p27^Kip1deficiency promotes prostate carcinogenesis but does not affect the efficacy of retinoids in suppressing the neoplastic process

Various risk factors, such as race (with black men having the highest risk), family history, and genetic predisposition appear to play principal roles in the development and progression of prostate cancer [1–3]. Over the last several years, increasing attention has been paid to the role of p27^Kip1 expression in the development and progression of various tumors, including prostate cancer. Human tumors lacking p27^Kip1 appear more malignant than those with high levels of the gene's expression [4, 5]. Normal prostate epithelial cells (PEC) exhibit abundant amounts of p27^Kip1 protein and mRNA, whereas in benign prostate hyperplasia (BPH) p27^Kip1 decreases to undetectable levels. In contrast to BPH, most prostate carcinomas contain p27 mRNA but low to undetectable levels of p27^Kip1 protein, suggesting post-transcriptional alterations in the gene's activity [6, 7]. Little is known regarding when in the course of prostate carcinogenesis disassociation between p27 mRNA and protein expression occurs or whether p27^Kip1 alone or in cooperation with other genes is involved in mediating the response of prostate pre-malignant and tumor cells to various chemopreventive and antitumor agents.

p27^kip1 is a cell cycle suppressor gene, whose protein product is a negative regulator of cyclin dependent kinases (CDKs) [8–10]. Cyclin dependent kinases-2/4/6 (CDKs) selectively bind to cyclin D1- D3, E, A, B, forming complexes that are variably expressed during the cell cycle. When inhibited by p27^Kip1, p21^Wef1/Cip1, or p16^Ink4a, CDKs can suppress cell cycle progression by modulating pRb phosphorylation, leading to inhibition of E2F transcription factors and further to suppression of DNA replication [11, 12]. p27^Kip1 may also cooperate with other cell cycle suppressor genes and thus further inhibit cell proliferation and carcinogenesis [13, 14]. For example, 100% of mice deficient in both p27^Kip1 and PTEN (phosphatase and tensin homolog deleted from chromosome 10) (PTEN +/-; p27-/-) developed prostate tumors within 4-6 months vs. 50% of those with the PTEN mutation only [15–17]. Recently, Besson et al. [18] discovered an oncogenic activity of p27^Kip1 that causes stem cell expansion and a multiple tumor phenotype. They generated a knock-in mouse in which four amino acid substitutions in the CDKN1b gene product prevented its interaction with cyclins and CDKs (p27^CK*) and found tumors in multiple organs, including: lung, pituitary, retina, adrenals, ovary, spleen, and lymphomas. No data has been published on the effects of p27^Kip1 deficiency on chemically-induced prostate carcinogenesis and on the sensitivity of PEC to retinoids.

Studies from the Fero et al. [19] group have shown that p27-/- and, to a lesser extent, p27+/- mice are more susceptible than p27+/+ mice to radiation and chemically induced carcinogenesis. They have found that partial reduction in p27 expression in p27+/- mice can also sensitize cells in a tissue specific manner to undergo malignant transformation. However, they did not examine prostate glands (personal communication). In human tumors haplo-insufficiency is not a frequent phenomenon. However, a moderate decrease in protein expression of certain tumor suppressors, including p27^Kip1, may also promote the neoplastic process [20].

Most chemopreventive and antitumor agents, including retinoids, affect normal and tumor cells by inhibiting cell proliferation, and this has been associated with increased expression of cell cycle suppressors [21]. It has been suggested that retinoids which induce cell differentiation and suppress cell proliferation can up-regulate p27^kip1 and thereby inhibit cell cycle progression [22]. Others [23] and we [24] recently observed that 9cRA can suppress prostate carcinogenesis when given to rats treated with a carcinogen or to transgenic mice that spontaneously develop prostate tumors [25]. Our previous studies on Noble rats revealed that, among various retinoids, 9cRA but not 4-(Hydroxyphenyl) retinamide (4-HPR) reduced the incidence and multiplicity of prostate intraepithelial neoplasia (PIN) [24]. The molecular mechanisms involved in inhibition of prostate carcinogenesis by 9cRA and other retinoids are poorly understood. 9cRA is a ligand of both retinoic acid receptors (RARs α,β,γ) and retinoid × receptors (RXRs α,β,γ), suggesting receptor dependent mechanisms of cell and tumor growth inhibition [22]. However, retinoids may also affect normal and tumor cells by receptor independent pathways [26, 27].

In this study we assessed the role of p27 deficiency on prostate cell proliferation, response to hormonal stimulation, and to treatment with retinoids. We also examined whether p27^Kip1 deficiency may affect MNU-induced prostate carcinogenesis and the efficacy of 9cRA in suppressing the neoplastic process. We found that p27^Kip1 deficiency promotes prostate cell proliferation and carcinogenesis but does not affect the efficacy of 9cRA in suppressing PIN and tumor development.

The main goal of this study was to determine whether p27 deficiency in p27+/- and p27-/- mice may affect PEC proliferation and carcinogenesis, as well as the efficacy of 9cRA in suppressing the neoplastic process. p27 heterozygous (+/-) and p27 null (-/-) mice were selected for the following reasons: a) the difference in p27^Kip1 expression between p27+/+, p27+/- and p27-/- mice simulates the gene's alterations in the course of prostate cancer development in humans, as p27^Kip1 is expressed in normal PEC, lacking in BPH, and differentially expressed in malignant tumors [4, 8]; b) p27-/- mice develop predominantly hyperplastic and premalignant (PIN) lesions, which are much easier to suppress by cancer preventive agents, like retinoids, than aggressively growing carcinomas such as those in PTEN mice [15–17]; c) previous studies have shown that p27 haplo-insufficiency promotes carcinogen- and radiation-induced tumors in various organs [19, 20]; d) retinoids are efficacious inhibitors of prostate carcinogenesis in animal models [23, 24]; and e) p27^Kip1 is involved in mediating the effect of retinoids on PEC both in vitro and in vivo [25–27]. Here, we confirmed two of our previous observations [28, 31] and those of others [32–35]. First, we confirmed that p27-/-mice are larger than p27+/+ mice, whereas heterozygous p27+/- mice have an intermediate body weight (see Table 2). The increase in body weight in p27-/- mice was due to generalized organomegaly with disproportional enlargement of the spleen, thymus, testis, ovaries, pituitary gland (data not shown). Second, we confirmed that old (> 7-12 months) p27-/- and p27+/- mice develop spontaneous hyperplasia and adenomas in the intermediate pituitary lobe, intestine, kidney and thymus that lead to the animals' death. The data generated in this study indicates that under normal physiological conditions, PEC proliferation in two-month-old mice is similar in all three p27 genotypes (Table 1). However, when the values of proliferating cells were compared in 9- to12-month-old mice, proliferating cells were much higher in p27-/- and p27+/- than in p27+/+ mice, suggesting that p27 absence or insufficiency promotes PEC proliferation in the aging animals where hyperplastic and pre-malignant lesions occur (Table 5). This data supports previous studies from Cardon-Cardo's group [7], which found higher PEC proliferation (Ki-67 labeling) in p27 null (-/-) mice as compared to p27+/+ mice. However, they did not provide information whether p27 heterozygous and null mice develop PIN and whether the frequency of PIN could be regulated by hormone stimulation or cancer prevention agents. Thus, it appears that the loss of p27^Kip1 keeps a high base level of PEC proliferation in old animals that leads to development of hyperplastic, pre-malignant and malignant lesions. When PEC proliferation was stimulated by testosterone, a much higher proliferative response was observed in p27+/- and p27-/- than in p27+/+ mice, indicating that p27^Kip1 lack or haplo-insufficiency increases the sensitivity of PEC to hormone stimulation. This information is important, because there are sufficient clinical data showing that an increased testosterone circulation level in men is associated with increased risk in developing prostate cancer [16]. Therefore, temporary or continuous down-regulation of p27^Kip1 expression in PEC may promote the effect of testosterone or other growth factors in stimulating prostate carcinogenesis.

In a previous study Fero et al. [19] reported that p27^Kip1 protein expression was lacking in p27-/- mice, decreased by 50% in p27+/- mice, and promoted radiation and chemical-induced (N-Ethyl-N-Nitrosourea-ENU) carcinogenesis. In p27-/- mice, tumors developed in the lung, intestine, pituitary, ovary, and endometrium earlier and at a higher frequency as compared to both p27+/- and p27+/+ mice. However, in Fero's study prostate and male reproductive organs were not examined at the tissue level (personal communication) and neither PEC proliferation was stimulated by testosterone. In another study, Di Cristofano et al. [15] reported on cooperation between p27^Kip1 and PTEN in developing PIN and prostate cancer. PTEN is a tumor suppressor and its activity leads to the induction of p27^Kip1; therefore, the lack or down-regulation of PTEN may suppress p27^Kip1 expression and thus promote cell proliferation and carcinogenesis. Concomitant inactivation of one PTEN allele and one or both p27 alleles in transgenic mice accelerates the development and progression of prostate and other malignancies [16, 17]. We did not use PTEN+/-, p27-/- mice in this study because tumors in the prostate and other organs develop very early, within 3 months after birth, and most animals die at the age of 16-20 weeks. This very aggressive neoplastic process is difficult to modulate with retinoids and other cancer prevention agents. We found that the progressive decrease of p27^Kip1 expression in p27+/- and further in p27-/- mice promotes MNU-induced prostate carcinogenesis (Table 3). It appears that there is a negative correlation between the level of p27^Kip1 expression and the incidence and multiplicity of PIN and prostate tumors. Since most animals were sacrificed between 9 and 12 months of age, this time period is probably not sufficient for PEC in p27+/+ mice to transform in PIN and progress in invasive carcinomas. For instance, studies on rats treated with MNU have shown that prostate pre-malignant lesions and carcinomas occur 12 to 18 months after carcinogen administration [23, 24]. To accelerate prostate carcinogenesis, we treated p27+/+, p27+/- and p27-/- mice with MNU and stimulated PEC proliferation by testosterone + estradiol releasing pellets (Table 4). Previous studies, including ours, have shown that testosterone + estradiol stimulate PEC proliferation [35] and carcinogenesis in rats [23, 24]. However, because of the severe prostate hypertrophy with complete urethra blockage in some animals and high death rate, the study continued with surviving animals only. Hormone stimulation was more efficacious in inducing PIN and tumors in p27+/- and p27-/- mice than in p27+/+ mice, suggesting that deregulation of p27^Kip1 expression in PIN may promote hormone mediated prostate carcinogenesis in men. The data for multiplicity of PIN in Table 3, where MNU was only used to induce prostate carcinogenesis, are close to those in Table 4, where MNU + testosterone + estradiol were employed. This can be explained by termination of hormone stimulation which occurs 60 days after pellet implantation and which may cause cell death and thereby eliminate transformed PEC and their potential progression towards PIN and tumors.

What is most interesting from this study was that 9cRA suppressed MNU-induced prostate carcinogenesis but was unable to suppress MNU+hormone-stimulated carcinogenesis. Even in p27+/- mice, a tendency for an increase of PIN by 9cRA is obvious, although the differences in values with placebo-treated animals were not significant. In all three p27 genotypes, the values of PIN incidence are close in placebo- and 9cRA-treated animals. This data supports the results from clinical trials where patients with an increased risk of prostate cancer have been treated with beta carotene, retinol or other antioxidants [36, 37]. No efficacy on prostate cancer reduction has been found, but an increase in lung cancer incidence has been observed that led to preliminary termination of a clinical trial. One potential hypothesis is that 9cRA or its metabolites are not efficacious in highly proliferating cell systems, like PEC stimulated by the above hormones. Our data from short-term experiments where mice have been treated with testosterone + 9cRA supports this hypothesis. Another alternative is that in long-term experiments 9cRA may protect PEC from cell death after expiration of testosterone + estradiol stimulation (60-day releasing pellets). However, in the long-term carcinogenesis study, 9cRA suppressed the incidence and frequency of PIN in MNU-treated mice, indicating that a lack of or decreased p27^Kip1 level does not affect the efficacy of 9cRA in suppressing prostate carcinogenesis. 9cRA is a ligand of both RARs α,β,γ and RXRs α,β,γ, suggesting that it may affect various cellular functions by receptor dependent mechanisms [22, 25, 38]. By ICH we found that RARα and RXRα, principal targets of retinoids and rexinoids, respectively, are expressed in PEC of all three p27 genotypes, suggesting a potential receptor-dependent mechanism of 9cRA-induced inhibition of PEC proliferation and carcinogenesis. An attempt was made to compare RARα and RXRα positive cells in control- and 9cRA-treated animals in a carcinogenesis study (Table 3). However, no difference in the percentage of receptor positive cells was found (data not presented), suggesting receptor independent mechanisms of 9cRA-induced inhibition of prostate carcinogenesis [39]. It has been shown that retinoids, including 9cRA, at pharmacological doses preferentially suppress cell proliferation and may also induce cell differentiation [40]. Some previous studies have also shown that retinoids may induce apoptosis as well, depending on the tissue type and the dose used [9, 15]. Our data obtained from the carcinogenesis study indicates that 9cRA suppresses cell proliferation but is unable to induce apoptosis. Interestingly, 9cRA did not significantly affect cell proliferation in non-PIN cells, which have low proliferative activity, but significantly reduced proliferating cells in PIN of both p27+/- and p27-/- mice, indicating that 9cRA may have the potential to suppress preferentially pre-malignant stages of prostate carcinogenesis in men independent of the levels of p27^Kip1 expression. Surprisingly, we found that in addition to inhibition of cell proliferation, 9cRA can also induce CS in both PIN and non-PIN associated PEC (Table 5 and Fig. 2E-H). In a previous study on mammary carcinogenesis, we compared the effect of 9cRA and 4-HPR on CS in MNU-induced mammary tumors and found that 9cRA has a superior role [24]. Data from human prostate also suggest that senescent cells are frequently detected in BPH and rarely in prostate cancer, suggesting that malignant transformation of prostate epithelial cells is associated with loss of the potential to senesce [8, 41].

Our data indicate that p27^Kip1 deficiency in PEC promotes cell proliferation in an age-dependent manner and increases cellular response to hormone stimulation. p27^Kip1 haplo-insufficiency and deficiency stimulate MNU-induced prostate carcinogenesis, suggesting that PIN in human prostate lacking partial or total p27^Kip1 expression may have a higher potential to progress and develop malignant phenotype than lesions with wild type p27. The efficacy of 9cRA in suppressing PIN is apparently not p27-dependent, indicating that in potential clinical trials, 9cRA may affect pre-malignant and tumor cells that differentially expressed p27^Kip1. The induction of CS by 9cRA in PIN suggests that this biomarker could be used as a potential biomarker of response in clinical trials for the prevention and treatment of prostate cancer.