DR-nm23 gene expression in neuroblastoma cells: relationship to integrin expression, adhesion characteristics, and differentiation

DR-nm23 Gene Expression in Neuroblastoma Cells: Relationship to Integrin Expression, Adhesion Characteristics, and Differentiation Roberto Amendola, Robert Martinez, Anna Negroni, Donatella Venturelli, Barbara Tanno, Bruno Calabretta, Giuseppe Raschella`*

Tumor cells of different histologic types often show specific patterns of gene expression that result in altered pathways of cell growth and differentiation. Knowledge of the mechanisms involved in maintaining the appropriate differentiation pathways in normal cells and how such mechanisms are subverted in malignant cells is becoming increasingly important, both for prognosis and for therapy. Neuroblastoma, a common childhood tumor derived from cells of the embryonic neural crest (1), 1300 ARTICLES

provides a good example. This tumor retains the ability to express multiple phenotypes [e.g., neuronal, epithelial-Schwannlike, and melanocytic (2)] and is characterized by widely variable clinical outcomes. The identification of nonrandom chromosomal changes in neuroblastoma cells has led to the suggestion that specific genetic alterations play an important role in maintaining and/or inducing an aggressive phenotype (3). Indeed, several genetic abnormalities can be used as prognostic indicators and are associated with different clinical outcomes. Among these abnormalities, N-myc (also known as MYCN) amplification (4), allelic loss of chromosome 1p (5), and the expression of trk receptors (6) are currently used in association with clinical parameters to help determine the appropriate treatment for patients with neuroblastoma. In spite of the importance of these genetic abnormalities, the mechanisms contributing to the onset and the progression of neuroblastoma are still largely unknown. The putative metastasis suppressor genes nm23-H1 (also known as HSNM23H1) (7) and nm23-H2 (also known as HSNM23H2G) (8), which encode proteins with nucleoside diphosphate kinase activity (9), are expressed in several types of tumors, but they show an uncertain association with prognosis. In melanoma (10), hepatocellular carcinoma (11), and breast carcinoma (12), elevated levels of nm23 gene expression have been correlated with a decreased tendency to form metastases. In contrast, higher nm23 gene expression in prostate carcinoma (13) and in squamous cell lung carcinoma (14) has been associated with poor prognosis. In neuroblastoma, nm23 expression has been associated with adverse outcome, amplification of the N-myc gene, and an increase in mitotic labeling index (15,16). However, point mutations, which may alter nm23 function, have been found both in the nm23-H1 gene and in the nm23-H2 gene (15,17,18), raising the possibility of dominant-negative effects of nonfunctional Nm23 proteins in the neoplasms that are characterized by an aggressive phenotype. On the other hand, several lines of evidence suggest a direct

*Affiliations of authors: R. Amendola, A. Negroni, B. Tanno, G. Raschella`, Enea, CR-Casaccia, Section of Toxicology and Biomedical Sciences, Rome, Italy; R. Martinez, D. Venturelli, B. Calabretta, Kimmel Cancer Institute, Department of Microbiology, Thomas Jefferson University, Philadelphia, PA. Correspondence to: Bruno Calabretta, M.D., Kimmel Cancer Institute, Department of Microbiology, Thomas Jefferson University, 10th St. and Locust St., Philadelphia, PA 19107; or Giuseppe Raschella`, Ph.D., Enea, CR-Casaccia, Section of Toxicology and Biomedical Sciences, via Anguillarese 301, 00060, Rome, Italy. See ‘‘Notes’’ following ‘‘References.’’ © Oxford University Press

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

Background: Neuroblastoma, a childhood tumor originating from cells of the embryonic neural crest, retains the ability to differentiate, yielding cells with epithelial-Schwann-like, neuronal, or melanocytic characteristics. Since nm23 gene family members have been proposed to play a role in cellular differentiation, as well as in metastasis suppression, we investigated whether and how DR-nm23, a recently identified third member of the human nm23 gene family, might be involved in neuroblastoma differentiation. Methods: Three neuroblastoma cell lines (human LAN-5, human SK-N-SH, and murine N1E-115) were used in these experiments; cells from two of the lines (SK-N-SH and N1E-115) were also studied after being stably transfected with a plasmid containing a full-length DR-nm23 complementary DNA. Cellular expression of specific messenger RNAs and proteins was assessed by use of standard techniques. Cellular adhesion to a variety of protein substrates was also evaluated. Results: DR-nm23 messenger RNA levels in nontransfected LAN-5 and SK-N-SH cells generally increased with time after exposure to differentiation-inducing conditions; levels of the other two human nm23 messenger RNAs (nm23-H1 and nm23-H2) remained essentially constant. Transfected SK-NSH cells overexpressing DR-nm23 exhibited some characteristics of differentiated cells (increased vimentin and collagen type IV expression) even in the absence of differentiationinducing conditions. Compared with control cells, DR-nm23transfected cells exposed to differentiation-inducing conditions showed a greater degree of growth arrest (SK-N-SH cells) and greater increases in integrin protein expression, especially of integrin b 1 (N1E-115 cells). DR-nm23transfected N1E-115 cells also showed a marked increase in adhesion to collagen type I-coated tissue culture plates that was inhibited by preincubation with an anti-integrin b1 antibody. Conclusions: DR-nm23 gene expression appears to be associated with differentiation in neuroblastoma cells and may affect cellular adhesion through regulation of integrin protein expression. [J Natl Cancer Inst 1997;89:1300–10]

role for nm23 genes in cellular differentiation. In Drosophila, reduced expression or mutation of the nm23 homologue, abnormal wing discs (awd), affects cell morphology (19). In mouse embryogenesis, the Nm23 protein has been detected in virtually all epithelial-Schwann-like tissues during physiologic differentiation, and it disappears in adult epithelia (20). Finally, ectopically expressed nm23-H1 in breast carcinoma cells triggers differentiation to reconstitute acinus-like spheres in a threedimensional tissue culture system (21), and overexpression of nm23-H1 induces neural differentiation in PC12 rat pheochromocytoma cells (22). Despite their relatively small size, Nm23 proteins have structural characteristics that suggest multiple roles. Kinase activity was the first biochemical function demonstrated for both Nm23H1 and Nm23-H2 (9). A putative leucine-zipper motif, involved in the dimerization of many transcriptional factors, is also present in Nm23 proteins. In fact, Nm23-H2 has been identified as the human c-myc gene transcription factor PuF, which binds the Pu-rich motif in the c-myc promoter (23,24). Moreover, the presence of the RGD (arginine–glycine–aspartic acid) integrinbinding domain suggests additional, as yet unidentified, roles for Nm23 proteins. The scenario is further complicated because it is still not clear whether these distinct functions operate in different cell types or in the same cell in a temporally regulated manner. Recently, a third member of the human nm23 gene family, DR-nm23, was identified through differential screening of a blast-crisis chronic myelogenous leukemia complementary DNA (cDNA) library (25). DR-nm23, which shares 70% homology with nm23-H1 and nm23-H2, retains the domains postulated to be important for nm23 function, and its overexpression induces apoptosis in myeloid precursor 32Dcl3 cells cultured in the presence of the differentiation inducer G-CSF (granulocyte colony-stimulating factor) and in the absence of interleukin 3. DR-nm23 expression increases in the initial stages of myeloid differentiation and declines thereafter, suggesting a role for this gene in hematopoietic differentiation (25). To determine whether and how DR-nm23 is involved in neuroblastoma differentiation, we evaluated the expression of DRnm23 messenger RNA (mRNA) during both neuronal and epithelial-Schwann-like differentiation of human neuroblastoma cells and compared the pattern of DR-nm23 expression with that of nm23-H1 and nm23-H2. We also examined the growth characteristics and the morphologic and biochemical features of epithelial-Schwann-like human SK-N-SH and neuronal murine N1E-115 neuroblastoma cells that had been transfected with and constitutively expressed a full-length human DR-nm23 cDNA.

bovine serum (Sigma Chemical Co.). The human neuroblastoma cell line SKN-SH (28) and its derivatives transfected with either pcDNA3 (Invitrogen) (pool A and pool B, each collected from at least 15 individual clones) or the pcDNA3/ DR-nm23 construct (pool A and pool B, each collected from at least 15 individual clones and expressing at least 10 times more DR-nm23 mRNA than the control vector clone pools) were grown in minimal essential medium (Life Technologies, Inc. [GIBCO BRL], Gaithersburg, MD), supplemented with 10% fetal bovine serum. Transfectant clones were selected in the presence of 400 mg/mL G418 (Sigma Chemical Co.). The murine neuroblastoma cell line N1E115 (29) and all of its derivatives were grown in minimal essential medium supplemented with 7.5% fetal bovine serum. Stable transfectants of N1E-115 were selected in the presence of 500 mg/mL G418. Cell differentiation experiments were carried out by inducing chemical differentiation in the human LAN-5 and SK-N-SH cell lines with all-trans-retinoic acid (RA) (Sigma Chemical Co.) and in murine N1E-115 cells with dimethyl sulfoxide (DMSO) (Sigma Chemical Co.). The cells were seeded at an initial density of 5 × 103 cells/cm2. After 16 hours, the culture medium was replaced with fresh medium containing 5 mM RA or 1.25% DMSO. The medium containing the differentiation inducer was replaced every 3 days.

Materials and Methods

Western Blot Analysis of Integrins

A human DR-nm23 full-length cDNA (23) was subcloned in the vector pcDNA3 (Invitrogen, San Diego, CA). It was also subcloned in the vector pGFP (CLONTECH Laboratories, Inc., Palo Alto, CA) in-frame with cDNA encoding the green fluorescent protein (GFP). This second construct produces a fusion protein in which the GFP sequence is located at the amino-terminus of DRnm23. The vector CMV-Rb1 was described by Claudio et al. (26).

Cell Culture The human neuroblastoma cell line LAN-5 (27) was grown in RPMI-1640 medium (Sigma Chemical Co., St. Louis, MO) supplemented with 15% fetal

The pcDNA3/DR-nm23 vector used in transfection experiments was transcribed and translated in vitro in the presence of [35S]methionine (Du Pont NEN, Bad Homburg, Germany) by use of a commercially available kit (Promega Corp., Madison, WI) and following the manufacturer’s instructions. The translation product was characterized by sodium dodecyl sulfate (SDS)– polyacrylamide gel electrophoresis (PAGE), and the size of the translated protein was compared with the size of molecular weight marker standards.

Indirect Immunofluorescence Analysis SK-N-SH cells and their transfectant derivatives were fixed in monolayer cultures by use of acetone–methanol (75:25 [vol:vol]) and then analyzed for expression of vimentin (by use of a monoclonal antibody from Sigma Chemical Co. [catalog No. V-6330]; 1:40 dilution), collagen type IV (by use of monoclonal antibody from Sigma Chemical Co. [catalog No. C-1926]; 1:500 dilution), and glial fibrillary acidic protein (GFAP) (by use of monoclonal antibody from Sigma Chemical Co. [catalog No.G-3893]; 1:400 dilution). A fluorescein isothiocyanate (FITC)-conjugated secondary antibody (Sigma Chemical Co. [catalog No. F-6257]) was also used in these assays. Control assays employed the FITC-labeled secondary antibody only. Independent photomicroscopy experiments were carried out by use of the same time-exposure and film-processing protocols to assess different levels of antigen expression.

Immunocytofluorometry of Integrin a1 and a2 Expression SK-N-SH cells and their transfectant derivatives were fixed with 70% ethanol and analyzed for modulation of integrins a1 and a2 by use of murine monoclonal antibody ACT-T-SET VLA1 (T Cell Science, Cambridge, MA) and murine monoclonal antibody clone P1E6 (Life Technologies, Inc.), respectively, and immunocytofluorometry as described (30). As a control, integrin a3 expression, which is invariant during neuroblastoma cell differentiation (30), was assessed by use of the anti-human integrin a3 monoclonal antibody clone P185 (Life Technologies, Inc.).

Parental and pcDNA3/DR-nm23-transfected cells (106) were lysed with 100 mL 50 mM Tris–HCl (pH 7.4), 250 mM NaCl, 5 mM EDTA, 50 mM sodium fluoride, 0.1 mM sodium vanadate, 0.1% Triton X-100, 1 mM phenylmethylsufonyl fluoride, and 10 mg/mL leupeptin, and the proteins were separated in a 6% SDS–PAGE gel. Western blot analysis was carried out as described (31), using an ECL (chemiluminescence) detection protocol kit (Amersham, Little Chalfont, U.K.) after probing with the following antibodies: a rabbit polyclonal antiintegrin b1 (32), a rabbit polyclonal anti-integrin a1 (33), a rabbit polyclonal anti-integrin a2 (34), and a monoclonal anti-HSP70 (StressGen Biotechnology Corp., Victoria, Canada). The anti-HSP70 antibody was used to control for equivalence of protein loading.

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

ARTICLES 1301

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

Expression Plasmids

In Vitro Transcription and Translation of DR-nm23

Northern Blot Analysis

Results

For northern blot analysis, total RNA was separated in 1% agarose– formaldehyde gels and transferred to nylon membranes (Stratagene, La Jolla, CA). The membranes were hybridized sequentially with specific radioactively labeled probes by use of the high-stringency procedure suggested by the membrane manufacturer. The relative amount and the integrity of the RNA loaded in each lane were assessed by means of ethidium bromide staining of the gels and by hybridization of the membranes with a b-actin probe. Hybridized probe was detected by means of autoradiography. DR-nm23 mRNA was detected through hybridization of a BamHI/Xho I restriction-fragment probe that was derived from the plasmid pcDNA3/DR-nm23, which, as noted above, contains a DR-nm23 full-length cDNA (25). Probes specific for nm23-H1 and for nm23-H2 were generated by means of DNA amplification, using the polymerase chain reaction (PCR) as described (8). Human b-actin mRNA was detected through hybridization of a 2.1-kilobase (kb) Xho I restriction fragment probe derived from the clone HFbA-1 (35).

Expression of DR-nm23, nm23-H1, and nm23-H2 mRNAs During Neuronal and Epithelial-Schwann-Like Neuroblastoma Cell Differentiation

Reverse Transcription (RT)–PCR Analysis

Growth Analysis Three independent experiments were carried out by seeding SK-N-SH cells and their transfected derivatives at an initial density of 5 × 103 cells/cm2. Thymidine labeling indices were determined after a 6-hour incorporation of [3H]thymidine (20 Ci/mmol; DuPont NEN). The labeled cells were counted in a scintillation counter (Beckman Instruments, Inc., Fullerton, CA).

Cell Adhesion Six aliquots of 2 × 105 pcDNA3- and pcDNA3/DR-nm23-transfected N1E115 cells were seeded separately into 9.6-cm2 wells of tissue culture plates coated with polylysine, fibronectin, laminin, collagen type I, or collagen type IV (Collaborative Research, Inc., Bedford, MA) in minimal essential medium that was supplemented with 2.5 mg/mL bovine serum albumin. After 1 hour at 37 °C in 5% CO2, the cells were fixed with 4% paraformaldehyde and stained with Giemsa (Sigma Chemical Co.). For cell-adhesion inhibition assays that used a specific anti-integrin b1 antibody, cells were incubated in minimal essential medium supplemented with 2.5 mg/mL bovine serum albumin and serial dilutions of the rabbit polyclonal anti-murine integrin b1 antibody 6-236 (gift of P. Bernardi; dilutions of 1:50, 1:100, 1:1000, and no antibody) for 15 minutes at 4 °C and then plated on collagen type I-, collagen type IV-, or fibronectin-coated plates and processed as described for the adhesion assay. Microphotographs from six independent experiments were analyzed for the number of attached cells on each substrate by use of the computer-assisted software Mocha (Jandel Scientific, San Rafael, CA).

Statistical Analyses Ninety-five percent confidence intervals (CIs) were used to assess variability in each analysis and differences between the various experimental groups. Student’s t test was also used in the data analysis. Reported P values are two-sided.

1302 ARTICLES

Epithelial-Schwann-Like Differentiation in Neuroblastoma Cells Constitutively Expressing DR-nm23 Human neuroblastoma SK-N-SH cells show a fibroblast-like morphology when cultured under basal (nondifferentiating) conditions. Induction of epithelial-Schwann-like differentiation causes a cohort of morphologic and biochemical changes, including an enlarged and rounded shape, enhancement of vimentin production, deposition of collagen type IV (2), and modulation of integrin patterns (30,37). To determine the effects of DR-nm23 overexpression in SK-N-SH cells, plasmid pcDNA3/ DRnm23, in which a full-length DR-nm23 cDNA is under control of the constitutively active cytomegalovirus (CMV) promoter, was used to transfect the cells. After transfection and selection in G418-containing medium, pools of pcDNA3- and pcDNA3/DR-nm23-transfected cells were collected to minimize the consequences of clonal variability and to mimic the heterogeneity of the parental cells. As expected, DR-nm23 mRNA levels were very abundant in the transfected cells, whereas the endogenous transcripts were barely detectable in the parental cells and in empty vector-transfected cells (Fig. 2). It is interesting that ectopic expression of DR-nm23 mRNA in SK-N-SH cells had no effect on either nm23-H1 or nm23-H2 mRNA levels (Fig. 2). Immunofluorescence staining of vimentin and collagen type IV in parental SK-N-SH cells, in pcDNA3-transfected cells, and in pcDNA3/DR-nm23 transfectants revealed a marked increase in vimentin expression and deposition of collagen type IV only in the pcDNA3/DR-nm23 transfectants (Fig. 3). This pattern of extracellular matrix protein (i.e., collagen type IV) expression in pcDNA3/DR-nm23 transfectants is consistent with the pheno-

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

Aliquots (500 ng each) of total RNA isolated from parental, pcDNA3- and pcDNA3/DR-nm23-transfected NIE-115 cells were reverse transcribed in the presence of Moloney murine leukemia virus reverse transcriptase (Promega Corp.) and random hexamer oligonucleotides (The Perkin-Elmer Corp., Emeryville, CA). The reactions were carried out in a 480-Thermal Cycler (The PerkinElmer Corp.) at 40 °C for 45 minutes. The reaction mixtures were subsequently heated at 95 °C for 5 minutes and then quickly chilled in an ice bath. Reactions run without added reverse transcriptase were used as negative controls for DNA carry-over contamination. Aliquots of the first-strand cDNAs were amplified separately in the presence of specific primers for human DR-nm23 and murine b-actin. The sequences of the sense and antisense primers spanning a 405-basepair (bp) region of the DR-nm23 cDNA were as follows: 58-ATCATGATCTGCCTGGT-38 and 58-AATCAGGTTCTTGCCAACCT-38, respectively. The sequences of the sense and antisense primers spanning a 539-bp region of b-actin cDNA were as follows: 58-GTGGGCCGCTCTAGGCACCAA-38 and 58-CTCTTTGATGTCACGCACGTATTC-38, respectively (36). Aliquots of the amplified products were transferred to nylon membranes and hybridized sequentially with internal, 32P-end-labeled probes specific for each amplification product.

Human neuroblastoma cell lines LAN-5 and SK-N-SH exhibit neuronal and epithelial-Schwann-like differentiation, respectively, when cultured in the presence of 5 mM RA. To compare the kinetics of mRNA expression of nm23-H1 and nm23-H2 with that of DR-nm23 during neuronal and epithelialSchwann-like differentiation processes, we analyzed by northern blots RNAs from LAN-5 and SK-N-SH cells treated with RA for 6 hours or for 1, 3, or 10 days (Fig. 1). The levels of nm23-H1 and nm23-H2 mRNA, both detected as single species of 0.8 kb, remained essentially constant during RA-induced differentiation of both cell lines and were higher than the levels of DR-nm23 mRNA (Fig. 1). The levels of DR-nm23 mRNA, detected as a main species of 0.8 kb, steadily increased during RA-induced differentiation of SK-N-SH cells (Fig. 1); in LAN-5 cells, however, an early decline in the abundance of the 0.8-kb DR-nm23 mRNA was followed by a steady increase (Fig. 1). A second, less abundant DR-nm23 transcript (approximately 1.1 kb) was also detected in LAN-5 cells; however, levels of this transcript were markedly reduced to the point of nondetection upon RA treatment (Fig. 1). Thus, DR-nm23, but not nm23-H1 or nm23H2, mRNA expression appears to be modulated during neuroblastoma cell differentiation.

Fig. 1. Northern blot analysis of nm23 gene expression following all-trans-retinoic acid (R.A.) induction of differentiation in human LAN-5 and SK-N-SH neuroblastoma cells. Twenty micrograms of total RNA was loaded in each gel lane. UN. 4 untreated cells; 6h (hours), 1d (day), 3d, and 10d indicate the duration of R.A. induction. The probes used to detect DR-nm23, nm23-H1, nm23-H2, and human b-actin messenger RNAs are described in the ‘‘Materials and Methods’’ section. The film exposure times were overnight for nm23-H1, nm23-H2, and human b-actin messenger RNAs and 4 days for DR-nm23 messenger RNA. kb 4 kilobase.

Fig. 2. Expression of nm23 messenger RNAs in control and pcDNA3/ DR-nm23-transfected human SK-N-SH neuroblastoma cells. Ten micrograms of total RNA from parental SK-N-SH cells (lane 1), pcDNA3 (empty vector) control transfectants (lane 2), and pcDNA3/DR-nm23 transfectants (lane 3) was subjected to electrophoresis, blotted, and hybridized to DR-nm23, nm23-H1, nm23-H2, and human b-actin probes as described in ‘‘Materials and Methods’’ section. Film exposure times were overnight for nm23-H1, nm23-H2, and human bactin messenger RNAs and 4 days for DR-nm23 messenger RNA.

neuroblastoma cell differentiation, but the changes were more pronounced in the DR-nm23 transfectants (especially the increase in the levels of integrin b1). For example, a1 expression increased somewhat in SK-N-SH cells after RA treatment (Fig. 4, A; compare panels a and c), but the a1 fluorescence intensity of RA-treated pcDNA3/DR-nm23 transfectants was twice as high as that of uninduced pcDNA3/DR-nm23 transfectants (Fig. 4, A; compare panels e and g). Flow cytometry analysis of a2 expression demonstrated a bimodal pattern in SK-N-SH cells grown under basal conditions (Fig. 4, A, panel b) but a unimodal pattern after RA treatment (Fig. 4, A, panel d). By contrast, pcDNA3/DR-nm23 transfectants showed a unimodal pattern of a2 expression under basal conditions (Fig. 4, A, panel f), indistinguishable from that induced by RA treatment in SK-N-SH cells (Fig. 4, A, panel d). RA treatment of pcDNA3/DR-nm23 transfectants further enhanced a2 expression (Fig. 4, A, panel h). Western blot analysis of integrin b1 demonstrated the expected increase in expression upon chemically induced differentiation of parental SK-N-SH cells (Fig. 4, B); it is interesting that an approximately threefold increase in integrin b1 levels was detected in pcDNA3/DR-nm23-transfected cells, regardless of the growth conditions (untreated or RA-treated) (Fig. 4, B). Integrin modulation in pcDNA3 control transfectants was similar to that induced by RA treatment in SK-N-SH parental cells (data not shown). Growth Analyses Growth inhibition is a well-known hallmark of terminal differentiation, although it is unclear whether inhibition of proliferation and the onset of differentiation are causally related. For example, myoblasts stop proliferating before entering into the differentiation pathway (38), whereas senescent melanocytes can remain nonproliferative without terminally differentiating (39). In neuroblastoma, the loss of cell proliferation is not sufficient for either epithelial-Schwann-like or neuronal differen-

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

ARTICLES 1303

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

type of differentiated cells. By contrast, immunofluorescence analysis of GFAP, an epithelial-Schwann-like marker of very late stages of differentiation, yielded negative results in all tested cells (data not shown). Flow cytometric analysis of expression of integrins a1 and a2 and western blot analysis of integrin b1 expression in SK-N-SH cells and in pcDNA3/DR-nm23 transfectants cultured in basal growth conditions or in the presence of RA were performed to assess the modulation of integrin expression. The results of these analyses indicate modulation of integrin expression typical of

Fig. 3. Immunofluorescence analysis of vimentin and collagen type IV expression in parental human SK-N-SH neuroblastoma cells (panel A and panel D, respectively), pcDNA3 (empty vector) control transfectants (panel B and panel E, respectively), and pcDNA3/DR-nm23 transfectants (panel C and panel F, respectively). See ‘‘Materials and Methods’’ section for details.

midine incorporation between the two cell types (pcDNA3-B pool and DR-nm23-B pool in Fig. 5). Under basal conditions, a mean of 95.93 (95% CI 4 ±22.54) cpm were incorporated per 103 pcDNA3/DR-nm23 transfectants compared with a mean of 186.00 (95% CI 4 ±7.74) incorporated per 103 control transfectants (P 4 .0034; 95% CI for the difference between the groups 4 77.99–102.14), suggesting a clear decrease in the proliferation (by 48.42%) of pcDNA3/DR-nm23-transfected cells. Under conditions of serum starvation (4-day culture in 1% fetal calf serum), which affects proliferative activity but does not trigger overt differentiation, a significantly decreased [3H]thy-

Fig. 4. Cytofluorometric analysis of integrins a1 and a2 and western blot analysis of integrin b1 in parental (control) human SK-N-SH neuroblastoma cells and in pcDNA3/DR-nm23 transfectants under basal culture conditions (untreated) and after a 6-day treatment with the differentiation inducer all-trans-retinoic acid (RA) (5 mM). A) Panels a and c show integrin a1 expression in untreated and RA-treated SK-NSH cells, respectively. Panels e and g show integrin a1 expression in untreated and RA-treated DR-nm23 transfectants, respectively. Panels b and d show integrin a2 expression in untreated and RA-treated SKN-SH cells, respectively. Panels f and h show integrin a2 expression in untreated and RA-treated DR-nm23 transfectants, respectively. Expression of integrin a3, which is not modulated during neuroblastoma cell differentiation (30), remained unchanged in untreated and RA-treated parental and DR-nm23 transfectants (data not shown). FITC 4 fluorescein isothiocyanate. Increases in integrin expression result in a rightward shift of the peak of fluorescing cells. See ‘‘Materials and Methods’’ section for details. B) Western blot analysis of integrin b1 expression in SK-N-SH cells (N) and in pcDNA3/DR-nm23 transfectants (DR). Antibodies used for integrin b1 and HSP70 detection are described in ‘‘Materials and Methods’’ section.

1304 ARTICLES

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

tiation. In LAN-5 cells, for example, overexpression of the proliferation-related B-Myb gene prevents RA-induced differentiation (40), but inhibition of c-myb mRNA expression by ectopic expression of c-myb antisense transcripts reduces proliferation and induces apoptosis instead of differentiation (41). However, treatment with RA inhibits growth of LAN-5 and SK-N-SH cells and triggers differentiation. [3H]Thymidine incorporation assays were performed in triplicate in pcDNA3- and DR-nm23-transfected SK-N-SH cells (Fig. 5). The two-tailed Student’s t test and 95% CIs were used to analyze the statistical significance of differences in [3H]thy-

midine incorporation was observed in pcDNA3/DR-nm23 transfectants (25.53 [95% CI 4 ±6.27] cpm per 103 cells; a 73.39% reduction in comparison with basal conditions), whereas [3H]thymidine incorporation in control transfectants (154.37 [95% CI 4 ±6.00] cpm per 103 cells; a 17.01% reduction in comparison with basal conditions) decreased only slightly (P 4 1.29E-06; 95% CI for the difference between the groups 4 124.42–133.26). Upon RA-induced differentiation, a similarly marked decrease in proliferation was observed for pcDNA3/DRnm23 transfectants (24.10 [95% CI 4 ±10.62] cpm per 103 cells; a 74.88% reduction in comparison with basal conditions), while control transfectants showed the expected decrease (81.93 [95% CI 4 ±5.90] cpm per 103cells; a 55.95% reduction in comparison with basal conditions) (P 4 .0005; 95% CI for the difference between the groups 4 51.63–64.03). Together, these results suggest that DR-nm23 overexpression inhibits SK-N-SH cell proliferation, possibly as a consequence of its ability to induce differentiation.

Fig. 6. A) Reverse transcription–polymerase chain reaction detection of DRnm23 and b-actin messenger RNA expression in pcDNA3 (empty vector) control transfectants of murine N1E-115 neuroblastoma cells (lanes 2 and 3) and in pcDNA3/DR-nm23 transfectants (lanes 4 and 5). Lane 1 shows the products of a negative control reaction in which reverse transcription was omitted. B) In vitro transcription and translation of the pcDNA3 control (empty) plasmid vector (lane 1) and the pcDNA3/DR-nm23 plasmid (lane 2). See ‘‘Materials and Meth-

To assess the role of DR-nm23 overexpression in neuronal differentiation, the murine neuroblastoma cell line N1E-115, which undergoes complete neuronal differentiation upon treatment with various inducers (42), was co-transfected with pcDNA3/DR-nm23 and pCMV/bGAL (expresses Escherichia coli b-galactosidase under control of the CMV promoter) at a 4:1 ratio. As indicated by RT–PCR analysis, expression of human DR-nm23 mRNA was detected only in DR-nm23transfected N1E-115 cells (Fig. 6, A). Since a specific antibody is not yet available for DR-Nm23 protein, the plasmid used for

ods’’ section for details. C) Subcellular localization of DR-nm23 that has been tagged at the amino-terminus with green fluorescent protein (GFP). Parental N1E-115 cells (a) show background fluorescence; pGFP/DR-nm23-transfected N1E-115 cells (b) show cytoplasmic green fluorescence. Images were prepared by use of a confocal microscope (MRC 600; Bio-Rad Laboratories, Hercules, CA), setting the excitation and emission filters as recommended by the filter manufacturer (CLONTECH Laboratories, Inc.).

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

ARTICLES 1305

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

Fig. 5. [3H]Thymidine uptake in pcDNA3 (empty vector) control transfectants of human SK-N-SH neuroblastoma cells and in pcDNA3/DR-nm23 transfectants. UN. 4 untreated cells; 1% FCS 4 cells grown in medium supplemented with 1% fetal calf serum for 4 days; RA (all-trans-retinoic acid) 4 cells treated with 5 mM RA for 6 days. Vertical bars indicate ±95% confidence intervals. The Student’s t test (two-sided) was used to determine the statistical significance (P) of comparing the less divergent pools of cells (pool B of pcDNA3 control transfectants and pool B of pcDNA3/DR-nm23 transfectants) to avoid overestimation of the effect. See ‘‘Materials and Methods’’ section for more information.

Neuronal Differentiation in Neuroblastoma Cells Constitutively Expressing DR-nm23

Fig. 7. Neuronal differentiation (A) and integrin expression (B, C) in pcDNA3/ DR-nm23-transfected murine N1E-115 neuroblastoma cells. A) Dark cells indicate b-galactosidase-producing N1E-115 cells co-transfected with pCMV/ bGAL and either the control (empty) pcDNA3 vector (negative control), the pcDNA3/DR-nm23 vector, or the pRb-1 vector (encodes the product of the Rb-1 [retinoblastoma] gene) (positive control). B) Western blot analysis of integrin a1, a2, and b1 expression. N 4 the integrin levels in parental N1E-115 cells; C 4

1306 ARTICLES

difference between the groups 4 33.24–41.02; n 4 6). As expected, b-tubulin expression was decreased in all differentiated cells (data not shown). Western blot analysis of integrin levels showed a marked up-regulation (i.e., increase) in integrin b1 in DR-nm23-expressing cells (a twofold to threefold increase) and a much more modest increase in both integrins a1 (approximately 30%) and a2 (approximately 10%) (Fig. 7, B and C). Cell Adhesion The growth and differentiation of cells can be affected via intracellular signals activated by interaction with components of the extracellular matrix. Such interactions are believed to be modulated by a complex network of integrins that bind to this matrix (44). Integrins b1, a1, and a2, which are modulated in neuroblastoma cells that have been chemically induced to differentiate (30,37), form the a1 b1 and a2 b1 heterodimers that mediate cell adhesion to laminin and collagen types I and IV in several cell types (45,46). Since neuroblastoma cells overexpressing DR-nm23 show both augmented levels of these integrins and of other markers of cell differentiation, it is possible that pcDNA3/DR-nm23-transfected N1E-115 cells might exhibit increased adhesion to extracellular matrix proteins. In adhesion assays to substrate-coated plates, the fraction of pcDNA3 control and pcDNA3/DR-nm23-transfected cells able to adhere

the levels in control (empty vector) pcDNA3-transfected cells; DR-A 4 the levels in pcDNA3/DR-nm23-transfected cells, pool A; DR-B 4 the levels in pcDNA3/DR-nm23-transfected cells, pool B (see ‘‘Materials and Methods’’ section). C) Increases in integrin expression represented as bar graphs of arbitrary units of densitometric analysis of the blots shown in B. HSP70 expression was used for normalization of overall protein levels.

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

the transfection experiments was transcribed and translated in vitro to demonstrate that it generates a protein with the expected size of 20 kd (Fig. 6, B). Moreover, a pGFP-DR-nm23 construct was transfected into N1E-115 cells to demonstrate synthesis and sublocalization of the chimeric protein, which were revealed by detection of the amino-terminus-tagged green fluorescence product in the cell cytoplasm (Fig. 6, C). Differentiation was assessed by morphologic criteria, such as neurite outgrowth and flattening and enlargement of the cellular body, and by biochemical changes, such as a decrease in b-tubulin levels (43) and modulation of integrin expression (30,37). N1E-115 cells were chosen for these experiments because they can be transfected far more efficiently than human LAN-5 cells, thus minimizing unwanted counterselection of nonexpressing clones versus proliferation-inhibited DR-nm23-expressing clones. Furthermore, use of the bGAL-plasmid allows assessment of neural differentiation only in transfected cells (Fig. 7, A). The percentage of fully differentiated bGAL-positive cells among the pcDNA3/DR-nm23-transfected cells was 32.27% (95% CI 4 ±0.075). In contrast, this value was only 8.43% (95% CI 4 ±0.0023) among the control pcDNA3 transfectants (P 4 .0011; 95% CI for the difference between the groups 4 15.96–31.71; n 4 6). Cells transfected with the vector pRb-1 were used as positive controls (42), yielding 45.56% (95% CI 4 ±0.0039) differentiated cells (P 4 2.01E-07; 95% CI for the

to either laminin or fibronectin was similar, and their morphology was also indistinguishable (Fig. 8, A; data not shown). The proportion of cells adhering to collagen type IV-coated plates was not statistically significantly different between control and pcDNA3/DR-nm23-transfected cells (Fig. 8, A), although pcDNA3/DR-nm23-transfected cells exhibited a more differentiated phenotype (Fig. 8, B). By contrast, pcDNA3/DR-nm23transfected cells exhibited a marked increase in the adhesion to collagen type I-coated plates, consistent with a1b1 and a2b1 binding specificity to collagen type I (45,46), and the increased expression of integrins a1, a2, and b1 in DR-nm23-transfected cells (Fig. 7); in six independent experiments, the mean number

of adherent cells/cm2 (±95% CI) was 2223 (±367.25) for pcDNA3/DR-nm23 transfectants compared with 366.80 (±109.55) for pcDNA3 transfectants (P 4 .009; 95% CI for the difference between the groups 4 1472.95–2239.45; n 4 6) (Fig. 8, A). As evidenced by Fig. 8, B, the increased adherence correlated with a more differentiated phenotype. To determine whether the increased levels of integrin b1 in pcDNA3/DR-nm23-transfected cells were responsible for the enhanced adhesion of these cells to collagen type I-coated plates, we performed adhesion assays in the presence of the antiintegrin b1 polyclonal antibody 6-236 (47). This antibody blocked the adhesion of N1E-115/pcDNA3- and pcDNA3/DR-

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

Fig. 8. Adhesion of pcDNA3 (empty vector) control transfectants of murine N1E-115 neuroblastoma cells and pcDNA3/DR-nm23 transfectants to various substrates. A) Bar graphs show the number of adherent cells/cm2. Error bars represent 95% confidence intervals (CIs). PolyLy 4 polylysine; LM 4 laminin; FN 4 fibronectin; Coll IV 4 collagen type IV; and Coll I 4 collagen type I. See ‘‘Materials and Methods’’ section for details. B) Morphologic evidence of a more differentiated phenotype of the pcDNA3/DR-nm23 transfectants on collagen substrates. Microphotographs were made by use of Zeiss Axioskop microscope (Zeiss, Oberkochen, Germany) (×40 objective). C) Bar graphs show the inhibition of cell adhesion to fibronectin, collagen type IV, and collagen type I mediated by the anti-mouse integrin b1 polyclonal antibody 6-236. Cells transfected with pcDNA3 or pcDNA3/DR-nm23 were incubated at 4 °C for 15 minutes in the presence of serial dilutions of the anti-integrin antibody. They were then plated for 1 hour at 37 °C on tissue culture plates coated with fibronectin, collagen type IV, or collagen type I. Subsequently, the cells were fixed and stained, and the adherent cells were counted. The percentage of inhibition was calculated on the basis of 0% inhibition with no added antibody. Error bars represent 95% CIs. A statistically significant difference in the inhibition of adhesion to collagen type Icoated plates was observed between the pcDNA3 and the pcDNA3/DR-nm23 transfectants at the higher antibody dilution tested (1:50) (two-sided P 4 .0059). See ‘‘Materials and Methods’’ section for more details.

ARTICLES 1307

nm23-transfected cells to fibronectin with similar efficiency (50% inhibition at the 1:50 dilution, 35% inhibition at the 1:100 dilution, and 15% inhibition at 1:1000 dilution; Fig. 8, C). Adhesion of both transfectants to collagen type IV-coated plates was also inhibited similarly, but the effect was lower than that obtained on fibronectin-coated plates (35% inhibition at the 1:50 dilution, Fig. 8, C). Consistent with the expectation that the enhanced expression of integrin b1 (and the increased levels of a1b1 and a2b1 heterodimers) in pcDNA3/DR-nm23-transfected cells is essential for the increased adhesion of these cells to collagen type I-coated plates, the anti-integrin b1 antibody 6-236 inhibited the adhesion of DR-nm23-expressing cells at the 1:50 dilution to a significantly greater extent than was observed for N1E-115/pcDNA3-transfected cells (52.70% [95% CI 4 ±5.13] inhibition versus 29.12% [95% CI 4 ±6.52] inhibition) (P 4 .0059; 95% CI for the difference between the groups 4 15.28– 31.87; n 4 3).

Cellular differentiation is accomplished by the loss of proliferative activity and the acquisition of specialized functions, which are often manifested in the context of a multicellular tissue organization. The proliferative activity of normal cells is also restrained by their interaction with neighboring cells, in contrast with tumor cells, which escape the proliferation inhibition associated with cell–cell communication in multicellular tissues. Therefore, it is important to understand the mechanisms underlying the inability of tumor cells to differentiate and, conversely, how differentiation-inducing agents can be used to restore a more normal phenotype to these cells (48). Neuroblastoma, the most common pediatric solid tumor, often retains the differentiative and developmental potential characteristics of primitive neural crest cells, making its clinical outcome extremely variable. Regardless of the presence of genetic abnormalities, such as N-myc amplification and chromosome 1p deletion, poorly differentiated neuroblastomas have an unfavorable prognosis; meanwhile, in localized and stage IVS (localized primary tumor with dissemination confined to liver, skin, or bone marrow) disease, large tumor masses can spontaneously regress if the cells enter a differentiative pathway (49). Unfortunately, the molecular mechanisms responsible for this crucial cellular decision are largely unknown. The nm23-H1 and nm23-H2 genes, initially isolated and characterized as metastasis suppressors, have also been associated with differentiation in numerous studies. Since both processes are critical for neuroblastoma outcome, the assessment of nm23 expression levels and the ability to alter nm23 activity might have practical implications at the prognostic level and, eventually, for gene-directed therapy. The expression of nm23H1 and nm23-H2 is constant during epithelial-Schwann-like and neuronal differentiation of human SK-N-SH and LAN-5 neuroblastoma cells, respectively. However, the expression of DRnm23 increases during the differentiation of both cell lines, suggesting that DR-nm23 is involved in determining and/or maintaining a differentiated phenotype. To provide direct evidence for this hypothesis, human (SK-N-SH) and murine (N1E115) neuroblastoma cells were transfected with a full-length DR-nm23 cDNA driven by a constitutively active CMV pro1308 ARTICLES

References (1) Abemayor E, Sidell N. Human neuroblastoma cell lines as models for the in vitro study of neoplastic and neuronal cell differentiation. Environ Health Perspect 1989;80:3–15.

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

Discussion

moter (i.e., with the vector pcDNA3/DR-nm23) and assessed for differentiation potential. DR-nm23 overexpression was associated with morphologic and biochemical changes typical of the differentiation program of each cell line. In SK-N-SH cells, an analysis of cellular modifications of the extracellular matrix (collagen type IV deposition), of the components of the cellular membrane that serve as collagen receptors (a1b1 and a2b1 integrin complexes) (46,50,51), and of the cytoskeleton (vimentin) clearly indicates an advanced epithelial-Schwann-like differentiative state similar to that characteristic of chemically induced differentiation in vitro. In agreement with these biochemical changes, the proliferative activity of pcDNA3/DR-nm23transfected cells was markedly decreased. In N1E-115 cells, morphologic features (neurite outgrowth), cytoskeleton reorganization (decrease in b-tubulin expression), and altered expression of integrins a1, a2, and b1 confirm the ability of DR-nm23 to induce differentiation and to cause changes in cellular adhesion properties, as indicated by an increased ability of transfected cells to bind collagen components of the extracellular matrix. Since DR-nm23 overexpression induces both Schwannlike and neuronal differentiation, suggesting the involvement of a common mechanism(s) that leads to distinct differentiation states, an intriguing explanation could be the requirement for integrin b1 up-regulation (30) in initiating neuroblastoma differentiation. DR-nm23 might enhance integrin b1 levels through its involvement in regulatory pathways leading to integrin b1 biosynthesis, or it may act as an inhibitor of the integrin b1 internalization/degradation pathway that occurs when adhesion to the extracellular matrix substrate is blocked (52). Because integrin b1 accumulates in the cytoplasm as part of a pre-existing intracellular pool and is normally in excess of the a1 and a2 subunits (37,53), the much larger increase in integrin b1 expression observed in pcDNA3/DR-nm23-transfected cells is not inconsistent with the moderately augmented levels of integrins a1 and a2. The slight enhancement of integrins a1 and a2 in DR-nm23 transfectants might also be caused by DR-nm23-dependent positive regulation or by inhibition of the internalization/degradation pathway of the a subunits (53). Regardless of which mechanisms are responsible for the DR-nm23-induced modulation of integrin expression, such changes correlate with the chemically induced differentiation of neuroblastoma cells (30,37). Of interest, down-regulation (i.e., a decrease) of integrin b1 expression related to N-myc overexpression was observed in neuroblastoma cell lines with higher metastatic propensities in athymic nude mice (54). In conclusion, our findings demonstrate that DR-nm23 expression is relevant to neuroblastoma proliferation and differentiation. Since the clinical outcome of neuroblastoma strongly depends on the ability of the tumor cells to acquire a differentiated phenotype, DR-nm23 could be a useful tool for assessing prognosis and for eventual gene-directed therapeutic approaches. Moreover, in light of recent data on the involvement of integrin b1 in cell motility (55), our results do not rule out the possibility that DR-nm23 is also linked to this process.

(23) Postel EH, Berberich SJ, Flint SJ, Ferrone CA. Human c-myc transcription factor PuF identified as nm23-H2 nucleoside diphosphate kinase, a candidate suppressor of tumor metastasis. Science 1993;261:478–80. (24) Postel EH, Weiss VH, Beneken J, Kirtane A. Mutational analysis of NM23H2/NDP kinase identifies the structural domains critical to recognition of a c-myc regulatory element. Proc Natl Acad Sci U S A 1996;93: 6892–7. (25) Venturelli D, Martinez R, Melotti P, Casella I, Peschle C, Cucco C, et al. Overexpression of DR-nm23, a protein encoded by a member of the nm23 gene family, inhibits granulocyte differentiation and induces apoptosis in 32Dcl3 myeloid cells. Proc Natl Acad Sci U S A 1995;92: 7435–9. (26) Claudio PP, Howard CM, Baldi A, De Luca, A, Fu Y, Condorelli G, et al. P130/pRb2 has growth suppressive properties similar to yet distinctive from those retinoblastoma family members pRb and p107. Cancer Res 1994;54:5556–60. (27) Seeger RC, Danon YL, Rayner SA, Hoover F. Definition of a Thy-1 determinant on human neuroblastoma, glioma, sarcoma, and teratoma cells with a monoclonal antibody. J Immunol 1982;128:983–9. (28) Biedler JL, Helson L, Spengler BA. Morphology and growth, tumorigenicity, and cytogenetics of human neuroblastoma cells in continuous culture. Cancer Res 1973;33:2643–52. (29) Kimhi Y, Palfrey C, Spector I, Barak Y, Littauer UZ. Maturation of neuroblastoma cells in the presence of dimethylsulfoxide. Proc Natl Acad Sci U S A 1976;73:462–6. (30) Rozzo C, Ratti P, Ponzoni M, Cornaglia-Ferraris P. Modulation of alpha 1 beta 1, alpha 2 beta 1, and alpha 3 beta 1 integrin heterodimers during human neuroblastoma cell differentiation. FEBS Lett 1993; 332:263–7. (31) Baldi A, De Luca A, Claudio PP, Baldi F, Giordano GG, Tommasino M, et al. The RB2/p130 gene product is a nuclear protein whose phosphorylation is cell cycle regulated. J Cell Biochem 1995;59:402–8. (32) Tomaselli KJ, Damsky CH, Reichardt LF. Purification and characterization of mammalian integrins expressed by a rat neuronal cell line (PC12): evidence that they function as alpha/beta heterodimeric receptors for laminin and type IV collagen. J Cell Biol 1988;107:1241–52. (33) Tarone G, Russo MA, Hirsch E, Odorisio T, Altruda F, Silengo L, et al. Expression of b1 integrin complexes on the surface of unfertilized mouse oocyte. Development 1993;117:1369–75. (34) Defilippi P, Silengo L, Tarone G. Alpha 6 beta 1 integrin (laminin receptor) is down-regulated by tumor necrosis factor a and interleukin-1 b in human endothelial cells. J Biol Chem 1992;267:18303–7. (35) Gunning P, Ponte P, Okayama H, Engel J, Blau H, Kedes L. Isolation and characterization of full-length cDNA clones for human a-, b-, and g-actin mRNAs: skeletal but not cytoplasmic actins have an aminoterminal cysteine that is subsequently removed. Mol Cell Biol 1983;3: 787–95. (36) Alonso S, Minty A, Bourlet Y, Buckingham M. Comparison of three actincoding sequences in the mouse: evolutionary relationship between the actin genes of warm-blooded vertebrates. J Mol Evol 1986;23:11–22. (37) Rossino P, Defilippi P, Silengo L, Tarone G. Up-regulation of the integrin a1/b1 in human neuroblastoma cells differentiated by retinoic acid: correlation with increased neurite response to laminin. Cell Regul 1991;2: 1021–33. (38) Miner JH, Wold BJ. c-myc inhibition of Myod and myogenin-initiated myogenic differentiation. Mol Cell Biol 1991;11:2842–51. (39) Medrano EE, Yang F, Boissy R, Farooqui J, Shah V, Matsumoto K, et al. Terminal differentiation and senescence in the human melanocyte: repression of tyrosine-phosphorylation of the extracellular signal-regulated kinase 2 selectively defines the two phenotypes. Mol Cell Biol 1994;5: 497–509. (40) Raschella G, Negroni A, Sala A, Pucci A, Romeo A, Calabretta B. Requirement of B-myb function for survival and differentiative potential of human neuroblastoma cells. J Biol Chem 1995;270:8540–5. (41) Piacentini M, Raschella G, Calabretta B, Melino G. c-myb downregulation is associated with apoptosis in human neuroblastoma cells. Cell Death Differ 1994;1:85–92. (42) Kranenburg O, Scharnhorst V, Van der Eb AJ, Zantema A. Inhibition of cyclin-dependent kinase activity triggers neuronal differentiation of mouse neuroblastoma cells. J Cell Biol 1995;131:227–34.

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

ARTICLES 1309

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

(2) Tsokos M, Scarpa S, Ross RA, Triche TJ. Differentiation of human neuroblastoma recapitulates neural crest development. Study of morphology, neurotransmitter enzymes, and extracellular matrix proteins. Am J Pathol 1987;128:484–96. (3) Brodeur GM, Green AA, Hayes FA, Williams KJ, Williams DL, Tsiatis AA. Cytogenetic features of human neuroblastomas and cell lines. Cancer Res 1981;41(11 Pt 1):4678–86. (4) Seeger RC, Brodeur GM, Sather H, Dalton A, Siegel SE, Wong KY, et al. Association of multiple copies of the N-myc oncogene with rapid progression of neuroblastomas. N Engl J Med 1985;313:1111–6. (5) Caron H, van Sluis P, de Kraker J, Bokkerink M, Egeler M, Laureys R, et al. Allelic loss of chromosome 1p as a predictor of unfavorable outcome in patients with neuroblastoma. N Engl J Med 1996;334:225–30. (6) Nakagawara A, Arima-Nakagawara M, Scavarda NJ, Azar CG, Cantor AB, Brodeur GM. Association between high levels of expression of the TRK gene and favorable outcome in human neuroblastoma. N Engl J Med 1993; 328:847–54. (7) Steeg PS, Bevilacqua G, Kopper L, Thorgeirsson UP, Talmadge JE, Liotta LA, et al. Evidence for a novel gene associated with low tumor metastatic potential. J Natl Cancer Inst 1988;80:200–4. (8) Stahl JA, Leone A, Rosengard AM, Porter L, King CR, Steeg PS. Identification of a second human nm23 gene, nm23-H2. Cancer Res 1991;51: 445–9. (9) Backer JM, Mendola CE, Kovesdi I, Fairhurst JL, O’Hara B, Eddy RL Jr, et al. Chromosomal localization and nucleoside diphosphate kinase activity of human metastasis-suppressor genes NM23-1 and NM23-2. Oncogene 1993;8:497–502. (10) Florenes VA, Aamdal S, Myklebost O, Maelandsmo GM, Bruland OS, Fodstad O. Levels of nm23 messenger RNA in metastatic malignant melanomas: inverse correlation to disease progression. Cancer Res 1992;52: 6088–91. (11) Nakayama T, Ohtsuru A, Nakao K, Shima M, Nakata K, Watanabe K, et al. Expression in human hepatocellular carcinoma of nucleoside diphosphate kinase, a homologue of the nm23 gene product. J Natl Cancer Inst 1992;84:1349–54. (12) Hennessy C, Henry JA, May FE, Westley BR, Angus B, Lennard TW. Expression of the antimetastatic gene nm23 in human breast cancer: an association with good prognosis. J Natl Cancer Inst 1991;83:281–5. (13) Igawa M, Rukstalis DB, Tanabe T, Chodak GW. High levels of nm23 expression are related to cell proliferation in human prostate cancer. Cancer Res 1994;54:1313–8. (14) Engel M, Theisinger B, Seib T, Seitz G, Huwer H, Zang KD, et al. High levels of nm23-H1 and nm23-H2 messenger RNA in human squamous-cell lung carcinoma are associated with poor differentiation and advanced tumor stages. Int J Cancer 1993;55:375–9. (15) Leone A, Seeger RC, Hong CM, Hu YY, Arboleda MJ, Brodeur GM, et al. Evidence for nm23 RNA overexpression, DNA amplification and mutation in aggressive childhood neuroblastomas. Oncogene 1993;8:855–65. (16) Hailat N, Keim DR, Melhem RF, Zhu XX, Eckerskorn C, Brodeur GM, et al. High levels of p19/nm23 protein in neuroblastoma are associated with advanced stage disease and with N-myc gene amplification. J Clin Invest 1991;88:341–5. (17) Chang CL, Zhu XX, Thoraval DH, Ungar D, Rawwas J, Hora N, et al. nm23-H1 mutation in neuroblastoma [letter]. Nature 1994;370:335–6. (18) Chang CL, Strahler JR, Thoraval DH, Qian MG, Hinderer R, Hanash SM. A nucleoside diphosphate kinase A (nm23-H1) serine 120—glycine substitution in advanced stage neuroblastoma affects enzyme stability and alters protein–protein interaction. Oncogene 1996;12:659–67. (19) Timmons L, Hersperger E, Woodhouse E, Xu J, Liu LZ, Shearn A. The expression of the Drosophila awd gene during normal development and in neoplastic brain tumors caused by Igl mutations. Dev Biol 1993;158: 364–79. (20) Lakso M, Steeg PS, Westphal H. Embryonic expression of nm23 during mouse organogenesis. Cell Growth Differ 1992;3:873–9. (21) Howlett AR, Petersen OW, Steeg PS, Bissell MJ. A novel function for the nm23-H1 gene: overexpression in human breast carcinoma cells leads to the formation of basement membrane and growth arrest. J Natl Cancer Inst 1994;86:1838–44. (22) Gervasi F, D’Agnano I, Vossio S, Zupi G, Sacchi A, Lombardi D. nm23 influences proliferation and differentiation of PC12 cells in response to nerve growth factor. Cell Growth Differ 1996;7:1689–95.

1310 ARTICLES

(53) Heino J, Ignotz RA, Hemler ME, Crouse C, Massague J. Regulation of cell adhesion receptors by transforming growth factor-b. Concomitant regulation of integrins that share a common b1 subunit. J Biol Chem 1989;264: 380–8. (54) Judware R, Lechner R, Culp LA. Inverse expressions of the N-myc oncogene and b 1 integrin in human neuroblastoma: relationships to disease progression in a nude mouse model system. Clin Exp Metastasis 1995;13: 123–33. (55) Felsenfed DP, Choquet D, Sheetz MP. Ligand binding regulates the directed movement of b1 integrins on fibroblasts. Nature 1996;383: 438–40.

Notes Supported in part by the Associazione Italiana Ricerca sul Cancro; Associazione Italiana per la Lotta al Neuroblastoma, Genoa, Italy; and a grant from the American Cancer Society. We thank Adelma Di Stefano and Vincenzo Cesi for technical assistance, Drs. Rita Falcioni and Carlo Gaetano (Istituto Regina Elena, Rome, Italy) for help in the preliminary experiments on integrin determination, Professor Ivan De Curtis (Ist. San Raffaele, Milan, Italy) for the rabbit integrin b1 polyclonal antibody, Professor Guido Tarone (University of Turin, Italy) for the rabbit anti-carboxycytoplasmic region of both human integrin a1 and integrin a2 polyclonal antibodies, and Professor Paolo Bernardi (University of Padua, Italy) for the rabbit anti-murine integrin b1 polyclonal antibody. We also thank Professor G. Tarone, P. Bernardi, and Professor Daniela Lombardi (University of L’Aquila, Italy) for helpful comments and suggestions. Manuscript received February 12, 1997; revised June 13, 1997; accepted July 3, 1997.

Journal of the National Cancer Institute, Vol. 89, No. 17, September 3, 1997

Downloaded from http://jnci.oxfordjournals.org/ by guest on February 21, 2013

(43) Portier MM, Croizat B, Gros F. A sequence of changes in cytoskeletal components during neuroblastoma differentiation. FEBS Lett 1982;146: 283–8. (44) Lin CQ, Bissel MJ. Multi-faceted regulation of cell differentiation by extracellular matrix. FASEB J 1993;7:737–43. (45) Saelman EU, Keely PJ, Santoro SA. Loss of MDCK cell alpha 2 beta 1 integrin expression results in reduced cyst formation, failure of hepatocyte growth factor/scatter factor-induced branching morphogenesis, and increased apoptosis. J Cell Sci 1995;108:3531–40. (46) Keely PJ, Fong AM, Zutter MM, Santoro SA. Alteration of collagendependent adhesion, motility, and morphogenesis by the expression of antisense alpha 2 integrin mRNA in mammary cells. J Cell Sci 1995;108(Pt 2):595–607. (47) Bernardi P, Patel VP, Lodish HF. Lymphoid precursor cells adhere to two different sites on fibronectin. J Cell Biol 1987;105:489–98. (48) Lotan R, Francis GE, Freeman CS, Waxman S. Differentiation therapy. Cancer Res 1990;50:3453–64. (49) Kogner P, Barbany G, Dominici C, Castello M, Raschella G, Persson H. Co-expression of messenger RNA for TRK protooncogene and low affinity nerve growth factor receptor in neuroblastoma with favorable prognosis. Cancer Res 1993;53:2044-50. (50) Takada Y, Wayner EA, Carter WG, Hemler ME. Extracellular matrix receptors, ECMRII and ECMRI, for collagen and fibronectin correspond to VLA-2 and VLA-3 in the VLA family of heterodimers. J Cell Biochem 1988;37:385–93. (51) Bhattacharyya-Pakrasi M, Santoro SA. a2b1 Integrin-dependent recognition of collagena1(I) c-terminal propeptide sequences [abstract]. Mol Biol Cell 1996;7:426a. (52) Dalton SL, Scharf E, Briesewitz R, Marcantonio EE, Assoian RK. Cell adhesion to extracellular matrix regulates the life cycle of integrins. Mol Biol Cell 1995;6:1781–91.