p27kip1 Expression in Rectal Cancer Correlates with Disease-Free Survival

Journal of Surgical Research 92, 78 – 84 (2000) doi:10.1006/jsre.2000.5871, available online at http://www.idealibrary.com on

p27 kip1 Expression in Rectal Cancer Correlates with Disease-Free Survival Klaus Gu¨nther, M.D.,* ,1 Andreas Jung, Ph.D.,† ,1 Uwe Vo¨lker, M.D.,† Martin Meyer, M.D.,‡ Thomas Brabletz, M.D.,† Klaus E. Matzel, M.D.,* Marc A. Reymond, M.D.,§ Thomas Kirchner, M.D.,† and Werner Hohenberger, M.D.* *Department of Surgery, †Department of Pathology, and ‡Bavarian Epidemiologic Cancer Registry, University of Erlangen–Nuremberg, Erlangen, Germany; and §Department of Surgery, University of Magdeburg, Magdeburg, Germany Submitted for publication December 15, 1999

of p27 kip1 remain unclear, and further investigation is needed. © 2000 Academic Press Key Words: rectal cancer; metastases; prognosis; disease-free survival; p27 kip1; immunohistochemistry.

Background. The cell-cycle inhibitor p27 kip1 is a potential tumor suppressor and might serve as a prognostic marker in rectal cancer, in particular with regard to patient selection for adjuvant therapy. Materials and methods. Immunohistochemical analysis was performed, using an anti-p27 kip1 monoclonal antibody, on paraffin sections of two matched [age, gender, UICC stage, year of operation (1982–1991)] groups of patients (n ⴝ 2 ⴛ 82) with rectal carcinoma curatively treated by surgery alone. The groups differed only in subsequent metachronous distant metastatic spread. All patients had to meet the selection criterion “free of local disease,” in order to exclude surgical influence. Follow-up was prospective (median of 74 months). The intensity of staining (ⴚ, ⴙ, ⴙⴙ, ⴙⴙⴙ) and rate of positive cells (as a percentage of total tumor volume) were judged separately for cytoplasms and nuclei. Results. On multivariate analysis, cytoplasmic staining intensity proved to be the best prognostic factor of disease-free survival and approached statistical significance (P ⴝ 0.0552, Cox regression). On univariate analysis, considering cytoplasmic staining alone, intensely stained (ⴙⴙⴙ) tumors showed significantly poorer disease-free survival (vs ⴙⴙ, ⴙ, ⴚ; Kaplan– Meier, logrank, P ⴝ 0.0185). Conclusions. The demonstrated correlation between cytoplasmic compartmentalization of p27 kip1 and increased metastatic spread as well as diseasefree survival underscores the role of p27 kip1 in rectal cancer. However, since other reports emphasize the importance of nuclear p27 kip1 expression, the mechanisms of steady-state and subcellular distribution 1

1

INTRODUCTION

A retrospective analysis of all patients with rectal cancer curatively treated by surgery alone at the Department of Surgery of the University of Erlangen between 1974 and 1991 revealed a decrease in the 5-year local recurrence rate from 39.4 to 9.8% (P ⬍ 0.0001), paralleling an improvement in the 5-year survival rate from 50 to 71% (P ⬍ 0.0001) [1]. However, despite this good surgical result, the overall rate of initial metachronous distant spread remained constant at about 25% (P ⫽ 0.44) [1]. This finding indicates that the ground is already prepared for subsequent spread by the time of primary surgery and cannot be reversed by surgery. Thus, this proposed minimal residual disease would be the main target for adjuvant treatment options. However, the selection of appropriate patients for additional treatment is still a major problem. Up to now the best prognostic criterion has been the UICC stage. After curative surgery for rectal cancer, the current policy of delivering additional treatment to all stage II (Dukes B) and stage III (Dukes C) patients appears too undifferentiated. This is also demonstrated by our analysis: 8% (stage I), 16% (stage II), and 40% (stage III) of patients remaining free from local recurrence develop initial metachronous distant metastases [1]. This shows that additional therapy administered routinely in stage II and III represents

These authors contributed equally to this work.

0022-4804/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

78

¨ NTHER ET AL.: p27 kip1 AND DISEASE-FREE SURVIVAL IN RECTAL CANCER GU

TABLE 1 Patient Characteristics No distant metastases

Distant metastases

61.6 1.2:1 4 16 62 82

60.5 1.5:1 5 14 63 82

Age (years) Gender (M:F) UICC I UICC II UICC III All cases a

ns a ns a ns a ns a ns a

Not significant.

overtreatment in 60% of stage III, and 84% of stage II, patients. Against this clinical background, a search for better prognostic factors, especially at the molecular level, is currently being undertaken. Unfortunately, no such factors have thus far attained clinical relevance. Recently, the prognostic value in terms of survival and metastazation of a new candidate factor, the cell cycle inhibitor p27 kip1, was demonstrated for several human tumors including colorectal cancer [2– 4]. The aim of our study was to investigate the role of p27 kip1 in terms of distant metastatic spread in our rectal cancer patients. MATERIALS AND METHODS Patient selection and tumor specimens. Between 1982 and 1991, at the Department of Surgery, University of Erlangen, Germany, 846 patients underwent curative surgery for rectal cancer. The adenocarcinoma was considered a primary rectal tumor if it was located within 16 cm of the anal verge. To exclude any influence of surgical technique, only the 737 patients who did not develop local recurrence were subsequently studied. No patient received either adjuvant chemotherapy or radiotherapy. All patients were investigated prospectively in the Tumor Center of the University of Erlangen for disease recurrence and death. One hundred sixty-nine patients developed distant metastases with no local recurrence. All resected specimens were analyzed in accordance with the TNM system [6]. All the specimens were reviewed under the supervision of the same senior pathologist. Eighty-two formalin-fixed, paraffin-embedded tumor samples obtained from the patients who developed distant spread were retrieved from the archives of the Department of Pathology of the University of Erlangen. These samples were matched for UICC stage, patient age, patient gender, and year of operation with 82 tumors from patients from the same population who remained disease-free (Table 1). These 164 paraffin-embedded samples formed the basis for the analysis in the present study. The median prospective follow-up was 74 months (range 2.5–190 months). Metachronous distant metastasation. The definition for metachronous metastasation is as follows: first tumor reappearance in any organ system or lymph nodes other than the primary tumor site or regional lymph nodes (i.e., those of the perirectal, superior rectal, inferior mesenteric, and internal iliac arteries [6]) after curative surgical treatment (R0 resection) of the primary tumor. To exclude synchronous spread, in addition to clinical examination all patients received abdominal sonography and CT scan, as well as chest X-ray before primary surgery. As in our patients metachronous distant lymph node metastases were not observed, the predominant sites of distant recurrence were liver, lung and bone, in descending fre-

79

quency. The diagnoses were made clinically via sonography (liver) or radiographically (lung and bone) and confirmed by CT scan in nearly all cases. In more than half of the patients a definitive histological diagnosis (autopsy, resected specimen, biopsy) could be made. Immunohistopathology. After being mounted on APES-covered slides (APES, tri-ethoxy-silyl-propylamine, Merck, Darmstadt, Germany) 4-␮m-thick sections were deparaffinized in xylol and ethanol using standard techniques. Antigen retrieval was performed by boiling for 10 min at 800 W and 10 min at 600 W, in 10 mM citrate, pH 6.0, using a microwave oven. Mouse anti-human p27 kip1 monoclonal antibody (Transduction Laboratories, Lexington, KY) was then applied in a 1:200 dilution for 60 min. After buffering, a secondary rabbit anti-mouse antibody (Dako, Hamburg, Germany) and then a Strept-AB-Complex AP (Dako, Hamburg, Germany) were applied for 30 min each. This was followed by an application of Fast Red TR (F2768, Sigma, Munich, Germany) for another 20 min. The slides were counterstained with Mayer⬘s solution, rehydrated, and mounted using Aquatex (Merck). Negative controls omitting the primary antibody were included. Normal rectal epithelial cells and tumor-infiltrating lymphocytes (TIL) were chosen as internal positive control, and all quantifications of staining were based on the staining of these cells. Analysis was performed independently in a blinded fashion by one and the same experienced pathologist. Nuclear and cytoplasmic staining was evaluated separately for intensity (none, weak, moderate, intense: ⫺, ⫹, ⫹⫹, ⫹⫹⫹) and rate of positive cells (as a percentage of total tumor volume). Statistics. Tumors with and without distant metastases were matched for UICC stage, patient age, patient gender, and year of operation. Primary outcome criteria were disease-free survival and distant spread. Comparisons were made with the ␹ 2 test for nominal data and with the Mann–Whitney U test for ordinal data. Survival distributions were estimated using the method of Kaplan–Meier with censoring of survivors. The logrank statistic with trend was used to compare survival distributions. A Cox regression analysis was made to assess the independent values of all investigated variables. To detect a possible time trend, the staining results of four subperiods (1982–1984; 1985–1987; 1988 –1989; 1990 –1991) were compared using the ␹ 2 test. A P value of less than 0.05 was considered statistically significant. All tests were two-tailed.

RESULTS

The characteristics of the 164 patients included in this study are summarized in Table 1. Immunohistochemical analysis of tissue sections obtained from these patients revealed that in normal rectal epithelium, p27 kip1 was expressed in the nuclei of the surface cells, but not in those from the crypt bases or walls (Fig. 1). This staining behavior “anti-parallels” the proliferation of the cells and confirms p27 kip1 as a marker for resting nonproliferating cells. Additionally, the nuclei of TIL also regularly expressed p27 kip1. Cytoplasmic or membranous staining was never observed in normal mucosa. Rectal adenocarcinomas showed three main categories of p27 kip1 immunostaining: cytoplasmic staining (Fig. 2a), nuclear staining (Fig. 2b), and absence of staining (Fig. 2c). For the first two categories, intensity (⫺, ⫹, ⫹⫹, ⫹⫹⫹) (Table 2) and rate of positive cells (as a percentage of total tumor volume) (Table 3) were considered. Rates of Positive Cells In brief, the rates of positive tumor cells (any ⫹, ⫹⫹, ⫹⫹⫹) were rather low (see Table 3). Only 22.6% (n ⫽

80

JOURNAL OF SURGICAL RESEARCH: VOL. 92, NO. 1, JULY 2000

FIG. 1. Staining pattern in normal rectal epithelial cells. Rectal crypt (three different magnifications, ⫻100, ⫻200, ⫻400, from left to right) demonstrating exclusive nuclear p27 kip1 expression in the superficial cells.

37) of the adenocarcinomas expressed p27 kip1 in more than 30% of the nuclei and 18.9% (n ⫽ 31) in more than 30% of the cytoplasms. No association was detected between any positivity rate and the occurrence of distant metastases or disease-free survival.

Time Trend The rates of intensely (⫹⫹⫹) stained cytoplasm were 22% (1982–1984), 23% (1985–1987), 23% (1988 –1989), and 35% (1990 –1991). The differences between these four time periods were not significant (P ⫽ 0.49; ␹ 2 test).

Staining Intensity DISCUSSION

Nuclear staining was absent (⫺) in 15.9% (n ⫽ 26) of the tumors, while in 51.8% (n ⫽ 85) p27 kip1 expression was intense (⫹⫹⫹) (see Table 2). The equivalent figures for cytoplasmic staining were 29.3% (n ⫽ 48) and 25.6% (n ⫽ 42), respectively. In multivariate analysis [(i) cytoplasmic staining intensity vs (ii) nuclear staining intensity vs (iii) rates of cytoplasmic positive (⫹, ⫹⫹, ⫹⫹⫹) cells vs (iv) rates of nuclear positive (⫹, ⫹⫹, ⫹⫹⫹) cells] cytoplasmic staining intensity was found to be the best prognostic factor for disease-free survival and approached statistical significance (P ⫽ 0.0552, Cox regression). With regard to cytoplasmic staining alone, the P value was 0.380 when each of the four categories (⫺, ⫹, ⫹⫹, ⫹⫹⫹) was compared with each of the others. In univariate analysis, tumors with intensely stained (⫹⫹⫹) cytoplasm showed significantly poorer diseasefree survival than the other three categories (⫹⫹, ⫹, ⫺; Kaplan–Meier, logrank, P ⫽ 0.0185) (Fig. 3).

p27 kip1 and Its Role in the Cell Cycle The cell cycle in mammalian cells comprises four main phases: gap1 (G1), synthesis (S), gap2 (G2) and mitosis (M). During the S-phase the DNA is doubled and in the M-phase the replicated chromosomes separate and cells divide. The processes preceding the Sand M-phases take place during the two gap phases. The phase transitions are marked by checkpoints (arrows in Fig. 4), which can be passed only when active complexes composed of cyclins and cyclin-dependent kinases (CDKs) are present. The activities of these complexes are downregulated—and the cell cycle thus decelerates— by two families of CDK inhibitors (CKIs): the INK4 group and the CIP/KIP group. The latter group, which is of interest here, comprises p21 waf1/cip1, p27 kip1, and p57 kip2 and has a more or less nonselective inhibitory effect on several cyclin–CDK complexes. Of these, p27 kip1—acting via the inhibition of cyclin

¨ NTHER ET AL.: p27 kip1 AND DISEASE-FREE SURVIVAL IN RECTAL CANCER GU

81

FIG. 2. Staining patterns in rectal adenocarcinomas (⫻400); (a) intense cytoplasmic staining; (b) intense nuclear staining; (c) no staining.

D–CDK4 and mainly cyclin E–CDK2 complexes— affects cell cycle arrest at the G1–S transition, the first checkpoint— called restriction (R) point (start point). In addition to this inhibitory effect on the cell cycle, p27 kip1 is also a promoter of apoptosis, via mechanisms that have yet to be specified. Thus, p27 kip1 acts as a potent depressor of cell proliferation (Fig. 4). Physiological Staining Pattern of p27 kip1 The staining pattern of normal rectal mucosa in our material is in good accord with the theoretical background and the results of staining reported by others [3, 5, 7]: p27 kip1 was expressed exclusively in the nuclei of the resting superficial cells, but not in those of the proliferating cells in the crypt bases or walls. Additionally, p27 kip1 was regularly expressed in the nuclei of the highly differentiated, nonproliferating TIL. In view of these results, we regard our staining procedure to be reliable.

Regulation and Subcellular Compartmentalization of p27 kip1 The regulatory mechanisms of p27 kip1 are not yet fully understood. An increasing number of different regulative pathways have been described on the basis of cell culture experiments, mice models, and investigations in human tumors [3, 8 –14]. Currently, it is generally accepted that the p27 kip1 gene does not behave like other tumor suppressor genes, since mutations in it have never been found to be a reason for human carcinogenesis [3, 15]. Thus, regulation of p27 kip1 should be regarded as a posttranscriptional event and the following mechanisms at or after translation have been shown: —stimulation of p27 kip1 via extracellular antiproliferative signals such as TGF-␤ [16]; —phosphorylation of p27 kip1 by cyclin E–CDK2 complexes [9, 10];

82

JOURNAL OF SURGICAL RESEARCH: VOL. 92, NO. 1, JULY 2000

TABLE 2 Staining Results: Intensities Nuclear staining

Cytoplasmic staining Total

Staining intensity No staining Slight staining Moderate staining Intense staining All cases a b

(⫺) (⫹) (⫹⫹) (⫹⫹⫹)

⫺met a n

⫹met b n

n

12 8 19 43 82

14 13 13 42 82

26 21 32 85 164

Total %

⫺met a n

⫹met b n

n

%

15.9 12.8 19.5 51.8 100.0

29 19 21 13 82

19 14 20 29 82

48 33 41 42 164

29.3 20.1 25.0 25.6 100.0

⫺met, without distant metastases. ⫹met, with distant metastases.

— ubiquitin/proteasome-dependent p27 kip1 degradation [3, 11]; —sequestration of p27 kip1 by unknown Myc-inducible proteins [8]; — binding of p27 kip1 to cyclin D–CDK4 complexes [13]; —inactivation of p27 kip1 by the viral E1A oncoprotein [12].

ries [17–19]. Of relevance to these findings is another discovery made by Tomoda and co-workers, who identified the export rate of p27 kip1 out of the nucleus to be another critical regulative mechanism of p27 kip1 expression [20]. In a mouse model, they identified the p38 protein— encoded by the Jab1 gene [21]—as being responsible for the p27 kip1 translocation from nucleus to cytoplasm and for decreasing the total amount of p27 kip1 in the cell by accelerating its degradation. Thus, p38 is another downregulator of p27 kip1. In conclusion, we believe that p27 kip1 expression is regulated by a complex system, based mainly on the import– export ratio, which is responsible for the subcellular compartmentalization of p27 kip1— cytoplasm or nucleus—and the stimulating or degrading influences of p27 kip1 in these areas. It is important to keep in mind that p27 kip1 is not automatically localized in the nucleus after synthesis and that other mechanisms may be responsible for the subcellular localization of the p27 kip1 protein. Surprisingly, our study showed cytoplasmic p27 kip1 expression to be a common event in rectal cancer and

Physiologically, p27 kip1 protein is first synthesized in the cytoplasm and then transferred via a nuclear localization signal into the nucleus, the main site of p27 kip1 action. As a consequence, the main thrust of p27 kip1 investigation originally focused on the nucleus. The result was that, for the first time in colorectal cancer, decreased nuclear p27 kip1 levels were shown to be related to metastatic behavior and a poorer prognosis [3, 11]. The authors considered the ubiquitin/proteasomedependent degradation to be responsible for the observed p27 kip1 downregulation [3, 11]. Over time, a number of publications have reported finding cytoplasmic p27 kip1 accumulation in various tumors—which did not seem to fit in with prior theo-

TABLE 3 Staining Results: Rates of Positive Cells (as a Percentage of Total Tumor Volume) Nuclei (any ⫹, ⫹⫹, ⫹⫹⫹)

Cytoplasm (any ⫹, ⫹⫹, ⫹⫹⫹)

Total % of total tumor volume

⫺met a n

⫹met b n

n

0% ⫽ no (⫺) staining 1% 5% 10% 20% 30–100% All cases

12 8 14 17 13 18 82

14 10 10 18 11 19 82

26 18 24 35 24 37 164

a b

⫺met, without distant metastases. ⫹met, with distant metastases.

Total %

⫺met a n

⫹met b n

n

%

15.9 11.0 14.6 21.3 14.6 22.6 100.0

29 0 2 21 21 9 82

19 3 5 23 10 22 82

48 3 7 44 31 31 164

29.3 1.8 4.3 26.8 18.9 18.9 100.0

¨ NTHER ET AL.: p27 kip1 AND DISEASE-FREE SURVIVAL IN RECTAL CANCER GU

83

FIG. 3. Disease-free survival as a function of cytoplasmic staining. Significantly decreased disease-free survival of patients with tumors showing intensive cytoplasmic p27 expression (⫹⫹⫹ vs ⫹⫹, ⫹, ⫺; Kaplan–Meier, logrank, P ⫽ 0.0185).

that this can be discussed only against the background of the above-mentioned theory. Of greater interest was the fact that in our series, not nuclear but cytoplasmic staining was correlated with distant metastases and disease-free survival: Tumors with intensive cytoplasmic p27 kip1 expression (⫹⫹⫹) showed a significantly poorer outcome than the other categories (⫺, ⫹, ⫹⫹), irrespective of their nuclear staining. We can offer no conclusive explanation for our observations. It can only be speculated that in primary tumors a progressive block of the transfer of p27 kip1 into the nucleus, or a higher export rate out of the nucleus, might lead to the cytoplasmic accumulation described. One would therefore expect a loss of CDK inhibition and subsequently a higher proliferation rate.

FIG. 4.

Prognostic Relevance of p27 kip1 The main aim of our study was to determine the prognostic value of p27 kip1 expression in primary adenocarcinoma of the rectum with regard to metachronous distant metastases and disease-free survival. To study this phenomenon exclusively with no confounding factors, we developed a highly selected patient population: To exclude any effects of radiation or chemotherapy, only patients undergoing surgery alone were studied. To rule out the prognostic factor “surgeon,” only patients who remained free of local disease were considered. In this way, optimal surgical quality was guaranteed. We were thus able to obtain two groups of 82 patients each, matched for age, gender, UICC stage,

p27 kip1 and its role in the cell cycle.

84

JOURNAL OF SURGICAL RESEARCH: VOL. 92, NO. 1, JULY 2000

and year of operation and differing only in the main objective of this study: occurrence of metachronous distant spread. Each group thus contained a respectable number of cases, made possible by the large population of 846 patients with rectal cancer curatively treated between 1982 and 1991. This, together with a median follow-up time of 74 months, provided us, we believe, with a reliable tool for investigating the potential prognostic value of p27 kip1 with regard to distant tumor spread. The quality of the paraffin-embedded specimens appeared to be consistent over the entire period 1982– 1991, as the rates of intensely stained (⫹⫹⫹) cytoplasms did not differ significantly among the four subperiods [22% (1982–1984), 23% (1985–1987), 23% (1988 –1989), 35% (1990 –1991)]. An influence of time on the p27 kip1 antigen stability can therefore be excluded. The central outcome of this study is that adenocarcinomas of the rectum that showed intensive cytoplasmic expression (⫹⫹⫹) of p27 kip1 are associated with significantly decreased disease-free survival (vs ⫹⫹, ⫹, ⫺). Only cytoplasmic expression of p27 kip1 appeared to be relevant for distant spread. This is further underscored by our finding that in multivariate analysis, not nuclear but cytoplasmic staining intensity was best linked with disease-free survival. In conclusion, the expression and distribution of p27 kip1 seem to play a role in regulative processes leading to the development of distant metastases in rectal cancer. However, the mechanisms of regulation and subcellular compartmentalization of p27 kip1 need further investigation. In particular, our findings of the cytoplasmic p27 kip1 staining association with distant spread, which are in contrast to the previously described importance of nuclear staining, must carefully be examined. In the meantime, no conclusive statements regarding the prognostic value of p27 kip1 in this tumor entity should be made.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15. 16.

17.

REFERENCES 1.

Ko¨ckerling, F., Reymond, M. A., Altendorf-Hofmann, A., and Hohenberger, W. Influence of surgery on metachronous distant metastases and survival in rectal cancer. J. Clin. Oncol. 16: 1, 1998. 2. Lloyd, R. V., Erickson, L. A., Jin, L., Kulig, E., Qian, X., Cheville, J. C., and Scheithauer, B. W. P27kip1: A multifunctional cyclin-dependent kinase inhibitor with prognostic significance in human cancers. Am. J. Pathol. 154: 313, 1999. 3. Loda, M., Cukor, B., Tam, S. W., Lavin, P., Fiorentino, M., Draetta, G. F., Jessup, J. M., and Pagano, M. Increased proteasome-dependent degradation of the cyclin-dependent kinase inhibitor p27 in aggressive colorectal carcinomas. Nat. Med. 3: 231, 1997.

18.

19.

20.

21.

Belluco, C., Esposito, G., Vieceli, G., Bertorelle, R., Fassina, A., Fedeli, U., Del Mistro, A., Chieco-Bianchi, L., Nitti, D., and Lise, M. Cell cycle inhibitors and prognosis in T2-3N0M0 colorectal cancer. Eur. Surg. Res. 30(Suppl): 74, 1998. Thomas, G. V., Szigeti, K., Murphy, M., Draetta, G., Pagano, M., and Loda, M. Down-regulation of p27 is associated with development of colorectal adenocarcinoma metastases. Am. J. Pathol. 153: 681, 1998. UICC. In Sobin, L. H., and Wittekind, C. (eds.), TNM Classification of Malignant Tumors. New York/Chichester: Wiley–Liss, 1997. ¨ berg, A., and Landberg, G. ProgPalmqvist, R., Stenling, R., O nostic significance of p27(Kip1) expression in colorectal cancer: A clinico-pathological characterization. J. Pathol. 188: 18, 1999. Vlach, J., Hennecke, S., Alevizopoulos, K., Conti, D., and Amati, B. Growth arrest by the cyclin-dependent kinase inhibitor p27Kip1 is abrogated by c-Myc. EMBO J. 15: 6595, 1996. Vlach, J., Hennecke, S., and Amati, B. Phosphorylationdependent degradation of the cyclin-dependent kinase inhibitor p27. EMBO J. 16: 5334, 1997. Sheaff, R. J., Groudine, M., Gordon, M., Roberts, J. M., and Clurman, B. E. Cyclin E-CDK2 is a regulator of p27Kip1. Genes Dev. 11: 1464, 1997. Pagano, M., Tam, S. W., Theodoras, A. M, Beer-Romero, P., Del Sal, G., Chau, V., Yew, P. R., Draetta, G. F., and Rolfe, M. Role of the ubiquitin-proteasome pathway in regulating abundance of the cyclin-dependent kinase inhibitor p27. Science 269: 682, 1995. Mal, A., Poon, R. Y., Howe, P. H., Toyoshima. H., Hunter, T., and Harter, M. L. Inactivation of p27kip1 by the viral E1A oncoprotein in TGF␤-treated cells. Nature 380: 262, 1996. Polyak, K., Kato, J. Y., Solomon, M. J., Sherr, C. J., Massague, J., Roberts, J. M., and Koff, A. P27kip1, a cyclin–Cdk inhibitor, links transforming growth factor-␤ and contact inhibition to cell cycle arrest. Genes Dev. 8: 9, 1994. Santoni-Rugiu, E., Jensen, M. R., Factor, V. M., and Thorgeirsson, S. S. Acceleration of c-myc-induced hepatocarcinogenesis by co-expression of transforming growth factor (TGF)-␣ in transgenic mice is associated with TGF-␤1 signaling disruption. Am. J. Pathol. 154: 1693, 1999. Steeg, P. S., and Abrams, J. S. Cancer prognostics: Past, present and p27. Nat. Med. 3: 152, 1997. Reynisdottir, I., Polyak, K., Iavarone, A., and Massague, J. Kip/Cip and Ink4 Cdk inhibitors cooperate to induce cell cycle arrest in response to TGF-␤. Genes Dev. 9: 1831, 1995. Singh, S. P., Lipman, J., Goldman, H., Ellis, F. H., Aizenman, L., Cangi, M. G., Signoretti, S., Chiaur, D. S., Pagano, M., and Loda, M. Loss or altered subcellular localization of p27 in Barrett’s associated adenocarcinoma. Cancer Res. 58: 1730, 1998. Orend, G., Hunter, T., and Ruoslahti, E. Cytoplasmic displacement of cyclin E– cdk2 inhibitors p21Cip1 and p27Kip1 in anchorage-independent cells. Oncogene 16: 2575, 1998. Capodieci, P., Gerald, W., Woodruff, J., Cordon-Cardo, C., and Koff, A. Cytoplasmic localization of p27kip1 in sarcoma. Proc. Am. Assoc. Cancer Res. 40: 687, 1999. Tomoda, K., and Kubota, Y. J. K. Degradation of the cyclindependent-kinase inhibitor p27kip1 is instigated by Jab1. Nature 398: 160, 1999. Claret, F. X., Hibi, M., Dhut, S., Toda, T., and Karin, M. A new group of conserved coactivators that increase the specificity of AP-1 transcription factors. Nature 383: 453, 1996.