High cytoplasmic expression of p27Kip1 is associated with a worse cancer-specific survival in clear cell renal cell carcinoma

Authors


Jens Bedke, Department of Urology, Eberhard-Karls-University Tuebingen, Hoppe-Seyler Strasse 3, 72076 Tuebingen, Germany. e-mail: bedke@live.com

Abstract

What's known on the subject? and What does the study add?

The loss of p27Kip1 correlates with poor prognosis in various human cancers, and was postulated as a biomarker in RCC. Up to now p27Kip1 analysis in RCC was focused on its nuclear localization.

We confirmed higher p27Kip1 expression in the nucleus and cytoplasm of RCC and correlated high cytoplasmic p27Kip1 with an unfavourable clinic and a reduced survival.

OBJECTIVES

  • • To analyse the cytoplasmic and nuclear differences of p27Kip1 expression and localization in benign and clear cell renal cell carcinoma (ccRCC) with regard to overall survival (OS) and cancer-specific survival (CSS).
  • • p27Kip1 is considered to contribute to the progression of ccRCC and is targeted by next generation dual-therapies.

PATIENTS AND METHODS

  • • In 140 patients, ccRCC and corresponding benign kidney tissue were analyzed for nuclear and cytoplasmic staining of p27Kip1 by immunohistochemistry using a tissue microarray technique.
  • • Staining intensity and percentage of positive stained cells are given as expression scores. p27Kip1 expression was categorized as high if ccRCC tissue stained stronger than the respective level of the corresponding benign tissue and categorized as low if ccRCC tissue stained less than or equal to the corresponding benign tissue.
  • • Differences in OS and CSS were analyzed by log-rank analysis and expression levels were correlated with tumour and patient characteristics using Fisher's exact test.

RESULTS

  • • Cytoplasmatic (mean [sd]: 13.8% [1.2%] vs 10.7% [1.7%]; P= 0.04) and nuclear (mean [sd]: 75.6% [2.7%] vs 13.6% [2.1%]; P < 0.001) staining of p27Kip1 were significantly stronger in ccRCC tissues compared to benign tissue.
  • • High cytoplasmic p27Kip1 expression was significantly associated with a higher T and N stage, Fuhrman grade and the presence of metastatic disease (P < 0.001).
  • • The median follow-up time was 38.2 months.
  • • There was no difference in OS between the low and high expression groups, although CSS was significantly lower in patients with high cytoplasmic p27Kip1 (P < 0.001) and CSS heavily tended to be lower in the nuclear low expression group (P= 0.069).

CONCLUSIONS

  • • High cytoplasmic p27Kip1 levels in renal cancer tissues are associated with advanced disease and reduced cancer specific survival, whereas low nuclear expression levels appear to be beneficial.
  • • The present study corroborates the consideration of cytoplasmic p27Kip1 for future diagnostic and targeted therapeutic approaches in RCC establishing a potential protective shift of p27Kip1 from the cytoplasm to the nucleus.
Abbreviations
ccRCC

clear cell RCC

CSS

cancer-specific survival

OS

overall survival

PKB

protein kinase B.

INTRODUCTION

RCC is known as the most frequent malignant kidney tumour, with ≈30 000 new diagnosed patients in Europe and over 38 000 new reported cases in the USA. To date, only complete surgical tumour excision offers patients a curative option for localized RCC disease. In patients with renal cancer metastases, the chemotherapeutic options are limited as a result of high chemotherapy and radiation resistance [1]. Recently, new targeted therapies were able to increase response rates and prolong both progression-free and overall survival (OS) [2]. However, accurate predictions of RCC aggressiveness and risks of progression are still difficult. Insight into the molecular mechanisms and pathways affected by the targeted therapy is still limited. The cyclin-dependent kinase inhibitor p27Kip1 is known to play a central role in the cell-cycle control of clear cell RCC (ccRCC) and various other human malignancies [3,4]. In addition to the ability of p27Kip1 to regulate cell progression from G0 into the G1-phase and from G1 into S-phase, p27Kip1 acts as an important mediator of different signalling pathways, such as the protein kinase B (PKB)/AKT pathway [5,6]. A wide variety of factors are known to induce (TGFs, serum starvation or contact inhibition) or decrease (platelet-derived growth factors, serum stimulation, oestrogen or interleukin-2) cellular p27Kip1 levels and activity [4]. Over-activation of RCC-relevant oncogenic kinases such as PKB can contribute to alterations of p27Kip1 function and potentially provides additional perspectives for new anticancer targets [7]. However, the regulation of the cell-cycle protein p27Kip1 in the human kidney and ccRCC tissue is not fully resolved. In the clinic, the loss of p27Kip1 is correlated with poor prognosis in various human cancers, and it is also proposed that p27Kip1 may act as a prognostic biomarker in RCC patients [8–14]. More recent studies suggest that the nuclear function of p27Kip1 may extend beyond the control of the cell cycle towards an inhibitory role in tumour progression. These results show the oncogenic relevance of p27Kip1 when the protein is located into the cytoplasm of the tumour cell [15–18]. In addition, the oncogenic importance of the cytoplasmic shift of p27Kip1 has been reported for breast, liver, thyroid and blood cancer [19–23]. Novel findings show that the cytoplasmic accumulation and degradation of p27Kip1 is of tremendous help in the evaluation of tumour progression and survival in human cancer [24,25]. In RCC, the analysis of p27Kip1 expression has focused on its nuclear localization and a loss of nuclear staining was found to be associated with unfavourable pathological features and reduced patient survival [8–12]. For RCC, there is still a lack of data concerning the relevance of the cytoplasmic export of p27Kip1, especially with regard to the survival of RCC patients. The present study aimed to analyse p27Kip1 expression in the different cellular compartments of the nucleus and cytoplasm. The results obtained were correlated with OS and cancer-specific survival (CSS) in a large cohort of ccRCC patients.

PATIENTS AND METHODS

PATIENTS AND TISSUE MICROARRAYS

A total of 140 ccRCC and corresponding normal paraffin tissue specimens from patients who underwent nephrectomy or nephron-sparing surgery at the University of Tuebingen, Germany, between May 1997 and October 2003 were included in the present analysis. Informed consent was obtained according to procedures approved by the institutional review board of the University of Tuebingen. Clinical and pathological data were collected by the treating physicians and by the tumour registry database of the University of Tuebingen. Samples from pathologically representative tumour regions and adjacent benign renal tissues were obtained. Specimens were classified according to the seventh edition of the Union Internationale Contre le Cancer/American Joint Committee on Cancer system (2009). After independent second histological evaluation of haematoxylin and eosin-stained slides, the tissue microarray slides were prepared as described previously with a core size of 0.6 mm; all probes were assembled as triplets [26].

IMMUNOHISTOCHEMICAL STAINING AND DATA ANALYSIS

Tissue specimens were deparaffinized by passing the specimens through xylene and rehydrated though serial dilutions of ethanol (100%, 96% and 70%). Endogenous peroxidase was blocked with 3% H2O2 in Aqua Dest (Sigma-Aldrich, Steinheim, Germany) for 30 min. Antigen retrieval was accomplished by incubating the sections for 3 × 5 min in 10 mM citrate buffer at 95 °C. To block non-specific binding sites, sections were incubated in antibody diluent with background reducing components (Dako S3022; DakoCytomation, Glostrup, Denmark). Expression of p27Kip1 was immunohistochemically detected by a commercially available antibody (dilution 1:75, clone SX53G8, monoclonal mouse; DakoCytomation). After 12 h of incubation at 4 °C, the sections were incubated with a biotinylated anti-mouse immunoglobulin G secondary antibody (Vectastain Elite ABC Kit; Vector Laboratories, Inc., Burlingame, CA, USA) for 60 min. A diaminobenzidine system (DAB Substrate Kit, SK- 4100; Vector Laboratories, Inc.) was used for visualization in accordance with the manufacturer's instructions. For the negative control, the primary antibody was replaced by a non-immune serum. Tissues from breast, colon and prostate carcinoma served as positive controls. The staining reaction shown in Fig. 1 was classified according to a semi-quantitative reference scale in the range 0–3 depending on the intensity of the p27Kip1 protein expression. The relative amount of tumour cells stained positive for p27Kip1, together with the rating of the staining intensity, resulted in a staining score (0–300, converted to percentage). Tissue microarray slides were evaluated in a blinded manner by two independent investigators.

Figure 1.

Representative immunohistochemical staining results (×160) from adjacent normal renal parenchyma and RCC. Higher expression levels were observed in RCC tissue samples for nuclear and cytoplasmic p27Kip1 staining compared to corresponding healthy renal tissue.

STATISTICAL ANALYSIS

Nuclear and cytoplasmic expression scores of p27Kip1 were compared between corresponding normal and tumour tissues using a non-parametric test (Wilcoxon). For analysis of clinical and pathological features, the survival analysis data were categorized as low and high expression. High expression of p27Kip1 was defined as ccRCC tissue that stained stronger than the respective corresponding benign tissue; low expression was defined as ccRCC tissue that stained less than or equal to the corresponding benign tissue. Both expression groups were correlated for clinical–pathological parameters, including T and N stage, Fuhrman grade and metastasis using Fisher's exact test. Survival analysis (OS and CSS) was conducted using a log-rank test. Patients alive were censored at the time of their last follow-up. For analysis of CSS, all patients who did not die cancer-specifically were censored at the time of death. All data were analyzed using SPSS software, version 12.0 (SPSS, Inc., Chicago, IL, USA). P < 0.05 was considered statistically significant.

RESULTS

PATIENT CHARACTERISTICS

The mean (sd) age of the 140 patients was 61.3 (1.1) years; 93 (66.4%) patients) were male. Patients were classified as pT1 (23.6%), pT2 (38.6%), pT3 (35%) and pT4 (2.8%). Lymph node involvement was present in 16 (11.4%) patients and 39 (27.9%) patients had evidence of distant metastatic disease. Nuclear grading according to the Fuhrman classification was G1 in 21 (15%) patients, G2 in 102 (72.9%) patients and G3 in the remainder (12.1%) (Table 1).

Table 1.  Clinical and pathological features of patients with high and low expression of p27Kip1 in the cytoplasm and nucleus of clear cell RCC cells
Clinical and pathological characteristicsAllCytoplasmic expressionNuclear expression
HighLowPHighLowP
Patients, n (%)14034 (24.3)106 (75.7) 118 (84.3)22 (15.7) 
 Age (years), mean (sd)61.34 (1.06)59.94 (2.10)61.78 (1.23) 61.11 (1.14)62.55 (2.94) 
 Range27–8827–8130–88 27–8840–85 
Sex, n (%)       
 Male93 (66.4)22 (64.7)71 (67.0)0.80778 (66.1)15 (68.2)0.850
 Female47 (33.6)12 (35.3)35 (33.0) 40 (33.9)7 (31.8) 
Histology parameters, n (%)       
 T133 (23.6)2 (5.9)31 (29.2)0.00428 (23.7)5 (22.7)0.790
 T254 (38.6)13 (38.2)41 (38.7) 46 (39.0)8 (36.4) 
 T349 (35.0)16 (47.1)33 (31.1) 40 (33.9)9 (40.9) 
 T44 (2.8)3 (8.8)1 (0.9) 4 (3.4)0 
 N0124 (88.6)20 (58.8)104 (98.1)<0.001104 (88.1)20 (90.9)0.707
 N116 (11.4)14 (41.2)2 (1.9) 14 (11.9)2 (9.1) 
 M0101 (72.1)14 (41.2)87 (82.1)<0.00184 (71.2)17 (77.3)0.560
 M139 (27.9)20 (58.8)19 (17.9) 34 (28.8)5 (22.7) 
 G121 (15.0)021 (19.8)<0.00119 (16.1)2 (9.1)0.084
 G2102 (72.9)17 (50.0)85 (80.2) 82 (69.5)20 (90.9) 
 G317 (12.1)17 (50.0)0 17 (14.4)0 
Follow-up time (months)       
 Median38.2240.2238.22 41.0514.61 
 Interquartile range12.9–77.421.2–71.410.8–78.8 14.3–80.15.1–46.5 
 Mean50.53 (4.12)51.52 (7.64)50.23 (4.87) 53.94 (4.54)32.22 (8.90) 
Total deaths, n (%)99 (70.7)29 (85.3)70 (66.0)0.03285 (72.0)14 (63.6)0.427
Tumour-related death, n (%)31 (22.1)16 (47.1)15 (14.2)<0.00125 (21.2)6 (27.3)0.528

IMMUNOHISTOCHEMISTRY

Nuclear p27Kip1 expression in normal kidney parenchyma of proximal and distal tubules was lower compared to ccRCC (mean [sd]: 13.6% [2.1%]; 95% CI, 10.8–16.4 vs 75.6% [2.7%]; 95% CI, 71.8–80.2; P < 0.001). Cytoplasmic staining of p27Kip1 was higher in ccRCC tissue compared to normal tubular tissue (mean [sd]: 13.8% [1.2%]; 95% CI, 4.1–8.7 vs 10.7% [1.7%]; 95% CI, 7.3–14.1; P= 0.04) (Fig. 1). High cytoplasmic expression of p27Kip1 was significantly associated with a higher T and N stage, Fuhrman grade and the presence of metastatic disease (P < 0.001) (Table 1). Additionally, patients with lymph node and distant metastasis were associated with high cytoplasmic p27Kip1 staining (P < 0.001) (Table 1).

SURVIVAL ANALYSIS

OS was not different between the groups of low and high p27Kip1 expression, neither for the cytoplasmic nor for the nuclear staining (Fig. 2A,C). CSS was significantly worse in patients with high cytoplasmic expression of p27Kip1 compared to patients with a low cytoplasmic expression (P < 0.001) (Fig. 2D). Patients with a high nuclear expression of p27Kip1 tended to have a better CSS compared to patients with a low nuclear p27Kip1 expression (P= 0.069) (Fig. 2B). The median (interquartile range) follow-up was 38.2 (12.9–77.4) months. The 5-year OS was not different between the groups with low and high p27Kip1 expression (cytoplasmic or nuclear). However, the 5-year CSS was reduced in patients with high cytoplasmic p27Kip1 levels (48.0%; 95% CI, 27.0–69.0) compared to patients with lower expression scores (79.7%; 95% CI, 69.5–89.9) (Table 2).

Figure 2.

Kaplan–Meier analysis of cancer-specific and overall survival of patients with high and low cytoplasmic and nuclear p27Kip1 expression. The different cytoplasmic and nuclear expressions levels did not influence patients overall survival when compared using a log-rank test (A,C). Cancer-specific survival was significantly different in high and low cytoplasmic ccRCC expression (log-rank test, P < 0.001) (D). In patients with high nuclear expression of ccRCC, cancer-specific survival tended to a better outcome compared to low expression (log-rank test P= 0.069) (B). n.s., not significant.

Table 2.  Overall and cancer-specific survival at 5 years
Variable5-year survival, % (95% CI)*
Overall survivalCancer-specific survival
  • *

    Kaplan–Meier survival analyses. ND, not determined.

All patients44.3 (34.6–53.7)70.5 (60.3–77.6)
P27 cytoplasmic expression  
 High30.2 (13.7–46.7)48.0 (27.0–69.0)
 Low48.7 (37.5–59.9)79.7 (69.5–89.9)
P27 nuclear expression  
 High48.7 (38.5–58.9)73.3 (61.7–83.9)
 Low16.7 (0–36.9)ND

DISCUSSION

The p27Kip1 protein functions as a cell-cycle regulator and suppressor of cell proliferation. The loss of p27Kip1 enhances carcinogenesis in knockout mouse models and decreased nuclear p27Kip1 expression is associated with an exaggerated proliferation of tumour cells [27,28]. Accordingly, a decreased nuclear p27Kip1 expression correlated with a poor prognosis in patients with malignant tumours [3,4]. The regulatory function of p27Kip1 in mediating the homeostasis of cell proliferation appears to be dependant on its localization. In particular, the phosphorylation of p27Kip1 is considered to regulate its shuttling between the nucleus and the cytoplasm. After cytoplasmic relocalization, p27Kip1 cannot exert its antiproliferative function. Inactivation also occurs as a result of proteolysis of p27Kip1 via the ubiquitin–proteasome pathway, mainly mediated by RING-finger containing ubiquitin ligases and the Skp, Cullin, F-box containing complex [18,29,30]. However, the exact role of p27Kip1 localization, either cytoplasmic or nuclear, and its prediction of tumour cell progression and patient survival remain unknown. For a variety of human malignancies, cytoplasmic expression of p27Kip1 was shown and could be identified as an important biological variable that influences tumorigenesis and cancer progression [19–23,25]. In breast cancer, the activation of the PKB/AKT pathway contributed to the resistance of tumour cells by cytoplasmic p27Kip1 accumulation and correlated with a poor patient prognosis [19,20]. The presence of the BCR/ABL fusion protein induced a shift of p27Kip1 into the cytoplasmic compartment in patients with a chronic myelogenous leukaemia [21]. In thyroid carcinomas, increased cytoplasmic p27Kip1expression was seen compared to healthy thyroid cells [22].

Expression of cytoplasmic p27Kip1 was found to be an independent negative prognostic parameter in patients with hepatocellular carcinoma [23]. In RCC so far, mainly the nuclear expression of p27Kip1 has been analyzed, showing a prognostic role for p27Kip1 and also that a loss of p27Kip1 results in a reduced patient outcome and is correlated with higher tumour stages and grades [8–12]. By contrast, other studies have observed even higher p27Kip1 expression in renal cancer tissue. A previous study investigated the influence of activated proteins of the PKB/AKT pathway in RCC patients and reported significant elevations of nuclear p27Kip1 in kidney tumours compared to corresponding healthy tissue. Higher p27Kip1 protein expression in immunohistochemistry was confirmed by western immunoblotting and real-time PCR analysis [13]. A recent study by Kim et al. [16] examined 20 cases of primary tumours, renal cell cancer and normal cells for p27Kip1 expression and localization. Higher levels of p27Kip1 were found in tumours compared to matched healthy control tissues. Furthermore, the cytoplasmic localization of p27Kip1 was correlated with a higher tumour grade [16]. Recently, a large trial in 375 patients with different histological RCC-subtypes showed positive nuclear p27Kip1 immunoreactivity in 78% of tumours and cytoplasmic staining in 46% of tumours. Furthermore, higher nuclear p27Kip1 expression was confirmed in ccRCC and inversely correlated with tumour size. In univariate analysis, metastatic RCC patients showed higher cytoplasmic p27Kip1 levels and cytoplasmic localization was identified as a predictive marker for negative disease-specific survival. However, cytoplasmatic p27Kip1 expression failed to show a significant difference in the multivariate analysis [24]. The present study confirmed higher p27Kip1 expression in the nuclear and cytoplasmic compartment of ccRCC samples compared to normal renal tissue. The results obtained support the hypothesis of the regulatory role of p27Kip1 in both the nucleus and the cytoplasm. We showed a differential expression of p27Kip1 in the subcellular compartments of the cytoplasm and nucleus, with a lower cytoplasmic p27Kip1 expression compared to the nucleus. Moreover, high cytoplasmic p27Kip1 levels showed a significant correlation with an unfavourable clinic and were associated with a reduced CSS during the median follow-up period of 38 months. These results affirm the potential impact of cytoplasmic p27Kip1 localization for enhanced cancer aggressiveness, as reflected by adverse cancer stage, grade and outcome. The present study also showed a strong trend towards a decreased CSS in association with lower nuclear protein expression. Such results imply that nuclear p27Kip1 loss as a result of a possible cytoplasmatic translocation may enable RCC cells to escape from the p27Kip1 controlled cell-cycle arrest to a progressive disease with a reduced CSS. This raises the question of whether therapeutic p27Kip1 stabilization may slow down cancer growth or induce apoptosis in RCC.

In this respect, small interfering RNA-mediated gene silencing has already shown promising results for preventing p27Kip1 degradation in various human cancer cell lines [17,31,32]. In RCC, small interfering RNA knockdown enabled the relocalization of p27Kip1 to the nucleus. This nuclear shift of p27Kip1 from the cytoplasm was reported to result in increased apoptosis. In addition, a mislocalization of p27Kip1 with enhanced apoptotic resistance was shown as a result of activation of the PKB/AKT pathway in RCC cell lines. Interestingly, only treatment with a phosphoinositide 3-kinase inhibitor, and not blockade of the mammalian target of rapamycin, resulted in nuclear relocalization of p27Kip1[16]. Accordingly, Cho et al. [15] confirmed a higher efficiency of dual phosphoinositide 3-kinase/mammalian target of rapamycin inhibition (NVP-BEZ235) compared to rapamycin alone in RCC lines. Treatment with NVP-BEZ235 not only inhibited PKB pathway activation, but also nuclear translocation of p27Kip1[15]. The potential advantage of a p27Kip1 relocalization effect in future targeted therapies requires a standardized interpretation of p27Kip1 tissue expression levels. Therefore, only strict discrimination between nuclear and cytoplasmic p27Kip1 can further substantiate the role of p27Kip1 for the identification of those patients who are suitable for new targeted therapies. In conclusion, the cytoplasmic shift of the cell-cycle regulator p27Kip1 and the loss of nuclear control appear to play important roles in cancer progression. A significant correlation of cytoplasmatic p27Kip1 levels with established parameters of ccRCC aggressiveness and, more importantly, with reduced cancer-specific outcome could be shown. The results reported in the present study substantiate the oncogenic potential of mislocated cytoplasmic p27Kip1 in recent cell culture and mouse models. Furthermore, higher levels of ‘properly located’ nuclear p27Kip1 raise new questions about the stronger impact of p27Kip1 localization rather than p27Kip1 expression in the progression of ccRCC. In the future, patients with high cytoplasmic p27Kip1 and with a high risk of death from renal cancer may benefit from new drugs that prevent cytoplasmic translocation of p27Kip1.

CONFLICT OF INTEREST

None declared.

Ancillary