Background
The putative tumor suppressor p27Kip1 (referred to as p27) controls the progression from G1 to the S phase by regulating the activity of cyclinE/and cyclinA/Cdk2 complexes [
1]. Several external signals regulate the intracellular level of p27 by either causing its increase (i.e. serum deprivation, TGFβ, contact inhibition) or its decrease (serum stimulation, estrogen, PDGF and others), thereby rendering p27 a central mediator of mitogenic and anti-mitogenic signals [
2]. In addition to its negative role in cell cycle progression, p27 is involved in cell migration, neuronal differentiation and apoptosis [
3‐
5]. Through studies of a mouse strain expressing a p27 protein impaired in cyclin/Cdk binding it has been demonstrated that p27 has a pro-oncogenic effect when it cannot bind to cyclin/Cdk complexes [
6].
The intracellular level of p27 is regulated at the transcriptional, translational and post-translational level [
7,
8], but the best known mechanism is ubiquitin-mediated proteasomal degradation. Two main pathways involved in p27 degradation have been identified. The first is mediated by the Skp2-dependent SCF (skp-cullin-f-box) E3 ligase: phosphorylation of p27 by cyclinE/Cdk2 at a conserved threonine (Thr187) creates a binding site for Skp2, which allows polyubiquitylation and subsequent proteasomal degradation of p27. This degradation pathway is active in the nucleus of G1-S and G2 phase cells [
3‐
5,
9]. The second pathway is mediated by the KPC ubiquitin ligase and is responsible for the degradation of p27 in the cytoplasm at the G0-G1 transition [
10].
Phosphorylation at specific residues regulates the activity of p27: phosphorylation at serine (Ser) 10 regulates its subcellular localization and stability [
11,
12]. Studies of p27S10A (serine 10 substituted by alanin) knock-in mice demonstrated that phosphorylation at Ser10 stabilizes p27 during quiescence by affecting its ability to bind to cyclin-CDK complexes [
13]. Ser10 phosphorylation also triggers the export of p27 from the cell nucleus to the cytoplasm upon mitogenic stimuli, thereby allowing the protein's degradation by the KPC ubiquitin ligase [
14]. Phosphorylation of p27wt at Thr187, as mentioned above, targets p27 for proteasomal-mediated degradation [
15], while phosphorylation at Thr198 prevents ubiquitin-dependent degradation of free p27 and regulates the stability of p27 in G0 phase [
16].
We recently identified a
Cdkn1b germline frameshift mutation as the cause of a recessive multiple endocrine neoplasia (MEN)-like syndrome (named MENX) in the rat [
17]. Rats affected by this syndrome (homozygous mutants) share phenotypic features with the p27 -/- knock-out mice (increase in size, pituitary tumors) but show additional neuroendocrine tumors (adrenals, thyroid, parathyroid). Interestingly, we and others identified
CDKN1B germline mutations in patients with a MEN type 1 (MEN1)-like features, thereby establishing a direct link between p27 alterations and tumor predisposition also in humans (MEN4; OMIM # 610755) [
17‐
19]. Germline
CDKN1B mutations are a rare event in patients with an MEN1-like phenotype (estimated to be around 1.5-2.8% of cases) [
18,
19], and some studies failed to identify them [
20,
21]. So far, five germline
CDKN1B mutations have been reported and the
in vitro functional studies performed on three of them show that these alterations affect the ability of the encoded protein to bind its usual protein partners (p27P95S variant) or affect its amount (ATG-7G>C; stop->Q) [
19]. Interestingly, the phenotype of the stop->Q germline mutation can be rescued by proteasomal inhibition
in vitro [
19].
In the MENX model, p27 protein expression is reduced or absent in the normal tissues of the mutant rats, but it becomes detectable in some advanced neoplastic lesions [
17]. Therefore, this mutation is not equivalent to a null allele and it may possess specific biological functions within this model system. The mechanisms causing reduced p27 expression in the MENX-affected rats are currently unknown.
In humans, neuroendocrine tumors are uncommon neoplasms and therefore comprehensive molecular studies are difficult to perform and clinical trials are challenging to set up. In addition, there are few suitable animal models of neuroendocrine tumors and this has hampered the development and preclinical testing of novel targeted therapeutic approaches against these tumors. In this scenario, the MENX model may be of relevance to identify novel molecular mechanisms in neuroendocrine tumorigenesis and to perform preclinical therapy-response studies. An in depth knowledge of the functional consequences of the underlying genetic mutation in Cdkn1b is a prerequisite to further exploit the MENX animal model in preclinical studies. Moreover, since germline CDKN1B mutations have been found in patients, studies of the molecular phenotype of the rat mutation may broaden our understanding of the role of p27 in tumor predisposition in humans as well. Therefore, we set out to characterize the MENX-associated p27 mutant protein in vitro and in the patients tissues. Our studies extend to the rats the observation that the amount of functional p27 is crucial for promoting neuroendocrine tumorigenesis, event that so far had been formally proven only in genetically engineered mice.
Discussion
We show here that the mutant p27 protein associated with MENX is very unstable due to the p27-unrelated C-terminal domain. Its rapid degradation is, at least in part, mediated by the Skp2-dependent ubiquitination pathway, just as p27wt. Interestingly, it has been reported that Skp2-dependent ubiquitination and subsequent degradation of p27 requires the phosphorylation of Thr187, which is then recognized and bound by the Skp2 ubiquitin ligase. Since p27fs177 lacks the Thr187 residue but its degradation is still regulated by Skp2 (as seen in siRNA knock-down experiments), we have to postulate that a Skp2-dependent but Thr187-independent mechanism for p27 degradation exists, as it had been previously hypothesized based on studies of cells (fibroblasts, thymocytes) from p27T187A (Thr substituted by Ala) knock-in mice [
26]. A mutant ubiquitin that inhibits ubiquitin chain elongation has no effect on the degradation of p27fs177, while it rescues wild-type p27 degradation, indicating that p27fs177 is also degraded by mechanisms different from those modulating the level of p27wt. However, what is relevant is that the expression of p27fs177 can be recovered by proteasome-inhibition. This finding, combined with the fact that p27fs177 retains the capacity to interact with cyclins and Cdks, suggests that if the expression of this protein can be rescued inhibiting the proteasome, it might resume its anti-proliferative activity.
The fact that p27fs177 retains the potential to bind to the usual partners of p27 is highly relevant: if this protein can be re-expressed in the tumor cells, it might then resume its inhibitory effect on the cell cycle and therefore prevent or reverse tumor growth in vivo.
Our observation that in some areas of very advanced tumors in MENX mutant rats p27fs177 is expressed at a detectable level might not be in contradiction to this hypothesis. In fact, re-expression of the protein is observed in rats whose general health status is irreversibly compromised by the detrimental effects related to tumor progression.
The slower nuclear dynamic of p27fs177 is therefore due to its being part of a multi-protein complex, likely together with the degradation machinery. This is in agreement with the finding that the nuclear Skp2 ubiquitin ligase plays a role in p27fs177 degradation.
p27 is frequently down-regulated in human tumors and in some of them increased proteasomal degradation has been identified as the cause of the reduced p27 level [
27,
28]. Because of these findings, efforts have been made to identify compounds that interfere with the protein turnover machinery in order to achieve p27 re-expression in cancer cells. An example is the novel compound, Argyrin A, which specifically prevents p27 degradation and holds great promise as anti-cancer drug [
29]. Due to the characteristics of p27fs177 here described, MENX rats represent a useful pre-clinical model in which to test the efficacy of targeted therapeutic approaches aiming at inhibiting p27 degradation. Interestingly, an established anticancer proteasome inhibitor compound, Bortezomib, can rescue p27fs177 expression in fibroblasts of MENX-affected rats. Therefore, MENX might be also employed to test
in vivo the efficacy of Bortezomib against neuroendocrine tumors and to study whether the potential effect of the drug is mediated by reexpression of p27 function. A Phase II clinical trial aiming at evaluating the effect of the proteasome inhibitor Bortezomib on metastatic neuroendocrine tumors failed to show any objective tumor response in this patient cohort [
30]. However, the patients had not been stratified for p27 expression, measurement of the increase in p27 expression following treatment was not among the biological end points of the study and Bortezomib induces a wide range of off-target effects [
27]. Thus, the efficacy of p27 re-expression in the treatement of neuroendocrine tumors is still an open question that can be addressed by studying MENX-affected rats.
Conclusions
Our findings suggest that the p27fs177 mutant protein does not acquire novel molecular properties, and triggers tumor formation in the spontaneous rat MENX model because it behaves as a hypoactive allele. So far, this had only been formally proven in various p27 knock-out and knock-in mouse models [
31‐
33].
The identification of the molecular phenotype of the few naturally-occurring p27 mutations so far identified in MENX and in human patients is important to better understand the link between p27 and neuroendocrine tumor predisposition (and tumorigenesis in general), and may provide clues for a targeted management of the families carrying such mutations.
Methods
Plasmid constructs and antibodies
The rat wild-type Cdkn1b cDNA was cloned in the pEGFP-C3 vector (BD Biosciences, Erembodegem, Belgium) in frame with the green fluorescent protein (GFP) tag. A rat Cdkn1b cDNA containing the MENX rat mutation (p27fs177) and a rat Cdkn1b cDNA containing a stop codon at position 177 (p27G177X) were generated by PCR.
The rat wild-type Cdkn1b cDNA was cloned in the pEGFP-C3 vector (BD Biosciences, Erembodegem, Belgium) in frame with the green fluorescent protein (GFP) tag. A rat Cdkn1b cDNA containing the MENX rat mutation (p27fs177) and a rat Cdkn1b cDNA containing a stop codon at position 177 (p27G177X) were generated by PCR.
The pUbr7 construct was gently provided by N. Malek
Primary antibodies used were: anti-p27 monoclonal antibody (BD Biosciences); anti-Skp2 monoclonal antibody; anti-Cyclin A polyconal antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA). To control for equal protein loading, α-tubulin immunostaining (Santa Cruz) was performed.
Cell culture, transfections, protein extraction and Western blotting
MCF7 (human breast cancer) cell line was maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum, 20 mM L-glutamine, 100 units/ml of penicillin G sodium, and 100 μg/ml streptomycin. When establishing stable transfectants, cells were selected by adding 0.4 mg/ml G418 to the culture medium.
Primary rat fibroblast cells were obtained from the epidermis of newborn rats having a p27 wt/wt (RNF wt) or a p27fs177/p27fs177 genotype (RNF mut) and were maintained in DMEM medium supplemented with 10% fetal bovine serum, 20 mM L-glutamine, 100 units/ml of penicillin G sodium, 100 μg/ml streptomycin, and Fungizone® antimycotic (Invitrogen, Darmstadt, Germany).
The cells were grown in a humidified atmosphere containing 5% CO2 at 37°C.Cells were growth arrested by decreasing the serum concentration in the medium to 0.1% for 72 h and then released by serum addition,
Asynchronously growing cells were transiently transfected with the different constructs when 70-80% confluent using the FuGene HD reagent (Roche Applied Bioscience, Basel, Switzerland). Cells were collected and lysed 24 h later in protein lysis buffer (10 mmol/l Tris-HCl pH 7·4, 5 mmol/l EDTA, 130 mmol/l NaCl, 1% Triton, and 1× Mini-Complete protease inhibitors cocktail, Roche).
Equal protein amounts (50 μg) were resolved on 4-15% SDS-PAGE pre-cast gels (Invitrogen, Darmstadt, Germany).
Drug treatments and RNA interference
The cells were treated with 25 μg/mL cycloheximide (Sigma-Aldrich Biochemie, Hamburg, Germany), 10 μM epoxomycin (Enzo Life Sciences, Lörrach, Germany), or 10 nM Bortezomib (LC Labs, Woburn, MA, USA).
For RNAi studies, short interfering RNA (siRNA) duplexes specific for Skp2 (n = 4) (1 μg; siGenome SMARTpool, Dharmacon, Lafayette, CO, USA) and KPC1 (1 μg each; ID 133486, 133487 Ambion-Applied Biosystems, Darmstadt, Germany) were obtained and transfected into MCF7 cells using X-tremeGENE reagent (Roche) together with expression plasmids for p27wt (30 ng) or p27fs177 (50 ng). To verify the specificity of the siRNA-mediated knock-down, a scrambled siRNA oligo was transfected in parallel.
Immunofluorescence
MCF7 cells were plated on dishes having a cover slip for microscopy (MatTek Corporation, Ashland, MA, USA) and transfected as above with p27fs177 and p27wt protein constructs. To re-create a more physiological condition (37°C and 5% CO2), dishes were put under a micro-incubation chamber (Zeiss, Jena, Germany) mounted on a confocal laser scanning microscope (Zeiss). Images were taken every 15 min using always the same settings.
Fluorescence activated cell sorting (FACS)
Cells were collected and fixed in 70% ice-cold ethanol and stored at -20°C until ready to proceed. Cells were rinsed and resuspended in 1× PBS containing 5 μg/ml propidium iodide and 150 μg/ml RNaseA. After incubation for 30 min at 37°C cells were analyzed in a FACS machine (Beckman Coulter, Krefeld, Germany) for DNA content.
Immunoprecipitations
500 μg of total protein extracted from the cells were incubated with a polyclonal anti-GFP antibody (BD Biosciences) in IP buffer (5 mM EDTA, 0.5% Triton X-100 in PBS 1×) and 20 μl of agarose resin overnight at 4°C. Unbound protein fraction was removed, the resin was washed five times with IP buffer and immunoprecipitated proteins were eluted from the agarose resin using Laemmli buffer and analyzed by Western blot.
Additional experimental procedures and associated reference are available in Additional File
8.
Acknowledgements
We thank Prof. N. Malek (Hannover Medical School, Hannover, Germany) for providing the Ubr7 construct. This work was supported in part by Award #107973 from the Deutsche Krebshilfe (NSP) and by a Fellowship from Centro per la Comunicazione e la Ricerca, Collegio Ghislieri, Pavia, Italy (SM).
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
SM carried out the molecular studies and drafted the manuscript. EK, CBJ, ML and EP carried out the molecular studies. HH and MJA participated in the design of the study. NSP conceived the study, and participated in its design and coordination and drafted the manuscript. All authors read and approved the final manuscript.