In our experiments, we used six human neuroblastoma cell lines obtained from cell collections, i.e., IMR-32 (ATTC, LCG Standards, Lomianki, Poland; cat no. CCL-127), LA-N-1 (ECACC, Sigma-Aldrich, Poznan, Poland; cat. no. 06041201), CHP-134 (ECACC, cat. no. 06122002), LA-N-5 (cat. no. ACC 673; Leibniz-Institut DSMZ-Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH), Kelly (cat. no. ACC 355; DSMZ) and SK-N-SH (cat. no. ATCC HTB11). Cells were cultured in 5% CO₂ atmosphere in media as previously described in detail (10).

The hybridoma cell line producing the 14G2a mAb, which is a mouse IgG2a that binds to GD2, was kindly provided by Dr R.A. Reisfeld (Scripps Research Institute, La Jolla, CA, USA). Cell culture and the antibody purification from fetal bovine serum (FBS) free-cell culture supernatant using HiTrap™ Protein G columns (cat. no. 17-0404-03; GE Healthcare Bio-Sciences AB, Uppsala, Sweden) were previously described in detail (10).

Analyses of integrin content of cells were performed with Alpha/Beta Integrin-Mediated Cell Adhesion Array Combo kit (cat. no. ECM532; Chemicon International, Merck Sp. z o.o., Warszawa, Poland) with protocol optimized based on the manufacturer's suggestions. Prior to the tests, neuroblastoma cells were detached from culture plastic vessels with 2 mM EDTA in phosphate-buffered saline (PBS, cat. no. PBS404.2; BioShop^® Canada, Lab Empire, Rzeszów, Poland), at room temperature and then centrifuged (for 10 min, at 4°C, at 315 × g). The cell pellets were resuspended in assay buffer provided by the kit manufacturer and cells were counted using the trypan blue exclusion method. Between 0.1×10⁶ and 0.2×10⁶ of cells in 100 µl of assay buffer (depending on the cell line tested) were added per well, followed by incubation for 2 h at 37°C in an incubator with 5% of CO₂. Single wells with antibodies binding respective integrin subunits or complexes (also single negative control wells, NC) were used per experiment. Then, the cells were processed according to the manufacturer's instructions. Signals of absorbance at 560 nm were collected with an Infinite M200 reader (Tecan Schweiz AG, Männedorf, Switzerland). To normalize signals between experiments, we calculated ratios between signals from wells coated with, respective, anti-integrin antibodies and signals of negative control wells (NC). Two experiments are shown for IMR-32, CHP-134, LA-N-5, Kelly (marked as ex-1, ex-2 in Fig. 4). One experiment was performed for LA-N-1 cells.

Cells were detached using 2 mM EDTA in PBS at room temperature, and 2% FBS in PBS (FBS, cat. no. 10270-106; Life Technologies, Warszawa, Poland) was used for cell staining and washing. Following antibodies were applied: the 14G2a mAb for GD2 staining, while PK136 mAb (mouse IgG2a) (10) was used as an isotype control; FITC-conjugated anti-CD49d (integrin α4) antibody (IgG1, cat. no. 304315; BioLegend, BioCourse.pl, Katowice, Poland), while FITC-conjugated mouse IgG1 mAb (cat. no. 400109; BioLegend) was used as an isotype control; mouse anti-human CD29 (integrin β1) mAb (IgG1, cat. no. 303001; BioLegend), mouse anti-human integrin αVβ3 mAb (IgG1, cat. no. MAB1976Z; EMD Millipore Corp.), mouse anti-human integrin αVβ5 mAb (IgG1, cat. no. MAB1961Z; EMD Millipore), while purified mouse IgG1 was used as an isotype control (cat. no. 554721; BD Biosciences, Diag Med, Warszawa, Poland). Were appropriate, mouse Ig-specific FITC-conjugated goat F(ab')₂ fragments (cat. no. 55526; Cappel, MP Biomedicals, LLC, Warszawa, Poland) were used to detect binding of the primary antibodies. Cells were analyzed using flow cytometer BD™ LSR II with BD FACSDiva software (BD Biosciences, Warszawa, Poland). Single cells were analyzed. Dead cells were excluded with 7-AAD viability staining solution (cat. no. 420403; BioLegend). For GD2 analyses, histograms showing levels of FITC fluorescence were analyzed. Pools of the positively stained cells were gated based on the signals from control cell samples, in which the PK136 mAb was used instead of the 14G2a mAb for staining (Table I). For measurements of integrin expression, scatter plots (showing levels of 7-AAD fluorescence vs. levels of FITC fluorescence) were analyzed. Gates to measure percentage of positively stained cells of the analyzed pools were set based on signals from cells stained with respective isotype control antibodies (Table II).

We used several types of 96- or 6-well plates for cell culture. The plates include: a standard cell culture plate (cat. no. 353072; BD Falcon, Immunogen Polska, Chorzów, Poland), white/clear tissue culture treated plates (cat. no. 353377; BD Falcon, Immunogen Polska), human vitronectin coated plates (BSA, bovine serum albumin-blocked by the manufacturer, with only BSA-blocked wells included on the same plate, cat. no. CWP003; R&D Systems, Biokom Sp.j, Janki, Poland), Nunclon™ Sphera™ flat bottom cellware (cat. no. 174927; Thermo Fisher Scientific, Biokom), collagen type IV plates (cat. no. 354429; BD BioCoat™, Immunogen Polska); and human fibronectin cellware (cat. no. 354409; Corning BioCoat™, Immunogen Polska, cat. no. 354402; Corning BioCoat™, Diag Med). For some attachment assays and for western blots, cells were grown on wells coated with human vitronectin (cat. no. G5381; Promega Gmbh, Mannheim, Germany) and blocked with BSA (cat. no. A9418; Sigma-Aldrich) based on the manufacturer's protocol.

#94-12-F/W-AAEGD peptide [GD2 MIMIC (RCNPNMEPPRCWAAEGD) (28,29)] and the scrambled peptide [SCR CONTROL (NPERCNRWMPCPAEADG)] were obtained from GenScript Biotech Corp. (Piscataway, NY, USA; dissolved in water). We used four integrin inhibitors: obtustatin, an inhibitor of α1β1 (dissolved in water, cat. no. 4664; Tocris Bioscience, Biokom); BIO 1211, an inhibitor of active α4β1 (dissolved in 20X PBS, cat. no. 3910, Tocris Bioscience, Biokom); cilengitide, an inhibitor of αVβ3 and αVβ5 (dissolved in DMSO, cat. no. S7077; Selleck Chemicals LLC, STI, Poznań, Poland); and SB273005, an inhibitor of αVβ3 and αVβ5 (dissolved in DMSO, cat. no. S7540; Selleck Chemicals LLC, STI).

Prior to the tests, neuroblastoma cells were detached from culture plastic with 2 mM EDTA in PBS, at room temperature, and then centrifuged (for 10 min, at 4°C, at 139 × g). The cell pellets were resuspended in respective complete media and counted using the trypan blue exclusion method. The seeding density for IMR-32 and LA-N-1 was 2×10⁴ cells/100 µl/well and for CHP-134, LA-N-5, Kelly, SK-N-SH was 5×10³ cells/100 µl/well. For 14G2a-treatment of cells, the antibodies in PBS were added (control samples were treated with PBS alone when necessary). All samples with cells were incubated on ice for 1 h with mixing after every 15 min of incubation, before transfer of 100 µl to wells of respective plates.

Where appropriate, inhibitors (see above) were added to cells, after incubation with the 14G2a mAb (controls treated with solvents of inhibitors were included where necessary). Obtustatin was tested at the concentrations of 0.012, 0.06, 0.3 and 1.5 µM; BIO 1211 was tested at the concentrations of 1, 2.5 and 5 µM; cilengitide was tested at the concentrations of 3 and 10 µM; and SB273005 was tested at the concentrations 0.3, 1, 3, 10 and 30 µM.

For testing of effects of #94-12-F/W-AAEGD or scrambled control peptides on cytotoxicity of the 14G2a mAb, the seeding density for IMR-32 and CHP-134 cell lines was 2×10⁴ cells/100 µl/well to better fit microscopic examinations. The peptides were added to cells after incubation with antibodies on ice (from 10 mM stock solution) to the final concentrations of 0.23 or 0.114 mM. Separate controls cells (when necessary) were also treated with peptides, and/or diluent of peptides, and/or PBS (diluent for 14G2a) to check for any effects on cellular ATP levels.

After the above treatments, cells were incubated for 72 h at 37°C in an incubator with 5% of CO₂. On the third day of incubation microscopic pictures were taken of some cell cultures using the Leica microscope DM IL LED Fluo equipped with the DFC450C camera, using 10x-(506263, HI PLAN I, 10x/0.22) or ×20-magnification objectives (506264, HI PLAN I, 20x/0.30) and collected using the Leica Application Suite Version 4.3.0 (Leica Microsystems GMBH, Wetzlar, Germany). Cellular survival (based on measurements of ATP levels) or caspase-3/-7 activities were determined as described below.

For attachment assays, the seeding density for all cell lines was 2×10⁴ cells/100 µl/well. Cells were treated with integrin inhibitors, 14G2a, or their combinations (see above), and then incubated for 2 h at 37°C in an incubator with 5% of CO₂. Control cells treated with diluents were included (were necessary). Next, unattached cells were removed by inversion of plates, while the remaining cells were analyzed using cellular ATP measurements (see below). For western blot analysis, LA-N-5, Kelly and SK-N-SH cells were treated with 3 µM SB273005 for 6 and 24 h, IMR-32 cells were treated with 5 µM BIO 1211 for 6 and 24 h. Control cells were treated with diluents.

We used an ATP assay as a sensitive method to estimate viable cell number in attachment experiments, and to determine effects of tested agents (the 14G2a mAb, the peptides and the integrin inhibitors) on neuroblastoma cell viability (proliferation and cytotoxicity). Cellular ATP contents were analyzed using ATPlite, luminescence ATP detection assay system (cat. no. 6016947; Perkin-Elmer, Warszawa, Poland) based on the protocol of the manufacturer to measure cell survival in experiments. For measurements of activities of apoptosis-executing caspase-3/-7, we used Caspase-Glo^® 3/7 assay (cat. no. G8091; Promega) according to the manufacturer's protocol. Where necessary, equal volumes of cell lysates were transferred to white plates and luminescence signals were collected using the Infinite M200 reader.

Cells were collected using trypsin-EDTA (cat. no. T3924; Sigma-Aldrich) and centrifuged for 5 min (2,500 × g at 4°C). Cell pellets were washed twice with ice-cold PBS (with centrifugation steps to recover cells as pellet). Next, cells were lysed with gentle mixing by inversion (for 45 min, at 4°C) using buffer containing: 25 mM Tris-HCl (pH 7.6, cat. no. TRS001.1; BioShop), 10 mM NaCl (cat. no. 794121116; Avantor Performance Materials Poland S.A., POCH, Gliwice, Poland), 1% NP-40 (cat. no. NON999.500; BioShop), 1% sodium deoxycholate (cat. no. D6570; Sigma-Aldrich), 0.1% SDS (cat. no. L4390; Sigma-Aldrich), Halt Phosphatase Inhibitor Cocktail (cat. no. 78420; Thermo Fisher Scientific, ALAB Laboratoria, Warszawa, Poland), and Complete Protease Inhibitor Cocktail (cat. no. 11697498001; Roche, Warszawa, Poland). Then, samples were centrifuged 14,000 × g, at 4°C, for 15 min. Samples were resolved using SDS-PAGE (12% gels) as described in detail (10). Western blot analyses were performed with application of the following antibodies (all from Cell Signaling Technology-Lab-JOT Ltd., Warszawa, Poland): anti-FAK (cat. no. 13009, dilution 1:1,00), anti-phospho-FAK (Tyr397) (cat. no. 8556, dilution 1:1,000), anti-α-tubulin (cat. no. 2125, dilution 1,000–2,000), anti-rabbit IgG, HRP-linked antibody (cat. no. 7074, 1:2,000). Chemiluminescent signals were developed with the Immobilon Western HRP Substrate (cat. no. WBKLS0500; Millipore, Warszawa, Poland) and collected using the MicroChemi System (DNR Bio Imaging Systems, ALAB Laboratoria, Warszawa, Poland).

Multivariate analysis was performed with one-way repeated measures ANOVA with a Greenhouse-Geisser correction. Then, we performed post hoc pairwise comparisons by Dunnett's test at the significance levels set at 0.05, 0.01, 0.001 were applicable. In multivariate analyses data were analyzed without normalization (30). Data shown on graphs were normalized to control set as 1 for clarity. Data on graphs (except for Fig. 4) are presented as means of experiments (n, numbers of experiments performed were from 2 to 5, performed in duplicate, triplicate or quadruplicate, as described in the figure legends) ± SEM (a standard error of the mean). Data in Table II are presented as means of two experiments ± SEM. Some normalized samples were analyzed using two-sample t-tests, comparing, e.g., values of control cells with values of treated cells to measure statistical significance [p-values: P<0.05, P<0.01, P<0.001]. Statistical analyses were performed with the OriginPro 2017 (b9.4.0.220; OriginLab Corp., Northampton, MA, USA; one-way repeated measures ANOVA) and Microsoft Excell (Microsoft Corp., Way Redmond, WA, USA; t-tests). Microsoft software was used for the preparation of the graphs.

Phage-displayed peptide libraries can be screened with monoclonal antibodies with various specificities and yield mimicking peptides binding to the screening molecules and inhibiting interactions with the cognate antigens. Previously, we reported optimization and characterization of binding to the 14G2a mAb of cyclic peptides mimicking GD2 ganglioside (28,29). In the present study, we show that peptide #94-12F/W-AAEGD (RCNPNMEPPRCWAAEGD, marked as GD2 MIMIC in Fig. 1), used at the concentration 0.23 mM (also 0.114 mM concentration; data not shown), can abolish the cytotoxic effects of the 14G2a mAb as compared to the control scrambled sequence (marked as SCR PEPTIDE, NPERCNRWMPCPAEADG). IMR-32 and CHP-134 cells grown in the presence of the 14G2a mAb exhibited morphological changes such as cell aggregation. The effects were abolished by the GD2 MIMIC (Fig. 1A, images taken 72 h after incubation). Also, for both cell lines, the ATP levels in the presence of the GD2 MIMIC and the 14G2a mAb were statistically significantly higher than in cells treated with the SCR PEPTIDE and the mAb (Fig. 1B; GD2 MIMIC + 14G2a vs. SCR PEPTIDE + 14G2a by Dunnett's test with the significance level set at 0.001; and Fig. 1C, GD2 MIMIC + 14G2a vs. SCR PEPTIDE + 14G2a by Dunnett's test with the significance level set at 0.05). There was no statistically significant difference between ATP cell levels in samples treated with 14G2a or 14G2a and SCR peptide (Dunnett's test), thus, the scrambled control peptide did not affect sensitivity of cells to the mAb. Moreover, there was no statistically significant difference between cells treated with SCR PEPTIDE and GD2 MIMIC (Dunnett's test). Therefore, the fact that the GD2 mimicking peptide, but not the control sequence, eliminated the cytotoxic effects of 14G2a confirms that the observed effects rely on binding of the mAb to GD2 ganglioside. In our previous study, we used the PK136 mAb as a negative control (10). Hence, the data obtained with the peptide provide additional evidence that the cytotoxic effects of 14G2a are specific and require binding between GD2 and 14G2a.

We have previously published data showing that selected neuroblastoma cell lines are sensitive to direct cytotoxicity of the anti-GD2 antibody 14G2a mAb (10,23). Thus, we measured that cellular ATP levels decreased in time- and dose-dependent manner in cultures of some neuroblastoma cells treated with the mAb (10). Additionally, for IMR-32 and LA-N-1 cell lines treatment with the mAb induced clear changes in appearance such as cell rounding, aggregation, detachment (10) (Fig. 2). For other cells, such as LA-N-5 and CHP-134 the changes were less pronounced (10) (Fig. 2). Kelly cells did not exhibit easily visible changes in their morphology and were only to a limited extent sensitive to the mAb treatment (10) (Fig. 2).

The above findings prompted us to hypothesize that modulation of attachment of cells may influence sensitivity of neuroblastoma cell lines to 14G2a. Hence, we used IMR-32, LA-N-1, LA-N-5, CHP-134 and Kelly neuroblastoma cell lines, as they are highly GD2-positive (Table I shows GD2 analyses on IMR-32, LA-N-5, CHP-134 and Kelly cells) and are to various extent sensitive to the antibody (10). We compared effects of attachment on survival of neuroblastoma cell lines, also in the presence of the mAb 14G2a (used in the concentration of 20 µg/ml). First, we used ultra-low binding plates (non-adherent conditions) and standard cell culture plates (adherent conditions) to compare effects of such culture plastic vessels on cell morphology by microscopic observation and cell survival by measurements of cellular ATP. We report that all tested cell lines formed large aggregates on non-adherent plates (Fig. 2, right panel).

Also, we compared levels of cellular ATP in cultures after 72 h of incubation, with or without treatment with the mAb, on both types of plates (normalized to control cells treated with PBS and grown in adherent conditions, set as 1; Fig. 3A–E). We report a statistically significant decrease in cellular ATP of cell cultures grown on non-adherent plates as compared to standard culture plates. Thus, the measured values were from 0.62±0.03 (for Kelly cells) to 0.75±0.045 (for LA-N-5 cells) of the values of cells grown on the standard plates (P<0.05 for CHP-134, LA-N-1, LA-N-5; P<0.01 for IMR-32 and Kelly). For the cells treated with 14G2a (for the great majority of the tests), we measured lower ATP levels in cultures grown on non-adherent surfaces as compared to standard plates (Fig. 3A–E). Out of five cell lines tested, for CHP-134 cells the ATP levels decreased approximately by additional 10% on non-adherent plates and were statistically meaningful for the 20 µg/ml concentrations of the 14G2a mAb (P<0.05). For Kelly cells, the levels of ATP were statistically significantly lower for the 20 µg/ml concentration of the 14G2a mAb used (P<0.01; Fig. 3E). For IMR-32, LA-N-1 and LA-N-5 cells the results of 14G2a treatment obtained compared between adherent and non-adherent conditions in the presence of the 14G2a mAb were not statistically meaningful.

Previously, we reported that treatment with the mAb caused a rise in caspase-3/-7 activities in IMR-32 cells after 24 and 48 h of treatment (10). Here, we provide more data, showing that culturing of the IMR-32 cells for 72 h on non-adherent plates resulted in an increase of activities of caspase-3/-7 ~20%, but without statistical significance. However, for both adherent and non-adherent plates in the presence of 14G2a (20 µg/ml), the increase of activities of caspase-3/-7 was ~60 and 70%, respectively (Fig. 3F, when compared to control cells treated with PBS grown in adherent conditions, set as 1). Also, for CHP-134, LA-N-1 and LA-N-5 the highest activities of caspase-3/-7 were measured for cells cultured on low-binding plates and in the presence of the 14G2a mAb (20 µg/ml), when compared to other conditions tested, but the data did not reach statistical significance due to variation between experiments (Fig. 3G–I).

We made an attempt to preliminary characterize presence of integrins on IMR-32, CHP-134, LA-N-1, LA-N-5 and Kelly cell lines with application of an α/β integrin-mediated cell adhesion array that uses antibody-coated wells to retain cells expressing given integrin chains or their complexes (Fig. 4, data from two experiments are shown, marked as ex-1 and ex-2 on the graphs). For the analyses, we assumed that a relative signal >2 is a positive sign that a given integrin subunit or complex is present on cells. For IMR-32 and CHP-134 cells clear signals of binding of α1, α4 and β1 chains were measured. Additionally, on CHP-134 cells α2 and α3 subunits were detected. On LA-N-5 cells α1, α2, and β1 but no α4 were detected. Also, for CHP-134, LA-N-5 signals of binding to αV, αVβ3 and αVβ5 were measured. For Kelly cells due to high background measured in wells of controls it is hard to draw unequivocal conclusions, hence the line was included in the further tests (described below). Also for LA-N-1 cells, in one experiment, we measured positive signals from α1, α2, α4, αV and β1 (data not shown). Further testing was performed using flow cytometry (Table II shows percentage of positively stained cells in the analyzed pools). We showed that in concordance with the aforementioned array, β1 integrin was expressed on majority of IMR-32, CHP-134, LA-N-5 and Kelly cells. Furthermore, α4 chain was unequivocally present in pools of analyzed IMR-32 cells (i.e., on ~30% cells). On CHP-134 and Kelly cells, expression of α4 chain was detectable, yet restricted only to small percentage of cells (i.e., ~3% of CHP-134 cells and ~6.5% of Kelly cells). Lack of αVβ3 and αVβ5 heterodimers was confirmed on IMR-32 cells. The presence of the αVβ3 integrin complex was detectable on CHP-134, LA-N-5 and Kelly cell cultures, yet restricted only to minority of cells (~2%). The αVβ5 complex was unequivocally present in Kelly cell pools (~60%), while it was detectable only on minority of LA-N-5 and CHP-134 cells. Notably, in six measurements on LA-N-5 cells, percentage of cells positive for αVβ3 ranged approximately from 2 to 6%, and for αVβ5 from 2.5 to 15% (data not shown). This suggests need for further characterization of levels of integrins and relationships between their expression and cell culture conditions (such as cell time in culture, culture cell density, or specific cell banks used). Additionally, we included GD2-negative SK-N-SH cells to the analyses (Table I) and showed that α4, β1 chains and αVβ3, αVβ3 complexes are present in majority of the analyzed cells (Table II). Hence, the cell line was included in some research described in following chapters.

The above data prompted us to investigate effects of extracellular matrix proteins, i.e., fibronectin, collagen IV on changes of ATP levels induced after treatment of cells with the 14G2a mAb.

We compared cellular ATP levels of cells cultured on fibronectin- and collagen IV-coated wells to control cells grown on standard cell culture plates, after treatment with the mAb in the concentrations of 20 µg/ml (PBS treated cells were also included). The effects were subtle; hence we calculated data as means ± SEM from three experiments (data not shown). For IMR-32, CHP-134 and LA-N-1 cells, we observed an increase in levels of cellular ATP on plates coated with fibronectin or collagen IV, when compared to cells grown on uncoated plates. However, the statistically meaningful increases in ATP levels (~10%) were measured only for the LA-N-1 cells treated with 14G2a and cultured on collagen IV plates (P<0.05). The smallest changes (and without statistical significance) were measured for CHP-134 cells. Thus, we find that interaction of neuroblastoma cells with the ECM proteins can lead to lower sensitivity of some cell lines to the mAb treatment.

We decided to use two inhibitors of integrin complexes α1β1 or α4β1, i.e., obtustatin or BIO 1211, to test their effects on IMR-32, CHP-134 and Kelly cells. The inhibitors were tested alone and in combination with the 14G2a mAb (20 µg/ml) on plates coated with collagen IV or fibronectin, respectively.

Obtustatin used in the concentrations 0.012, 0.06, 0.3 and 1.5 µM did not affect ATP levels of IMR-32, CHP-134, Kelly cells (data not shown). Moreover, no change in sensitivity to treatment with 14G2a was observed when the inhibitor (applied in the above mentioned concentrations) was combined with 20 µg/ml of 14G2a (after 72 h of treatment; data not shown). This is in line with the observations that the inhibitor did not induce changes of the cell morphology and attachment of the three lines, when investigated on the third day of culture by microscopic observation. Some cell rounding was observed 2 h after obtustatin (e.g., 1.5 µM) addition to IMR-32 cells (data not shown).

BIO 1211 used in the concentrations 1, 2.5 and 5 µM affected appearance of IMR-32 cells, as visible after 72 h of culture, especially for 2.5 and 5 µM BIO 1211 (Fig. 5A for 5 µM BIO 1211), while no changes were observed for CHP-134 and Kelly cells. This was further analyzed with application of attachment assays on fibronectin-coated wells. The signals of cells remaining in wells were detected using ATP measurements. Results of multivariate analyses did not yield statistically significant P-values for the three cell lines, hence, further pairwise tests were not conducted (Fig. 5B and data not shown). For IMR-32, when mean values from three experiments were compared, the 14G2a mAb (20 µg/ml) reduced signal to 0.7±0.06, 5 µM BIO 1211 reduced signal to 0.62±0.1, and both agents reduced signal to 0.43±0.1 of control cells set as 1 (Fig. 5B). For Kelly cells, the 14G2a mAb (20 µg/ml) did not affect attachment, 5 µM BIO 1211 reduced signal to 0.93±0.4, and both agents reduced signal to 0.89±0.04, when compared to control cells (set as 1, data not shown). For CHP-134 cells, the 14G2a mAb (20 µg/ml) and 5 µM BIO 1211 did reduce signals, and both agents reduced signal to 0.86±0.05, when compared to control cells (set as 1, data not shown). Also after 72 h of culture, the inhibitor reduced cellular ATP levels of the IMR-32 cells to 0.9±0.02 (of control cells treated with diluent, set as 1), 0.84±0.03 (of control) and 0.86±0.04 (of control); when used at the concentrations of 1, 2.5 and 5 µM, respectively (Fig. 5C).

When compared to IMR-32 cells treated with mAb alone, lower levels of ATP were observed for cells treated with combinations of 1, 2.5 and 5 µM of BIO 1211 and 20 µg/ml of 14G2a (after 72 h of treatment), i.e., 0.45±0.03 (of control cells treated with diluent); 0.42±0.02 (of control cells treated with diluent); 0.44±0.03 (of control cells treated with diluent); respectively, yet the values did not reach statistical significance when compared to signals from samples treated only with 14G2a by Dunnett's test (Fig. 5D, also data not shown). For CHP-134 and Kelly cells treated with BIO 1211 in the concentrations 1, 2.5 and 5 µM the reduction levels were either not observed (e.g., CHP-134 cells treated with 1 µM) or <10% (Fig. 5C). No additional statistically meaningful effects were observed when the 14G2a mAb (20 µg/ml) was combined with 2.5 and 5 µM BIO 1211 for the CHP-134 cells [when compared to the antibody (20 µg/ml) used alone, data not shown]. For Kelly cells, in the settings where BIO 1211 (1 or 2.5 or 5 µM) and 14G2a (20 µg/ml) were used alone or the inhibitor and the mAb were used in combinations, multivariate analyses did not yield statistically significant P-values (data not shown). As a result, further pairwise tests were not conducted. Using western blot analysis, we showed that 5 µM BIO 1211, added for 6 and 24 h to IMR-32 cells, did not change levels of p-FAK (Tyr397) and total FAK (Fig. 5E).

We decided to use two inhibitors of integrin complexes αVβ3 or αVβ5, i.e., cilengitide (CGT) or SB273005 (SB), to test their effects on LA-N-5, CHP-134 and Kelly cells. The inhibitors were tested alone and in combination with the 14G2a mAb (20 µg/ml) on plates coated with vitronectin.

In cell cultures grown on vitronectin (VN), SB37005 (Fig. 6) and cilengitide (data not shown) induced changes: in shape (in Kelly cells), cell spreading (of CHP-134 cells), and aggregation (of LA-N-5 cells). Aggregation was evident for LA-N-5 cells, but for Kelly cells the cell rounding was quite subtle. CHP-134 cells tend to cover vitronectin-coated surface more evenly without SB273005 suggesting that their migration may be affected by it. Effects were observed for 3 and 10 µM concentrations of both inhibitors (Fig. 6 and data not shown). Microscopic observations have been extended with attachment assays on vitronectin-coated wells. First, we report that vitronectin-coated wells retained more cells, ~80% more for LA-N-5, ~40% more for Kelly, and approximately 30% more for CHP-134, as compared to control, BSA blocked wells (after 2 h of incubation, P<0.01 for LA-N-5, by two sample t-test, data not shown). Moreover, 3 and 10 µM SB273005 used alone or in combination with 14G2a statistically meaningfully reduced attachment of LA-N-5, CHP-134 (as compared to controls set as 1, Fig. 7C and G) after 2 h of incubation. For Kelly cells, the inhibitor used alone or combined with 14G2a reduced cell attachment by ~40% (Fig. 7E). However, multivariate analysis did not yield statistically significant P-values (data not shown). As a result, further pairwise tests were not conducted.

Also, we tested SB273005 on uncoated plates in the range of concentrations from 0.3 to 30 µM. GD2-negative SK-N-SH cells known from the literature to express αV (19) and from our cytometry data to express αVβ3 and αVβ5 were also included (Table II). When ATP levels were measured 72 h after addition of the inhibitor to cell cultures grown on standard uncoated wells, for SK-N-SH a clear cytotoxic effect was observed, as the cells were decimated for all conditions tested (Fig. 7A). The ATP levels of SH-N-SH dropped by ~70% (Dunnett's test with the significance level set at 0.001, as compared to control cells). In case of LA-N-5 cells, the ATP levels decreased ~15% (but no significant P-value was obtained from multivariate analysis). For Kelly, ATP levels decreased <10% (with statistical significance by Dunnett's test; Fig. 7A). No effects on cellular ATP levels were observed for CHP-134 incubated with SB273005 for 72 h. Additionally, we analyzed levels of phosphorylated FAK, pFAK (Tyr397), and total FAK in SK-N-SH, LA-N-5 and Kelly treated with 3 µM SB273005 for 6 and 24 h (Fig. 7B, D and F). No changes of levels of total FAK were observed. A clear drop in the level of phosphorylated FAK (Tyr397) was observed in SK-N-SH cell treated with the inhibitor for 24 h (Fig. 7B), which correlates with high toxicity of the inhibitor against the cell line.

Clear changes in cellular appearance, i.e., cell rounding, partial or complete aggregation and detachment, were observed for CHP-134, LA-N-1 and Kelly when cilengitide or SB273305 (10 and 3 µM) were combined with the 14G2a mAb (for 72 h, data not shown). Importantly the combination of cilengitide or SB273005 (10 and 3 µM) and 14G2a (20 µg/ml) resulted in lower ATP levels, when compared to cells treated with the single drug, although the effects were rather small, possibly due to restricted expression of αVβ3 in cultures of CHP-134, LA-N-5, Kelly cells, and αVβ5 in cultures of CHP-134, LA-N-5 (Fig. 8 and Table II). For all cells effects of 14G2a where statistically significant when compared to control cells (Dunnett's test, not marked on Fig. 8). For LA-N-5 the effects of the mAb were statistically significantly potentiated by 3 and 10 µM SB273005 (Fig. 8D, Dunnett's test) and for Kelly for 3 µM cilengitide and 10 µM SB273005 (Fig. 8E and F, Dunnett's test). Also, SB273005 increased caspase-3/-7 activities in CHP-134, LA-N-5, Kelly cells (Fig. 9), and the effects were statistically meaningful for the last two cell lines mentioned (Fig. 9B and C; Dunnett's test). Importantly, for LA-N-5 and Kelly cells, the effects of the mAb were statistically significantly potentiated by 3 and 10 µM SB273005 (Fig. 9B and C, Dunnett's test).