• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Dec 7, 2004; 101(49): 17204–17209.
Published online Nov 29, 2004. doi:  10.1073/pnas.0407693101
PMCID: PMC535400
Medical Sciences

A critical role for p27kip1 gene dosage in a mouse model of prostate carcinogenesis


In human prostate cancer, the frequent down-regulation of p27kip1 protein expression is correlated with poor clinical outcome, yet p27kip1 rarely undergoes mutational inactivation. Here, we investigate the consequences of reducing or eliminating p27kip1 function for prostate carcinogenesis in the context of a mouse modeling lacking the Nkx3.1 homeobox gene and the Pten tumor suppressor. Unexpectedly, we find that triple mutant mice heterozygous for a p27kip1 null allele (Nkx3.1+/– or –/–; Pten+/–; p27+/–) display enhanced prostate carcinogenesis, whereas mice that are homozygous null for p27kip1 (Nkx3.1+/– or –/–; Pten+/–; p27–/–) show inhibition of cancer progression. Expression profiling reveals that Cyclin D1 is highly up-regulated in compound p27kip1 heterozygotes, but is down-regulated in the compound p27kip1 homozygous mutants. Using RNA interference in prostate cancer cell lines with distinct p27kip1 gene doses, we show that prostate tumorigenicity depends on levels of p27kip1 and that the consequences of p27kip1 gene dosage can be attributed, in part, to altered levels of Cyclin D1. Our findings suggest that p27kip1 possesses dosage-sensitive positive as well as negative modulatory roles in prostate cancer progression.

Keywords: Cyclin D1, expression profiling, p27, prostate cancer

Prostate cancer now represents a serious health problem worldwide (1). Despite intense investigations, it has remained a daunting task to elucidate the mechanisms underlying disease progression. However, difficulties in studying the disease in humans can be circumvented by using mouse models, which can provide mechanistic insights into prostate carcinogenesis as well as access to all disease stages (25). Indeed, recent studies have led to the generation of several relevant models, including those based on the loss-of-function of genes implicated in human prostate cancer, such as the NKX3.1 homeobox gene and the PTEN tumor suppressor (e.g., refs. 69). In our previous work, we found that Nkx3.1;Pten compound mutants recapitulate stages of prostate cancer progression, including its prostatic intraepithelial neoplasia (PIN), adenocarcinoma, and metastatic disease (7, 8). Given the predictable time course of disease progression in these mice, as well their phenotypic similarities to human prostate cancer, we reasoned that Nkx3.1;Pten compound mutants would provide a useful model to investigate the contributions of other genetic factors for cancer progression.

We have now investigated the role of the cell cycle regulator p27kip1 in prostate carcinogenesis in the context of Nkx3.1;Pten compound mutant mice. As a tumor suppressor, p27kip1 is somewhat atypical; although it rarely undergoes mutational inactivation, its frequent down-regulation in human prostate cancer is associated with poor clinical outcome (reviewed in refs. 10 and 11). In mutant mice, homozygosity as well as heterozygosity for a p27kip1 mutant allele can lead to tumorigenesis (1216), and these tumorigenic phenotypes are exacerbated in collaboration with loss of other genes, such as Pten (6). p27kip1 was identified as a negative regulator of cell cycle progression through its ability to inhibit Cyclin E-Cdk2 complexes; however, its functions in vivo are likely to be considerably more complex (e.g., ref. 17) and are not restricted to cell cycle control (e.g., ref. 18).

In our analyses of Nkx3.1;Pten;p27kip1 triple mutant mice, we have found that mice lacking one wild-type p27kip1 allele display an enhancement of the prostate cancer phenotype, whereas those lacking both alleles display a reduction in carcinogenesis specifically in the prostate. These phenotypic differences are correlated with levels of Cyclin D1 and tumor growth depends on the levels of p27kip1 and/or Cyclin D1. Our study reveals a critical role for p27kip1 gene dosage for prostate carcinogenesis, and underscores the importance of Cyclin D1 for disease progression.

Materials and Methods

Nkx3.1, Pten, and p27kip1 mutant mice and analyses of their prostate phenotypes have been described (16, 19, 20); note that the p27kip1 mutant allele generates a truncated protein corresponding to the first 51 amino acids of the protein. Tissue recombinants were generated using adult prostatic epithelium from wild-type or mutant mice and mesenchyme from rat embryonic urogenital sinus (7, 21). Statistical analyses of prostate phenotype were performed by using cumulative logistic regression models with histological grade as the response variable, and age and genotype as explanatory variables. Immunohistochemical analyses were performed by using paraffin sections of mouse dorsolateral prostate, as described (7, 8, 21). Antibodies were as follows: Ki67 (NovoCastra, New Castle, U.K., 1:2,000); Akt and p-Akt (Ser-473) (Cell Signaling Technology, Beverly, MA, 1:200 and 1:50, respectively); p27kip1 (Transduction Laboratories BD, Lexington, KY, 1:200); Cyclin D1 (Zymed, 1:500 for Western blotting; Santa Cruz Biotechnology, 1:100 for immunohistochemistry); Cdk4 (Sigma, 1:2,000); Cyclin E (Sigma, 1:1,000); Pten (Neomarker, 1:200); and pol II (Covance, 1:500). The Nkx3.1 antibody was published previously (8).

Laser-capture microdissection (LCM; Arcturus, Mountain View, CA, Pixcell IIE) was performed on cryosections of snap-frozen prostate tissues to isolate epithelial cells from wild-type prostate, PIN, or cancer lesions. RNA was prepared by using the PicoPure RNA isolation kit (Arcturus), followed by one round of amplification with the RiboAmp RNA amplification kit (Arcturus). Samples were labeled by using a BioArray High Yield transcript labeling kit (Enzo Life Scientific) and hybridized to Affymetrix GeneChips (MOE430A). For statistical analyses, initial data acquisition and normalization were performed by using Affymetrix microarray suite 5.0 software. For validation of gene expression differences, quantitative RT-PCR analyses were performed by using the Mx4000 Multiplex Quantitative PCR system (Stratagene). Expression levels of p27kip1 or Cyclin D1 were evaluated relative to that of glyceraldehyde-3-phosphate dehydrogenase (GAPDH); all validation analyses used at least three independent RNA samples.

RNA interference (RNAi) was performed in prostate epithelial cells made from triple mutant mice by using a retroviral strategy as described (22); infection efficiency was 80% or better, as determined by GFP expression. Cyclin D1 was overexpressed by retroviral gene transfer by using a vector expressing the mouse gene. Anchorage-independent growth was performed by plating 1 × 103 cells per 35-mm dish in 0.35% agarose (21). Orthotopic tumor assays were performed by implanting 5 × 105 cells (10 μl) unilaterally into the dorsolateral prostate of nude male mice. Following 1 month of growth, tumor weights were determined by comparing the injected and noninjected sides.

Additional methods are provided in Supporting Material and Methods, which is published as supporting information on the PNAS web site.


Prostate Cancer Progression Is Enhanced in Compound Mutant Mice Lacking One Wild-Type p27kip1 Allele, but Inhibited in Mice Lacking both Alleles. To investigate the consequences of combined inactivation of Nkx3.1, Pten, and p27kip1 for prostate carcinogenesis, we generated mutant mice carrying targeted germ-line deletions of all three genes. A comprehensive phenotypic analysis was performed on mice from all 18 viable genotypic combinations (n = 375 mice) including histopathological analyses of each prostatic lobe, followed by detailed immunohistochemical analyses of marker expression in representative cases from each genotype (see Supporting Materials and Methods). Our discussion herein focuses on the triple mutant mice; the phenotypes of other genotypic combinations are described in Fig. 6 and Tables 2 and 3, which are published as supporting information on the PNAS web site.

We found that compound p27kip1 heterozygotes (Nkx3.1+/– or –/–;Pten+/–;p27+/–) displayed a more severe prostate cancer phenotype as compared with Nkx3.1;Pten compound mutants having two wild-type p27kip1 alleles (Nkx3.1+/– or –/–;Pten+/–;p27+/+) (Fig. 1). In particular, the compound p27kip1 heterozygotes were 2.4-fold more likely to develop high-grade PIN and/or adenocarcinoma than compound mutants having both wild-type p27kip1 alleles (P = 0.0290; Tables Tables11 and 4, which is published as supporting information on the PNAS web site). Although high-grade PIN/adenocarcinoma occurred earlier and was more prevalent in the compound p27kip1 heterozygotes, the histopathology and other features associated with cancer progression were similar to those previously reported for Nkx3.1;Pten compound mutants (7, 8, 22). These include (i) robust activation of phosphorylated Akt; (ii) loss of heterozygosity of Pten within the lesions; (iii) heterogeneous expression of epithelial markers such as E-cadherin; (iv) attenuation of the stroma; (v) loss of Nkx3.1 protein expression in heterozygotes; (vi) metastases to the lymph nodes; and (vii) androgen-independence after androgen-ablation (Figs. (Figs.11 and 7, which is published as supporting information on the PNAS web site, and data not shown). However, a notable distinction between the prostate phenotype of the compound p27kip1 heterozygotes and that of compound mutants having both wild-type p27kip1 alleles was a greatly elevated rate of cellular proliferation, as evident by Ki67 immunostaining (Fig. 1 MP). Indeed, the proliferation index within the high-grade PIN/cancer lesions of the compound p27kip1 heterozygotes was >20%, compared with <5% in lesions from mice with both p27kip1 alleles (Fig. 3B).

Fig. 1.
Prostate cancer progression is enhanced in compound mutant mice lacking one p27kip1 allele, but inhibited in mice lacking both alleles. Hematoxylin/eosin (AH) or immunohistochemical (IT) staining of paraffin sections from the dorsolateral ...
Fig. 3.
Compound p27kip1 heterozygotes display elevated levels of Cyclin D1.(A) Expression profiling of PIN or cancer lesions was performed by using RNA obtained by LCM of compound mutants of the indicated genotypes. Shown is hierarchical clustering of selected ...
Table 1.
Comparison of p27kip1 status in the context of combined Nkx3.1 and Pten loss-of-function

In striking contrast to our findings for the compound p27kip1 heterozygotes, we observed an opposite phenotypic outcome in triple mutant mice with zero wild-type p27kip1 alleles (Fig. 1). In particular, the compound p27kip1 homozygous mutants (Nkx3.1+/– or –/–;Pten+/–;p27–/–) were 4-fold less likely to develop high-grade PIN and/or adenocarcinoma than the compound p27kip1 heterozygotes (P = 0.0074; Tables Tables11 and 4). Despite the fact that the compound p27kip1 homozygous mutants expressed phospho-Akt (Figs. (Figs.1L1L and 7), indicative of Pten inactivation, the prostate phenotype of the compound p27kip1 homozygous mutant mice was relatively mild, rarely progressing beyond low-grade PIN (PIN grades I and II; refs. 22 and 23) and lacking many of the histopathological features associated with the high-grade PIN or cancer phenotype in mutants having one or two wild-type p27kip1 alleles. Notably, our findings contrast with a previous study that reported that 100% of Pten+/–;p27–/– compound mutant mice develop prostate cancer by 3 months of age (6); however, this previous study antedated the introduction of a consensus nomenclature for prostate cancer phenotypes in mice (23).

Interestingly, other than the prostate, the compound p27kip1 homozygous mutants were highly prone to other neoplasms, particularly lymphomas, and they rarely survived past 6 months of age, compared with the compound p27kip1 heterozygotes, which often survived to at least 17 months (W.B.-P and C.A.-S, unpublished observations, and ref. 6). Our observations that prostate cancer progression is enhanced by loss of one wild-type p27kip1 allele, but inhibited by loss of both alleles, suggest that carcinogenesis of the prostate, but not most other tissues, is sensitive to p27kip1 gene dosage.

Sensitivity to p27kip1 Dosage Is an Intrinsic Property of the Prostatic Epithelium. To consider the possibility that the relatively mild prostate phenotype of the compound p27kip1 homozygous mutants reflected their tendency to succumb to other tumors before having a chance to develop prostate cancer, and to address whether the observed phenotypic differences between the prostatic tissues from mice with two, one, or zero p27kip1 alleles were intrinsic to the prostatic epithelium, we performed tissue recombination assays (Fig. 2). We have shown previously that the phenotype of such tissue recombinants closely resembles that of the prostate tissue from which they are derived, and that serial transplantation can result in neoplastic progression, depending on the genotype (7, 24). Thus, we used this approach to evaluate the neoplastic potential of the prostatic epithelium from the triple mutant mice independent of other factors, such as age and/or interactions with the stroma, which might influence the phenotype in the intact mice.

Fig. 2.
Phenotypic differences among compound p27kip1 mutant prostates are retained in tissue recombinants. Hematoxylin/eosin analyses of initial-round tissue recombinants made from prostatic epithelium of the indicated genotypes. (A) Normal prostate histology. ...

We found that tissue recombinants from compound mutants having two (Nkx3.1+/– or –/–;Pten+/–;p27+/+) or one (Nkx3.1+/– or –/–;Pten+/–;p27+/–) wild-type p27kip1 alleles tended to develop high-grade PIN/carcinoma (n = 7/9 and 5/8, respectively), whereas those from mutants lacking any wild-type p27kip1 alleles (Nkx3.1+/– or –/–;Pten+/–;p27–/–) rarely developed high-grade PIN or cancer (n = 1/17), and instead displayed low-grade PIN (n = 7/17) or were within normal limits (n = 9/17) (Fig. 2). Importantly, tissue recombinants made from compound p27kip1 homozygous mutants also did not progress to more severe phenotypes after serial transplantation (n = 7), in contrast to those having one or two wild-type p27kip1 alleles (n = 15) (ref. 7 and data not shown). These tissue recombination studies indicate that phenotypic differences observed among compound mutants having two, one, or zero wild-type p27kip1 alleles are stable and intrinsic to the prostatic epithelium, rather than a reflection of the poor survival of the compound p27kip1 homozygous mutants.

Gene Dosage of p27kip1 Is Correlated with Levels of Its Expression. To ascertain whether the phenotypic differences in the prostate of compound mutants having two, one, or zero wild-type p27kip1 alleles correlated with the actual levels of p27kip1 RNA, we performed LCM to isolate prostatic epithelial cells from PIN or cancer lesions of these mice followed by real-time RT-PCR. We found that expression levels of p27kip1 RNA were correlated with the number of wild-type p27kip1 alleles (Fig. 3C). Furthermore, the compound p27kip1 heterozygotes also expressed reduced levels of p27kip1 protein relative to mice having both p27kip1 alleles (Fig. 7), and sequence analysis of genomic DNA isolated from prostate lesions by LCM confirmed that the remaining p27kip1 allele in the compound heterozygotes was not mutated (data not shown). Therefore, the differences in p27kip1 gene dosage in the compound mutants are directly correlated with levels of p27kip1 expression in the PIN and cancer lesions from these mice.

Cyclin D1 Expression Is Up-Regulated in the Compound p27kip1 Heterozygotes but Down-Regulated in the Compound Homozygous Mutants. To investigate the molecular basis of the observed phenotypic differences between compound mutants with two, one, or zero wild-type p27kip1 alleles, we performed gene expression profiling by using RNA isolated from prostate epithelial cells by LCM. We found that the expression profile of the compound p27kip1 heterozygotes was distinct from those of mice having either two or zero wild-type p27kip1 alleles and included several genes and/or pathways whose deregulation has been implicated in human prostate carcinogenesis (Fig. 3A and Table 5, which is published as supporting information on the PNAS web site). Among these deregulated genes was Cyclin D1, which was noteworthy given that increased cellular proliferation is a distinguishing feature of the compound p27kip1 heterozygotes (Figs. (Figs.11 and and3B).3B). Real-time RT-PCR analyses confirmed that Cyclin D1 was significantly up-regulated in cancer lesions from the compound p27kip1 heterozygotes, but barely detectable in the compound p27kip1 homozygous mutants (Fig. 3D), and expression of Cyclin D1 protein was also up-regulated in the compound p27kip1 heterozygotes (Figs. 1 QT and 7). Therefore, the phenotypic differences observed among the triple compound mutants were correlated with differential expression of Cyclin D1. Furthermore, by using cell lines that differ in p27kip1 gene dosage (see below), we found that Cyclin D1 promoter activity in transient transfection assays was also dependent on levels of wild-type p27kip1 (Fig. 3E), suggesting that this differential expression is mediated at the level of Cyclin D1 transcription.

Levels of p27kip1 and Cyclin D1 Are Directly Correlated with Prostate Tumorigenicity. To directly assess the relationship between p27kip1 dosage and prostate tumorigenicity, and to evaluate whether Cyclin D1 is a mediator of this effect, we established a series of prostate cancer cell lines from compound mutant mice having two, one, or zero wild-type p27kip1 alleles, which we have named the CASP series (see Supporting Materials and Methods). One notable feature of these cell lines is that they are derived from primary tumors rather than metastases, unlike most human prostate cancer cell lines (discussed in ref. 4). The CASP cells retain prostate epithelial properties and form prostatic ducts in cell recombination assays, but also exhibit anchorage-independent growth in soft agar and form tumors when implanted orthotopically into the prostate (Fig. 8, which is published as supporting information on the PNAS web site). Notably, levels of p27kip1 protein expression were correlated with its gene dosage, and levels of Cyclin D1 were barely detectable in CASP cells lacking p27kip1 (Fig. 8).

To investigate the relationship between p27kip1 levels and prostate tumorigenicity, we “knocked-down” the expression of p27kip1 in CASP cells by ≈50% by using short-hairpin-mediated RNAi (Fig. 4A). The consequences for tumorigenicity were evaluated by assaying anchorage-independent growth in soft agar (Fig. 4D), or tumor growth after orthotopic implantation in the prostate (Fig. 4E). We performed these experiments in CASP cells having two (p27+/+), one (p27+/–), or zero (p27–/–) wild-type p27kip1 alleles, with the latter providing an ideal control for RNAi specificity.

Fig. 4.
Levels of p27kip1 and Cyclin D1 are directly correlated with prostate tumorigenicity. (AE) Knock-down of p27 or Cyclin D1 in mouse prostate cancer cells. (A) Schematic representation of the overall strategy to knock-down of p27kip1 or Cyclin ...

Although RNAi for p27kip1 had virtually no effect on the p27kip1 null cells (p27–/–) as expected, in the p27kip1 wild-type cells (p27+/+) it resulted in ≈50% reduction in the p27kip1 levels (Fig. 4B). Notably, this partial knock-down of p27kip1 in the wild-type cells was coincident with a moderate increase in tumorigenicity in vitro and in vivo (Fig. 4 D and E; P = 0.0092 and 0.0038, respectively). In the p27kip1 heterozygous cells (p27+/–), RNAi of p27kip1 also resulted in an ≈50% reduction in p27kip1 expression levels to ≈25% of the wild-type levels (Fig. 4B), which corresponded to a significant reduction in tumorigenicity both in vitro and in vivo (Fig. 4 D and E; P < 0.0001 and 0.0027, respectively). These observations are concordant with our histopathological findings in the compound mutant mice and establish a direct relationship between p27kip1 gene dosage, p27kip1 expression levels, and prostate tumorigenicity.

Finally, to assess the significance of Cyclin D1 for prostate tumorigenicity, we used RNAi to knock-down its expression in the CASP cells, without affecting expression of p27kip1 (Fig. 4C). After RNAi for Cyclin D1, we observed a significant reduction in tumorigenicity of cells that are wild-type or heterozygous for p27kip1 in soft agar assays (P = 0.0079 and P < 0.0001, respectively), as well as in the orthotopic implantation assay (P = 0.0228 and 0.0061, respectively), but virtually no effect on the p27kip1 homozygous mutant cells, which already have low levels of Cyclin D1 (Fig. 4 D and E). Conversely, to address whether restoration of Cyclin D1 in the p27kip1 homozygous mutant cells affected their tumorigenicity, we used retroviral gene transfer to express exogenous Cyclin D1 in these cells (Fig. 4F). This resulted in increased anchorage-independent growth (P = 0.0002) and increased tumor growth after orthotopic implantation in the prostate (P = 0.05; Fig. 4 G and H). Taken together, these loss- and gain-of-function studies suggest that the sensitivity of prostate carcinogenesis to p27kip1 dosage is mediated, in part, through Cyclin D1 and demonstrate the significance of Cyclin D1 for prostate tumorigenicity.


Our analysis of mouse models of prostate cancer has uncovered a critical relationship between gene dosage of p27kip1 and prostate tumorigenicity. Surprisingly, a 2-fold reduction in wild-type p27kip1 dosage greatly enhances prostate carcinogenesis, whereas its complete loss impedes cancer progression specifically in the prostate. We have demonstrated that these differences in the severity of the cancer phenotype are intrinsic to the prostatic epithelium, and that they are directly related to the levels of wild-type p27kip1 expression. Furthermore, we show that these consequences of p27kip1 gene dosage for prostate tumorigenesis are at least partially attributed to altered levels of Cyclin D1, underscoring a role for this Cyclin gene in prostate carcinogenesis. We propose that a reduction of wild-type p27kip1 gene dosage promotes prostate carcinogenesis in part by up-regulating Cyclin D1 expression, whereas elimination of p27kip1 leads to down-regulation of Cyclin D1 expression and reduced prostate carcinogenesis (Fig. 5). This finding may explain why p27kip1 expression is often down-regulated in prostate cancer, yet is rarely eliminated.

Fig. 5.
Model for the relationship of p27kip1 gene dosage, Cyclin D1 expression, and prostate tumorigenicity. See Discussion.

Although this complex relationship between p27kip1 gene dosage and prostate carcinogenesis may seem counterintuitive, our findings agree with previous observations reported in studies of mouse models of breast cancer. In particular, Arteaga and colleagues (13) have shown that heterozygosity for p27kip1 enhances cancer progression in the context of an MMTV-c-neu transgene, whereas nullizygosity for p27kip1 inhibits cancer progression. Notably, in their studies, the p27kip1 homozygous mutants were found to be deficient for Cyclin D1/Cdk4 function in mammary epithelial cells (25), comparable to our findings in the prostatic epithelium. Because p27kip1 nullizygosity does not generally inhibit carcinogenesis of other tissues, it is intriguing to speculate that the dosage of wild-type p27kip1 may be particularly relevant for hormonally regulated cancers. Indeed, a threshold level of p27kip1 is required for response to anti-estrogens in the breast (26), whereas p27kip1 plays a key role in androgen-stimulated proliferation in the prostate (27).

Previous studies have defined a haploinsufficient role of p27kip1 in carcinogenesis in mutant mice, based on the observed effects of p27kip1 heterozygosity in enhancing carcinogenesis in a range of tissues. Notably, this previous work demonstrated haploinsufficiency for p27kip1 in lung, intestine, and pituitary carcinogenesis, and found that homozygous inactivation leads to increased tumorigenicity relative to heterozygosity (14). However, our demonstration that homozygous deletion of wild-type p27kip1 retards cancer progression in the prostate is distinct from a simple model of haploinsufficiency, because we find that heterozygosity and nullizygosity for p27kip1 result in opposite phenotypes.

There are few genetic precedents for this relationship of p27kip1 gene dosage and phenotypic outcome. One example is provided by studies of Fgf8 dosage sensitivity in embryonic brain development, where cell survival within the telencephalon is decreased when Fgf8 expression is either up-regulated or eliminated, but is increased when Fgf8 expression is down-regulated (28). This unusual type of dosage effect can potentially be explained by a proportionate response of a feedback inhibitory pathway, as in the case of Fgf8 activity in the brain, or by distinct functional activities that prevail at high versus low protein concentrations, as is likely to be the case for p27kip1.

Our data are consistent with a complex mechanistic relationship between p27kip1 and Cyclin D1 that has previously been inferred based on biochemical and genetic analyses. Despite its original identification as a cell cycle inhibitor, p27kip1 may also have a positive effect on the cell cycle, because it can interact with Cyclin D1–Cdk4 protein complexes, thereby stabilizing these complexes and promoting progression through the G1 phase of the cell cycle (29). Moreover, p27kip1 also has cell cycle-independent functions that may contribute to the need for its residual expression for prostate carcinogenesis. Although our findings indicate that altered levels of Cyclin D1 represents a principal outcome of varying p27kip1 dosage, we cannot exclude the possibility that other functions of p27kip1 are also important, such as the recent identification of its role in cell migration (18). Indeed, our expression profiling analyses have revealed that the compound heterozygotes display alterations in expression of Rho family genes (see Table 4).

Our findings highlight a hitherto unappreciated role for Cyclin D1 in prostate carcinogenesis. However, Cyclin D1 has been reported to be up-regulated in advanced prostate cancer in humans, particularly at stages associated with androgen independence and metastases to bone (30). Although our findings in mutant mice are consistent with the idea that up-regulation of Cyclin D1 is associated with increased cellular proliferation, it is conceivable that the consequences of its up-regulation in the prostatic epithelium are not solely mediated through its effects on cell cycle progression. Indeed, Cyclin D1 has been shown to function independently of Cdk4 as a corepressor of androgen receptor (31); this finding may partially explain its particular relevance for hormonally regulated cancers, because Cyclin D1 can also regulate the transcriptional activity of estrogen receptor (e.g., ref. 32).

In summary, our findings indicate that prostate cancer progression is highly sensitive to wild-type p27kip1 expression levels, because a 50% reduction of p27kip1 activity due to heterozygous inactivation enhances prostate tumorigenesis, whereas a further 50% reduction (to ≈25% of wild-type levels) is sufficient to reverse this effect. Thus, a broad spectrum of functional consequences can be observed over a 4-fold range of p27kip1 activity. This narrow dosage window suggests that down-regulation of p27kip1 expression levels or activity may represent an effective route for therapeutic intervention, and that our mutant mice may provide a valuable preclinical model for testing these potential therapies.

Supplementary Material

Supporting Information:


We thank Xiaohui Sun, Hodan Ali, and Jayshree Rao for assistance; E. Lynette Wilson for help with orthotopic assays; Gregg Hannon and Scott Lowe (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY) for the RNAi retroviral vectors; Ed Ziff (New York University School of Medicine, New York) for the Cyclin D1 promoter plasmids; and Andy Koff (Sloan–Kettering Institute, New York) for the p27kip1 mutant mice and sharing unpublished data. This work was supported by National Institutes of Health Grants DK60887 (to M.M.S.) and RO1 CA76501 and UO1 CA84294 (to C.A.-S.) and Department of Defense Postdoctoral Fellowship DAMD17-01-1-0755 (to H.G.). C.A.-S., M.M.S, R.D.C., and A.D.B. are investigators of the National Cancer Institute Mouse Models of Human Cancer Consortium.


Author contributions: H.G., X.O., W.B.-P., M.K., M.M.S., and C.A.-S. designed research; H.G., X.O., W.B.-P., M.K., and H.L. performed research; M.K. and H.L. contributed new reagents/analytic tools; H.G., X.O., A.D.B., Y.L., W.J.S., R.D.C., M.M.S., and C.A.-S. analyzed data; and H.G., X.O., A.D.B., R.D.C., M.M.S., and C.A.-S. wrote the paper.

Abbreviations: PIN, prostatic intraepithelial neoplasia; RNAi, RNA interference; LCM, laser-capture microdissection.


1. Nelson, W. G., De Marzo, A. M. & Isaacs, W. B. (2003) N. Engl. J. Med. 349, 366–381. [PubMed]
2. Huss, W. J., Maddison, L. A. & Greenberg, N. M. (2001) Semin. Cancer Biol. 11, 245–260. [PubMed]
3. Abate-Shen, C. & Shen, M. M. (2002) Trends Genet. 18, S1–S5. [PubMed]
4. Abate-Shen, C. & Shen, M. M. (2000) Genes Dev. 14, 2410–2434. [PubMed]
5. Powell, W. C., Cardiff, R. D., Cohen, M. B., Miller, G. J. & Roy-Burman, P. (2003) Curr. Drug Targets 4, 263–279. [PubMed]
6. Di Cristofano, A., De Acetis, M., Koff, A., Cordon-Cardo, C. & Pandolfi, P. P. (2001) Nat. Genet. 27, 222–224. [PubMed]
7. Abate-Shen, C., Banach-Petrosky, W. A., Sun, X., Economides, K. D., Desai, N., Gregg, J. P., Borowsky, A. D., Cardiff, R. D. & Shen, M. M. (2003) Cancer Res. 63, 3886–3890. [PubMed]
8. Kim, M. J., Cardiff, R. D., Desai, N., Banach-Petrosky, W. A., Parsons, R., Shen, M. M. & Abate-Shen, C. (2002) Proc. Natl. Acad. Sci. USA 99, 2884–2889. [PMC free article] [PubMed]
9. Wang, S., Gao, J., Lei, Q., Rozengurt, N., Pritchard, C., Jiao, J., Thomas, G. V., Li, G., Roy-Burman, P., Nelson, P. S., Liu, X. & Wu, H. (2003) Cancer Cell 4, 209–221. [PubMed]
10. Macri, E. & Loda, M. (1998) Cancer Metastasis Rev. 17, 337–344. [PubMed]
11. Bloom, J. & Pagano, M. (2003) Semin. Cancer Biol. 13, 41–47. [PubMed]
12. Nakayama, K., Ishida, N., Shirane, M., Inomata, A., Inoue, T., Shishido, N., Horii, I. & Loh, D. Y. (1996) Cell 85, 707–720. [PubMed]
13. Muraoka, R. S., Lenferink, A. E., Simpson, J., Brantley, D. M., Roebuck, L. R., Yakes, F. M. & Arteaga, C. L. (2001) J. Cell Biol. 153, 917–932. [PMC free article] [PubMed]
14. Fero, M. L., Randel, E., Gurley, K. E., Roberts, J. M. & Kemp, C. J. (1998) Nature 396, 177–180. [PubMed]
15. Fero, M. L., Rivkin, M., Tasch, M., Porter, P., Carow, C. E., Firpo, E., Polyak, K., Tsai, L. H., Broudy, V., Perlmutter, R. M., et al. (1996) Cell 85, 733–744. [PubMed]
16. Kiyokawa, H., Kineman, R. D., Manova-Todorova, K. O., Soares, V. C., Hoffman, E. S., Ono, M., Khanam, D., Hayday, A. C., Frohman, L. A. & Koff, A. (1996) Cell 85, 721–732. [PubMed]
17. Cheng, M., Olivier, P., Diehl, J. A., Fero, M., Roussel, M. F., Roberts, J. M. & Sherr, C. J. (1999) EMBO J. 18, 1571–1583. [PMC free article] [PubMed]
18. Besson, A., Gurian-West, M., Schmidt, A., Hall, A. & Roberts, J. M. (2004) Genes Dev. 18, 862–876. [PMC free article] [PubMed]
19. Bhatia-Gaur, R., Donjacour, A. A., Sciavolino, P. J., Kim, M., Desai, N., Young, P., Norton, C. R., Gridley, T., Cardiff, R. D., Cunha, G. R., et al. (1999) Genes Dev. 13, 966–977. [PMC free article] [PubMed]
20. Podsypanina, K., Ellenson, L. H., Nemes, A., Gu, J., Tamura, M., Yamada, K. M., Cordon-Cardo, C., Catoretti, G., Fisher, P. E. & Parsons, R. (1999) Proc. Natl. Acad. Sci. USA 96, 1563–1568. [PMC free article] [PubMed]
21. Hemann, M. T., Fridman, J. S., Zilfou, J. T., Hernando, E., Paddison, P. J., Cordon-Cardo, C., Hannon, G. J. & Lowe, S. W. (2003) Nat. Genet. 33, 396–400. [PubMed]
22. Park, J. H., Walls, J. E., Galvez, J. J., Kim, M., Abate-Shen, C., Shen, M. M. & Cardiff, R. D. (2002) Am. J. Pathol. 161, 727–735. [PMC free article] [PubMed]
23. Shappell, S. B., Thomas, G. V., Roberts, R. L., Herbert, R., Ittmann, M. M., Rubin, M. A., Humphrey, P. A., Sundberg, J. P., Rozengurt, N., Barrios, R., et al. (2004) Cancer Res. 64, 2270–2305. [PubMed]
24. Kim, M. J., Bhatia-Gaur, R., Banach-Petrosky, W. A., Desai, N., Wang, Y., Hayward, S. W., Cunha, G. R., Cardiff, R. D., Shen, M. M. & Abate-Shen, C. (2002) Cancer Res. 62, 2999–3004. [PubMed]
25. Muraoka, R. S., Lenferink, A. E., Law, B., Hamilton, E., Brantley, D. M., Roebuck, L. R. & Arteaga, C. L. (2002) Mol. Cell. Biol. 22, 2204–2219. [PMC free article] [PubMed]
26. Arteaga, C. L. (2004) Clin. Cancer Res. 10, 368S–371S. [PubMed]
27. Waltregny, D., Leav, I., Signoretti, S., Soung, P., Lin, D., Merk, F., Adams, J. Y., Bhattacharya, N., Cirenei, N. & Loda, M. (2001) Mol. Endocrinol. 15, 765–782. [PubMed]
28. Storm, E. E., Rubenstein, J. L. & Martin, G. R. (2003) Proc. Natl. Acad. Sci. USA 100, 1757–1762. [PMC free article] [PubMed]
29. Sherr, C. J. & Roberts, J. M. (1999) Genes Dev. 13, 1501–1512. [PubMed]
30. Drobnjak, M., Osman, I., Scher, H. I., Fazzari, M. & Cordon-Cardo, C. (2000) Clin. Cancer Res. 6, 1891–1895. [PubMed]
31. Knudsen, K. E., Cavenee, W. K. & Arden, K. C. (1999) Cancer Res. 59, 2297–2301. [PubMed]
32. Zwijsen, R. M., Buckle, R. S., Hijmans, E. M., Loomans, C. J. & Bernards, R. (1998) Genes Dev. 12, 3488–3498. [PMC free article] [PubMed]
33. Yan, G. Z. & Ziff, E. B. (1997) J. Neurosci. 17, 6122–6132. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...