Aberrant nucleocytoplasmic localization of the retinoblastoma tumor suppressor protein in human cancer correlates with moderate/poor tumor differentiation

Abstract

Inactivation of the retinoblastoma (RB) tumor suppressor pathway, via elevated cyclin-dependent kinase (CDK) activity, is observed in majority of human cancers. Since CDK deregulation is evident in most cancer cells, pharmacological CDK inhibition has become an attractive therapeutic strategy in oncology. We recently showed that an oncogenic CDK4R24C mutation alters the subcellular localization of the normally nuclear RB phosphoprotein. Here, using 71 human cancer cell lines and over 300 primary human cancer tissues, we investigated whether changes in RB subcellular localization occur during human cancer progression. We uncover that diverse human cancers and their derived cell lines, particularly those with poor tumor differentiation, display significant cytoplasmic mislocalization of ordinarily nuclear RB. The nucleocytoplasmically distributed RB was derived via CDK-dependent and Exportin1-mediated nuclear export. Indeed, cytoplasmically mislocalized RB could be efficiently confined to the nucleus by pharmacologically reducing CDK activity or by inhibiting the Exportin1-mediated nuclear export pathway. Our observations uncover a post-translational CDK-dependent mechanism of RB inactivation and suggest that cytoplasmically localized RB may harbor a tumor promoting function. We propose that RB inactivation, via aberrant nucleocytoplasmic transport, may disrupt normal cell differentiation programs and accelerate the cancer process. These results are evidence that tumor cells modulate the protein transport machinery thereby making the protein transport process a viable therapeutic target.

Introduction

Mutations in cell cycle components that allow cells to bypass quiescence or cellular senescence pathways are important hallmarks of a cancer cell (Hanahan and Weinberg, 2000). The universal targeting of the cell-cycle machinery makes it an important target for anticancer therapeutic strategies (Malumbres and Barbacid, 2001; Schwartz and Shah, 2005; Shapiro, 2006). The retinoblastoma (RB) tumor suppressor pathway is inactivated in a majority of human cancers by overexpression of cyclins, aberrant activation of cyclin-dependent kinases (CDKs), inactivation of CDK inhibitors (CKIs) or loss of RB expression. RB is active in its hypo or underphosphorylated state whereas sequential CDK-mediated hyperphosphorylation of RB on serine/threonine phosphorylation sites leads to inactivation of RB tumor suppressor function (Knudsen and Wang, 1997). Phosphorylation by CDKs alters association of RB with its myriad interacting proteins that regulate cell cycle progression and transformation potential (Classon and Harlow, 2002).

Members of the INK4 family of proteins, chiefly p16Ink4a, are specific inhibitors of the cyclin D/CDK4 complexes (Malumbres and Barbacid, 2001) and mutations in CDK4 and p16INK4A are implicated in the genesis and progression of human cancer (Haluska and Hodi, 1998; Liggett and Sidransky, 1998; Ortega et al., 2002). The importance of the CDK4 locus in human cancer was further emphasized upon identification of a germline CDK4-Arg24Cys (R24C) mutation in patients with familial melanoma that abolishes the ability of the mutant CDK4R24C kinase to interact with the p16INK4A inhibitor (Wolfel et al., 1995; Zuo et al., 1996). Using mouse models, we and others showed that inheritance of the p16Ink4a-insensitive Cdk4R24C allele results in increased Cdk4 kinase activity, thereby increasing the transformation potential of cells and predisposing mice harboring this mutation to cancer due to loss of RB tumor suppressor function (Rane et al., 1999, 2002; Sotillo et al., 2001a, 2001b).

RB is regarded to be a nuclear phosphoprotein and its nuclear localization is facilitated by (1) a bipartite nuclear localization signal (NLS) in the C terminus (Zacksenhaus et al., 1993) and (2) association of its N terminus with nuclear matrix proteins (Durfee et al., 1994). Phosphorylation of RB by CDKs during the G1/S phase of the cell cycle results in decreased affinity of hyperphosphorylated RB for the nuclear compartment (Mittnacht and Weinberg, 1991; Stokke et al., 1993; Mittnacht et al., 1994). Transition through the G1/S boundary of the cell cycle leads to conversion of RB from a low salt resistant to low salt extractable hyperphosphorylated species (Mittnacht and Weinberg, 1991). Recently, we showed that RB is confined to the nucleus in normal human and mouse fibroblast cells during all stages of the cell cycle (Jiao et al., 2006). In contrast, nucleocytoplasmically localized RB was observed in fibroblasts harboring the oncogenic Cdk4R24C mutation. We demonstrated that the aberrant nucleocytoplasmic localization is facilitated by CDK-dependent nuclear export of RB by Exportin1 (CRM1). Further, we showed that phosphorylation residues in the C-domain of RB are critical in determining its nucleocytoplasmic localization (Jiao et al., 2006).

These results suggested that altered subcellular localization of RB may be a cancer-specific process, and here we investigated this possibility. Specifically, we inquired whether the deregulated CDK activity observed in many human cancers, often associated with hyperphosphorylation of RB, might alter RB subcellular localization and thereby compromise its tumor suppressor function. We present evidence of cytoplasmic mislocalization of RB in diverse human cancer cell lines and primary tumor specimens that exhibit poor tumor differentiation. Inhibition of CDK activity or the Exportin1-mediated nuclear export pathway resulted in retention of RB in the nucleus. Our results reveal that RB-mediated tumor suppression can be subverted during human cancer progression by CDK-phosphorylation-dependent enhancement of nuclear export.

Results

CDK-mediated phosphorylation regulates RB subcellular localization and function

The human prostate cancer cell line PC-3 is a metastatic androgen-independent cell line that was established from a human prostatic adenocarcinoma metastatic to bone. The functional and morphologic characteristics of PC-3 are those of a poorly differentiated adenocarcinoma (Kaighn et al., 1979). Higher basal levels of CDK2, CDK4 and p16 genes are constitutively expressed in PC-3 cells (Lu et al., 1997) and these cells proliferate extensively in culture (Figure 1a). Interestingly, immunofluorescence assays revealed that RB was dispersed over the nucleocytoplasmic compartments in PC-3 cells (Figure 1b). The observation of cytoplasmic RB is intriguing since RB is widely believed to be a nuclear protein and nuclear retention is presumed to be important for its tumor suppressor function. Previously, we elucidated the occurrence of nucleocytoplasmically distributed RB in Cdk4R/R mouse embryonic fibroblasts (Jiao et al., 2006). Here, we investigated the requirement for CDK activity in determining the nucleocytoplasmic distribution of RB by treating PC-3 cells with the potent CDK inhibitor flavopiridol (Sedlacek, 2001). Since previous studies have suggested that high concentration of flavopiridol can lead to reduced cell viability, we first performed experiments to examine the dose-dependent effect of flavopiridol on PC-3 cells. Treatment of PC-3 cells with increasing concentration of flavopiridol (up to 1.0 μM) resulted in an expected growth inhibition, although we observed greater than 80% cell viability (Supplementary Figure 1a). Higher doses of flavopiridol (2 and 4 μM) caused increased cell death with less than 30% viability. We therefore used a concentration of less than 1.0 μM flavopiridol for all experiments in this study. At similar concentrations of flavopiridol treatment, others have shown minimal effect on PC-3 cell viability and RB protein levels and an expected reduction in RB phosphorylation was observed (Camphausen et al., 2004). Treatment with flavopiridol resulted in reduced viability of PC-3 cells (Figure 1a). Importantly, similar to our observations with flavopiridol-treated Cdk4R24C cells (Jiao et al., 2006), flavopiridol treatment of PC-3 cells resulted in nuclear retention of RB (Figure 1b). A dose-dependent increase in nuclear retention of RB was observed with flavopiridol concentration between 0.1 and 0.5 μM (data not shown). In agreement with this, treatment of PC-3 cells with CDK2 siRNA expression vectors (Sui et al., 2002) lead to efficient nuclear retention of RB (Figure 1b), further validating the importance of CDK activity in regulating the subcellular distribution of RB.

Figure 1
figure1

Nucleocytoplasmic distribution of retinoblastoma (RB) in PC-3 cells. (a) Viability of vehicle-treated (untreated; open bars) or flavopiridol (0.5 μM) treated (closed bars) PC-3 cells for the indicated days. (b) RB (red) localization in vehicle-treated (PC-3), flavopiridol-treated (PC-3+0.5 μM Flavo) and siCDK2 treated (PC-3+siCDK2) PC-3 cells after 48 h treatment. (c) Promoter activity of the E2F-target gene cyclin E. Relative activity of the cyclin E promoter (pCE) was determined after treatment with increasing doses of flavopiridol (0.05 and 0.1 μM) for 24 h. (d) Vehicle-treated (closed bars) or flavopiridol (10 mg kg−1 per dose) treated (open bars) PC-3 xenografts were analysed for tumor regression. Total of five tumor measurements were performed (x axis), once during every four injections and at final necropsy. (e) RB localization in vehicle-treated (left panel) or flavopiridol (10 mg kg−1 per dose) treated (right panel) PC-3 xenografts. Arrowheads indicate cytoplasmic localization of RB in vehicle-treated PC-3 xenografts. Brown staining indicates RB immunoreactivity. Note the exclusive nuclear retention of RB in flavopiridol treated PC-3 xenografts (right panel).

In the nucleus, underphosphorylated RB associates with E2F transcription factors and preclude E2Fs from activating or repressing their target genes (Stevens and La Thangue, 2003). Elevated CDK activity, which leads to phosphorylation of RB, results in the release of E2Fs, which in turn activate downstream targets such as the cyclin E gene (Muller et al., 2001). We next examined whether the abnormal nucleocytoplasmic localization of RB in PC-3 cells compromised RB's negative regulation of E2F transcriptional activity. Since high concentration of flavopiridol could potentially induce a global inhibition of transcription, we first analysed the effect of flavopiridol treatment on promoters that are not activated by E2Fs. No significant change in reporter activity was observed in flavopiridol-treated PC-3 cells transfected with luciferase reporters expressing β-galactosidase, topoisomerase-2β, insulin and histone H1 (Supplementary Figures 1b and c). In addition, we observe no significant reduction in the RNA and protein levels of RB and cyclin D1 in flavopiridol-treated PC-3 cells (Supplementary Figures 1d and e). Together, these observations suggest that treatment of PC-3 with flavopiridol does not cause a global inhibition of transcription. Next, to monitor E2F transactivation potential in flavopiridol-treated PC-3 cells, we studied the promoter response of the E2F-target gene cyclin E. We observed a reduction in cyclin E promoter activity in flavopiridol-treated PC-3 cells (Figure 1c), which is consistent with increased RB-mediated repression of E2F transactivation potential.

Next, we examined the effects of flavopiridol on the tumorigenesis potential of PC-3 cells and the plausible role of RB subcellular localization in the process. In vivo, PC-3 cells exhibit enhanced tumorigenesis potential in xenograft-tumor assays and treatment with flavopiridol resulted in regression of the xenografted PC-3 tumors (Figure 1d). Untreated PC-3 xenografts, similar to PC-3 cells in culture, showed evidence of nucleocytoplasmically localized RB (Figure 1e). Interestingly, flavopiridol-induced regression of PC-3 xenografted tumors was accompanied by enhanced nuclear re-localization of RB (Figure 1e). Together, these results indicate that RB cytoplasmic mislocalization can be effectively reversed by pharmacological inhibition of CDK activity and the nuclear re-confinement of RB is associated with enhancement of cellular apoptosis, reduction in E2F promoter activity and regression of tumor burden in mice.

Exportin1-mediated nuclear export of RB in human cancer cells

Nuclear export of the tumor suppressor proteins p27Kip1, APC, Smad4(DPC4) and p53 is known to inactivate their tumor suppressor function (Boyd et al., 2000; Geyer et al., 2000; Henderson, 2000; Rosin-Arbesfeld et al., 2000; Blain and Massague, 2002; Inman et al., 2002; Kastan and Zambetti, 2003; Xu and Massague, 2004). The karyopherin family of nuclear export receptors (Weis, 2003), notably Exportin1 (CRM1) (Stade et al., 1997), are involved in trafficking of diverse substrates across the nuclear membrane. Our recent studies uncovered an association between the RB and Exportin1 proteins in Cdk4R/R cells, where Exportin1 preferentially recognizes and associates with RB phosphorylated on C-terminal residues and thereby mediates export of this phosphorylated RB species (Jiao et al., 2006). Consistent with the immunofluorescence experiments (Figure 1b), western blot analysis of nucleocytoplasmic fractions from PC-3 cells revealed nucleocytoplasmic localization of RB (Figure 2a). Furthermore, similar to our prior observation of RB–Exportin1 association in Cdk4R/R cells (Jiao et al., 2006), co-immunoprecipitation experiments revealed an interaction of RB and Exportin1 in PC-3 cells (Figure 2b) suggesting that RB could be subject to Exportin1-mediated nuclear export. In agreement, treatment of PC-3 cells with leptomycin B (LMB), an inhibitor of Exportin1-mediated nuclear export (Stade et al., 1997; Weis, 2003), resulted in a shift of RB from the cytoplasm to the nucleus (Figure 2a). To verify the role of Exportin1 in the nuclear export of RB we designed siRNA targeting Exportin1. Treatment of PC-3 cells with Exportin1 siRNA leads to growth arrest although cell viability was retained at near 90% (Figure 2c). Importantly, treatment of PC-3 cells with Exportin1 siRNA resulted in increased nuclear confinement of RB (Figure 2d). Taken together, these observations are evidence of Exportin1-mediated nuclear export of RB in human cancer cells.

Figure 2
figure2

Nuclear export of retinoblastoma (RB) via association with Exportin1. (a) Nuclear and cytoplasmic fractions from vehicle-treated (−) or leptomycin B-treated (+) PC-3 cells were analysed by western blot analysis using total RB antibodies. Lamin A/C and tubulin antibodies are controls for nuclear and cytoplasmic fractions, respectively. (b) Total protein from Cdk4R24C and PC-3 cells was immunoprecipitated (IP) with control IgG or total RB antibody followed by immunoblot (IB) with anti-Exportin1 antibodies (Exp-1). Control lysate (Lys) from both cell lines is shown. (c) Cell viability of PC-3 cells treated with control siRNA (open bars; control Si) or with siRNA against Exportin1 (closed bars; Exp1-si) for the indicated days. (d) RB (red) localization in PC-3 cells treated with control siRNA (control Si) and treated with siRNA against Exportin1 (siExportin1).

RB nucleocytoplasmic distribution in human cancer cells of diverse origin

Interestingly, SK-Mel-29 human melanoma cells, which were used to report the initial identification of the Cdk4R24C mutation in human familial melanoma (Wolfel et al., 1995), also present nucleocytoplasmically distributed RB (Figure 3). Furthermore, as with PC-3 cells, treatment of SK-Mel-29 cells with flavopiridol resulted in increased nuclear retention of RB (Figure 3). In both PC-3 and SK-Mel-29 cells, nuclear retention of RB was observed within 48 h of treatment with flavopiridol when the cells have a high degree of viability. Importantly, treatment of SM-Mel-29 cells with LMB resulted in increased nuclear confinement of RB (Figure 3). Taken together, these observations suggested that the RB nucleocytoplasmic localization in SK-Mel-29 cells, like in PC-3 cells, is dependent on CDK activity and regulated via Exportin1-mediated nuclear export.

Figure 3
figure3

Nucleocytoplasmic localization of retinoblastoma (RB) in SK-Mel-29 cells and its relocalization to the nucleus upon treatment with flavopiridol and leptomycin B. RB (red) and phospho-RB (green) localization in vehicle-treated (control), flavopiridol-treated (flavopiridol; 0.5 μM for 48 h) and leptomycin-B-treated (LMB; 30 nM for 3 h) SK-Mel-29 human melanoma cells. 4′-6-diamidino-2-phenylindole (DAPI) nuclear fluorescence is shown in bottom panels.

Since inactivation of the RB pathway via aberrant CDK activation is seen in virtually every human cancer, we hypothesized that nucleocytoplasmically dispersed RB may be observed in human cancer of diverse origin. To explore that proposition, we wanted to characterize RB localization in a set of cancer cells whose molecular characteristics have been cataloged and whose responses to a large number of anticancer compounds are known. The natural choice was a panel of 60 human cancer cell lines (NCI-60) used by the Developmental Therapeutics Program (DTP) of the National Cancer Institute (NCI) to screen >100 000 chemical compounds since 1990. Included among the 60 cell lines are leukemias, melanomas and cancers of ovarian, breast, prostate, lung, renal, colon and central nervous system origin. Patterns of drug activity across the cell lines and patterns of cell sensitivity across the set of tested drugs have been shown to contain detailed information on mechanisms of action and resistance (Paull et al., 1989; Weinstein et al., 1992). In addition to this pharmacological characterization, the NCI-60 cells have been extensively profiled at the DNA, mRNA, protein and functional levels. Therefore, if we were to measure their localization of RB expression, it would be possible to link RB localization and its tumor suppressor function to a variety of other molecular, physiological and pharmacological features of the cells. To this end, we performed immunohistochemistry analysis of the NCI-60 cell line array (CMA) and an additional panel of 11 human cancer cell lines representing melanoma (SK-Mel-28, SK-Mel-29, SK-Mel-39, SK-Mel2), prostate cancer (PC-3, LnCap), colon cancer (RKO, HT29, HCT116) and breast cancer (MDA-MB-468, MDA-MB-231) for a total of 71 human cancer cell lines. Two distinct monoclonal antibodies to total RB and two polyclonal antibodies against RB phosphorylated on serines 807/811 and serine 780 were used. Strikingly, 63% (45 out of 71 cell lines) cell lines exhibited dispersed nucleocytoplasmic distribution of total RB (Table 1). We also observed that a substantial number of cell lines demonstrated predominantly cytoplasmic distribution of total RB (25% or 18 out of 71 of cell lines). In contrast, predominant/exclusive nuclear localization of total RB was featured in only 12% (8 out of 71 cell lines). We observed a co-relation between localization of RB and phospho-RB in the human cancer cell lines. Phospho-RB was detected in the nucleocytoplasmic location in 81% cell lines, whereas, 13% of the cell lines showed predominantly cytoplasmic and 6% cell lines exhibited predominantly nuclear phospho-RB.

Table 1 Nucleocytoplasmic distribution of retinoblastoma (RB) in human cancer cell lines

To investigate if nucleocytoplasmic RB might correlate with cell proliferation we performed immunohistochemistry on CMA using antibodies against a proliferation-associated antigen Ki-67. CMA demonstrates a measurable proliferative rate with Ki-67 staining, which is largely related to the rapid cell cycle time. The populations of cells of CMA are homogeneous and provide a representative sampling for comparison of parallel immunohistochemistry experiments. Moreover, by performing immunohistochemistry on CMA, it allows us to directly compare the proliferative rate of the cell population that was stained for RB (and phospho-RB). As expected, the analysis using a Spearman's rank test revealed that total RB and phospho-RB correlate (P=0.0026). Total RB also correlated with Ki-67 expression (P=0.0062) and importantly the majority of Ki-67 expressing cells exhibited nucleocytoplasmic localization of total RB (51/54 cell lines) or phospho-RB (42/52 cell lines).

Nucleocytoplasmic distribution of RB correlates with moderate/poor tumor differentiation

To further examine the physiological relevance of RB subcellular localization we performed immunohistochemical staining of a panel of human tumors of breast, colon, lung, ovarian, prostatic, brain, melanoma and lymphoma origins presented in a tissue microarray (TMA) format. Immunohistochemistry analysis using total RB and phospho-RB antibodies revealed that an unexpectedly high proportion of tumors (128 out of 313; 41%) exhibited predominantly nucleocytoplasmic localization of total RB (>75% cells with nucleocytoplasmic localization) (Table 2). A high percentage of tumors presented with predominantly cytoplasmic localization of total RB (70 out of 313; 22%). In contrast, only 39 out of 313 (13%) tumors presented predominantly nuclear localization (>75% cells with nuclear localization) of total RB. Moreover, we failed to detect RB in 76 out of 313 (24%) tumors. As anticipated, normal tissues demonstrated low level of RB with predominantly nuclear staining. Normal prostate corresponding to the prostatic tumors was additionally examined and demonstrated the same findings (data not shown). Further, we independently profiled the expression and localization of phospho-RB in 197 tissues representing breast, lung, colon, ovarian and prostatic origin. RB phosphorylated on serine 807 and serine 811 was observed in a great majority (134 out of 197; 68%) of primary tumors. A high percentage of tumors presented with predominantly cytoplasmic localization of phospho-RB (45 out of 197; 23%). In contrast, only 18 out of 197 (9%) tumors presented predominantly nuclear localization of phospho-RB.

Table 2 Nucleocytoplasmic distribution of retinoblastoma (RB) in primary human cancer tissues

Importantly, we found a correlation between nuclear retention and tumor differentiation. The majority of tumors with exclusively nuclear RB were well differentiated (9 out of 14; 64% tumors). In contrast, intermediate or high-grade tumors with moderate (32 out of 45; 71%) or poor (26 out of 36; 72%) differentiation presented predominant nucleocytoplasmic localization of RB (Table 3). This observation is in agreement with our data with the human cancer cell lines that also show a high incidence of nucleocytoplasmically distributed RB. Since majority of established cancer cell lines proliferate extensively and exhibit characteristics of poorly differentiated cells, the observations presented here further support the association of altered nucleocytoplasmic RB distribution with poor tumor cell differentiation.

Table 3 Nucleocytoplasmic distribution of retinoblastoma (RB) correlates with moderate to poor tumor differentiation

Discussion

We observe that diverse human cancers and their derived cell lines, particularly those with poor tumor differentiation, display significant cytoplasmic mislocalization of ordinarily nuclear RB. The cytoplasmic RB population is generated by CDK-dependent Exportin1-mediated nuclear export. Cytoplasmically mislocalized RB could be efficiently confined to the nucleus by (1) inhibiting the Exportin1 pathway and (2) reducing CDK activity by treatment with the CDK inhibitor flavopiridol. Re-localization of RB to the nucleus in turn is correlated with restoration of RB tumor suppressor function. It is plausible that some of the flavopiridol-mediated effects are RB-independent since there is evidence for both CDK mediated RB-dependent and RB-independent mechanisms of action involving the antitumoral effect of flavopiridol (Sedlacek, 2001).

Altered RB subcellular localization: a cancer-specific phenomenon

The function of tumor suppressor proteins APC, p53, p27Kip1and Smad4 (DPC4) can be inactivated by nuclear export mechanisms (Boyd et al., 2000; Geyer et al., 2000; Henderson, 2000; Rosin-Arbesfeld et al., 2000; Blain and Massague, 2002; Inman et al., 2002; Kastan and Zambetti, 2003; Xu and Massague, 2004; Ziegler and Ghosh, 2005). Similarly, we illustrate here the possibility that RB-mediated tumor suppression can be effectively subverted during human cancer progression by CDK phosphorylation-dependent enhancement of nuclear export. These observations unravel an additional layer of regulatory control on RB tumor suppressor function that is influenced by CDK activity. Interestingly, p130 localization is also regulated by nucleocytoplasmic shuttling (Chestukhin et al., 2002), which elucidates a common mechanism that regulates subcellular localization of RB and p130.

An RB mutant that cannot be phosphorylated by CDKs on its seven C-terminal phosphoresidues remains confined to the nucleus of Cdk4R/R cells, in contrast to wild-type RB that is subject to nucleocytoplasmic shuttling in Cdk4R/R cells (Jiao et al., 2006). Furthermore, cytoplasmic RB is not detected in normal cells (WI-38 human lung fibroblasts and normal mouse embryonic fibroblasts) at any stage of the cell cycle (Jiao et al., 2006) suggesting that (1) CDK-mediated phosphorylation may be necessary but not sufficient for nuclear exclusion and (2) cytoplasmic RB localization may be specific to pathological conditions, such as cancer. The postulate that altered RB subcellular localization is a cancer-specific phenomenon is in agreement with our observations of nucleocytoplasmic RB in Cdk4R/R cells (Jiao et al., 2006) and, as reported here, in a majority of human cancer cell lines and human tumor tissues of diverse origin. We and others (Yen et al., 1997) found nucleocytoplasmic RB in HL-60 cells (one of the NCI-60 cell lines) where the ratio of nuclear versus cytoplasmic RB was found to be stable. Although not tested here, it is plausible that the cytoplasmic RB performs tumor-promoting functions that are facilitated by deregulated CDK-mediated phosphorylation and further detailed studies are needed to examine such a hypothesis. It will be of interest to know (1) the fate of cytoplasmic RB in cells that exhibit altered RB localization and (2) if cytoplasmic RB is degraded or gets imported into the nucleus. Tumor cell lines, notably the osteosarcoma cell line SAOS-2, contain truncated forms of RB that are localized to the cytoplasm due to lack of a NLS signal. Therefore, cytoplasmic localization of RB in human cancer may occur either via aberrant nuclear import due to a nonfunctional NLS or, as presented here, due to accelerated nuclear export. Further, analysis of human RB tumor samples showed that while some RB tumors present strong nuclear RB staining many tumors exhibit predominantly cytoplasmic or dispersed nucleocytoplasmic localization of RB (Nork et al., 1994). These observations indicate that (1) RB nuclear localization per se may not determine inhibition of tumorigenesis and (2) if nuclear localization is not sufficient then the possibility of a cytoplasmic function for RB formally exists.

Differentiation and cancer: the role of RB

RB regulates cellular differentiation and survival (Goodrich, 2006; Khidr and Chen, 2006) and promotes differentiation and development of neural cells, skeletal muscle cells, retinal cells and adipocytes (Khidr and Chen, 2006). It is plausible that inactivation of RB during cancer progression may inhibit normal differentiation processes and allow cells to aberrantly remain in the cell cycle and undergo cell division. Our observation of nucleocytoplasmic localization of RB in majority of moderately/poorly differentiated tumors is consistent with such a possibility. Interestingly, the link between altered subcellular localization of a cell cycle regulator (the p27Kip1 tumor suppressor protein) and tumor differentiation grade has been established (Alkarain et al., 2004). While the P27KIP1 gene is rarely mutated in human cancer, the action of p27Kip1 protein is impaired in breast and other human cancers. Interestingly, similar to our observations with RB, it was demonstrated that p27Kip1 mislocalization in the tumor cell cytoplasm occurs in a majority of breast cancer tissues (Blain and Massague, 2002; Liang et al., 2002; Shin et al., 2002; Viglietto et al., 2002) where reduced p27Kip1 protein is strongly associated with high histopathological tumor grade, reflecting a lack of tumor differentiation. Similarly, we observe that RB is nucleocytoplasmically localized in the majority of human cancer cell lines and primary tumor tissue specimens from diverse tissue origin—majority of which retain enhanced proliferation potential and are poorly differentiated. These observations allow us to suggest that altered nucleocytoplasmic localization of RB may be a characteristic of poorly differentiated cancer cells. However, further detailed analyses are needed to validate this possibility and to determine whether RB subcellular localization has predictive and/or prognostic value in human cancer. Mutations in the nuclear transport machinery via truncations or overexpression of the export receptor for karyopherin-α in colon, breast and liver neoplasms and chromosomal rearrangements in the loci coding for nucleoporins in acute myelogenous leukemia, chronic myelogenous leukemia, T-cell acute lymphoblastic leukemia and myelodysplastic syndrome emphasize the relevance of the nuclear transport apparatus in human cancer (Kau et al., 2004). The findings that critical tumor suppressor proteins like p53, Smad4, APC, p27Kip1 and RB are direct targets of the nuclear transport apparatus are consistent with the theory that tumor cells may modulate the protein transport apparatus to evade growth regulatory constraints making the protein transport machinery a viable therapeutic target.

Materials and methods

Cell culture and drug treatments, xenograft assays, plasmids and human tumor tissues

Mouse embryo fibroblasts (MEFs) (Rane et al., 1999, 2002) were cultured in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS). Human multi-tumor tissue arrays (TARP TMAs), NCI-60 human cancer CMA and human cancer cell lines representing melanoma (SK-Mel-28, SK-Mel-29, SK-Mel-39, SK-Mel-2; prostate cancer (PC-3, LnCap); colon cancer (RKO, HT29, HCT116); breast cancer (MDA-MB-468, MDA-MA-231) were obtained from the Developmental Therapeutics Program and the Tissue Array Research Program of the NCI. Although derived from independent sources and propagated in the laboratory SK-Mel28, SK-Mel2, PC3, HT29, HCT116, MDA-MB-231 were also part of CMA.

MEFs, PC-3 or SK-Mel-29 cells were cultured in 10% FBS containing DMEM and treated with flavopiridol where indicated and cells were monitored for viability using the WST-1 assay (Roche, Indianapolis, IN, USA) or processed for luciferase assay. Results presented are average of triplicate with standard deviations. Nuclear and cytosolic fractions of cells either untreated or treated with 10 or 30 nM LMB (L2913, Sigma, Saint Louis, MO, USA) for 5 h were prepared using NE-PER reagent (Pierce, Rockford, IL, USA) and 30 μg aliquots were used for immunoblot analysis. Cells were treated with 10 and 30 nM LMB and incubated for 2–5 h followed by immunofluorescence assay. Control and Exportin1 siRNAs were from Dharmacon Inc., Chicago, IL, USA. Transfection was performed with Lipofectamine Reagent (Invitrogen, Carlsbad, CA, USA).

For xenografts, 5 × 106 PC-3 cells were injected subcutaneously into bilateral flanks of athymic BALB/c nu/nu male mice (6–8 weeks of age, six mice in each group, Taconic Laboratories, Hudson, NY, USA). One week later, mice were treated intraperitoneally with flavopiridol (10 mg kg−1 per dose) or 0.1% dimethyl sulfoxide (DMSO) saline once every 4 days for a total of four treatments. Tumors were measured every 4 days and at necropsy for a total of five times and mean tumor volume was determined according to the formula (4/3)πr3 where r=(diameter1+diameter2)/4. Tumors were excised and processed for immunohistochemistry using RB antibodies (G3-245 and 4H1 clone) and counterstained with hematoxylin.

Transient transfection and reporter assays

Cells were seeded at 30–40% confluence into six-well plates one day before transfection. Treatment with flavopiridol (0.05–0.1 μM) was for 24 h. A 1 μg portion of luciferase driven reporters together with 0.2 μg pH1-β-galactosidase was co-transfected into cells using FuGENE 6 (Roche) or Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Empty vectors were used to supplement equal amounts of DNA in each transfection. Forty-eight hours after transfection, cells were rinsed with phosphate-buffered saline (PBS) and either lysed with reporter lysis buffer (Promega, Madison, WI, USA) for luciferase assay. Lysates were assayed for luciferase activity and normalized to total protein concentration and β-galactosidase activity. All experiments were performed in duplicate and repeated 2–3 times. The results shown are the mean±s.e.

Immunoblotting, immunoprecipitation, immunofluorescence and immunohistochemistry

Western blots, immunoprecipitation and immunohistochemistry assays were performed using standard methods. For immunoprecipitation, 300 μg of total protein was incubated with indicated antibodies and protein G beads (Amersham, Piscataway, NJ, USA). For immunofluorescence, cells were seeded at 80% confluence onto collagen I coated eight-well Biocoat culture slides. Drug treatments for immunofluorescence were as follows—flavopiridol (0.5 μM for 48 h) and leptomycin B (30 nM for 3 h). After culture the cells were fixed in 1 or 4% paraformaldehyde for 10 min and permeabilized with either acetone (2 min) or methanol (2 min). After antibody incubation and mounting in 4′-6-diamidino-2-phenylindole (DAPI) containing medium cells were observed on an Olympus IX70 inverted microscope with a Photometric CCD camera. Immunohistochemistry was performed by routine methods and results were analysed for staining intensity and staining localization (Braunschweig et al., 2005). The analysis of immunohistochemistry of the cell and tissue arrays was performed for each core, counting the amount of cells with a distinct pattern (nuclear, cytoplasmic and nucleocytoplasmic) in percentages. Additionally, the intensity of the labeling was registered as an estimated average from 0 (=no stain) up to 3 (=dark brown, max. stain) for each cellular pattern for each sample. Localization was determined by scoring nuclear/cytoplasmic and nuclear profiles from at least 100 cells of each sample treatment. Antibody sources are monoclonal anti-pRB antibody (clone G3-245; BD/Pharmingen; epitope amino acids (aa) 332–344 of human RB protein); monoclonal anti-RB antibody (clone 4H1; Cell Signaling Technology, Danvers, MA, USA, which recognizes aa 701–928 of human RB protein), monoclonal underphosphorylated-RB, phospho-RB780, phospho-RB807/811 polyclonal antibodies from Cell Signaling; monoclonal anti-α-tubulin (Sigma); monoclonal anti-Ki-67 (DAKO); polyclonal anti-Exportin1 (sc5595; clone H300); Lamin A/C polyclonal antibodies from Santacruz, Santa Cruz, CA, USA; and Alexa-fluor secondary antibodies from Molecular Probes, Carlsbad, CA, USA. The association of the expression rate with RB and phospho-RB was assessed by χ2-test. A P-value less than 0.05 was regarded as statistically significant.

References

  1. Alkarain A, Jordan R, Slingerland J . (2004). p27 deregulation in breast cancer: prognostic significance and implications for therapy. J Mammary Gland Biol Neoplasia 9: 67–80.

  2. Blain SW, Massague J . (2002). Breast cancer banishes p27 from nucleus. Nat Med 8: 1076–1078.

  3. Boyd SD, Tsai KY, Jacks T . (2000). An intact HDM2 RING-finger domain is required for nuclear exclusion of p53. Nat Cell Biol 2: 563–568.

  4. Braunschweig T, Chung JY, Hewitt SM . (2005). Tissue microarrays: bridging the gap between research and the clinic. Expert Rev Proteomics 2: 325–336.

  5. Camphausen K, Brady KJ, Burgan WE, Cerra MA, Russell JS, Bull EE et al. (2004). Flavopiridol enhances human tumor cell radiosensitivity and prolongs expression of gammaH2AX foci. Mol Cancer Ther 3: 409–416.

  6. Chestukhin A, Litovchick L, Rudich K, DeCaprio JA . (2002). Nucleocytoplasmic shuttling of p130/RBL2: novel regulatory mechanism. Mol Cell Biol 22: 453–468.

  7. Classon M, Harlow E . (2002). The retinoblastoma tumour suppressor in development and cancer. Nat Rev Cancer 2: 910–917.

  8. Durfee T, Mancini MA, Jones D, Elledge SJ, Lee WH . (1994). The amino-terminal region of the retinoblastoma gene product binds a novel nuclear matrix protein that co-localizes to centers for RNA processing. J Cell Biol 127: 609–622.

  9. Geyer RK, Yu ZK, Maki CG . (2000). The MDM2 RING-finger domain is required to promote p53 nuclear export. Nat Cell Biol 2: 569–573.

  10. Goodrich DW . (2006). The retinoblastoma tumor-suppressor gene, the exception that proves the rule. Oncogene 25: 5233–5243.

  11. Haluska FG, Hodi FS . (1998). Molecular genetics of familial cutaneous melanoma. J Clin Oncol 16: 670–682.

  12. Hanahan D, Weinberg RA . (2000). The hallmarks of cancer. Cell 100: 57–70.

  13. Henderson BR . (2000). Nuclear-cytoplasmic shuttling of APC regulates beta-catenin subcellular localization and turnover. Nat Cell Biol 2: 653–660.

  14. Inman GJ, Nicolas FJ, Hill CS . (2002). Nucleocytoplasmic shuttling of Smads 2, 3, and 4 permits sensing of TGF-beta receptor activity. Mol Cell 10: 283–294.

  15. Jiao W, Datta J, Lin HM, Dundr M, Rane SG . (2006). Nucleocytoplasmic shuttling of the retinoblastoma tumor suppressor protein via Cdk phosphorylation-dependent nuclear export. J Biol Chem 281: 38098–38108.

  16. Kaighn ME, Narayan KS, Ohnuki Y, Lechner JF, Jones LW . (1979). Establishment and characterization of a human prostatic carcinoma cell line (PC-3). Invest Urol 17: 16–23.

  17. Kastan MB, Zambetti GP . (2003). Parc-ing p53 in the cytoplasm. Cell 112: 1–2.

  18. Kau TR, Way JC, Silver PA . (2004). Nuclear transport and cancer: from mechanism to intervention. Nat Rev Cancer 4: 106–117.

  19. Khidr L, Chen PL . (2006). RB, the conductor that orchestrates life, death and differentiation. Oncogene 25: 5210–5219.

  20. Knudsen ES, Wang JY . (1997). Dual mechanisms for the inhibition of E2F binding to RB by cyclin-dependent kinase-mediated RB phosphorylation. Mol Cell Biol 17: 5771–5783.

  21. Liang J, Zubovitz J, Petrocelli T, Kotchetkov R, Connor MK, Han K et al. (2002). PKB/Akt phosphorylates p27, impairs nuclear import of p27 and opposes p27-mediated G1 arrest. Nat Med 8: 1153–1160.

  22. Liggett Jr WH, Sidransky D . (1998). Role of the p16 tumor suppressor gene in cancer. J Clin Oncol 16: 1197–1206.

  23. Lu S, Tsai SY, Tsai MJ . (1997). Regulation of androgen-dependent prostatic cancer cell growth: androgen regulation of CDK2, CDK4, and CKI p16 genes. Cancer Res 57: 4511–4516.

  24. Malumbres M, Barbacid M . (2001). To cycle or not to cycle: a critical decision in cancer. Nat Rev Cancer 1: 222–231.

  25. Mittnacht S, Lees JA, Desai D, Harlow E, Morgan DO, Weinberg RA . (1994). Distinct sub-populations of the retinoblastoma protein show a distinct pattern of phosphorylation. EMBO J 13: 118–127.

  26. Mittnacht S, Weinberg RA . (1991). G1/S phosphorylation of the retinoblastoma protein is associated with an altered affinity for the nuclear compartment. Cell 65: 381–393.

  27. Muller H, Bracken AP, Vernell R, Moroni MC, Christians F, Grassilli E et al. (2001). E2Fs regulate the expression of genes involved in differentiation, development, proliferation, and apoptosis. Genes Dev 15: 267–285.

  28. Nork TM, Millecchia LL, Poulsen G . (1994). Immunolocalization of the retinoblastoma protein in the human eye and in retinoblastoma. Invest Ophthalmol Vis Sci 35: 2682–2692.

  29. Ortega S, Malumbres M, Barbacid M . (2002). Cyclin D-dependent kinases, INK4 inhibitors and cancer. Biochim Biophys Acta 1602: 73–87.

  30. Paull KD, Shoemaker RH, Hodes L, Monks A, Scudiero DA, Rubinstein L et al. (1989). Display and analysis of patterns of differential activity of drugs against human tumor cell lines: development of mean graph and COMPARE algorithm. J Natl Cancer Inst 81: 1088–1092.

  31. Rane SG, Cosenza SC, Mettus RV, Reddy EP . (2002). Germ line transmission of the Cdk4(R24C) mutation facilitates tumorigenesis and escape from cellular senescence. Mol Cell Biol 22: 644–656.

  32. Rane SG, Dubus P, Mettus RV, Galbreath EJ, Boden G, Reddy EP et al. (1999). Loss of Cdk4 expression causes insulin-deficient diabetes and Cdk4 activation results in beta-islet cell hyperplasia. Nat Genet 22: 44–52.

  33. Rosin-Arbesfeld R, Townsley F, Bienz M . (2000). The APC tumour suppressor has a nuclear export function. Nature 406: 1009–1012.

  34. Schwartz GK, Shah MA . (2005). Targeting the cell cycle: a new approach to cancer therapy. J Clin Oncol 23: 9408–9421.

  35. Sedlacek HH . (2001). Mechanisms of action of flavopiridol. Crit Rev Oncol Hematol 38: 139–170.

  36. Shapiro GI . (2006). Cyclin-dependent kinase pathways as targets for cancer treatment. J Clin Oncol 24: 1770–1783.

  37. Shin I, Yakes FM, Rojo F, Shin NY, Bakin AV, Baselga J et al. (2002). PKB/Akt mediates cell-cycle progression by phosphorylation of p27(Kip1) at threonine 157 and modulation of its cellular localization. Nat Med 8: 1145–1152.

  38. Sotillo R, Dubus P, Martin J, de la Cueva E, Ortega S, Malumbres M et al. (2001a). Wide spectrum of tumors in knock-in mice carrying a Cdk4 protein insensitive to INK4 inhibitors. EMBO J 20: 6637–6647.

  39. Sotillo R, Garcia JF, Ortega S, Martin J, Dubus P, Barbacid M et al. (2001b). Invasive melanoma in Cdk4-targeted mice. Proc Natl Acad Sci USA 98: 13312–13317.

  40. Stade K, Ford CS, Guthrie C, Weis K . (1997). Exportin 1 (Crm1p) is an essential nuclear export factor. Cell 90: 1041–1050.

  41. Stevens C, La Thangue NB . (2003). E2F and cell cycle control: a double-edged sword. Arch Biochem Biophys 412: 157–169.

  42. Stokke T, Erikstein BK, Smedshammer L, Boye E, Steen HB . (1993). The retinoblastoma gene product is bound in the nucleus in early G1 phase. Exp Cell Res 204: 147–155.

  43. Sui G, Soohoo C, Affar el B, Gay F, Shi Y, Forrester WC . (2002). A DNA vector-based RNAi technology to suppress gene expression in mammalian cells. Proc Natl Acad Sci USA 99: 5515–5520.

  44. Viglietto G, Motti ML, Bruni P, Melillo RM, D'Alessio A, Califano D et al. (2002). Cytoplasmic relocalization and inhibition of the cyclin-dependent kinase inhibitor p27(Kip1) by PKB/Akt-mediated phosphorylation in breast cancer. Nat Med 8: 1136–1144.

  45. Weinstein JN, Kohn KW, Grever MR, Viswanadhan VN, Rubinstein LV, Monks AP et al. (1992). Neural computing in cancer drug development: predicting mechanism of action. Science 258: 447–451.

  46. Weis K . (2003). Regulating access to the genome: nucleocytoplasmic transport throughout the cell cycle. Cell 112: 441–451.

  47. Wolfel T, Hauer M, Schneider J, Serrano M, Wolfel C, Klehmann-Hieb E et al. (1995). A p16INK4a-insensitive CDK4 mutant targeted by cytolytic T lymphocytes in a human melanoma. Science 269: 1281–1284.

  48. Xu L, Massague J . (2004). Nucleocytoplasmic shuttling of signal transducers. Nat Rev Mol Cell Biol 5: 209–219.

  49. Yen A, Coder D, Varvayanis S . (1997). Concentration of RB protein in nucleus vs cytoplasm is stable as phosphorylation of RB changes during the cell cycle and differentiation. Eur J Cell Biol 72: 159–165.

  50. Zacksenhaus E, Bremner R, Phillips RA, Gallie BL . (1993). A bipartite nuclear localization signal in the retinoblastoma gene product and its importance for biological activity. Mol Cell Biol 13: 4588–4599.

  51. Ziegler EC, Ghosh S . (2005). Regulating inducible transcription through controlled localization. Sci STKE 2005: re6.

  52. Zuo L, Weger J, Yang Q, Goldstein AM, Tucker MA, Walker GJ et al. (1996). Germline mutations in the p16INK4a binding domain of CDK4 in familial melanoma. Nat Genet 12: 97–99.

Download references

Acknowledgements

We thank Yang Shi (siCDK2), Krishnendu Roy (Flavopiridol), Susan Holbeck (NCI-60 cell lines); Tatiana Karpova (LRBGE imaging facility) for microscopy training. Jashodeep Datta was supported by the Colgate University-NIH internship. This research was supported by the Intramural Research Program of the NIH.

Author information

Correspondence to S G Rane.

Additional information

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Jiao, W., Lin, H., Datta, J. et al. Aberrant nucleocytoplasmic localization of the retinoblastoma tumor suppressor protein in human cancer correlates with moderate/poor tumor differentiation. Oncogene 27, 3156–3164 (2008). https://doi.org/10.1038/sj.onc.1210970

Download citation

Keywords

  • RB
  • CDK
  • cancer
  • export
  • localization
  • differentiation

Further reading