C. J. Ciudad, Departamento de Bioquímica y Biología Molecular, Division IV, Facultad de Farmacia, Universidad de Barcelona, Avenue Diagonal 643, E-08028 Barcelona, Spain. Fax: + 34 93 402 4520, Tel.: + 34 93 403 4455, E-mail: firstname.lastname@example.org
We examined the effect of suboptimal concentrations of cyclin-dependent kinase inhibitors, which do not interfere with cell proliferation, on retinoblastoma expression in hamster (Chinese hamster ovary K1) and human (K562 and HeLa) cells. To achieve this, we used the chemical inhibitors roscovitine and olomoucine (which inhibit CDK2 preferentially), UCN-01 (which also inhibits CDK4/6) and p21 (as an intrinsic inhibitor). All chemical inhibitors and overexpression of p21 strongly induced retinoblastoma protein expression. UCN-01-mediated retinoblastoma expression was caused by an increase in both the levels of retinoblastoma mRNA and the stability of the protein. The expression of the transcription factor Sp1, a retinoblastoma-interacting protein, was also enhanced by all the cyclin-dependent kinase inhibitors tested. However, Sp1 expression was caused by an increase in the levels of Sp1 mRNA without modification in the stability of the protein. By using luciferase experiments, the transcriptional activation of both retinoblastoma and Sp1 promoters by UCN-01 was confirmed. Bisindolylmaleimide I, at concentrations causing a similar or higher inhibition of protein kinase C than UCN-01, provoked a lower activation of retinoblastoma and Sp1 expression. Finally, the effects of cyclin-dependent kinase inhibitors on dihydrofolate reductase gene expression were evaluated. Treatment with UCN-01 increased cellular dihydrofolate reductase mRNA levels, and dihydrofolate reductase enzymatic activity was enhanced by UCN-01, roscovitine, olomoucine and p21, in transient transfection experiments. These results support a mechanism for the self-regulation of retinoblastoma expression, and point to the need to establish the appropriate dose of cyclin-dependent kinase inhibitors as antiproliferative agents in anticancer treatments.
Cyclin-dependent kinases (CDKs) are key regulators of cell cycle progression. They constitute the catalytic subunits of holoenzymes formed in combination with regulatory subunits named cyclins. Thirteen CDKs [1,2] and at least 25 cyclins  have been reported to date. Cyclin expression varies during the cell cycle and the cyclin/CDK holoenzyme is activated by phosphorylation of specific residues in the CDK catalytic subunit by the cdk-activating kinase [3,4]. CDKs are involved in transcriptional control , mitotic progression , DNA repair (CDK7) , differentiation of brain neurons (CDK5)  and play a crucial role in the progression of cells from G1 to S phase by regulating the phosphorylation state of the retinoblastoma gene product (Rb). The tumor suppressor Rb is a nuclear protein of 928 amino acids  that is present in distinct phosphorylation states depending on the phase of the cell cycle [10,11]: it is nonphosphorylated when newly synthesized; hypophosphorylated in early G1; and hyperphosphorylated in late G1, S, and G2/M phases. In mitosis, a protein phosphatase 1-like protein removes all phosphates from phosphorylated Rb to reset the phosphorylation status of Rb in early G1. The hypophosphorylated form is involved in the growth inhibitory potential of Rb [12,13], which has been related to its capacity to bind and block the activity of the family of transcription factors, E2F [14,15], thus inhibiting the expression of genes that contain the E2F response element in their promoters, e.g. dihydrofolate reductase (DHFR) [16–18], DNA polymerase alpha , thymidine kinase [20,21], histone H2A , proliferating cell nuclear antigen, B-myb [24,25], cyclin D [26,27], cyclin E , cyclin A [29,30] and cdc2 [31,32]. Rb is phosphorylated by the action of various combinations of cyclin/CDK complexes, such as cyclin D/CDK4-CDK6 in early G1 and cyclin E/CDK2 in late G1 and G1/S phases. After mitosis, Rb returns to its nonphosphorylated state [10,11]. Cyclin/CDK complexes can be regulated by small inhibitory proteins, known as intrinsic CDK inhibitors, which suppress cell growth. The INK4 CDK inhibitors (p15, p16, p18 and p19) inhibit CDK4 and CDK6, whereas the family of p21, p27 and p57 inhibit or sequester the different known cyclin/CDK complexes [33–35]. Numerous human cancers present abnormalities in some components of the Rb pathway as a result of CDK hyperactivation, decrease in endogenous CDK inhibitors or Rb gene mutations. Therefore, the use of pharmacologic CDK inhibitors offers great potential for the treatment of many neoplasms . In this regard, chemical CDK inhibitors, such as UCN-01, roscovitine and olomoucine, have been developed. Roscovitine and olomoucine are more specific towards CDK2, whereas UCN-01 shows a similar 50% inhibitory concentration (IC50) for CDK2 and CDK4/6. Flavopiridol (another CDK inhibitor), UCN-01 and roscovitine are currently used in clinical studies on cancer therapy [2,37]. CDK inhibition leads to cell cycle arrest [38–41], apoptotic cell death [39,41,42], differentiation [43–45] and inhibition of angiogenesis . CDK inhibitors also modify the transcript levels of 2–3% of the genes in Saccharomyces cerevisiae, as measured by array methods .
Given that the modulation of CDK activity is an attractive target for cancer chemotherapy, we studied the changes produced by low concentrations of different CDK inhibitors at the molecular level, with a special focus on their natural substrate, retinoblastoma.
In addition to the primary function of Rb as a transcriptional co-repressor in cell cycle regulation, this tumor suppressor protein can function as a transcription co-activator through its physical interaction with selective transcriptional factors such as hBRm, C/EBP, AP2 and Sp1 [48–51]. Rb activates a set of gene promoters, e.g. c-fos, c-myc, transforming growth factor-β1 [52,53], transforming growth factor-β2 , c-jun , Cyclin D1 , thymidine kinase and dihydrofolate reductase (DHFR)  that control the cell cycle through stimulation of Sp1 mediated transcription. Furthermore, the Rb promoter contains potential binding sites for transcription factors such as ATF-2, Sp1 and RBF-1, through which these proteins may regulate Rb expression.
Sp1 and Rb are especially inter-related at the transcriptional level and in their degradation fate. Sp1 is a ubiquitous transcription factor involved in the activation of a large number of genes. Its activity can be modulated during differentiation [56,57], cell growth [58,59], and development . Sp1 and Rb interact physically, forming a complex that enhances the transcriptional activation of Sp1 . Rb has also been described as a transcriptional activator of the p21 gene in epithelial cells through Sp1 and Sp3 transcription factors [61,62]. The transcriptional interaction of Rb, Sp1 and p21 implies an auto-loop of regulation between Rb and CDK activities.
Sp1 can be phosphorylated by a cyclinA/CDK complex that probably includes CDK2, as phosphorylation is inhibited by olomoucine. Dephosphorylated Sp1 shows decreased DNA binding and transcriptional activity [63,64]. Moreover, Rb and Sp1 proteins are degraded by the same proteolytic enzyme, SPase, a nuclear and cytosolic protease with cathepsin B- and L-like proteolytic activity . The levels of SPase vary along the cell cycle, correlating with Rb degradation, suggesting that SPase regulates Rb .
Taken together, these data prompted us to analyze the changes in the expression (transcription, mRNA and protein levels), of both Rb and Sp1 after CDK inhibition by the chemical inhibitors UCN-01, roscovitine and olomoucine, and by the intrinsic inhibitor p21. As DHFR activity is enhanced by the association between Rb and Sp1, we used DHFR as a target model to study the final effects of CDK inhibitors.
We report that the expression of retinoblastoma and Sp1 is increased by low concentrations of CDK inhibitors in Chinese hamster ovary (CHO) K1 and human cells by a mechanism involving transcriptional activation and, also in the case of Rb, by an increase in its stability.
Materials and methods
UCN-01 was kindly provided by H. Nakano (Kyowa Hakko Co., Tokyo, Japan). Roscovitine and olomoucine were purchased from Calbiochem. Bisindolylmaleimide I (BSM-I) was obtained from Sigma-Aldrich. Stock solutions were prepared in dimethylsulfoxide and maintained at −20 °C.
Conditions for the monolayer culture of CHO cells were as described previously . CHO K1 and CHO DG44 cells  were grown in Ham's F12 medium supplemented with 7% (w/v) fetal bovine serum (both from Gibco) and maintained at 37 °C in a humidified 5% (v/v) CO2-containing atmosphere. Human K562 and HeLa cells were grown under the same culture conditions. When determining the activity of DHFR by the deoxyuridine method, cells were incubated in F12 selective DHFR medium (–GHT) lacking glycine, hypoxanthine and thymidine, the final products of DHFR activity.
Flow cytometry analysis
Cell cycle phase distribution upon incubation with CDK inhibitors was monitored by flow cytometry. Nuclei were stained with 25 µg·mL−1 propidium iodide (Sigma-Aldrich) and analyzed on a Becton-Dickinson flow cytometer.
mRNA levels were determined by quantitative RT-PCR using total cell lysates as the starting material for the RT reaction, as described in Noéet al. . K1 cells were plated in 35 mm-diameter dishes and, after several periods of incubation with 5 × 10−8m UCN-01, they were collected in 500 µL of F-12 medium. The cells were then centrifuged and washed with NaCl/Pi, and the final pellet was resuspended in 11.25 µL of diethylpyrocarbonate-treated water. The cells were lysed at 80 °C for 5 min. cDNA was synthesized in 20 µL of reaction mixture containing 125 ng of random hexamers (Roche), 10 mm dithiothreitol, 20 U of RNasin (Promega), 0.5 mm dNTPs (AppliChem), 4 µL of 5× RT buffer, 200 U of M-MLV reverse transcriptase (BRL-Gibco) and the cell lysate. The reaction mixture was incubated at 37 °C for 60 min. Five microlitres of the cDNA mixture was used for PCR amplification.
PCR reactions were typically carried out as follows. A standard 50 µL mixture contained 5 µL of the cDNA mixture, 4 µL of 10× PCR buffer (Mg2+-free), 1.5 mm MgCl2, 0.2 mm dNTPs, 2.5 µCi of [32P]dATP[αP] (3000 Ci·mmol−1; Amersham Ibérica), 1.5 U of Taq polymerase (Ecogen) and 500 ng of each of the four primers.
For the determination of mRNA levels, the primers were: 5′-CGCCAAACTTGGGGGAAGCA-3′ and 5′-GAACCAGGTTTTCCGGCCCA-3′ for DHFR; 5′-GTGCCAATGGCTGGCAGATCA-3′ and 5′-ACCATCCTGCTGCACTTGGGC-3′ for Sp1; 5′-CTCCACACACTCCAGTTAGGA-3′ and 5′-CTGATTTAAGCATGGATTCCA-3′ for Rb; and 5′-CGCAGTTTCCCCGACTTCCC-3′ and 5′-GGCAGCGCACATGGTTCCTC-3′ for adenine phosphoribosyltransferase (APRT), which was used as an internal control.
The reaction mixture was separated into two phases by a solid paraffin wax layer (melting temperature 58–60 °C; Fluka), which prevents complete mixing of PCR reactants until the reaction has reached a temperature that minimizes nonspecific annealing of primers to nontarget DNA. The lower solution contained the cDNA, the MgCl2, the dNTPs, the [32P]dATP[αP] and one half of the buffer, and the upper solution contained the four primers, the Taq enzyme and the remaining buffer.
PCR was performed for 30 cycles, in the case of DHFR, and for 22 cycles, in the case of Sp1 and Rb, after a 1 min denaturation step at 94 °C. Each cycle consisted of denaturation at 92 °C for 30 s; primer annealing at 59 °C for 1 min for DHFR and Sp1, and at 55 °C for 1 min for Rb; and primer extension at 72 °C for 1 min. Five microlitres of each PCR sample was electrophoresed in a 5% (w/v) polyacrylamide gel. The gels were dried and the radioactive bands visualized by autoradiography. Results were quangified by image analysis using the bio-1d, version 99.03, software from Vilbert-Lourmat. The DHFR, Rb or Sp1 mRNA levels were expressed as the ratio of the intensities of the DHFR, Rb or Sp1 signals and APRT signals.
Nuclear extracts from K1 cells were prepared as described in Ciudad et al. . Protein concentrations were determined by the Bio-Rad protein assay based on the Bradford method , using bovine serum albumin as a standard (Sigma), and extracts were frozen in liquid N2 and stored at −80 °C.
Whole extracts were obtained from K1 or K562 cells according to the method of Kraus et al. . Cells were collected in ice-cold F-12 medium and centrifuged at 1000 g for 5 min. The cell pellet was gently resuspended in 5 mL of hypotonic buffer (15 mm NaCl, 60 mm KCl, 0.5 mm EDTA, 1 mm phenylmethanesulfonyl fluoride, 1 mm 2-mercaptoethanol, 15 mm Tris/HCl, pH 8). After centrifugation (1000 g, 5 min), the cell pellet was resuspended in 100 µL of a buffer containing deoxycholate (100 mm NaCl, 10 mm NaH2PO4, pH 7.4, 0.5% sodium deoxycholate, 0.1% SDS, 1% Triton X-100) and centrifuged at 13 000 g for 10 min. The resulting supernatant corresponded to the whole extract. The entire procedure was carried out at 4 °C. Five microlitres of the extract was used for determining the protein concentration by using the Bradford assay (Bio-Rad). The extracts were frozen in liquid N2 and stored at −80 °C.
Western blot analysis
Twenty micrograms of nuclear extract from CHO K1 or human K562 cells was resolved on SDS/5% polyacrylamide gels  and transferred to poly(vinylidene difluoride) membranes (Immobilon P; Millipore) using a semidry electroblotter. The membranes were probed with anti-Rb [C-15, against amino acids 914–928; IF-8, against amino acids 300–380 (both from Santa Cruz Biotechnology); and G3-425, against amino acids 332–344 (BD-Pharmingen)] or anti-(Sp1 PEP 2) (Santa Cruz Biotechnology). Detection of p21 was by using anti-SX118 (BD-Pharmingen). Signals were detected by secondary horseradish peroxidase-conjugated antibody and enhanced chemiluminescence, as recommended by the manufacturer (Amersham). Blots were reprobed with C-21 antibody against Oct-1, or with antibody A2066 against actin (Sigma), in the case of experiments with p21, to normalize the results.
Half-life of retinoblastoma and Sp1
The stability of Rb and Sp1 proteins was assessed by calculating their half-life from the concentration of protein remaining at different time-points after addition of cycloheximide to the cell culture. CHO K1 control cells, or those treated with UCN-01, 5 × 10−8m for 48 h, were incubated with 50 µg·mL−1 cycloheximide for various periods of time. Total protein extracts were prepared and analyzed by Western blot using Rb and Sp1 specific antibodies. The results are expressed as the number of cells collected.
Cloning of the retinoblastoma promoter
Human genomic DNA from HeLa cells was used to isolate, by PCR, a clone containing a 630 bp fragment corresponding to the Rb promoter region. The PCR fragment was generated using two Rb-specific primers whose sequences were taken from GenBank accession number L11910. For the forward primer, the specific sequence (shown below in upper case text) was preceded by an arbitrary sequence (shown below in lower case text) that included an NheI restriction site (underlined). The reverse primer followed a similar structure but contained a XhoI restriction site (underlined) in the arbitrary sequence. The numbers indicated after the primer sequences correspond to the distance, in nucleotides (nt), from the translational start site: Rbprm-for, 5′-tcaagtcaggctagcGTTCCGCACCTATCAGCGCTCC-3′ (630 nt); Rbprm-rev, 5′-cagtgctgcctcgagGACGCCTTTCGCGGCGGGAGC-3′ (1 nt). The PCR product was sequenced using Big Dye v2.0 (PE Biosystems). After digestion with NheI and XhoI, the PCR fragment from the Rb promoter was cloned into the NheI/XhoI sites of the reporter luciferase vector pGL3-basic (Promega). The resulting construct was named prmRB-luc.
Transfections and luciferase assay
HeLa or K562 cells were seeded into six-well plates, the day before transfection, at a density of 7.5 × 104 cells/well in HAM F-12 containing 5% fetal bovine serum. Eighteen hours later, the CDK inhibitors were incubated with the cells and, 18 h later, transfections were performed using FuGENE6 (Roche Molecular Biochemicals). For each well, 3 µL of FuGENE6 in 100 µL of serum free HAM F-12 medium was incubated at room temperature for 5 min. The mixture was added to 125 ng of the Rb promoter (prmRB-luc) or 250 ng of Sp1 promoter (prmSp1-luc) constructs. The DNA–lipid mixture was incubated at room temperature for 15 min. The mixture was added to the cells for 30 h.
Luciferase activity was assayed 48 h after treatment with the CDK inhibitors and 30 h after transfection of the constructs. Cells were lysed in 500 µL of freshly diluted 1× reporter lysis buffer (Promega). The lysate was centrifuged at 10 000 g for 2 min to pellet the cell debris and the supernatants were transferred to a fresh tube. A 10 µL aliquot of the cell extract was added to 25 µL of the luciferase assay substrate (Promega) and the luminescence of the samples was read immediately on a TD-20/20 Luminometer, in which light production (relative light units) was measured for 10 s. Luciferase activity was normalized to cellular protein concentration, determined using the Bio-Rad protein assay reagent in accordance with the manufacturer's protocol.
Transfections, cotransfections and the DHFR transient activity assay
CHO K1 cells were cotransfected with increasing amounts (1, 3 and 5 µg) of a eukaryotic expression vector for p21 (pCMV-Cip1), together with 0.4 µg of BPV-Neo, by using the calcium phosphate method . After 24 h of expression, selection with Geneticin® (800 µg·mL−1) was applied. Three weeks later, the surviving colonies were pooled.
Transient expression experiments were carried out in dhfr-deficient cells (CHO-DG44) by transfecting a dhfr minigene in the presence and absence of CDK inhibitors. When using p21, the expression vector corresponding to this protein was co-transfected together with the dhfr minigene. All transient transfections were also performed by the calcium phosphate method. The plasmid providing basal DHFR activity was p410, corresponding to a dhfr minigene driven by its minimal promoter . After 24 h of expression, the medium was replaced with –GHT medium (DHFR selective medium) and the resulting DHFR activity was determined by the incorporation of radioactive deoxyuridine to cellular DNA, as described by Noéet al. .
The chemical CDK inhibitors UCN-01, roscovitine and olomoucine were added to the medium immediately after transfection and maintained during the expression and labeling times of the assay.
Effects of the CDK inhibitor UCN-01 on retinoblastoma and Sp1 proteins
CHO K1 cells were incubated with increasing concentrations (10 to 100 nm) of UCN-01, for 48 h. Nuclear extracts were then prepared and analyzed by Western blot. UCN-01 resulted in an increase of the total amount of Rb protein in a dose-dependent manner, with a maximum (≈ 10-fold) at 5 × 10−8m(Fig. 1A). The results were normalized using the signal obtained upon reprobing the same blots with an antibody against Oct-1.
To help define the phosphorylation states of Rb in these cells, we performed Western blot analysis with nuclear extracts from CHO cells in the different phases of the cell cycle. Three bands were observed: a nonphosphorylated form of Rb, which is the only band present in starved cells; a hypophosphorylated form, with intermediate mobility, which appears when cells are in G1; and a low-mobility hyperphosphorylated form, which is present mainly in S and G2/M phases (Fig. 1C). Only the nonphosphorylated form of Rb was detected upon incubation with high concentrations of UCN-01 (10−6m) (Fig. 1C), revealing that this compound inhibits the CDK–cyclin complexes that phosphorylate Rb in these cells. At this concentration, K1 cells were arrested. However, 50 nm UCN-01, at which the maximum increase of Rb was observed, did not affect cell proliferation (data not shown). As UCN-01 is also able to inhibit protein kinase C (PKC) activity (IC50 = 7 × 10−9m) we investigated the different effects of UCN by using the PKC inhibitor, BSM-I (IC50≈ 10−8m). In CHO cells, 10−6m BSM-I produced an increase in Rb expression that represented 40% of the increase caused by 50 nm UCN-01 (Fig. 1D).
We also aimed to determine the levels of Sp1 protein in UCN-01-treated cells. To achieve this, the blots used for determining Rb protein levels were reprobed with a specific antibody (PEP 2) against Sp1. Three bands of Sp1 were detected in nuclear extracts from CHO K1 cells, as reported previously . The levels of this transcription factor increased, in a dose-dependent manner, when K1 cells were incubated for 48 h with UCN-01 (Fig. 1B), peaking at 5 × 10−8m UCN-01.
Mechanisms explaining the increase in Rb and Sp1 protein caused by CDK inhibitors in CHO cells
The increase in total levels of Rb and Sp1 proteins caused by UCN-01 may be a result of the enhanced transcription and stability of Rb and Sp1.
To assess transcription, we determined Rb and Sp1 mRNA levels after incubation of CHO cells with 50 nm UCN-01 for different periods of time. The CDK inhibitor increased Rb mRNA levels by 3.5-fold at 20 h of incubation (Fig. 2A) and Sp1 mRNA levels by fivefold at 6 h (Fig. 2B). In both cases, the signal obtained for APRT mRNA was used to normalize the results.
To test whether the effect of UCN-01 was also caused by an increase in the stability of Rb and Sp1 proteins, we determined the decay of these proteins after inhibiting protein synthesis by cycloheximide. CHO cells were incubated with 50 nm UCN-01 for 48 h, which yields maximal expression of Rb and Sp1, and control and UCN-01-treated cells were then incubated with 50 µg·mL−1 cycloheximide for different periods of time. Whole protein extracts were prepared and used to determine the levels of Rb and Sp1 by Western blot, as described in the Materials and methods. The half-life of the Rb protein increased from 8.7 h to 14.2 h upon treatment with UCN-01 (Fig. 3A,B), which corresponds to a 62% increase in the stability of the protein, whereas the difference in Sp1 stability between control and UCN-01 treated cells was not significant (data not shown).
Therefore, the effect of UCN-01 on the levels of Rb protein may be caused by an increase in the synthesis of Rb and by a decrease in its degradation. However, the effect on Sp1 could be accounted for by the increase observed in Sp1 mRNA.
Effects of roscovitine, olomoucine and p21 in CHO cells
To examine whether the effects caused by UCN-01 were shared by other CDK inhibitors, we extended the analysis to roscovitine and olomoucine. These inhibitors belong to the C2,N6,N9-substituted adenine family and mainly inhibit CDK2 activity, as their IC50 values for CDK4/6 are ≈ 100 and 1000 times higher than for CDK2, respectively. Cells were incubated with increasing concentrations of these two chemical inhibitors, and the levels of Rb and Sp1 protein were determined. The range of concentrations of olomoucine used was higher than for roscovitine, given its higher IC50 for CDK2. Both inhibitors resulted in an increased total amount of the proteins Rb and Sp1 (Fig. 4A,B).
We also tested the effect of overexpression of the intrinsic CDK inhibitor p21 in pooled permanent transfectants. CHO K1 cells were co-transfected with increasing amounts of an expression vector for p21 together with BPV-Neo and, upon selection with Geneticin, the pools were used to determine the protein levels of Rb and Sp1. The levels of both proteins increased, even in the transfectants obtained with 1 µg of p21 (Fig. 4C). The overexpression of p21 was confirmed in these transfectants. (Fig. 4D).
These results extended the original observations for UCN-01 and confirmed that the inhibition of CDK activity increases the expression of Rb and Sp1.
Effects of CDK and PKC inhibition on Rb and Sp1 expression in human cells
Next, we determined whether the increased expression of Rb and Sp1 protein upon incubation with UCN-01, roscovitine and olomoucine were also produced in human cells. It was observed that low concentrations of UCN-01 (50 nm), roscovitine (100 nm) or olomoucine (500 nm) caused an increase in the expression of Rb protein, both in nuclear and total extracts from K562 cells (Fig. 5A,C,E). Sp1 expression was also increased by low concentrations of these three inhibitors in K562 cells (Fig. 5B,D). The changes in transcriptional activity caused by UCN-01, roscovitine and olomoucine in K562 cells were also determined by using luciferase assays. The three CDK inhibitors caused an increase in transcription upon transfection of reporter constructs Rb-luc and Sp1-luc in K562 (Fig. 6A,B,C,D). Transient transfection with a p21 expression vector also caused an activation ofRb- and Sp1 promoters (data not shown). In addition, UCN01 also increased Rb and Sp1 transcription in HeLa cells (Fig. 6E,F). Cell cycle distribution was determined in K562 to study whether the effect of the three compounds on Rb expression was related to CDK inhibition. UCN-01, at 50 nm, caused a change in the distribution of the cell cycle towards the G1 phase, at the 24 h time-point, without affecting cell proliferation. At higher concentrations, this effect persisted, whereby an increase of cells in G2/M was observed (Fig. 5F). Roscovitine produced a displacement of the distribution of the cell cycle towards the G2/M phase that started at the low concentration of 100 nm. Roscovitine and olomoucine caused similar effects, and the results corresponding to roscovitine are shown in Fig. 5F. The CDK-independent effect of UCN01 was studied upon incubation with BSM-I, a PKC inhibitor. In K562 cells, concentrations of BSM-I that have a similar PKC inhibitory ability as UCN-01, based on their IC50 values, increased Rb and Sp1 expression to a lower level than UCN-01 (Fig. 5A,B). These concentrations of BSM-I increased Rb and Sp1 promoter activity in luciferase assays (Fig. 6A,B) to ≈ 30% of the increase observed with UCN-01.
Effects of UCN-01 on DHFR activity
Taking into account that Sp1 is a powerful activator of DHFR transcription and that Rb stimulates Sp1 transcriptional activity, we aimed to assess whether the increases in Rb and Sp1 proteins caused by CDK inhibitors affected DHFR expression. To achieve this, we determined the levels of DHFR mRNA upon incubation of CHO K1 cells with UCN-01, and DHFR activity in transient transfections after incubation with UCN-01, roscovitine and olomoucine and in co-transfections with p21. After incubation with 50 nm UCN-01 for different periods of time, the levels of DHFR mRNA transiently increased to a maximum at 36 h (Fig. 7A). The effect of UCN-01 on DHFR activity was analyzed in transiently transfected dhfr-deficient cells (CHO DG44) using a dhfr minigene (p410). DHFR activity increased to a maximum at 50 nm UCN-01; at higher concentrations of the inhibitor, the activity decreased (Fig. 7B). Roscovitine and olomoucine also enhanced DHFR activity, to a maximum at 100 nm roscovitine and 500 nm olomoucine, but the activity decreased thereafter (Fig. 7C,D). In addition, the effect of overexpression of p21 on DHFR activity was also analyzed in transient co-transfection experiments. DHFR activity increased, depending on the amount of co-transfected p21 (Fig. 7E).
We aimed to explore the effects of the inhibition of CDKs on the expression of their natural substrate, retinoblastoma. This has special interest given that some CDK inhibitors, like UCN-01 and roscovitine, are undergoing clinical trials for use in anticancer treatment, as high concentrations of CDK inhibitors have antiproliferative effects.
We used the chemical CDK inhibitors UCN-01, roscovitine and olomoucine, as each shows a distinct specificity toward CDKs. p21 was also used as an intrinsic CDK inhibitor. UCN-01 inhibits CDK4/6 and CDK2 with IC50 values of 0.032 µm and 0.030 µm, respectively. However, roscovitine and olomoucine have an IC50 for CDK4/6 that is ≈ 100-fold higher than for CDK2 (0.7 µm for CDK2 and > 100 µm for CDK4/6, for roscovitine, and 7 µm for CDK2 and > 1000 µm for CDK4/6, for olomoucine, respectively) [76,77]. Thus, CDK2 can be selectively inhibited by using low concentrations of these inhibitors, according to their IC50 values.
A first conclusion of this work is that upon incubation with submaximal concentrations of CDK inhibitors, the total amount of Rb protein increases in a dose-dependent manner. In the case of UCN-01, the maximal effect was observed at 50 nm, a concentration that did not interfere with cell proliferation. However, this inhibitor was able to arrest the cells and to prevent Rb phosphorylation (Fig. 1C) when used at high concentrations (10−6m), in agreement with previous observations . The amount of Rb protein was also increased by the CDK intrinsic inhibitor, p21, with the same broad spectrum of action as UCN-01, and by roscovitine and olomoucine, which inhibit CDK2 activity.
UCN-01 was originally described as a PKC inhibitor and shows a low IC50 for this kinase (0.007 µm). Therefore, we explored the possible contribution of the PKC inhibitiory activity of UCN-01 to the increase in Rb expression, to characterize the extent of the CDK-independent effect. Indeed, inhibition of PKC by using BSM-I, a more selective inhibitor of PKC, also triggered the expression of Rb protein and Rb transcription. Therefore, the action of UCN01 is caused by overlapping effects. However, this increase in Rb expression is lower than that caused by UCN-01, when similar inhibition of PKC was achieved. In addition, 100 nm roscovitine and 500 nm olomoucine, which increase Rb expression, do not inhibit PKC (IC50 values of roscovitine and olomoucine for PKC are > 100 µm and > 1000 µm, respectively [76,77]).
Regarding roscovitine and olomoucine, the levels of Rb protein increase at concentrations that inhibit CDK2 but not CDK4/6. Thus, inhibition of CDK2 alone, which prevents hyperphosphorylation of Rb, is sufficient to trigger the increase in expression of Rb.
The effects on Rb and Sp1 have also been demonstrated in human cells both at the level of protein expression and transcriptional activity. The changes in the cell cycle distribution in human K562, produced by the low concentrations of these compounds, show an effect on CDK activity, and it is precisely at these submaximal concentrations where the clearest effects on RB and Sp1 expression are seen.
Given that Rb has various phosphorylation states depending on the phase of the cell cycle, two mechanisms may explain the increase in the total amount of Rb upon CDK inhibition. Rb expression may be enhanced by a negative effect caused by the hyperphosphorylated form, by a positive effect caused by accumulation of the hypophosphorylated form or by a combination of the two mechanisms, in keeping with the self-regulation of the Rb gene by its own gene product, as proposed elsewhere. On the one hand, Hamel et al. , using transfections of Rb in differentiated P19 cells, thus overexpressing nonphosphorylated Rb, and Gill et al. , using transfections of a mutant form of Rb (Δp34) refractory to phosphorylation, demonstrated that nonphosphorylated Rb represses Rb transcription. On the other hand, Park et al.  described that the Rb promoter is positively self-regulated by its own gene product when Rb is overexpressed in exponentially growing cells, in which the hypophosphorylated form is the major species. Sandig et al.  reported that overexpression of the CDK inhibitor, p16, which prevents the phosphorylation of Rb by CDK4/6, down-regulates Rb. The hypophosphorylated form of Rb may thus activate Rb transcription. Our results are in keeping with those observations, as low concentrations of CDK inhibitors trigger a response of increased Rb transcription. Of particular interest are the results obtained upon incubation with specific CDK2 inhibitors, as their action would partially decrease the step of phosphorylation towards hyperphosphorylated RB. Thus, a partial increase of the hypophosphorylated form may increase Rb transcription, and a decrease of the hyperphosphorylated form may serve as a signal to restart Rb transcription.
From the experiments performed in UCN-01 treated cells, we conclude that the increased level of total Rb protein is caused by two mechanisms. On the one hand, there is an increase in transcription, as shown by the luciferase experiments and in the levels of Rb mRNA; and, on the other hand, the stability of Rb protein is increased.
As stated in the Introduction, Rb interacts with a variety of proteins. We selected Sp1 to study the possible effect of CDK inhibitors for the following reasons, namely that (a) the presence of GC-boxes in the Rb promoter may allow regulation through Sp1, (b) Sp1-mediated transcription is stimulated by Rb through a physical complex between the two proteins, (c) Sp1 can be phosphorylated by a cyclinA/CDK complex and (d) Sp1 is degraded by SPase, a protease that is also active on Rb and regulates this protein in the cell cycle.
There are remarkable similarities in the protein expression of Sp1 and Rb in response to low concentrations of CDK inhibitors. First, the expression of Sp1 strongly increases with the chemical inhibitors UCN-01, roscovitine and olomoucine, and with the overexpression of the intrinsic inhibitor p21. Second, the levels of Sp1 mRNA increase upon incubation with UCN-01, reflecting an elevated transcription, as demonstrated in the transient transfection with Sp1-luc in human cells. However, there are also some differences, namely (a) in the case of Rb, there is an increase in both the levels of mRNA and the stability of the protein, while the degradation rate of Sp1 does not vary and thus the increase in the protein levels seems to depend only on the increase in transcription, and (b) the time dependency of the increase in mRNAs is shorter for Sp1 than for Rb.
As CDK2 phosphorylates Sp1 and enhances Sp1-mediated transcription , and high concentrations of olomoucine reduce Sp1 transcriptional activity , the phosphorylated form of Sp1 may be needed for the transcriptional activation of a variety of Sp1-controlled genes. A slight decrease in the phosphorylated form of Sp1, caused by CDK inhibitors, may trigger Sp1 synthesis to re-establish its normal levels. Given that Sp1 mRNA increases earlier than Rb mRNA, and that the Rb promoter contains GC boxes, Sp1 may contribute to the rise in Rb expression.
Finally, we used the dhfr gene as a target model to study the effect of CDK inhibitors, on a late response gene, upon stimulation to proliferate. The dhfr gene was selected because it is activated mainly by Sp1, especially when the latter is associated with Rb . DHFR mRNA was increased by UCN-01 in a time-dependent manner in CHO K1 cells. In addition, DHFR activity was increased by UCN-01, roscovitine, olomoucine and p21 in assays of transient transfections. This increase takes place at concentrations of the inhibitors where the expression of Rb and Sp1 starts to increase. At higher concentrations of these inhibitors, the effect on DHFR is no longer noticeable.
In summary, we describe that CDK inhibitors, when used at submaximal concentrations, enhance expression of the CDK natural substrate, retinoblastoma, and the Rb-interacting protein, the transcription factor Sp1. These results uncover intriguing aspects of Rb regulation, e.g. that Rb may be self-regulated through its phosphorylation status in combination with Sp1, also subjected to changes in phosphorylation. Additionally, these findings emphasize the need for caution when adjusting the dose of CDK inhibitors to be used in anticancer treatments, alone or in combination with other chemotherapeutic agents. Lower than optimal concentrations of CDK inhibitors can increase the activity of potential target proteins, such as DHFR, which is inhibited by treatment with methotrexate. Moreover, overexpression of Rb decreases the susceptibility of cells to therapy with agents that increase apoptosis . Thus, the dose adjustment of potential chemotherapy combinations is required. In this regard, the results presented in this study may prove useful.
This work was supported by grant SAF99-0120, SAF02-0363 (from Comisión Interdepartamental de Ciencia y Tecnología, Spain) and 2001SGR00141 (from Direcció General de Recerca, Catalunya). We thank Mr Robin Rycroft (from the Language Advisory Service) for correcting the English manuscript.