cancer tissues express reduced levels of p27 mRNA. (A) Staining ofrepresentative tissue sections of wild-type p27 lung tissue and tumor tissue fromK-ras* and p27T187A�K-ras* mice with an antibody specific for the cks1 proteinshowed reduced expression of the cks1 protein in tumor tissue. (B) p27 geneexpression in K-ras* normal lung, K-ras* tumors, and p27T187A�K-ras* tumorswas determined by real-time PCR using probe and primers at the border of exons1 and 2. The amount of RNA in each sample was normalized to HPRT1 geneexpression. The fold change observed is relative to p27 transcript levels in theK-ras* normal lung sample. (C) p27 protein staining in representative tumorsections demonstrates high nuclear staining (a) and low nuclear staining (b). (b)Tumor tissue has low p27 staining, but infiltrating lymphocytes are positive forp27 protein. (D) p27 mRNA levels in the high and low p27 protein tumors wereassayed by quantitative real-time PCR. The quantity of RNA in each sample wasnormalized to HPRT1 gene expression.

Fig. 3. Expression of p27T187A interferes with intestinal tumor progression.(A) Survival curve of wild-type and p27T187A mice injected with ENU at 12days of age. No significant differences in survival were observed betweenwild-type and p27T187A mice. Mice were killed when clear signs of morbiditywere present. (B) Incidence of different tumor types that were identified atnecropsy and after histological examination of all organs. A significant dif-ference was observed in the numbers of intestinal adenomas and carcinomasthat arose in the p27T187A strain (P � 0.05) compared with p27 wild-typecontrols.

Timmerbeul et al. PNAS � September 19, 2006 � vol. 103 � no. 38 � 14011

CELL

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

May

23,

202

1

spectrum of tumors that arose in wild-type p27 and mutantp27T187A mice.

Interestingly, we did however observe a significant differencein the frequency of intestinal carcinomas with wild-type miceshowing double the number of carcinomas of the intestine at thetime of death (P � 0.05) compared with p27T187A (Fig. 3B).The development of intestinal carcinomas has been studiedextensively in humans and in transgenic mouse models, and it isgenerally agreed that adenomatous lesions represent an early,nonmalignant stage that can progress through a series of mo-lecular steps to invasive carcinomas (27). Several previousstudies that examined primary human colon cancer tissue as wellas experimental mouse models of intestinal tumorigenesis (28)showed a correlation between a more malignant phenotype andreduced p27 expression (29, 30). The observation that p27T187Amice developed fewer intestinal carcinomas after treatment withENU suggested that a T187-dependent degradation mechanismmight indeed represent a key step in the progression of thistumor.

Stabilization of p27 Affects Tumor Development in the Colon by aMechanism Independent of Controlling Cell Proliferation. As shownin Fig. 4C, untreated p27T187A mice had reduced numbers ofproliferating cell nuclear antigen (PCNA)-positive cells in theepithelial lining of the bowel compared with wild-type controls.

This reduction in the number of proliferating cells in theintestinal epithelium of p27T187A mice correlated with in-creased expression of p27. Fig. 4A e and f shows representativesections through p27 wild-type and p27T187A intestinal epithe-lia after staining with a p27-specific antibody. Fig. 4B shows aquantification of the number of cells that stained positive for p27in normal, adenomatous, and carcinoma tissues derived from thedifferent mouse strains. Interestingly, this baseline difference incell proliferation did not result in a decreased incidence ofENU-induced intestinal adenomas in p27T187A mice (31%)compared with wild-type (21%) mice, probably because of thefact that the intestinal tumors arising in the p27T187A strain hada tendency to accumulate at the adenoma stage rather thanprogressing to carcinoma. We then determined the expressionlevels of p27 in adenomas and carcinomas from wild-type andp27T187A mice. p27 staining was clearly detectable in the tissuesurrounding adenomatous lesions in wild-type mice (Fig. 4Ac),whereas the adenomas expressed p27 at much reduced levels.p27T187A mice expressed significantly more p27 protein in bothadenomas and surrounding normal tissue (Fig. 4Ad). However,the increased levels of p27 in adenomas from p27T187A mice didnot lead to a reduced proliferative activity, because p27 wild-typeand p27T187A mice showed similar levels of PCNA staining(Fig. 4C) and no significant differences in the number ofapoptotic cells (data not shown). This observation is in agree-ment with previous studies that showed correlations between thehistological differentiation and expression levels of p27 butfailed to detect a correlation between p27 expression levels andproliferative indices in human tissue samples representing var-ious stages of colon carcinogenesis (29, 30).

We concluded that the SCFskp2 pathway for p27 down-regulation is essential for the normal degradation of p27 inintestinal epithelia and facilitates the progression from intestinaladenoma to carcinoma. Moreover, in contrast to what weobserved in lung tumors, we found only a 20% reduction in thelevels of p27 mRNA in intestinal tumors arising in the p27T187Astrain compared with surrounding wild-type tissue (data notshown).

DiscussionBy comparing the mechanisms of p27 down-regulation in lungand intestinal cancers, we discovered that these pathways aremore complex than previously anticipated. Although phospho-rylation-induced degradation of the p27 protein is part of theprogression of intestinal adenomas to carcinomas, inhibition ofthis pathway did not prevent p27 down-regulation in lung tumors.We and others have previously shown that degradation of p27 bythe Skp2-dependent SCF complex controls the size, cell number,and ploidy of lung epithelial cells (31, 32). However, the resultsof this study point to an entirely different mechanism of p27regulation in lung epithelial cells under conditions of tumori-genesis, namely decreased p27 mRNA abundance. Our resultsunderscore the fact that p27 levels can be regulated by severalindependent mechanisms and suggest that lung cancer cells mayswitch between different modes of regulation during the courseof tumor development.

Previously, we and others demonstrated that the RAF-MEK-ERK and phosphatidylinositol 3-kinase-PDK1-AKT signaltransduction pathways downstream of Ras control p27 mRNAlevels (17, 33). We tested whether these signaling pathways wereactivated in lung tumors from K-ras* or p27T187A�K-ras* miceusing phosphoepitope-specific antibodies. AKT activation andphosphorylation of the downstream FoxO1–3 transcription fac-tors were not observed (data not shown). However, we detectedactivation of ERK1�2 in most of the lung tumors from bothmouse strains (data not shown), and the tumors expressing thelowest levels of p27 mRNA had strong phosphorylation ofERK1�2. Additionally, in human lung tumors we observed a

Fig. 4. The p27T187A protein is stabilized in intestinal tumors. (A) Charac-teristic sections through normal tissue (e and f), adenomas (c and d), andcarcinoma tissue (a and b) from wild-type and p27T187A mice after stainingwith a p27-specific antibody showing strong p27 expression in p27T187A-derived intestinal epithelium and wild-type noncancerous tissue but not inwild-type tumor tissue. (B) Percentage of cells that stained positive with a p27antibody in normal intestinal epithelium, adenomas, and carcinomas fromwild-type and p27T187A mice. (C) Percentage of PCNA-positive cells in normalintestinal epithelium, adenomas, and carcinomas from wild-type andp27T187A mice shows a significant (**, P � 0.05) difference in proliferativeactivity.

14012 � www.pnas.org�cgi�doi�10.1073�pnas.0606316103 Timmerbeul et al.

Dow

nloa

ded

by g

uest

on

May

23,

202

1

correlation between loss of p27 protein and phosphorylation ofERK1�2 (data not shown). In vascular smooth muscle cells ERKdecreases p27 transcript levels posttranscriptionally through the3� UTR (17), and c-myc, one of the transcription factors func-tioning downstream of ERK, represses p27 transcription (34).Therefore, ERK may regulate p27 transcript abundance in lungtumors by decreasing gene expression and�or RNA stability.

In addition to analyzing p27 mRNA regulation in lung tumors,we examined human breast cancers with nuclear p27 stainingwhere the protein levels are known to be predictive of outcome.Reduced survival for breast cancer patients with the lowestquartile of p27 transcript has been reported, but the p27 proteinlevels were not assessed. In our pilot study, all of the tumors inthe lowest quartile for p27 mRNA levels had low p27 protein,suggesting that p27 mRNA was regulated in a subset of humanbreast cancers. In our analysis low p27 transcript level was a goodpredictor of low p27 protein abundance, but the reciprocal wasnot true. No activation of ERK1�2 was detected in thesespecimens; however, other transcription factors known to regu-late p27 gene expression or different mechanisms affecting p27mRNA abundance may function in breast tumors, and the modeof p27 regulation may depend on the tissue types. Despiteincreasing evidence that p27 mRNA levels can be regulated bydiverse transcriptional and posttranscriptional mechanisms, ourin vivo studies represent the first demonstration that p27 mRNAmisregulation is important during tumorigenesis.

Our observation regarding the role of p27 turnover in intes-tinal tumor progression is also relevant to the choice of clinicalsettings in which inhibition of p27 degradation may have ther-apeutic benefit. Previous studies using purified extracts fromprimary colon cancer tissue suggested that increased degrada-tion of p27 by the SCFskp2-dependent mechanism correlated withincreasing aggressiveness of the respective tumor. Our previousanalysis of the p27T187A mouse strain had not shown majordifferences in p27 levels in most tissues compared with wild-typemice (13). However, in this study we directly show that theexpression of p27T187A decreases turnover of p27 in intestinalepithelia, thereby pointing to a specific role of the T187�SCFskp2

pathway in the regulation of intestinal cell proliferation. The highdegree of tissue specificity is also reflected by the observationthat although ENU treatment induced a wide variety of differentepithelial cancers only intestinal tumorigenesis was significantlyaffected in the p27T187A mouse as compared with wild type.Moreover, our observation that the progression of the adeno-matous polyp to invasive intestinal cancers is impaired when p27is stabilized points to a stage-specific role of p27 turnover in thedevelopment of colon cancers. In agreement with previousstudies in human colon cancer samples we did not find asignificant difference in the proliferative activity of tumors thatexpress high levels of p27 (T187A) and low levels of p27 (wildtype). Our results therefore suggest that p27 turnover increasesthe tendency of intestinal tumors to progress to a more malignantstate but that this function is not directly connected to a changein cell proliferation rates. Studies in mice that express a cyclin�cdk nonbinding form of p27 recently provided evidence for a roleof p27 in regulating the stability of a tissue-specific transcriptionfactor that controls cell differentiation (35). Whether similarmechanisms exist in intestinal epithelial cells and control tumorprogression dependent on p27 expression levels is currentlyunknown.

Thus, interference with the Skp2-mediated p27 turnover mighthave a role in disorders associated with high risk of developingmalignant intestinal carcinomas, like ulcerative colitis or the ad-enomatous polyposis syndrome. In light of the tissue-specific dif-ferences we observed in p27 misregulation, it will be equallyimportant to address whether pharmacological approaches aimedat preventing p27 down-regulation should target proteolytic, tran-scriptional, or posttranscriptional mechanisms.

Materials and MethodsMice and Treatment. The animals studied for ENU mutagenesiswere C57B6�J 129�Sv hybrid littermates produced from het-erozygous crosses. The animals for the K-ras study were all 129littermates produced from heterozygous crosses. The p27T187Aand p27Kip1 knockout mutations were recombined on to theK-rasLA2 chromosome. All mice were genotyped by PCR asdescribed (36) and were weaned at 3–4 weeks. The p27 wild-typeand p27T187A mice were i.p. injected with ENU (0.5 mmol pergram of mouse) at the age of 15 � 2 days (3). Mice were killedat the first sign of morbidity, which included excessive weightloss, rectal bleeding, abdominal swelling, hunched posture, andrapid breathing, or anemia. Complete necropsies of all internalorgans were performed for ENU-injected mice, and focusednecropsies of lungs were performed for K-ras*, p27�/��K-ras*,and p27T187A�K-ras* mice. For histological and immunohisto-chemical analysis of intestinal tumors longitudinal cross-sections(5 �m thick) through intestinal rolls spanning the entire intestinewere used. All hematoxylin and eosin-stained sections wereassessed by a mouse pathologist (K.K.). To distinguish adenomafrom intestinal carcinoma the following morphological criteriawere used: malignant cellular morphology, i.e., nuclear hyper-chromatism and pleomorphism; increased number of mitoticfigures, especially atypical mitosis; and irregularly shaped glandsand invasion through the lamina muscularis mucosae by atypi-cal glands. Statistical analysis was done on K-ras* (n � 65),p27�/��K-ras* (n � 13), p27T187A�K-ras* (n � 38), p27T187A(n � 19), and p27 wild-type (n � 14) mice after ENU treatment.The R program algorithm (www.R-project.org) was used todetermine statistical significance among mortality rates forK-ras*, p27�/��K-ras*, and p27T187A�K-ras* mice. Differ-ences in the occurrence of intestinal adenomas and carcinomaswere tested by using Fisher’s exact test and Pearson’s �2 Test.Kaplan–Meier survival curves were used to display the time totumor mortality. All experiments were done according to therules and regulations of the animal health authorities of the stateof Lower Saxony, Germany, and the Fred Hutchinson CancerResearch Center.

Immunostaining. Lung tissues from wild-type, p27���, K-ras*,and p27T187A�K-ras* mice were fixed in 10% formalin, pro-cessed as paraffin sections, and deparaffinized. p27 and Cks1immunostaining were performed according to the methodsprovided in the ImmunoCruz Staining Systems Kit (Santa CruzBiotechnology, Santa Cruz CA). Briefly, rehydrated tissue sec-tions were steamed for 30 min in 0.1% citrate acid and blockedin serum block for 20 min at room temperature. Samples weresubsequently incubated with anti-p27 antibody (C19, 1:100 di-lution; Santa Cruz Biotechnology) or anti-Cks1 antibody (C-term, 1:100 dilution; Zymed Laboratories, San Francisco, CA)for 1 h and biotinylated secondary antibody for 30 min, andvisualized by the chromogen 3�3�-diaminobenzidine. Intestinaltissue was stained by using the Histostain-Plus kit (ZymedLaboratories), p27 monoclonal antibody (1:50 dilution; Trans-duction Laboratories), and PCNA antibody (DAKO) followingthe manufacturers’ instructions. Samples were counterstainedwith hemalaun solution. The percentage of p27- and PCNA-positive cells was determined by counting the number of positivecells in 20 low-power (�10) magnification fields chosen atrandom and expressing the number of positive cells as a per-centage of all cells counted. At least three sections and threevisual fields of each collected sample were quantified.

RNA Isolation and RT-PCR. cDNA synthesis and semiquantitativeRT-PCR in intestinal cancer tissue were done as described previ-ously (32). For semiquantitative RT-PCR 1 �l of cDNA wasamplified by using the following cycle profile: denaturation at 94°C

Timmerbeul et al. PNAS � September 19, 2006 � vol. 103 � no. 38 � 14013

CELL

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

May

23,

202

1

for 30 s, annealing at 57°C for 30 s, and extension at 72°C for 45 sby using PlatinumR Taq Polymerase (Invitrogen) for 30 cycles. Forquantitative real-time PCR, RNA was extracted from pieces offrozen lung tumors (K-ras* and p27T187AK-ras*) and normal lungs(K-ras* and p27T187A) or sections of frozen human breast tumorsfollowing the TRIzol protocol (Invitrogen). cDNAs were generatedby reverse transcribing 1 �g of total RNA by using oligo dT andthe TaqMan reverse transcription kit (Applied Biosystems). ThecDNAs were diluted 1:10, and 5 �l was added to each reactioncontaining TaqMan master mix at a 1� concentration and the p27Mm00438167�g1 and Hs or Hprt1 Mm00446968�m1 andHs99999909�m1 Assay on Demand primers and probe (AppliedBiosystems). The results were verified by using a second p27 primerand probe set and a GAPDH control primer and probe (AppliedBiosystems). The sequences for the p27 primers are 5�-AGGAGAGCCAGGATGTCAGC-3� and 5�-CAGAGTTTGC-CTGAGACCCAA-3�, and the probe sequence is 5�-AGCCGC-CAGGCGGTGCCT-3�. Each 50-�l reaction was done in triplicate.

TaqMan real-time PCRs were performed by using an ABI PRISM7900HT sequence detector (Applied Biosystems) and analyzed bySDS2.2 software. Sequences for the Assay on Demand probeswere as follows: mouse p27, AGGAAGCGACCTGCTGCAG-AAGATT; mouse HPRT, AGGTTGCAAGCTTGCTGGTGA-AAAG; human p27, ACCTGCAACCGACGATTCTTCTACT;and human HPRT1, AGATGGTCAAGGTCGCAAGCTTGCT.

Laser microdissection of stained histological tissue sectionsand subsequent RNA isolation were performed essentially asdescribed (37).

We thank Tyler Jacks (Massachusetts Institute of Technology Center forCancer Research, Cambridge, MA) for providing K-rasLA2 transgenicmice and Holly Sundberg Malek, Lenhard Rudolph, and Matthew Ferofor valuable discussion and comments on the manuscript. Work in theN.P.M. laboratory is funded by the Deutsche Krebshilfe (Max EderProgram) and by the Deutsche Forschungsgemeinschaft. C.M.G.-E. isfunded by a National Cancer Institute postdoctoral fellowship.

1. Sherr CJ, Roberts JM (1999) Genes Dev 13:1501–1512.2. Fero ML, Rivkin M, Tasch M, Porter P, Carow CE, Firpo E, Polyak K, Tsai

LH, Broudy V, Perlmutter RM, et al. (1996) Cell 85:733–744.3. Fero ML, Randel E, Gurley KE, Roberts JM, Kemp CJ (1998) Nature

396:177–180.4. Di Cristofano A, De Acetis M, Koff A, Cordon-Cardo C, Pandolfi PP (2001)

Nat Genet 27:222–224.5. Kang-Decker N, Tong C, Boussouar F, Baker DJ, Xu W, Leontovich AA,

Taylor WR, Brindle PK, van Deursen JM (2004) Cancer Cell 5:177–189.6. Park MS, Rosai J, Nguyen HT, Capodieci P, Cordon-Cardo C, Koff A (1999)

Proc Natl Acad Sci USA 96:6382–6387.7. Slingerland J, Pagano M (2000) J Cell Physiol 183:10–17.8. Loda M, Cukor B, Tam SW, Lavin P, Fiorentino M, Draetta GF, Jessup JM,

Pagano M (1997) Nat Med 3:231–234.9. Blain SW, Scher HI, Cordon-Cardo C, Koff A (2003) Cancer Cell 3:

111–115.10. Carrano AC, Eytan E, Hershko A, Pagano M (1999) Nat Cell Biol 1:193–199.11. Sheaff RJ, Groudine M, Gordon M, Roberts JM, Clurman BE (1997) Genes

Dev 11:1464–1478.12. Vlach J, Hennecke S, Amati B (1997) EMBO J 16:5334–5344.13. Malek NP, Sundberg H, McGrew S, Nakayama K, Kyriakides TR, Roberts JM

(2001) Nature 413:323–327.14. Kullmann M, Gopfert U, Siewe B, Hengst L (2002) Genes Dev 16:3087–3099.15. Medema RH, Kops GJ, Bos JL, Burgering BM (2000) Nature 404:782–787.16. Kamura T, Hara T, Matsumoto M, Ishida N, Okumura F, Hatakeyama S,

Yoshida M, Nakayama K, Nakayama KI (2004) Nat Cell Biol 6:1229–1235.17. Sakakibara K, Kubota K, Worku B, Ryer EJ, Miller JP, Koff A, Kent KC, Liu

B (2005) J Biol Chem 280:25470–25477.18. Shin I, Yakes FM, Rojo F, Shin NY, Bakin AV, Baselga J, Arteaga CL (2002)

Nat Med 8:1145–1152.19. Johnson L, Mercer K, Greenbaum D, Bronson RT, Crowley D, Tuveson DA,

Jacks T (2001) Nature 410:1111–1116.20. Kawana H, Tamaru J, Tanaka T, Hirai A, Saito Y, Kitagawa M, Mikata A,

Harigaya K, Kuriyama T (1998) Am J Pathol 153:505–513.

21. Spruck C, Strohmaier H, Watson M, Smith AP, Ryan A, Krek TW, Reed SI(2001) Mol Cell 7:639–650.

22. Ganoth D, Bornstein G, Ko TK, Larsen B, Tyers M, Pagano M, Hershko A(2001) Nat Cell Biol 3:321–324.

23. Inui N, Kitagawa K, Miwa S, Hattori T, Chida K, Nakamura H, Kitagawa M(2003) Biochem Biophys Res Commun 303:978–984.

24. Porter PL, Malone KE, Heagerty PJ, Alexander GM, Gatti LA, Firpo EJ,Daling JR, Roberts JM (1997) Nat Med 3:222–225.

25. Liang J, Zubovitz J, Petrocelli T, Kotchetkov R, Connor MK, Han K, Lee JH,Ciarallo S, Catzavelos C, Beniston R, et al. (2002) Nat Med 8:1153–1160.

26. Viglietto G, Motti ML, Bruni P, Melillo RM, D’Alessio A, Califano D, VinciF, Chiappetta G, Tsichlis P, Bellacosa A, et al. (2002) Nat Med 8:1136–1144.

27. Weitz J, Koch M, Debus J, Hohler T, Galle PR, Buchler MW (2005) Lancet365:153–165.

28. Philipp-Staheli J, Kim KH, Payne SR, Gurley KE, Liggitt D, Longton G, KempCJ (2002) Cancer Cell 1:355–368.

29. Palmqvist R, Stenling R, Oberg A, Landberg G (1999) J Pathol 188:18–23.30. Ciaparrone M, Yamamoto H, Yao Y, Sgambato A, Cattoretti G, Tomita N,

Monden T, Rotterdam H, Weinstein IB (1998) Cancer Res 58:114–122.31. Nakayama K, Nagahama H, Minamishima YA, Miyake S, Ishida N,

Hatakeyama S, Kitagawa M, Iemura S, Natsume T, Nakayama KI (2004) DevCell 6:661–672.

32. Kossatz U, Dietrich N, Zender L, Buer J, Manns MP, Malek NP (2004) GenesDev 18:2602–2607.

33. Mirza AM, Gysin S, Malek N, Nakayama K, Roberts JM, McMahon M (2004)Mol Cell Biol 24:10868–10881.

34. Yang W, Shen J, Wu M, Arsura M, FitzGerald M, Suldan Z, Kim DW, HofmannCS, Pianetti S, Romieu-Mourez R, et al. (2001) Oncogene 20:1688–1702.

35. Nguyen L, Besson A, Heng JI, Schuurmans C, Teboul L, Parras C, Philpott A,Roberts JM, Guillemot F (2006) Genes Dev 20:1511–1524.

36. Malek NP, Sundberg H, McGrew S, Nakayama K, Kyriakides TR, Roberts JM,Kyriakidis TR (2001) Nature 413:323–327.

37. Lehmann U, Celikkaya G, Hasemeier B, Langer F, Kreipe H (2002) Cancer Res62:6634–6638.

14014 � www.pnas.org�cgi�doi�10.1073�pnas.0606316103 Timmerbeul et al.

Dow

nloa

ded

by g

uest

on

May

23,

202

1