Skip Navigation


Carcinogenesis Advance Access originally published online on January 19, 2008
Carcinogenesis 2008 29(3):560-567; doi:10.1093/carcin/bgn010
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
29/3/560    most recent
bgn010v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Cucciolla, V.
Right arrow Articles by Ragione, F. D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cucciolla, V.
Right arrow Articles by Ragione, F. D.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2008. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Histone deacetylase inhibitors upregulate p57Kip2 level by enhancing its expression through Sp1 transcription factor

Valeria Cucciolla, Adriana Borriello, Maria Criscuolo, Antonio A. Sinisi1, Debora Bencivenga, Annunziata Tramontano, Anna Chiara Scudieri, Adriana Oliva, Vincenzo Zappia and Fulvio Della Ragione*

Department of Biochemistry and Biophysics ‘F. Cedrangolo’, Second University of Naples, Via Costantinopoli 16, 80138, Naples, Italy
1 Dipartimento Medico-Chirurgico di Internistica Clinica e Sperimentale ‘F. Magrassi e A. Lanzara’, Istituto di Endocrinologia, Second University of Naples, 80131, Naples, Italy

* To whom correspondence should be addressed. Tel: +39 081 5665812; Fax: +39 081 5665812; Email: fulvio.dellaragione{at}unina2.it


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 Funding
 References
 
Histone deacetylase inhibitors (HDACIs) represent a new class of targeted anticancer agents. Here, we evaluate the effects of butyrate (BuA) and other HDACIs on p57Kip2, a cyclin-dependent kinase inhibitor (cki). We observed that inhibitors of class I/II histone deacetylases (HDACs), but not of class III HDACs, induce a remarkable accumulation of p57Kip2 in several cells. The cki upregulation is associated with an increased gene expression that was not prevented by cycloheximide, indicating that HDACIs affect directly p57Kip2 transcription. The characterization of p57Kip2 promoter indicates that the first 165 bp are mostly involved in the BuA effects. Chromatin immunoprecipitation studies demonstrated that the BuA treatment causes the recruitment of Sp1 transcription factor. The Sp1 importance was confirmed by the reduction of BuA effects by mithramycin A (an Sp1 antagonist) and, most stringently, by Sp1 downregulation due to Sp1 siRNA. Moreover, both the treatments reduce the p57Kip2 transcription in untreated cells, suggesting that Sp1 is required for the constitutive cki expression. Studies employing plasmids containing parts of the 165 bp of p57Kip2 promoter indicate that the promoter region between –87 and –113 bp, which includes two putative Sp1 consensus sequences, plays a critical role in the response to HDACIs. Since this p57Kip2 promoter region also embraces the consensus sequence for the transcriptional repressor chicken ovalbumin upstream promoter transcription factor-interacting protein 2 (CTIP2), we evaluated whether this factor is involved into the BuA effect. When CTIP2 was downregulated by a specific siRNA, we observed the enhancement of BuA activity on p57Kip2 expression suggesting that CTIP2 might also be involved in HDACIs effects.

Abbreviations: ChIP, chromatin immunoprecipitation; cki, cyclin-dependent kinase inhibitor; CTIP2, chicken ovalbumin upstream promoter transcription factor-interacting protein 2; HDAC, histone deacetylase; HDACI, histone deacetylase inhibitor; PCR, polymerase chain reaction; TSA, trichostatin A


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 Funding
 References
 
Acetylation of histones plays a pivotal role in the regulation of genome transcription, modulating the expression of about 4–8% of human genes (1–3).

Modification of the {epsilon}-amino group of lysine residues changes the configuration of nucleosomes. The positive charge on unacetylated lysines of histones H3 and H4 is attracted by the negatively charged DNA yielding a compact chromatin state that is repressive for transcription. Acetylation of lysines by histone acetylase removes their positive charge resulting in an open chromatin structure, which facilitates gene expression. Histone deacetylases (HDACs) remove the acetyl groups from lysines, thus reversing this process and silencing gene expression. There are two families of HDACs: the so-called ‘classical’ HDAC family and the recently discovered SIR2 family of NAD+-dependent HDACs (class III HDACs). Members of the classical HDAC family fall into two different phylogenetic classes (4). The class I HDACs (HDAC1, 2, 3 and 8) are most closely related to the yeast transcriptional regulator RPD3. Class II HDACs (HDAC4, 5, 6, 7, 9 and 10) share domains with similarity to HDA1, another deacetylase found in yeast (4). HDACs of class I are expressed in most cell types, whereas the tissue distribution of class II HDACs is more restricted, suggesting that these enzymes might be involved in cellular differentiation and developmental processes (4).

The function of class III HDACs (corresponding to the sirtuin family in high eukaryotes) is still unclear, although pieces of evidence correlate the deacetylases to the ageing process (5).

The reprogramming of the histone acetylation/deaceylation process by changes of environment or addition of drugs might result in marked phenotypical changes. This is well illustrated by the antiproliferative and anticancer effects of histone deacetylase inhibitors (HDACIs), which have been demonstrated both in cellular and animal models and in clinical trials (6).

One of the most interesting HDACI is the short-chain fatty acid butyric acid (BuA). The molecule and its derivatives (particularly phenylbutyrate, PBuA) induce important phenotypical changes, which include cell cycle arrest, differentiation and apoptosis. The remarkable in vitro efficacy and the lack of in vivo toxicity of BuA prompted the use of the compound in the treatment of a number of human diseases, including cancers, bowel inflammatory pathologies, β-thalassemia, sickle cell anemia, cystic fibrosis and X-linked adrenoleukodystrophy (7–9). However, the clinical use of BuA is hampered by its in vivo very short half-life.

Currently, many powerful and stable HDACIs are available, including trichostatin A (TSA), trapoxin, suberoylanilide hydroxamic acid, MS-275, apicidin and others. Their use in cancer treatment is under evaluation in phase I and II clinical trials (10 and references therein). Additionally, it has been demonstrated that valproic acid, a drug employed in the therapy of epilepsy and bipolar syndrome, inhibits HDACs, thus suggesting its use in chemotherapy (11).

Here, we report a study on the effect of HDACIs on the transcription of p57Kip2, a tight-binding cyclin-dependent kinase inhibitor (cki), belonging to the Cip/Kip family.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 Funding
 References
 
Materials
BuA, PBuA, valproic acid, nicotinamide and cycloheximide were supplied by Sigma Chemical Company, St Louis, MO. TSA and mithramycin A were furnished by BioMol Research Laboratories, Plymouth Meeting, PA). MS-275 was from Calbiochem EMD Biosciences, La Jolla, CA. Splitomicin was supplied by Alexis Biochemicals, San Diego, CA. FuGENE 6 transfection reagent was from Roche Applied Science, Monza, Italy, and lipofectamine 2000 by Invitrogen Life Technology, Carlsbad, CA. siRNAs directed against Sp1 and chicken ovalbumin upstream promoter transcription factor-interacting protein 2 (CTIP2) were furnished by Santa Cruz Biotechnologies, Santa Cruz, CA.

Cell line and treatment
The following cell lines were employed in our experiments: K562 (from an erythroleukemia); HL-60 (from a promyelocytic leukemia); HT-29 and CaCO2 (from colon cancers); Lan-5, SK-N-SH and SK-N-BE (from neuroblastomas); HeLa (from a cervical cancer) and HEK293 (from human embryonic kidney). The culture conditions of these cell lines were previously reported (12). EPN is a cell line derived from the normal prostate epithelium as described in (13). Cell treatments were described in (14).

Reverse transcription–polymerase chain reaction
Reverse transcription–polymerase chain reaction (PCR) analyses for glyceraldehyde 3-phosphate dehydrogenase, p21Cip1, p27Kip1, p57Kip2, {gamma}-globin and GpIIb expression were performed as in (14). Glyceraldehyde 3-phosphate dehydrogenase amplification was used for normalization purposes.

Antibodies used and immunoblotting
Monoclonal antibodies to poly (ADP-ribose) polymerase were from BioMol Research Laboratories. Monoclonal antibodies against p27Kip1 were from Transduction Laboratories, Lexington, KY. The polyclonal antibodies to Sp1, Sp3 and p57Kip2 were from Santa Cruz Biotechnologies. Polyclonal antibodies to actin were from Sigma Chemical Company. Antiacetyl-histone H4 and antiacetyl-histone H3 were obtained from Upstate (Upstate Biotechnology, Charlottesville, VA). Immunoblotting analyses were performed as described (12,14).

Firefly luciferase reporter gene plasmid transfection and luciferase activity assay
Some of the firefly luciferase reporter gene plasmids employed was kindly given by Dr Olivier Delattre (INSERM U509, Laboratoire de Pathologie Moleculaire des Cancers, Institut Curie, Paris, France) and their preparation and features were described in (15). Particularly, one plasmid contains the full-length (2191 bp) p57Kip2 promoter, whereas the other three plasmids contain –1550 bp, –595 bp or –165 bp of the promoter region relative to the transcription start site. Starting from the plasmid containing 165 bp of p57Kip2 promoter, we constructed the following plasmids: S1 (from +1 to –87), S2 (from +1 to –113), S3 (from –59 to –151) and S4 (from –59 to –165). In particular, in order to prepare plasmids S1, S3 and S4, the relative DNA fragments were amplified by PCR employing as template the plasmid containing 165 bp of p57Kip2 promoter. The primers employed for the PCRs include a KpnI linker at the 5' end of the sense primers and a HindIII linker at the 3' end of the antisense primer. The PCR products were digested with KpnI and HindIII and then subcloned into the KpnI–HindIII site of pGL3-basic. To generate plasmid S2, we amplified the DNA fragment from +1 to –113 of p57Kip2 promoter by using primers that have a HindIII linkers at the 5' end of the sense and antisense primers. The PCR product was digested with HindIII and then subcloned into the HindIII site of pGL3-basic vector. Details on the primers employed are available on request. The sequence of the prepared plasmids was confirmed by direct sequencing.

The plasmids (0.5 µg per 1.0 x 106 cells) were transiently transfected in K562 cells as described previously (14). Briefly, 5 h following the transfection, the medium was changed, and 2–3 h later, the transfected cells were treated as reported in the text. The luciferase activities of the cell lysates were measured and normalized as reported (14).

The treatments with siRNA were performed as follows: 24 h before siRNA addition, HeLa or EPN cells were seeded in six-wells tissue culture plates and cultured in Dulbecco modification of Eagle's medium/F12-HAM (Sigma Chemical Company) supplemented with 5% fetal bovine serum without antibiotics. Transfection was performed by lipofectamine 2000 following manufacturer's instructions and using a 40 nM siRNA final concentration. After 72 h, the cells were treated as reported in the text.

Chromatin immunoprecipitation assays
Chromatin immunoprecipitation (ChIP) assays were performed exactly as reported (14). Briefly, treated and control K562 cells (1.2 x 107 cells) were collected by low-speed centrifugation, re-suspended in phosphate-buffered saline and cross-linked with 1% formaldehyde at room temperature. After 10 min, cells were washed twice with ice-cold phosphate-buffered saline, then placed on ice, lysed and sonicated. Insoluble materials were removed and the supernatants pre-cleared by 30 µl of packed protein A agarose beads (pre-treated with sheared DNA salmon sperm). The supernatants were immunoprecipitated with the specific antiserum plus protein A agarose beads. After extensive washing, the beads were mixed with 100 µl of 1% sodium dodecyl sulfate/10mM Tris-HCl, 1.0mM EDTA and incubated at 65°C for 10 min, centrifuged and the eluates were transferred to fresh tube. The beads were then washed with 150 µl of 0.67% sodium dodecyl sulfate/10mM Tris-HCl, 1.0mM EDTA, centrifuged again and the washes were added to eluates. To reverse the formaldehyde cross-links, the samples were then incubated overnight at 65°C in 0.05 M NaCl (final concentration). After this step, the samples were treated with 50 µg Proteinase K solution and incubated for 2 h at 37°C. Finally, DNA was extracted, precipitated and recovered in 20 µl of sterile H2O. Five microliters of DNA samples were then subjected to amplification by employing the following primers, which amplify the promoter regions of p57Kip2 (from –160 to +87): 5'-CGCCAATCGCCGTGGTGTTG-3' (sense) and 5'-GACTCTTCTGCGTCGGGT-TC-3' (antisense), hot start of 5 min at 95°C, 30–35 cycles composed of steps at 95°C for 1 min, 60°C for 1 min, 72°C for 1 min and a final elongation step at 72°C for 7 min. For the amplification of the negative control region (exon XI of the urokinase-type plasminogen activator gene), the following primers were used: 5'-TTGTATCTTTGGCGTCACAGG-3' (sense) (position 5262 of the gene) and 5' CATTCTCTTCCTTGGTGTGAC-3' (antisense) (position 5444). For the negative control region, PCR was performed as follows: hot start 95°C, 3 min; 95°C, 1 min; 58°C, 1 min; 72°C, 1 min (repeated for 33 cycles) and 72°C, 7 min (final extension) (16).


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 Funding
 References
 
HDACIs inhibit K562 cells growth, induce G1 phase arrest and upregulate p21Cip1 and p57Kip2 levels
As reported in Figure 1A, 1 mM BuA or 2 mM PBuA inhibited the growth of K562 cells. Higher concentration of the molecules (5 mM) resulted in a progressive decrease of cell number associated to clear morphological changes (cell shrinkage and lysis), which represent putative signs of apoptotic or necrotic events. Thus, we choose to employ in our study the HDACIs amounts that exert only an antiproliferative activity and do not induce complex phenotypes.


Figure 1
View larger version (42K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Effect of BuA and PBuA on K562 cells proliferation and poly (ADP-ribose) polymerase proteolysis. (A) K562 cells were incubated without control, (Con) and with BuA or PBuA. Cells were then counted daily. (B) Effect of 1 mM BuA and 2 mM PBuA on cell cycle progression of K562 cells. Flow cytometry analyses were performed, as described in ref. 14, on cells treated for 24, 48 and 72 h with the reported HDACIs. The control sample corresponds to untreated cells at 24 h. (C) K562 cells were incubated without (Con) and with BuA or PBuA for 72 h. Cell extracts were then prepared and analyzed by a specific antiserum to poly (ADP-ribose) polymerase. The cleavage of poly (ADP-ribose) polymerase was demonstrated by the accumulation of the 85 kDa fragment with respect of the intact protein (signal at 115 kDa). (D) The expression of {gamma}-globin and GpIIb genes was evaluated by reverse transcription–PCR after 3 days incubation, as reported in Materials and Methods. In this experiment, K562 cells were treated with 1 mM BuA.

 
Flow cytometry analyses showed minor profile changes after 24 h treatment, whereas a G1 phase accumulation of cells was gradually observed after 48 and 72 h (Figure 1B). Moreover, a low percentage of apoptotic cells occurred at all time intervals (Figure 1B). The absence of poly (ADP-ribose) polymerase cleavage, a marker of programmed cell death, after 72 h of BuA or PBuA treatment, confirmed the lack of apoptosis (Figure 1C).

Since K562 cells differentiate towards two different phenotypes, we analyzed the expression of genes specific of erythroid and megakaryocytic phenotypes, namely {gamma}-globin and GpIIb, respectively (14). BuA and PBuA upregulated the expression of {gamma}-globin gene, whereas GpIIb gene transcription was completely unmodified (Figure 1D).

The accumulation of K562 cells in G1 phase prompted us to investigate the levels of pivotal ckis, namely p21Cip1, p27Kip1 and p57Kip2, involved in the control of G1 -> S transition. When their expression was evaluated by semi-quantitative reverse transcription–PCR after 8 h treatment (Figure 2A), we observed a significant increase of both p21Cip1 and p57Kip2 transcription, whereas a small decrease of p27Kip1 expression was evidenced. The upregulation of p21Cip1 and p57Kip2 expression was not modified by the addition of cycloheximide (Figure 2B), thus indicating that the effect was not mediated by the synthesis of new proteins. Analysis by immunoblotting confirmed, at protein level, the increase of p21Cip1 (data not showed) and p57Kip2 (Figure 2C). Interestingly, after 18–24 h of BuA treatment, p27Kip1 showed a strong increase (data not shown). Preliminary data suggest that such an effect is due to a decreased amount and activity of p27Kip1 ubiquinylation complex (data not shown). Since the effect of BuA on p21Cip1 has been investigated previously by others, we focused our interest on p57Kip2 gene modulation.


Figure 2
View larger version (48K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2. Effect of HDACIs on p57Kip2 expression. (A) K562 cells were incubated with 1 mM BuA and 2 mM PBuA for 8 h. p21Cip1 (p21), p27Kip1 (p27), p57Kip2 (p57) and glyceraldheyde 3-phosphate dehydrogenase (GAPDH) gene expression was determined by reverse transcription–PCR as reported in Materials and Methods. (B) As in the (A) except that p21Cip1 and p57Kip2 gene expression was determined in K562 cells in the presence or absence of 36 µM cycloheximide (Cx). (C) K562 cells were incubated for the reported time intervals with 1 mM BuA and the content of p57Kip2 was determined by immunoblotting. (D) K562 cells were incubated with various inhibitors of class I and II HDACs, i.e. TSA, 50 nM; valproate (VPA, 2 mM) and 1 µM MS-275 (on the left side) and class III HDAC (nicotinamide, NAM, 10 mM; and splitomicin, Splito, 100 µM) (on the right side). The content of p57Kip2 and actin was evaluated by western blotting. (E) SK-N-SH, SK-N-BE and Lan-5 cells (neuroblastoma cell lines) were incubated for 24 h with BuA, Splito and NAM at the concentration reported in (D). The content of p57Kip2 was evaluated by western blotting. (F) HT-29, CaCo2 and EPN cells were incubated for 24 h with 1 mM BuA. The content of p57Kip2 and actin was evaluated by western blotting. (G) CaCo2 cells were incubated for the reported time intervals with 1 mM BuA and the content of p57Kip2 was determined by immunoblotting.

 
The induction of p57Kip2 expression by HDACIs is associated with histone hyperacetylation
We tested the effect of various HDACIs acting either on class I/II HDACs or on class III HDACs. TSA, MS-275 and valproic acid, which inhibit the activity of HDACs of class I and II, significantly upregulated p57Kip2 level in K562 cells (Figure 2D). Conversely, the addition of splitomicin or nicotinamide, two powerful inhibitors of SIR2 family of NAD+-dependent HDACs, did not modify p57Kip2 cellular content (Figure 2D).

In order to assess the general value of the findings obtained in K562 cells, we extend our study to other cell lines. As shown in Figure 2E and F, BuA caused the cki accumulation in HT-29, CaCo2, EPN, Lan-5 and SK-N-SH cells. Similar results were also obtained in HeLa and HEK293 cells (data not shown). Conversely, in one cell line, i.e. SK-N-BE, butyric acid was unable to induce p57Kip2 expression (Figure 2F). In all the investigated cells, nicotinamide and splitomicin were unable to significantly upregulate the cki (Figure 2F and data not shown).

When, we compared the kinetics of BuA-dependent p57Kip2 induction in K562 and CaCo2 cells, we observed a significant difference that may reflect distinct transcriptional (or translational) machineries (Figure 2G versus Figure 2C).

It has been reported that the methylation of CpG islands localized at the p57Kip2 promoter region causes a strong downregulation of its transcription (17). Particularly, HL-60 cells, which has a fully methylated p57Kip2 promoter, do not express the cki (17). Accordingly, we observed in HL-60 cells, both at mRNA (data not shown) and protein level (Figure 3A), the complete absence of p57Kip2. The addition of BuA and other HDACIs caused an early re-expression of p57Kip2 in HL-60 cells (Figure 3A and B). The demethylation of the promoter, obtained by 5-azacytidine treatment, also allowed the re-expression of the cki but at time intervals later than that observed with HDACIs (Figure 3C).Then, we investigated in K562 cells the effect of BuA on the luciferase expression driven by constructs containing different elements of p57Kip2 gene promoter. As shown in Figure 3D, the minimal region of promoter that is strongly activated by BuA included 165 bp before the starting point of transcription. Interestingly, we did not observe significant differences between the expression of the various constructs in untreated cells suggesting that the first 165 bp probably include consensus sequences pivotal for p57Kip2 expression.


Figure 3
View larger version (34K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3. Effect of HDACIs on p57Kip2 expression and gene promoter activity. (A) HL-60 cells were incubated for 8 h with BuA (1 mM). The content of p57Kip2 and actin was determined by immunoblotting. The blot was exposed for different time periods (1 and 5 min) in order to confirm the absence of p57Kip2 in HL-60 cells. (B) As in (A) except that HL-60 cells were incubated for 8 h with various HDACIs. (C) HL-60 cells were incubated for the reported time intervals with 1 mM BuA and 5-AzaCytidine (AzaCyt, 10 µM). The content of p57Kip2 and actin was determined by immunoblotting. (D) Firefly luciferase reporter gene plasmids were transiently transfected into K562 cells, and luciferase activities were analyzed after 24 h treatment with 1 mM BuA. Luciferase activity is reported as shown. On the left, the dimension of the cloned promoter sequence is reported. Data are the mean (bars, standard deviation) of at least four different experiments. White box, control; black box, BuA-treated cells. (E) K562 cells were treated for 24 h with or without 1 mM BuA. Then, ChIP experiments were performed employing different antisera (anti-acetylated H4 antibodies, H4-Ac; anti-acetylated H3 antibodies, H3-Ac). Details of the method are described in Materials and Methods. Input, PCR of DNA from cells (diluted 100-fold).

 
We also evaluated whether the short-chain fatty acid increased histone acetylation at the proximal region of p57Kip2 promoter by means of ChIP approach. Particularly, in the experiment reported in Figure 3E, we evaluated, by means of a specific PCR, the occurrence of the 165 bp of p57Kip2 promoter in the DNA precipitated with antibodies against antiacetyl-histone H4. As shown in Figure 3E, the addition of BuA increased the amount of immunoprecipitated DNA. This pattern was also observable, although with a minor intensity, by employing in the ChIP experiment antiacetyl-histone H3 antibodies.

Sp1 but not Sp3 or Egr1 modulates p57Kip2 promoter activity
A computer-aided analysis (by the AliBaba 2 software) of the proximal region of p57Kip2 promoter showed the presence of several putative consensus regions for transcriptional factors localized in the initial 165 bp (Figure 4A). We investigated the role of some of them that are particularly frequent in this region and are also important in the control of cell growth, i.e. Egr1 and Sp1/Sp3. Recently, we demonstrated that resveratrol, a polyphenol endowed with antiproliferative activity, strongly upregulates Egr1 level (14). Thus, we evaluated in K562 cells the effect of resveratrol on the expression of p57Kip2 gene. Although the polyphenol remarkably increased Egr1 content, the molecule did not affect p57Kip2 cellular level (Figure 4B) and did not increase the luciferase expression driven by the p57Kip2 gene promoter (Figure 4C). These findings suggest that Egr1 is not involved in p57Kip2 gene expression.


Figure 4
View larger version (26K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4. Role of Egr1 and Sp1 in the BuA-dependent p57Kip2 activation. (A) Sequence of the first 165 bp of p57Kip2 gene promoter. The TATA box and the putative Sp1, Egr1 and CTIP2 consensus sequences are underlined. The transcription start point and the positions –59, –87, –113, –151 and –165 are marked by the arrows. (B) K562 cells were incubated for 24 h with resveratrol (Res, 100 µM) and the Egr1, p57Kip2 and actin content were determined by western blot. (C) Firefly luciferase reporter gene plasmids containing (165 bp) were transiently transfected into K562 cells, and the luciferase activity was analyzed after 24 h treatment with 100 µM resveratrol. Data are the mean (bars, standard deviation) of at least four different experiments. (D) K562 cells were treated for 24 h with or without 1 mM BuA. Then, chromatin was cross-linked with nuclear proteins and broken by ultrasounds into small fragments. After immunoprecipitation with Sp1 antibodies, DNA was released from proteins and the occurrence of p57Kip2 promoter region was evaluated by PCR analysis employing primers which amplify a region (from –160 to +87 of p57Kip2 gene) that includes the Sp1 consensus sequences (see also Materials and Methods). (E) Exactly as in (D) except that the cross-linked chromatin was immunoprecipitated either with anti-Sp1 or anti-Sp3 antisera. IgG, samples immunoprecipitated by control antiserum. (F) Exactly as in (D) except that the cross-linked chromatin immunoprecipitated with anti-Sp1 was used for amplifying p57Kip2 promoter region or a negative Sp1-containing region, i.e. exon XI of the urokinase-type plasminogen activator (uPA) gene (14). The values reported on the sides of the gel correspond to the dimension of p57Kip2 and uPA-amplified fragments.

 
Figure 4A shows that the 165 bp region of p57Kip2 promoter embraces nine putative Sp1/Sp3 consensus sequences. Therefore, we evaluated the importance of these transcription factors in the control of the cki expression. ChIP experiments demonstrated that the DNA fragments immunoprecipitated by an anti-Sp1 antiserum contained the proximal region of p57Kip2 promoter (Figure 4D and E). Importantly, the amount of the immunoprecipitated promoter increased after BuA addition (Figure 4D and E). In addition, an anti-Sp3 antiserum did not precipitate significant amount of p57Kip2 promoter (Figure 4E). Anti-Sp1 antibodies did not precipitate the exon XI of urokinase-type plasminogen activator gene, a genomic region that does not contain Sp1 consensus sequences (Figure 4F) (16). This finding demonstrates the absence of non-specific DNA in the genomic fragments precipitated by anti-Sp1 antibodies.

In order to confirm the role of Sp1 in the HDACI activity on p57Kip2 promoter, they were treated with mithramycin A, a molecule that prevents the activation of Sp1 consensus sequences. As shown in Figure 5A, mithramycin A hampered the BuA-dependent activation of the luciferase expression driven by 165 bp of the p57Kip2 gene promoter. Moreover, the addition of mithramycin A prevented p57Kip2 protein accumulation in K562 cells treated with BuA (Figure 5B). Intriguingly, the molecule downregulates the expression of the gene in the untreated cells (Figure 5A and B).


Figure 5
View larger version (31K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 5. Effect of mithramycin A and Sp1 siRNA on the BuA-dependent p57Kip2 activation. (A) Luciferase reporter gene plasmids (165 bp) were transiently transfected into K562 cells, and the enzymatic activity was analyzed after 24 h treatment with or without 1 mM BuA, 10 µM mithramycin A (Myt) or both the molecules. Relative luciferase activity is expressed as reported. Data are the mean (bars, standard deviation) of at least three different experiments. (B) K562 cells were incubated for 8 h with 1 mM BuA, 10 µM mithramycin A (Myt) or both the molecules. The content of p57Kip2 and actin was then determined by immunoblotting. (C) HeLa (top) and EPN (bottom) cells were transfected with siRNA directed against Sp1 mRNA. After 72 h, the amount of Sp1 and p57Kip2 were determined by immunoblotting. (D) K562 cells were treated with or without siRNA directed against Sp1 mRNA. After 72 h, firefly luciferase reporter gene plasmids (containing 165 bp of p57Kip2 promoter) were transiently transfected and the luciferase activity was analyzed after 24 h treatment with or without 1 mM BuA. Luciferase activity is reported as shown. On the left, the types of treatment. Data are the mean (bars, standard deviation) of at least four different experiments. White box, control; black box, BuA-treated cells.

 
In order to definitely establish the role of Sp1 in the control of p57Kip2 expression, we treated cells with Sp1 siRNA. However, as also reported by others (18), K562 cells appear difficult to transfect with a good efficiency. Thus, we decided to use two different model systems, i.e. HeLa and EPN cells, in the experiment employing interfering RNAs. We observed, in both the cell lines, that Sp1 siRNA induces a remarkable transcription factor decrease (Figure 5C) paralleled by a downregulation of p57Kip2 level and a diminished activation by BuA (Figure 5C). Accordingly, the decrease of Sp1 protein (by siRNA treatment) downregulated about 10- to 15-fold the luciferase expression driven by the 165 bp of p57Kip2 promoter in untreated cells (Figure 5D). The reduction of Sp1 content also strongly hindered the BuA stimulation of the luciferase expression (Figure 5D).

Sp1 sequences putatively involved in the HDACI activity on p57Kip2 promoter
In order to define the region of the promoter mostly involved in the HDACI response, we constructed four deleted mutants starting from the plasmid containing the 165 bp of promoter (Figure 6A). Two mutants contained 89 bp (from +1 to –87, S1 plasmid) or 115 bp (from +1 to –113, S2 plasmid). The two other deleted mutants included p57Kip2 promoter region from –59 to –151 (excluding the region containing the last Sp1 site, S3 plasmid) or from –59 to –165 (S4 plasmid). The starting point of these two mutants (–59) was chosen in order to include the TATA box.


Figure 6
View larger version (27K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 6. Effect of BuA on the luciferase activity of deleted plasmids. (A) Structure of the four (S1, S2, S3 and S4) deleted firefly luciferase reporter gene plasmids constructed on the wild-type plasmid including the initial 165 bp of p57Kip2 promoter. The nine Sp1 consensus sequences are also shown as well as the TATA box. (B) The four deleted plasmids as well as the wild-type plasmid described in (A) were transiently transfected into K562 cells and luciferase activities were analyzed after 24 h treatment with 1 mM BuA. Luciferase activity is reported as shown. On the left, the dimension of the cloned promoter sequence is reported. Data are the mean (bars, standard deviation) of at least four different experiments. White box, control; black box, BuA-treated cells. (C) HeLa (top) and EPN (bottom) cells were transfected with (+) or without (–) siRNA directed against CTIP2 mRNA. After 72 h, the cells were treated for 24 h with or without 1 mM BuA. Finally, the amount of p57Kip2 was determined by immunoblotting.

 
Then, we evaluated the luciferase activation by BuA treatment. As shown in Figure 6B, although all the mutants were similarly expressed, the S1-deleted plasmid was the unique plasmid not activated by BuA. This result suggests that the region localized between the –87 and –113 bp plays an important role in the HDACI regulation of p57Kip2 gene.

Recently, it has been demonstrated that the sequence –102/–107 of the p57Kip2 promoter might represent the consensus for the CTIP2 factor (19). This protein functions as a repressor of transcription and its activity is hindered either by inhibitor of HDAC of class I/II or of class III, depending on the promoter type. In the case of p57Kip2 gene, the activity of CTIP2 is reverted by TSA, an inhibitor of HDAC of class I/II, whereas the effect of sirtuin inhibitors was not investigated (19). Our finding that the region of p57Kip2 promoter required for HDACIs response includes not only two putative Sp1 sequences but also the CTIP2 binding sequence appears of interest.

Therefore, we evaluated the effect of CTIP2 downregulation on the BuA activity. As shown in Figure 6C, the treatment of HeLa and EPN cells with CTIP2 siRNA results in a remarkable increase of BuA response. Moreover, a decrease of p57Kip2 content was observed in HeLa control cells after the treatment with the CTIP2 siRNA.


   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
Funding
 References
 
In this paper, we investigated the effects of HDACIs on the expression of p57Kip2, a powerful inhibitor of cyclin-dependent kinases. HDACIs represent a new class of promising targeted anticancer agents. Several of these compounds are in clinical trials with significant activity against a spectrum of both hematological and solid tumors at doses that are well tolerated by the patients (10 and references therein).

The characterization of HDACI mechanism of action represents a main goal of the studies on these molecules, particularly in view of identifying novel targets for therapy. The most widely demonstrated molecular effect is the regulation of gene transcription. Therefore, the analysis of expression profiles is thought as the key methodology to identify the HDACIs cellular effectors. The modulation of gene expression appears due to the relaxation of nucleosome structure and/or the acetylation of non-histone proteins. Indeed, the function of a number of transcription factors (including p53, GATA-1, FOXO proteins, estrogen and androgen receptors and others) is controlled by their acetylation status (20–22).

p57Kip2 is a cki strongly involved in differentiation. Differently from p21Cip1 and p27Kip1 gene ablated mice, which show scarce phenotypic defects, p57Kip2 null mice die after birth and display severe developmental abnormalities with varying degree of penetrance (23). p57Kip2 is also required for myogenesis, cardiomyogenesis, osteoblast differentiation, lens morphogenesis, retina maturation and podocyte growth maturation (24–26).

In addition, a clear correlation exists between the decrease of p57Kip2 content and human tumor progression. Particularly, low levels of the cki have been demonstrated in gastric cancers, esophageal squamous cell carcinomas, human bladder cancers, head and neck carcinomas, hepatocellular carcinomas, pancreatic adenocarcinomas, duct carcinomas and intrahepatic cholangiocellular carcinomas, leukemias and thyroid tumors (27,28).

Although these observations point to p57Kip2 as a protein vital in the process of differentiation and cancerogenesis, few information is available on the mechanisms controlling the cki gene expression. p57Kip2 promoter silencing by hypermethylation has been demonstrated in cancer cells (17) but the importance of DNA methylation in the physiological control of p57Kip2 gene expression is still unclear. Moreover, the fusion protein EWS-FLI-1 has been reported to downregulate p57Kip2 expression in Ewing tumor (15). Finally, CTIP2, a transcriptional repressor, has been recently reported to hinder p57Kip2 gene transcription (19).

So far, only scarce and descriptive data on the effects of HDACIs on p57Kip2 expression are available in literature (29), whereas the precise mechanism of their action has not been investigated. Here, we reported that all the tested inhibitors of class I and II HDACs strongly and rapidly induce the upregulation of p57Kip2. Conversely, we demonstrated that the addition of inhibitors of SIR2 family of NAD+-dependent HDACs, i.e. splitomicin and nicotinamide, does not affect the cki expression. These findings have been observed in numerous cell lines of different origin, thus representing a general response to HDACIs addition.

A detailed analysis of p57Kip2 promoter demonstrates, for the first time, that the minimal region responsive to HDACIs includes the initial 165 bp and that the treatment with the molecules causes an increased acetylation of the histones H4 and H3 interacting with this region. The 165 bp region of the promoter embraces putative consensus sequences for a number of transcription factors including several presumed binding sites for Egr1 and Sp1/Sp3. The role of Egr1 has been ruled out since a substantial upregulation of this transcription factor does not modify p57Kip2 level. On the other hand, ChIP experiments indicate that HDACIs addition cause a remarkable recruitment of Sp1 (but not Sp3) on the 165 bp region of the promoter, thus suggesting a role of this transcription factor in the response to the inhibitors of HDACs.

The finding was strongly confirmed by the observation that the forced decrease of Sp1 levels, obtained by Sp1 siRNA, results in a significant loss of HDACIs activity (Figure 5C). Interestingly, in the untreated cells, the addition of the mithramycin A and Sp1 siRNA also downregulates the level of p57Kip2 expression arguing that Sp1 is also vital for the constitutive cki gene transcription.

The role of Sp1 has also been demonstrated in the control of the expression of other two ckis, p21Cip1 (30) and CDKN2D (31), thus indicating that the transcription factor might act as a critical regulator of cell cycle and phenotype by influencing contemporaneously several cki genes.

Experiments with different deleted constructs of the 165 bp promoter demonstrated that the region between –87 and –113 is necessary for the activating effect of HDACIs. Intriguingly, this region includes not only two putative Sp1 consensus sequences but also the putative binding site of CTIP2 on p57Kip2 promoter (19). CTIP2 is a novel C2H2 zinc finger protein that was first isolated and identified as a chicken ovalbumin upstream promoter transcription factor-interacting factor (32,33). The protein acts as a transcriptional repressor that is recruited to the template either by the interaction with other transcriptional factors, including chicken ovalbumin upstream promoter transcription factors and Sp1 (34,35) or by direct, sequence-specific DNA binding activity (36). In both cases, CTIP2 mediates transcriptional repression. In initial investigations, it was found that only inhibitors of class III HDAC (i.e. NAM) hinder the activity of CTIP2, whereas the protein function was largely insensitive to reversal by TSA (36). However, very recently, it has been reported that CTIP2 downregulates p57Kip2 expression being such an activity prevented by TSA (19). In this study, the putative binding region of the transcriptional repressor was localized from –102 to –107 position of the cki promoter (19).

On the basis of the data reported above and of our findings, an interesting picture emerges which can be summarized as follows: (i) the region between –87 and –113 of p57Kip2 promoter is required for HDACI response; (ii) this region includes Sp1 and CTIP2 binding sites and (iii) downregulation of Sp1 hinders BuA stimulation, whereas CTIP2 downregulation enhances BuA response.

Thus, we might propose a mechanism that could explain the effect of HDACIs on p57Kip2 expression and that might furnish putative insights on the transcription of the cki gene in basal condition. First, it appears clear that HDACIs increased the acetylation of histones (H3 and H4) localized in the 165 bases proximal to the starting site of p57Kip2 transcription, thus allowing a local nucleosome relaxation and the recruitment of Sp1 proteins. This event, in turn, increases the cki expression. Sp1 protein is also required for p57Kip2 basal expression since its removal decreases the level of the cki in untreated cells.

Second, our data are consistent with the hypothesis that the region between –87 and –113 plays a major role in the activity of HDACIs. In this area, it has been reported to be localized the transcription repressor CTIP2 associated with the NurD complex that includes HDAC2 protein (19). The inhibition of the HDAC activity by HDACI abolishes CTIP2 transcriptional inhibition (described in 19) and enhances the recruitment of Sp1 (this study). These two events suggest that the inhibition of CTPI2/HDAC2/NurD2 complex might result in a change of the interplay between CTIP2 and its DNA-interacting region, thus allowing the binding of Sp1 to the relative consensus sequences. Alternatively, since it has been reported that CTIP2 might hinder Sp1 activity by a direct binding (34,35), the HDACI-dependent inhibition of CTIP2 activity might enhance Sp1 interaction with the p57Kip2 promoter and, in turn, increase the cki transcription. Future studies appear necessary to shed light on these putative intriguing mechanisms.

In conclusion, the present investigation demonstrates that (i) solely inhibitors of class I/II HDACs (and not of class III) increase p57Kip1 gene transcription, (ii) the minimal promoter region modulated by HDACIs includes the initial 165 bp, (iii) HDACIs treatment increases the acetylation of H3 and H4 interacting with this promoter sequence, (iv) Sp1 (but not Sp3) factor plays an important role in the cki expression and in the response to HDACIs. Moreover, we demonstrate that the region between –87 and –113 that includes 2 Sp1 binding sites and the CTIP2 interacting region probably represents a pivotal site of HDACIs regulation of p57Kip2 gene.

Due to the importance of HDACIs in new therapeutic approaches, our data contribute to unravel their mechanism of action identifying p57Kip2 as a major molecular target.


   Funding
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Funding
References
 
Associazione Italiana per la Ricerca sul Cancro (AIRC); Progetti di Rilevante Interesse Nazionale (PRIN); Fondo degli Investimenti della Ricerca di Base (FIRB).


   Acknowledgments
 
We sincerely thank Dr Olivier Delattre (INSERM U509, Laboratoire de Pathologie Moleculaire des Cancers, Institut Curie, 26 rue d'Ulm, 75248 Paris Cedex 05, France) for the luciferase reporter plasmids. We also thank Dr Antonio Menna and Dr Luigi Buononato for their skilful experimental assistance.

Conflict of Interest Statement: None declared.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Funding
 References
 

  1. Mariadason J, et al. Genetic reprogramming in pathways of colonic cell maturation induced by short chain fatty acids: comparison with trichostatin A, sulindac, and curcumin and implications for chemoprevention of colon cancer. Cancer Res. (2000) 60:4561–4572.[Abstract/Free Full Text]
  2. Della Ragione F, et al. Genes modulated by histone acetylation as new effectors of butyrate activity. FEBS Lett. (2001) 499:199–204.[CrossRef][Web of Science][Medline]
  3. Mitsiades CS, et al. Transcriptional signature of histone deacetylase inhibition in multiple myeloma: biological and clinical implications. Proc. Natl Acad. USA (2004) 101:540–545.[Abstract/Free Full Text]
  4. De Ruijter AJM, et al. Histone deacetylases (HDACs): characterization of the classical HDAC family. Biochem. J. (2003) 370:737–749.[CrossRef][Web of Science][Medline]
  5. Tissenbaum HA, et al. Increased dosage of a sir-2 gene extends lifespan in Caenorhabditis elegans. Nature (2001) 410:227–230.[CrossRef][Medline]
  6. Marks P, et al. Histone deacetylases and cancer: causes and therapies. Nat. Rev. Cancer (2001) 1:194–202.[CrossRef][Medline]
  7. Perrin P, et al. An interleukin 2/sodium butyrate combination as immunotherapy for rat colon cancer peritoneal carcinomatosis. Gastroenterology (1994) 107:1697–1708.[Web of Science][Medline]
  8. Harig JM, et al. Treatment of diversion colitis with short-chain-fatty acid irrigation. N. Engl. J. Med. (1989) 320:23–28.[Abstract]
  9. Perrine SP, et al. A short-term trial of butyrate to stimulate fetal-globin-gene expression in the beta-globin disorders. N. Engl. J. Med. (1993) 328:81–86.[Abstract/Free Full Text]
  10. Dokmanovic M, et al. Prospects: histone deacetylase inhibitors. J. Cell. Biochem. (2005) 96:293–304.[CrossRef][Web of Science][Medline]
  11. Gurvich N, et al. Histone deacetylase is a target of valproic acid-mediated cellular differentiation. Cancer Res. (2004) 64:1079–1086.[Abstract/Free Full Text]
  12. Della Ragione F, et al. 5'-Deoxy-5'-methylthioadenosine phosphorylase and p16INK4 deficiency in multiple tumor cell lines. Oncogene (1995) 10:827–833.[Web of Science][Medline]
  13. Sinisi AA, et al. EPN: a novel epithelial cell line derived from human prostate tissue. In Vitro Cell. Dev. Biol. Anim. (2002) 38:165–172.[CrossRef][Web of Science][Medline]
  14. Della Ragione F, et al. p21Cip1 gene expression is modulated by Egr1: a novel regulatory mechanism involved in the resveratrol antiproliferative effect. J. Biol. Chem. (2003) 278:23360–23368.[Abstract/Free Full Text]
  15. Dauphinot L, et al. Analysis of the expression of cell cycle regulators in Ewing cell lines: eWS-FLI-1 modulates p57KIP2 and c-Myc expression. Oncogene (2001) 31:3258–3265.
  16. Ibanez-Tallon I, et al. Binding of Sp1 to the proximal promoter links constitutive expression of the human uPA gene and invasive potential of PC3 cells. Blood (2002) 100:3325–3332.[Abstract/Free Full Text]
  17. Li Y, et al. Aberrant DNA methylation of p57(KIP2) gene in the promoter region in lymphoid malignancies of B-cell phenotype. Blood (2002) 100:2572–2577.[Abstract/Free Full Text]
  18. Scandura JM, et al. Transforming growth factor β-induced cell cycle arrest of human hematopoietic cells requires p57KIP2 up-regulation. Proc. Natl Acad. Sci. USA (2004) 101:15231–15236.[Abstract/Free Full Text]
  19. Topark-Ngarm A, et al. CTIP2 associates with the NurD complex on the promoter of p57KIP2, a newly identified CTIP2 target gene. J. Biol. Chem. (2006) 281:32272–32283.[Abstract/Free Full Text]
  20. Barlev NA, et al. Acetylation of p53 activates transcription through recruitment of coactivators/histone acetyltransferases. Mol. Cell (2001) 8:1243–1254.[CrossRef][Web of Science][Medline]
  21. Faiola F, et al. Dual regulation of c-Myc by p300 via acetylation-dependent control of Myc protein turnover and coactivation of Myc-induced transcription. Mol. Cell. Biol. (2005) 25:10220–10234.[Abstract/Free Full Text]
  22. Matsuzaki H, et al. Acetylation of Foxo1 alters its DNA-binding ability and sensitivity to phosphorylation. Proc. Natl Acad. Sci. USA (2005) 102:11278–11283.[Abstract/Free Full Text]
  23. Zhang P, et al. Altered cell differentiation and proliferation in mice lacking p57KIP2 indicates a role in Beckwith–Wiedemann syndrome. Nature (1997) 387:151–158.[CrossRef][Medline]
  24. Zhang P, et al. p21(CIP1) and p57(KIP2) control muscle differentiation at the myogenin step. Genes Dev. (1999) 13:213–224.[Abstract/Free Full Text]
  25. Urano T, et al. p57(Kip2) is degraded through the proteasome in osteoblasts stimulated to proliferation by transforming growth factor beta1. J. Biol. Chem. (1999) 274:12197–12200.[Abstract/Free Full Text]
  26. Lovicu FJ, et al. Spatial and temporal expression of p57(KIP2) during murine lens development. Mech. Dev. (1999) 86:165–169.[CrossRef][Web of Science][Medline]
  27. Shin JY, et al. Mutation and expression of the p27KIP1 and p57KIP2 genes in human gastric cancer. Exp. Mol. Med. (2000) 32:79–83.[Web of Science][Medline]
  28. Matsumoto M, et al. Immunohistochemical characterization of p57KIP2 expression in human esophageal squamous cell carcinoma. Anticancer Res. (2000) 20:1947–1952.[Web of Science][Medline]
  29. Shin J-Y, et al. Mechanism for inactivation of the KIP family cyclin-dependent kinase inhibitor genes in gastric cancer cells. Cancer Res. (2000) 60:262–265.[Abstract/Free Full Text]
  30. Nakano K, et al. Butyrate activates the WAF1/Cip1 gene promoter through Sp1 sites in a p53-negative human colon cancer cell line. J. Biol. Chem. (1997) 272:22199–22206.[Abstract/Free Full Text]
  31. Yokota T, et al. Histone deacetylase inhibitors activate INK4d gene through Sp1 site in its promoter. Oncogene (2004) 23:1–10.[CrossRef][Web of Science][Medline]
  32. Avram D, et al. Isolation of a novel family of C2H2 zinc finger proteins implicated in transcriptional repression mediated by COUP-TF orphan nuclear receptors. J. Biol. Chem. (2000) 275:10315–10322.[Abstract/Free Full Text]
  33. Leid M, et al. CTIP1 and CTIP2 are differentially expressed during mouse embryogenesis. Gene Expr. Patterns (2004) 4:733–739.[CrossRef][Medline]
  34. Marban C, et al. COUP-TF interacting protein 2 represses the initial phase of HIV-1 gene transcription in human microglial cells. Nucleic Acids Res. (2005) 33:2318–2331.[Abstract/Free Full Text]
  35. Marban C, et al. Recruitment of chromatin-modifying enzymes by CTIP2 promotes HIV-1 transcriptional silencing. EMBO J. (2007) 26:412–423.[CrossRef][Web of Science][Medline]
  36. Avram D, et al. COUP-TF (chicken ovalbumin upstream promoter transcription factor)-interacting protein 1 (CTIP1) is a sequence-specific DNA binding protein. Biochem. J (2002) 368:555–563.[CrossRef][Web of Science][Medline]
Received July 31, 2007; revised November 26, 2007; accepted January 4, 2008.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
29/3/560    most recent
bgn010v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Cucciolla, V.
Right arrow Articles by Ragione, F. D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cucciolla, V.
Right arrow Articles by Ragione, F. D.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?