Suppression of Survivin Induced by a BCR-ABL/JAK2/STAT3 Pathway Sensitizes Imatinib-Resistant CML Cells to Different Cytotoxic Drugs Free

A reciprocal translocation between chromosomes 9 and 22 generates the abnormal Philadelphia (Ph) chromosome that, in turn, encodes for the chimaeric BCR-ABL oncogene (1). The ensuing BCR-ABL oncoprotein displays constitutive tyrosine kinase activity thereby activating multiple signaling pathways that lead to the development of chronic myeloid leukemia (CML; ref. 2). BCR-ABL–dependent activation of the phosphoinositide 3-kinase and the upregulation of Bcl-xL (3, 4) contribute to block cytochrome C release from the mitochondria and inhibit caspase activity, thus hampering the apoptotic machinery of CML cells and favoring the survival of leukemic myeloid progenitors. Treatment of CML cells with the ABL inhibitor imatinib mesylate (IM) suppresses BCR-ABL kinase activity reverting the antiapoptotic phenotype of the Ph-positive cells and resulting in the induction of cell death (5).

Survivin is a member of the inhibitor of apoptosis proteins (IAP) that is undetectable in normal tissues but highly expressed in most forms of human cancer and functions as both an inhibitor of cell death and a mitotic regulator (6–9). It has been reported that survivin overexpression reduces the sensitivity of neoplastic cells to death stimuli induced by gamma radiation (10) or chemotherapeutic agents (11, 12). The design of shepherdin, a cell-permeable peptidomimetic molecule that reduces survivin half-life by disrupting its binding with the molecular chaperone Hsp90, represents an attempt to increase the killing of cancer cells while preserving the viability of their normal counterpart (13).

We previously reported a significant correlation between BCR-ABL levels and the expression of survivin in a series of 44 patients with CML (14), suggesting that BCR-ABL could directly regulate survivin expression. Indeed, subsequent evidence indicated that BCR-ABL upregulates survivin at the mRNA and protein levels (15). A further study by Carter and colleagues suggested that the RAS/RAF/MAPK cascade was involved in BCR-ABL–dependent induction of survivin in a human CML cell line and that suppression of survivin expression increased responsiveness to IM therapy (16).

The aims of this study were 3-fold. We initially wanted to ascertain the signaling pathway(s) triggered by the BCR-ABL kinase to induce survivin expression. We then wanted to establish whether survivin overexpression contributed to the failure of IM treatment in a representative panel of CML cell lines with well-established mechanisms of IM resistance. Finally, we wished to determine whether reduced survivin levels increased the antiproliferative activity of different cytotoxic compounds.

We found that JAK2/STAT3 signaling modulates BCR-ABL–dependent induction of survivin and that survivin downregulation enhances the apoptotic effect of IM on CML cells that are sensitive to this compound. Lowering survivin expression increased the cytotoxic effect of hydroxyurea (HU) and doxorubicin (Doxo) on multiple cell lines unresponsive to IM. Likewise, exposure of IM-resistant cells to shepherdin reduced survivin expression and induced higher cell killing by both cytotoxic drugs and other ABL tyrosine kinase inhibitors (TKI).

Human CML cell lines K562, KCL22, KYO-1, LAMA84, K562 S, K562 R, KCL22 S, KCL22 R, LAMA84 S, LAMA84 R, and murine Ba/F3 cells expressing either wild-type BCR-ABL (Ba/F3p210) or different BCR-ABL mutants (Ba/F3p210^Y253F and Ba/F3p210^T315I) were grown in RPMI-1640 (Sigma-Aldrich Corp) supplemented with 10% heat-inactivated FBS (Lonza), 2 mmol/L l-glutamine (Sigma), and penicillin/streptomycin (100 μg/mL each, also from Sigma). Human cell lines were obtained from the German Collection of Microorganisms and Cell Cultures DSMZ and used for less than 6 months after receipt with the exception of K562 S and R, KCL22 S and R, LAMA84 S and R that were provided by Carlo Gambacorti-Passerini (University of Milano Bicocca, Italy). The latter cells have not been authenticated in our lab. The K562, KCL22, and LAMA 84-IM–sensitive cell lines were maintained in parallel cultures without IM for the time required to obtain the IM-resistant clone. Ba/F3BCR-ABLp210, Ba/F3p210^Y253F, and Ba/F3p210^T315I were generated in our lab after viral transduction of the Ba/F3 pro-B cell line (purchased from DSMZ) with lentiviral vectors encoding for the specified BCR-ABL constructs.

Primary CML cells were obtained from the peripheral blood of patients with CML after receiving written informed consent. Total leukocytes were separated by red cell lysis.

The following drug treatments were used: 1 μmol/L IM (a gift from Novartis), 12.5 nmol/L nilotinib (a gift from Novartis), 50 μmol/L PD98059 (Calbiochem), 20 μmol/L LY294002 (Sigma), 7.5 nmol/L TG101209 (Symansis), 1 mmol/L HU (Sigma), 1 μmol/L doxorubicin (Sigma), 5 nmol/L dasatinib (a gift from Bristol-Myers Squibb), 17-Allylamino-17-demethoxygeldanamycin (17-AAG, InvivoGen; Fig. 1).

In further experiments, we used 30 μmol/L of cell-permeable shepherdin [biotin-X-KKWKMRRNQFWVKVQRLFACGSSHK-CONH2; a gift from Dario Altieri, Wistar Cancer Institute, Philadelphia, PA] or an equimolar concentration of the corresponding scrambled peptide (biotin-X-KKWKMRRNQFWVKVQRGHSFCALKS-CONH2). The Antennapedia homeodomain sequence is underlined. X = EAHX, hexanoic acid spacer.

Cell pellets were resuspended in isotonic buffer [25 mmol/l Trizma base (pH 8.5), 100 mmol/l NaCl, 7 mmol/l β-mercaptoethanol, 1× protease inhibitor cocktail (Roche)], sonicated and harvested by centrifugation at 14,000 rpm for 10 minutes at 4°C.

For immunoprecipitation experiments, 2 μg of antiphosphotyrosine antibody (clone 4G10, Millipore) were incubated with protein G-Sepharose (Amersham Biosciences) for 2 hours and then with 1 mg of cell lysates for an additional 2 hours. The sepharose beads were then washed in isotonic buffer and resuspended in Laemmli buffer.

For immunoblotting experiments, 40 μg of protein were separated by SDS-PAGE, transferred to nitrocellulose membranes, and blocked with 5% nonfat dry milk or with 10% bovine serum albumin (Sigma) in TBS with 0.2% Tween 20 (Sigma).

Monoclonal antibodies used were: antisurvivin (Abcam); antiphosphotyrosine (Millipore); anti-actin and anti-FLAG (Sigma). Polyclonal sera used were: anti-MAPK, anti-pMAPK, anti-AKT and anti-pAKT (Cell Signaling Technology); anti-JAK2, anti-STAT3 and anti-c-abl (clone K12; Santa Cruz Biotechnology). Appropriate horseradish peroxidase-conjugated secondary antibodies (Amersham Biosciences) were added and proteins were detected using the enhanced chemiluminescence reagent Lite Ablot (Euroclone).

Chemiluminescence images were captured and densitometric quantification of protein bands was carried out using the ImageJ 1.36b software (NIH, Bethesda, MD).

To evaluate cell viability, 2 × 10⁴ cells were incubated, for 24 hours in 96-well plates with different drugs, siRNAs, or peptides and cell proliferation was subsequently determined using the luminescence ATP detection assay system ATPlite 1 step (Perkin-Elmer), following the manufacturer's instructions. To assess cell death, 5 × 10⁵ cells were treated with different drugs or peptides, alone or in combination, for 24 hours. Cells were harvested and triplicate samples were selected to determine apoptosis using the FITC Annexin V Apoptosis Detection Kit I (BD Pharmingen) following the manufacturer's instructions. Apoptotic events were recorded as a combination of Annexin V+/PI− (early apoptosis) and Annexin V+/PI+ (late apoptosis) events and results were expressed as percentage of Annexin V+ cells.

For the production of the lentiviral supernatant, transfer vectors pGIPZ nonsilencing shRNA (Cat n° RHS4346) or pGIPZ shRNA anti-JAK2 (NM_004972, Cat n° RHS4430) were cotransfected with the packaging plasmids in the HEK293T cell line, using the calcium phosphate transfection method (Trans-Lentiviral shRNA Packaging Kit, Open Biosystems). Viral supernatants were harvested 48 hours after transfection. CML cells were then transduced by a double round of spin-infection. Cells were centrifuged at 1,200 × g for 90 minutes at 32°C in the presence of 1 mL of viral supernatant and 8 μg/mL polybrene (Sigma-Aldrich). Transduced cells were finally selected with 6 μg/mL puromycin (Sigma-Aldrich) for 72 hours.

Lentiviruses containing the pLEX empty vector or pLEX STAT3C were produced using the calcium phosphate transfection method (Trans-Lentiviral ORF Packaging Kit, Open Biosystems) as previously described.

For antisense oligonucleotides, 1 × 10⁶ cells were resuspended in 0.8 mL of culture media mixed with 500 nmol/L antisense oligonucleotides against STAT3 (AS STAT3) or with a control oligonucleotide (AS Scr) and then electroporated in a 0.4 cm cuvette using the Gene Pulser Electroporation apparatus (Bio-Rad) with a double pulse protocol (25 μF, 500 V). Both the STAT3 antisense (5′-GCTCCAGCATCTGCTGCTTC-3′), and the control mismatch oligonucleotide (5′-GCTCCAATACCCGTTGCTTC-3′) were synthesized with a phosphorothioate backbone and with a 2′-O-methoxyethyl modification of the first and last 5 nucleotides (Eurofins MWG Operon).

For siRNA electroporation of Ba/F3 cells, cells were resuspended as previously described and mixed with 100 nmol/L of si Surv or with the corresponding scrambled control (si Scr; Dharmacon). si RNA electroporation was carried out as described above using different electrical settings (960 μF, 250 V).

STAT3 was amplified by PCR using the indicated forward (FLAG-tag sequence underlined) 5′-ATAAGAATGCGGGCCGCGCCACCATGGATTACAAGGATGACGACGATAAGATGGCCCAATGGAATCAGCTAC-3′ and reverse 5′-CCGCTCGAGTCACATGGGGGAGGTAGCGC-3′ primers and cloned NotI and XhoI in the expression vector pcDNA3.1 (Life Technologies). STAT3 was then mutagenized using the Quik-Change II XL Site-Direct Mutagenesis Kit (Agilent Technologies). Primers used to generate point mutations were 5′-CATGGGCTATAAGATCATGGATTGTACCTGTATCCTGGTGTCTCCACTGGTC-3′ and 5′-GACCAGTGGAGACACCAGGATACAGGTACAATCCATGATCTTATAGCCCATG-3′. Incorporation of 2 point mutations was verified by direct sequencing. STAT3C was then cloned in the NotI and XhoI unique sites of the pLEX lentiviral vector (Open Biosystems).

To evaluate the effect of different treatments on the colony-forming potential of CML myeloid progenitors, 5 × 10⁴ primary cells were treated for 24 hours with shepherdin alone or in combination with IM, nilotinib, or dasatinib. Drug-treated cells were then seeded in Methocult H4435 methylcellulose medium (StemCell Technologies). Formation of colonies was assessed 14 days after plating.

Statistical analysis was conducted using GraphPad Prism 5.0 (GraphPad Software Inc). Unpaired, single-tail t tests with 95% confidence intervals were used to compare cell viability in different experimental conditions. The one-way ANOVAs according to Bonferroni's posttest were used to compare the effect of the combination of shepherdin with different tyrosine kinase inhibitors.

We have previously reported that, in primary hematopoietic cells isolated from patients diagnosed with CML, the amount of BCR-ABL transcript significantly correlates with the expression levels of survivin (14). Our findings suggested that BCR-ABL might directly regulate survivin expression in CML cells and indeed, subsequent manuscripts have shown that BCR-ABL upregulates survivin at both the transcriptional and protein levels (15–17). However, these reports used single immortalized cell lines and attributed BCR-ABL–dependent induction of survivin to different signaling pathways.

To clarify these contrasting observations, we conducted an antisurvivin immunoblot on a wide panel of human CML cells before and after IM or nilotinib treatment. BCR-ABL inactivation produced a striking decrease in survivin levels in CML cells, confirming that BCR-ABL constitutive tyrosine kinase activity induces survivin expression (Fig. 2A and B). We next wanted to establish the BCR-ABL–induced pathway(s) responsible for this effect. Carter and colleagues had previously suggested that the RAS/RAF/MAPK cascade was involved in BCR-ABL–dependent upregulation of survivin (16). In a different cell line, Fang and colleagues identified JAK2/PI3K/c-myc signaling as the main regulator of BCR-ABL–dependent induction of survivin (17). However, when we suppressed mitogen-activated protein kinase (MAPK) activity with PD98059 in a set of 4 human CML cells, we detected a reduction in survivin levels only in the K562 line (Fig. 2C). Likewise, suppression of phosphoinositide 3-kinase (PI3K) activity by LY294002 induced a decrease in survivin expression only in LAMA84 cells (Fig. 2D). We therefore turned our attention to JAK2, as previous evidence had shown that BCR-ABL binds to JAK2 and phosphorylates it on tyrosine 1007, thereby activating its kinase activity (18). In turn, catalitically active JAK2 phosphorylates different STAT monomers inducing their dimerization, nuclear translocation, and transcriptional activity (18, 19). After blocking JAK2 signaling with the selective inhibitor TG101209, we observed a reduction in survivin expression in all CML lines tested (Fig. 2E). Prior evidence had established that JAK2 inhibition reduced survivin levels in AML cells (20). Furthermore, Fang and colleagues proposed that survivin upregulation was largely controlled at the transcriptional level through a JAK2-dependent mechanism (17). Our findings suggested a role for JAK2 in modulating survivin levels in CML cells.

To confirm that the BCR-ABL–dependent induction of survivin involves JAK2, we stably transduced CML cell lines with lentiviral vectors expressing a JAK2-specific shRNA or a nonsilencing control (Fig. 3A) and detected a reduction in survivin in all cell lines. Interestingly, a constitutive activation of the JAK2/STAT3 pathway by the BCR-ABL tyrosine kinase has been previously described in CML progenitors (21) and several reports have shown that STAT3 induces survivin expression in both solid and hematologic malignancies (22–24). Moreover, Bewry and colleagues have suggested that STAT3 might contribute to the development of IM resistance (25). To determine whether STAT3 was the pivotal downstream mediator of the BCR-ABL/JAK2 pathway, we electroporated STAT3-targeted antisense oligonucleotides or their scrambled counterpart in 4 CML lines. Immunoblot analyses showed that STAT3 downregulation reduced survivin levels in each leukemic cell line, whereas this was not the case with the scrambled control (Fig. 3B). To further confirm the role of STAT3 in the regulation of survivin, we generated an activated STAT3 (STAT3C) as previously reported by Bromberg and colleagues (Fig. 3C; ref. 26). We found that STAT3C promoted an increase in survivin levels in all 4 CML cells tested, implying that the JAK2/STAT3 cascade is a major regulator of survivin expression in CML cells.

Resistance to IM has emerged as a major caveat in the long-term treatment of CML (27). BCR-ABL amplification or point mutations in the oncoprotein kinase domain represent 2 major mechanisms underlying IM failure. We wanted to establish whether survivin downregulation by specifically targeted siRNAs would increase the apoptotic effect of HU on CML cell lines displaying different mechanisms of IM resistance. To this end, we initially used LAMA84 R cells that are unresponsive to IM because of BCR-ABL gene amplification (28). Suppression of survivin in LAMA84 R failed to sensitize them to either HU or IM (Fig. 4B), indicating that resistance caused by BCR-ABL amplification is independent from survivin expression. However, survivin silencing in IM-responsive LAMA84 S increased the cytotoxic effect of both drugs (Fig. 4A). We next repeated these experiments using Ba/F3p210 (murine Ba/F3 cells lentivirally transduced with wild-type p210 BCR-ABL), Ba/F3p210^Y253F, and Ba/F3p210^T315I cell lines that are completely unresponsive to IM (>10-fold increase in the IC₅₀ for the drug; refs. 28–30). In Ba/F3p210, survivin silencing caused a significant increase in both IM- and HU-induced death (Fig. 4C). Likewise, suppression of survivin in Ba/F3p210 cells expressing the Y253F and T315I mutants significantly increased the response to HU (Fig. 4D and E, middle). As anticipated, a reduction in survivin levels failed to restore responsiveness to IM (Fig. 4D and E, right), as point mutations induce conformational changes of the BCR-ABL kinase domain that prevent IM binding. These findings support the notion that lowering survivin expression can increase HU efficacy on CML cells unresponsive to IM because of BCR-ABL point mutations. These results also confirm that survivin does not affect the structural modifications of the BCR-ABL kinase domain induced by different amino acidic substitutions. Hence, reducing survivin levels fail to resensitize BCR-ABL mutants to the effect of IM.

The design of shepherdin, a compound that reduces survivin expression by disrupting its interaction with Hsp90, represents an attractive opportunity to determine the possible therapeutic potential of survivin downregulation in CML cells. We therefore incubated IM-resistant cell lines K562 R, KCL22 R, LAMA84 R, and their IM-sensitive counterpart with shepherdin alone or in combination with HU or doxorubicin. As expected, addition of shepherdin, but not of the scramble control peptide, reduced survivin expression in all cell lines (Fig. 5A, D, and G). When we assayed cell proliferation and survival in the presence of shepherdin and HU or doxorubicin, we found significant reductions in the growth and viability of K562 and KCL22, regardless of their IM sensitivity (Fig. 5B, C, E, and F). Likewise, LAMA84 S cells were highly sensitive to the combination of shepherdin and different chemotherapeutic drugs (Fig. 5H). However, in agreement with previous silencing data (Fig. 4B), shepherdin failed to sensitize LAMA84 R to either HU or doxorubicin (Fig. 5I), confirming that resistance caused by BCR-ABL amplification is independent from survivin. These results were further validated by using the Ba/F3 cell model transformed by either wild-type (Ba/F3p210) or mutant BCR-ABL (Ba/F3p210^Y253F, Ba/F3p210^T315I). In the context of these cell lines, shepherdin efficiently suppressed survivin expression (Fig. 6A) and its association with either HU or doxorubicin significantly reduced cell proliferation and cell viability (Fig. 6B, C, and D).

To further ascertain the efficacy of shepherdin on BCR-ABL–expressing clones, we isolated primary cells from the peripheral blood of 3 different patients with CML: 2 individuals with newly diagnosed disease and a subject that had failed treatment with IM, nilotinib, and dasatinib. We then assessed the colony-forming potential of these cells after treatment with shepherdin alone or in combination with one of the 3 TKIs. Interestingly, we observed a significant decrease in overall colony number when shepherdin was associated with nilotinib and dasatinib, regardless of the TKI sensitivity of the primary cells (Fig. 7A, B, and C). These findings further reinforced the conclusion that downregulation of survivin expression may represent a potential therapeutic strategy to eliminate CML cells both responsive and unresponsive to the TKI-dependent inhibition of BCR-ABL kinase activity.

Although IM has changed the natural course of CML achieving 89% overall survival rates after 8 years of follow-up (31), approximately 50% of patients with CML fail to obtain an optimal response as defined by the current European Leukemia Net (ELN) recommendations (32). To address these critical issues, second-generation TKIs (i.e., dasatinib and nilotinib) have been generated, tested, and approved for the treatment of patients with CML that fail IM (33). Despite the early success of these compounds, 50% of patients receiving either dasatinib or nilotinib will not obtain long-term benefit from these drugs and will progress to the advanced phases of the disease (34). Thus, the need for additional therapeutic strategies that will benefit this selected patient population (34, 35).

Carter and colleagues have previously shown that BCR-ABL induces the expression of the IAP survivin through the MAPK and PI3K pathways, and that reduction of survivin expression increases both spontaneous and IM-dependent death of a CML cell line (16). However, it is unclear whether these results could: (i) be extended to a more representative panel of CML cells both sensitive and resistant to IM; (ii) be used to potentiate the antiproliferative activity of other pharmacologic compounds; (iii) be exploited to devise a possible therapeutic strategy aimed at reducing survivin levels in patients with CML.

We report here that BCR-ABL kinase activity induced survivin expression in CML cell models via the JAK2/STAT3 pathway. Silencing of survivin reduced the viability of CML cells and enhanced the efficacy of IM. Moreover, survivin downregulation significantly increased the cytotoxic activity of HU, doxorubicin, nilotinib, and dasatinib on cells unresponsive to IM because of either point mutations (including T315I) in the BCR-ABL catalytic domain or other yet uncharacterized mechanisms. These results were obtained with the use of survivin-targeted siRNAs or by using shepherdin, a peptidomimetic compound that reduces survivin half-life by abrogating its interaction with the Hsp90 chaperone (13). These findings were further confirmed by experiments on human myeloid progenitors derived from both IM-sensitive and IM-resistant patients with CML.

Our observations lead to several considerations. First, and at variance with the data reported by Carter and colleagues, we found that only MAPK and PI3K signaling cascades modestly contribute to the regulation of survivin expression in CML. Conversely, we identified the JAK2/STAT3 pathway as a major contributor of BCR-ABL–mediated induction of survivin. This discrepancy could be due to the different cells used in the 2 studies. It is also possible that, in certain cell types, both JAK2 and MAPK contribute to survivin overexpression, as published evidence suggests that JAK2 and MAPK stimulate phosphorylation of STAT3, thus enhancing its biologic activity (36).

We have also found that suppression of survivin is unable to restore IM activity regardless of the mechanisms responsible for drug failure. These results are not unexpected as IM resistance attributable to BCR-ABL gene amplification is due to the enhanced signaling of a higher absolute number of BCR-ABL molecules (37). Intuitively, this phenomenon cannot be constrained by lowering survivin expression. Likewise, point mutations associated with IM failure induce alterations in the 3-dimensional structure of the BCR-ABL catalytic domain that either mimic an active conformation or affect specific amino-acidic residues required for the physical interaction with IM (38, 39). Thus, it is evident that suppression of survivin per se fails to restore sensitivity to IM, as survivin does not associate with BCR-ABL and is therefore incapable of modulating the conformation of its kinase domain.

However, suppression of survivin levels significantly increased the cytotoxic activity of different compounds on CML cells failing IM because of yet uncharacterized mechanisms or point mutations in the BCR-ABL kinase domain. In these contexts, suppression of survivin may represent a potential therapeutic strategy to reduce the viability of these leukemic clones. To this end, we have conducted several experiments using the peptidomimetic shepherdin to lower survivin expression in IM-sensitive and IM-resistant CML cells. Previous studies investigating the use of shepherdin have described strong antitumoral activity in human models of both solid and hematologic malignancies (13, 40). Our findings indicate that shepherdin-dependent reduction of survivin levels significantly increased the cytotoxic activity of HU and doxorubicin on all cell lines expressing BCR-ABL mutants. In addition, as further compounds (ISIS 23722, SPC3042, and YM155) are currently being developed as inhibitors of survivin expression (41), any pharmacologic strategy aimed at reducing survivin levels could be of significant value in patients with IM-resistant clones carrying point mutations in the BCR-ABL kinase domain.

Finally, the reduction in cell viability that we observed after shepherdin treatment was superior to that detected after siRNA-mediated silencing, suggesting that shepherdin may increase CML cell death by both survivin-dependent and survivin-independent mechanisms. Kang and colleagues have previously shown that shepherdin disables a mitochondrial chaperone network that cancer cells use to protect themselves from multiple apoptotic stimuli (42). Hence, it is possible that part of shepherdin's efficacy on CML cells depends on this mechanism. Indeed, when we treated Ba/F3 cells expressing different BCR-ABL mutants with HU or doxorubicin in association with 17-AAG (a cytosolic inhibitor of Hsp90), we found no significant increases in cell death (Fig. 8). These preliminary findings support the hypothesis that shepherdin kills CML cells through a double mechanism: (i) a reduction in survivin expression and (ii) the mithochondrial inhibition of Hsp90 chaperones.

In summary, our results suggest that pharmacologic reduction of survivin expression may provide a new therapeutic approach for the molecular targeting of IM-insensitive CML cells.

No potential conflicts of interest were disclosed.

Conception and design: S. Stella, E. Tirrò, E. Conte, P. Vigneri

Development of methodology: S. Stella, E. Tirrò

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Stella, E. Conte, F. Stagno, F.D. Raimondo

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Stella, E. Tirrò, F. Stagno, F.D. Raimondo, P. Vigneri

Writing, review, and/or revision of the manuscript: S. Stella, E. Tirrò, E. Conte, F. Stagno, F.D. Raimondo, L. Manzella, P. Vigneri

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Stella, P. Vigneri

Study supervision: S. Stella, P. Vigneri

This work was supported by a fellowship (E. Tirrò) and a national grant from Associazione Italiana per la Ricerca sul Cancro (P. Vigneri).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.