Anticancer peptide CIGB-300 binds to nucleophosmin/B23, impairs its CK2-mediated phosphorylation, and leads to apoptosis through its nucleolar disassembly activity

CK2-mediated phosphorylation has been regarded as a druggable target to develop anticancer drugs. The CK2 signal is uniformly dysregulated 3-fold to 7-fold in different cancer types (1), and it has also been associated with aggressive tumor behavior in human squamous cell carcinoma of head and neck cancer (2). Different groups around the world have tried to manipulate this biochemical event by targeting the ATP-binding site of CK2 or its gene transcription using antisense oligonucleotides (3, 4). Otherwise, CIGB-300 peptide, formerly know as P15-tat, was developed following the innovative approach to target the phosphoaceptor site on the CK2 substrates rather than the enzyme per se (5).

CIGB-300 is a proapoptotic peptide with established antiproliferative activity in vitro (5), and antitumoral activity in vivo both in syngeneic and xenograft mouse models (5, 6). However, the actual in vivo target(s) on human cancer cells and subsequent events that lead tumor cells to apoptosis thus far remain to be determined.

In this work, we identified a major target for CIGB-300 and provided some clues about the disturbed biological process which ultimately leads tumor cells to die. First, using pull-down experiments and in situ colocalization, we uncovered NPM1/B23 as a major target for CIGB-300 in vivo. Subsequently, the impairment of its CK2-mediated phosphorylation both in vitro and in vivo was shown. Finally, using established markers for nucleolar structure, we showed that the disturbed biological process was ribosomal biogenesis, and that it was further correlated with the rapid and massive onset of apoptosis in tumor cells.

The Mycoplasma-free small cell lung cancer line NCI-H82 was routinely cultured in RPMI 1640 (Life Technologies) supplemented with 10% fetal bovine serum (FBS; PAA, Canada) and 100 μg/mL of gentamicin (Sigma). Metabolic labeling experiments were done in DMEM phosphate-free medium (Invitrogen) supplemented with the indicated percentage of FBS. Cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂.

The peptide chimeras used in this work were synthesized as previously described (5). The peptide chimera F20-2, used as negative control in several experiments, is composed of the same CPP (Tat) linked to the HPV-16 E7 acidic domain in which the two phophorylatable residues have been substituted by alanine (5).

Cells were collected by centrifugation, washed with PBS, and fixed with ice-cold methanol/acetone (4:1) for 1 h. Subsequently, cells were stained with 100 μg/mL of propidium iodide solution in PBS containing 10 μg/mL of DNase-free RNase for 20 min at 37°C in the dark. After gating out cellular aggregates and debris, the cell cycle distribution analysis was done on FACSCalibur flow cytometer using CellQuest software (Becton Dickinson).

Cells were seeded in appropriate vessels at 4 × 10⁵ cells/mL and cultured for 18 to 20 h. The next day, the CIGB-300 peptide conjugated to biotin (CIGB-300-B) was added to the cell cultures at a final concentration of 100 μmol/L and incubated for 30 min. Subsequently, cells were collected by centrifugation, washed twice with cold PBS, and lysed in hypotonic PBS solution (0.1×) containing 1 mmol/L of DTT (Sigma) and Complete protease inhibitor (Roche) by five freeze-thaw (37°C) cycles. Then, cellular lysate was cleared by centrifugation at 12,000 rpm at 4°C for 15 min and 300 μg of total protein, as determined by the Bradford assay (Bio-Rad), were added to 50 μL of pre-equilibrated streptavidin-sepharose matrix (Sigma). After 1 h of incubation at 4°C, the matrix was collected by short spin, extensively washed with PBS 1 mmol/L DTT, and treated for Western blotting analysis or stored to −70°C in double-distilled water for mass spectrometry analysis.

Affinity-purified proteins were separated by 12% SDS-PAGE and subsequently Coomassie blue–stained. Each visible band was subjected to in-gel digestion with porcine trypsin (Promega). The resulting peptide mixtures were extracted and desalted with ZipTips columns. Purified samples were analyzed in a quadrupole time-of-flight mass spectrometer (QTOF2 Micromass) equipped with a nanoelectrospray source. Acquired data was searched against the human proteins in the UniProt Database with the search engine Mascot (Matrix Science, UK). Search parameters were set to a mass tolerance of 0.2 Da for the precursor ions and 0.1 Da for the fragment ions. One missed cleavage site was allowed. Carbamidomethyl cysteine was set as fixed and oxidized methionine was searched as variable modification. All spectra were manually inspected and the identification of a protein or peptide was considered positive using the consensus of several criteria: the peptide score was >20, the assignment of four consecutive y'' ion fragments in the tandem mass spectrum (MS/MS), and the most intense signal must be explained from the sequence.

For Western blotting experiments, 10 μL of 5× Laemli buffer was added to 25 μL of previously incubated streptavidin-sepharose matrix and heated at 95°C for 10 min. After short centrifugation, supernatants were loaded into individual SDS-PAGE wells, electrophoretically resolved, and transferred to a nitrocellulose membrane. The membrane was blocked in PBS 4% skimmed milk for 18 to 20 h at 4°C, and subsequently, 1 μg/mL of either anti-B23 monoclonal antibody (Zymed) or anti-C23 rabbit polyclonal antibodies (Sigma) were added and incubated for 1 h at 37°C. After washing thrice with PBS, the membrane was incubated with anti-mouse IgG (1:1,000) or anti-rabbit polyclonal (1:1,000) secondary antibodies conjugated to horseradish peroxidase (Sigma). Finally, enhanced chemiluminescence (Amersham Life Science) was done according to the instructions of the manufacturer.

Treated or nontreated tumor cells were collected by centrifugation, washed with PBS, and fixed in 3% paraformaldehyde for 30 min at room temperature. After permeabilization with 0.1% Triton X-100, the cells were blocked by incubation with 3% bovine serum albumin (Sigma) for 30 min at 4°C, washed again and incubated for 1 h at 37°C with one the following reagents: FITC-streptavidin (Dako Cytomation), rabbit anti-fibrillarin polyclonal antibody (Abcam), or mouse anti-B23 monoclonal antibody (Zymed). Incubation with FITC-conjugated anti-rabbit IgG or rhodamine-conjugated anti-mouse IgG was carried out at 37°C for 1 h (both from Chemicon Temecula). Using glass coverslips, cells were mounted in 80% glycerol containing 1 mg/mL of paraphenylenediamine in 0.2 mol/L of Tris-HCl buffer (pH 8.5) and analyzed using an Olympus FV300 laser confocal fluorescent microscope (Olympus Fluoview FV300; Tokyo, Japan). FITC was excited using the 488 nm line from an argon-ion laser, and the emission collected with a 510 nm to 530 nm band-pass filter. Rhodamine was excited using the 543 nm line from a helium-neon laser, and the emission collected with a 605 nm long-pass filter. To reduce interchannel cross-talk, a sequential technique was used. Image acquisition was done using a 60× 1.4 numerical aperture oil objective lens. Images were taken at a resolution of 1,024 × 1,024 pixels. Confocal scanning variables were set up so that the cells without the compounds had no fluorescent signal. The cells with FITC only displayed a green signal, and the cells with rhodamine only displayed a red signal. We then used these variables to scan the cells treated with both compounds. Images acquired with a 60× objective were processed using FluoView 3.3 software.

For [³²P]orthophosphate labeling, NCI-H82 cultures were grown in RPMI at 10% FBS for 20 h (nonarrested condition) or 0.2% FBS for 48 h (arrested conditions). Serum-deprived cultures were further grown in RPMI 10% FBS for 2 h to commit cells to re-enter the cell cycle. Subsequently, cells were incubated in phosphate-free medium supplemented with 10% FBS for 1 h. Fresh medium containing 1 mCi/mL of [³²P]orthophosphate (Amersham) was added, and 30 min afterwards, cells were treated with different concentrations of selected peptides or TBB inhibitor for an additional 30 min. Then, cells were washed twice with cold PBS and lysed with radioimmunoprecipitation assay buffer containing 2 mmol/L of each NaF, Na₃VO₄, and B-glycerophosphate. Lysates were clarified at 12,000 × g 4°C for 15 min and supernatants collected for immunoprecipitation. For each reaction, 200 μg of total proteins were mixed with 3 μg of anti-B23 monoclonal antibody. Subsequent steps were done as recommended in the protein G immunoprecipitation kit (Sigma). Finally, immunoprecipitates were analyzed by SDS-PAGE and gels were blue-stained, dried, and further exposed. Image analysis from the X-ray films (FUGI, Japan) and gels was done with ImageJ 1.37v (NIH). The phosphorylation inhibition was calculated relative to the ³²P signal from nontreated cells (100%) at each condition. Immunoprecipitation efficiency and gel loading were normalized by dividing the ³²P signal by their corresponding blue-stained protein signal for each lane.

The experiments were done essentially as previously described (5). Recombinant B23-GST protein was obtained by direct cloning of the B23 cDNA from NCI-H82 cells into BamHI/EcoRI sites of the bacterial expression vector pGEX-6P-2. Expression and semipurification of the recombinant protein was carried out following recommendations from MicroSpin GST Purification module (Amersham Biosciences). The identity of the B23-GST fusion protein was verified by Western blot using the abovementioned anti-B23 monoclonal antibody.

Radioactive DNA-laddering assays were done essentially as described in the DNA Fragmentation Assay web site.⁴

Apoptosis was also quantified using the Annexin V-FITC apoptosis detection kit (BD PharMingen) according to instructions from the manufacturer. Briefly, NCI-H82 treated or untreated cells were resuspended in Annexin V-binding buffer at a final concentration of 1 × 10⁶ cells/mL. Subsequently, 5 μL of both FITC-conjugated Annexin V and propidium iodide reagents were added to cells and incubated for 15 min at room temperature in the dark. Analysis was carried out on a FACscan flow cytometer (Becton Dickinson) using the software WinMDI v. 2.8. Unstained cells were classified as live (Q3), cells stained for Annexin V only were early apoptotic (Q4), cells stained for both Annexin V and propidium iodide were late apoptotic (Q2), and cells stained for propidium iodide only were dead (Q1).

To identify the actual intracellular target(s) for CIGB-300, we carried out in vivo pull-down experiments on NCI-H82 cells. Taking into consideration that previous experiments showed that a maximum peptide uptake was reached 30 minutes postincubation (data not shown), we selected this period of time to perform the in vivo pull-down experiments. Major protein bands from SDS-PAGE–separated pull-down fractions were analyzed by mass spectrometry and 20 interacting proteins were consistently identified (Fig. 1A; Supplementary Table S1).⁵

To further evidence that B23 and C23 are in vivo–interacting targets for CIGB-300, we did Western blot experiments with the pull-down fractions using suitable antibodies against these two proteins (Fig. 1B). A protein band above 30 kDa was clearly visible when the anti-B23 monoclonal antibody was used (Fig. 1B, top) whereas in the anti-C23 immunoblot, a major protein band above 97 kDa was observed in agreement with its reported molecular weight (Fig. 1B, bottom; refs. 7, 8). Moreover, we also analyzed if such interactions are direct peptide protein(s) interactions and/or RNA-mediated, considering that B23 and C23 are usually found associated with rRNA or small nucleolar RNA complexes in the ribosomal biogenesis pathway (7). Treatments of pull-down fractions with excess amounts of RNase showed that CIGB-300–C23 but not CIGB-300–B23 interaction is mediated by RNA, hence, suggesting that B23 but not C23 is the actual in vivo target for CIGB-300 (Fig. 1B, top and bottom).

To corroborate the physical proximity between CIGB-300 and B23 at the subcellular level, we carried out in situ colocalization experiments. As early as 5 minutes after the treatment of NCI-H82 cells, we observed a clear colocalization pattern for CIGB-300 and B23 at the cell nucleoli (Fig. 2). Such colocalization seems to be reinforced at the nucleolar periphery in correspondence with the fact that B23 is considered as an established marker for the granular components (9). Moreover, for CIGB-300, a more diffuse pattern throughout the cell nucleoplasm was also observed at all times tested. Altogether, these results are in accordance with the fast kinetics of peptide internalization and the identified CIGB-300–B23 interaction.

To determine if the identified CIGB-300–B23 interaction leads to an impairment of the CK2-mediated phosphorylation on the substrate, we carried out metabolic labeling experiments in the NCI-H82 cell line.

Considering that B23 is also phosphorylated in vivo by at least three different protein kinases which exert their function at defined cell cycle phases (10), in one experimental setting, the metabolic labeling was done in cell cultures in which 89% of the cells showed typical G₁-S DNA content, and therefore, such kinase activity was minimal (11). Alternatively, we evaluated the CK2-mediated phosphorylation of B23 in nonarrested conditions in which only 60% of cells displayed G₁-S DNA content (Fig. 3A, top). A significant dose-response inhibition of B23 phosphorylation was consistently observed on CIGB-300–treated cells, whereas the use of identical doses of the F20-2 negative control peptide produced a partial inhibition only at the higher dose tested (Fig. 3A, bottom). However, the inhibitory effect over B23 phosphorylation using 100 μmol/L of CIGB-300 was unexpectedly similar (∼40%) in both arrested and nonarrested cell cultures.

To further verify the observed in vivo inhibition on B23 phosphorylation arising from direct CIGB-300–B23 interaction, we used an in vitro assay. A clear dose-response inhibition of recombinant B23-GST protein phosphorylation was observed after the addition of CIGB-300 to the CK2-containing reaction, whereas control peptide showed 3-fold to 6-fold less inhibition at equivalent concentrations (Fig. 3B). These experimental results indicate that the observed effect is attributable to the cargo (P15) rather than to the CPP (tat) in the chimera because the control peptide (F20-2) composed of the same CPP, but with an unrelated cargo, did not show comparable patterns of inhibition.

To evaluate if the impairment of the CK2-mediated phosphorylation of B23 could affect the nucleolar structure, we did a confocal analysis on CIGB-300–treated and untreated cell cultures. Data on Fig. 4 (top) showed that CIGB-300–treated but not F20-2–treated NCI-H82 cells displayed morphologic features typical of nucleolar disassembly as verified using both an anti-fibrillarin polyclonal antibody and anti-B23 monoclonal antibody. Selecting individual but representative cells from both treated and untreated conditions, we attempted to construct a putative sequence of events that might occur at the cell nucleus following the addition of CIGB-300 (Fig. 4, right). In untreated cells, B23 and fibrillarin are mainly located at the cell nucleolus, although a diffuse fluorescence pattern was also observed in the nucleoplasm for both markers (Fig. 4, arrow 1). Following CIGB-300 treatment, an initial reinforcement of B23-derived fluorescence at the cell nucleolus was observed (arrow 2), with a subsequent loss of nucleolar structures as shown by the delocalization of both nucleolar markers (arrows 3 and 4).

Further evidences of the nucleolar breakdown were obtained from the CIGB-300–B23 colocalization experiments done at the same peptide concentrations (Fig. 4, bottom). Using B23 staining as a marker for the granular component of the nucleolus, a red punctuate pattern within the cell nucleoplasm was observed after 1 hour of incubation with CIGB-300, but not with F20-2, probably reflecting dramatic changes in the nucleolar architecture.

To gain insight into the kinetics and the magnitude of apoptotic cell death, and to correlate it with the above presented data, we used two established apoptotic markers and evaluated them in a quantitative fashion. Radioactive DNA fragmentation assays revealed that a fast and dose-dependent DNA laddering was induced on tumor cells following CIGB-300 treatment (Fig. 5A). Up to 68% of DNA fragmentation was achieved in the first hour after treatment with the high peptide dose of 200 μmol/L (Fig. 5A, left), whereas treatment with 100 μmol/L for 2 hours produced 60% DNA laddering (Fig. 5A, right). By contrast, tat peptide alone did not induce any effect at the higher concentration tested.

On the other hand, using the early apoptosis marker Annexin V, we could verify that indeed CIGB-300 induces a fast death in treated cell populations which ranges from 46% at 10 minutes up to 63% at 60 minutes (Fig. 5B, top). Such an effect was in agreement with the observed changes in cell morphology shown by SSC versus FSC plotting (Fig. 5B, bottom). Analysis of treated and untreated NCI-H82 cell populations reveals that CIGB-300 induces an increase in cell granularity, a morphologic change related to apoptosis.

In this article, we uncovered the putative mechanism of action for the novel proapoptotic peptide CIGB-300 (5, 6, 12). This peptide was developed to impair the tumor-exacerbated CK2 pathway by targeting the phosphoaceptor site on the substrate rather than the enzyme per se (3, 4). Indeed, the preclinical (5, 6, 12) and preliminary clinical findings (12) accumulated thus far support the suitability of such idea.

To identify the actual in vivo CK2 substrates targeted by CIGB-300 and give some clues about the mechanistic links that leads tumor cells to apoptosis, we studied the NCI-H82 cell line because of its marked sensitivity to CIGB-300 in vitro (5). Two well-described CK2 substrates with critical nucleolar functions, nucleophosmin/B23 and nucleolin/C23, were identified to bind CIGB-300 in vivo. However, previous reports showed that B23 and C23 remains associated during all phases of the cell cycle (13), and that such association depends of rRNA (7). Those findings support the formal possibility that CIGB-300 could actually interact with one such target and that the other could be pulled-down indirectly by the B23–RNA–C23 association. In line with this, pretreatment of pull-down fractions with RNase showed that B23, but not C23, interacts directly (i.e., protein-protein) with CIGB-300.

To further validate B23 as a major CIGB-300 in vivo target, we conducted metabolic labeling experiments. In both arrested and nonarrested cell cultures, we observed a significant inhibition of B23 phosphorylation following CIGB-300 treatment. Serum starvation with subsequent re-feeding (i.e., arrested cells) was carried out to obtain G₁-S–enriched cell cultures and minimize the potential misleading results from other CK2-unrelated phosphorylation events (10). Interestingly, CIGB-300 abrogated the B23 phosphorylation at a similar extent both in arrested and nonarrested NCI-H82 cells; hence, suggesting that in our experimental model, CK2 is the main kinase that phosphorylated B23 irrespective of the cell cycle distribution. On the other hand, a conventional CK2 inhibitor such as TBB only produced ∼20% inhibition in spite of the relatively high drug concentrations used (14).

Moreover, using an in vitro CK2 phosphorylation assay, we also showed that B23 phosphorylation inhibition was produced by the direct binding of CIGB-300 to this target. Certainly, the effect could only be attributable to the P15 cargo peptide rather than to tat moiety because the F20-2 control peptide did not show comparable inhibitory effects both in vitro and in vivo.

Compelling evidence that CIGB-300, through its inhibitory effect over the CK2-mediated phosphorylation of B23, could lead to a nucleolar disassembly were obtained by confocal microscopy. The relevance of CK2 for nucleolar structure and hence for its master function on the ribosomal biogenesis was first denoted by Louvet et al. (15) using the CK2 inhibitor DRB. Subsequently, using a genetic approach, the same group (16) showed that the single mutation of the CK2 phosphorylatable residue in one particular CK2 substrate, the B23 protein (Ser¹²⁵ residue), leads to nucleolar breakdown. In line with this, we showed that following CIGB-300 administration, the cell nucleoli seems to adopt a structure that resembles in some way the transition from the loose package nucleolus state proposed by Louvet et al. (nucleolar network; ref. 15) to the final complete disconnection of fibrillar and granular components of the cell nucleolus. Both molecular markers, fibrillarin and B23, have been widely used to document such nucleolar breakdowns (15–17). Thus, fibrillarin is considered a marker for the early pre-RNA processing step of ribosome biogenesis that occurs at the dense fibrillar component of the nucleolus, whereas B23 is for late pre-RNA processing at the granular components (17). Changes in the typical distribution of these markers, such as those induced by the CIGB-300 peptide in the NCI-H82 cells show an impaired nucleolar function. On the other hand, considering that the nucleolus is formed and maintained by processes necessary to build ribosomes (18), such changes may reflect an impaired ribosomal production.

Altogether, the fast kinetics of the events triggered by CIGB-300, which comprises peptide internalization, target binding, phosphorylation inhibition, and subsequent nucleolar disassembly, are in agreement with the early onset of apoptosis shown in this tumor cell line both at the molecular and cellular level. According to the inhibition of B23 phosphorylation in vivo, a dose-dependent DNA laddering was evident within the first 2 hours after treatment. Interestingly, as soon as 10 minutes after CIGB-300 treatment, we observed that nearly 50% of cells displayed early signs of apoptosis. Collectively, these results showed that the integrity of the nucleolus, and hence, the ribosome biogenesis may be significantly affected by the CIGB-300–mediated disruption of the CK2-B23 partnership, and that such impairment ultimately leads tumor cells to die by apoptosis.

B23 overexpression has been correlated with increased cell growth and proliferation, inhibition of differentiation, and apoptosis, typical features of neoplastic transformation (reviewed in ref. 19). On the other hand, nucleoli of cancer cells are pleiomorphic and hyperactive (20), underscoring the relevance of ribosomal biogenesis for neoplastic transformation. Considering the critical roles of B23 in such demanding biological processes, it is conceivable that the high levels of B23 in malignant cells might support aberrant cell growth by sustaining the ribosome machinery (20), and hence, suitable compounds targeting B23 could be promissory for cancer therapy.

At least four different functions have thus far been ascribed to B23 in ribosomal biogenesis: rRNA transcription regulation (21), RNase activity (22), molecular chaperone (23), and shuttle protein (24). These findings involve B23 throughout the entire pathway of ribosome biogenesis, from the early events of rRNA transcription to RNA processing, subunit assembly, and nuclear export. However, the CK2-mediated phosphorylation event targeted by the CIGB-300 has only been directly implicated with B23 protein chaperone activity (25). The authors propose that subsequent cycles of phosporylation-dephosphorylation events could regulate the proper folding of structural ribosomal proteins and/or the sequential addition of such proteins to ribosomal subunits in the assembly steps of ribosomal biogenesis. Interestingly, in our in vivo pull-down experiments, we identified a group of structural ribosomal proteins that may represent such proteins or complexes that interact with B23, considering that the CIGB-300 could impair the proposed chaperoning model through its inhibitory effect (25).

At present, how the nucleolar breakdown connects at the molecular level with the onset of apoptosis is still unclear. Although p53 has been recently involved in a ribosomal biogenesis checkpoint (26), our attempts to evaluate it on CIGB-300–treated cells failed because we could not detect such proteins in the NCI-H82 cell line (data not shown). However, considering that the nucleolus has been regarded as a cellular stress sensor (27), it is probable that several released proapoptotic factors or their modulators rapidly shift the molecular balance towards cell death.

Finally, based on the suitability of B23 as a molecular target for cancer therapy, at least two compounds that target different B23 domains have been tested in the preclinical setting (28, 29). However, CIGB-300 has become the first compound that targets the CK2 phosphoaceptor site on B23, and provides compelling evidence that such impairment could be a promising therapeutic approach to treating cancer.

No potential conflicts of interest were disclosed.

We thank Euyeni Diaz for her technical assistance.