Abstract
Background/Aim: Treatment with retinoic acid (RA) often promotes neuroblastoma differentiation and growth inhibition, including the suppression of the expression of the MYCN oncogene. However, RA also targets protumoral chemokines, such as CCL2, which may contribute to the development of resistance. The present study aimed to investigate the regulation and function of CCL2 and N-Myc in RA-treated neuroblastoma cells. Materials and Methods: In Kelly or SH-SY5Y cells, viability was quantified by cell fitness assays. Expression was analyzed using quantitative PCR and the regulation of proteins using enzyme-linked immunoabsorbent assays (ELISA) or western blots. Results: In MYCN-amplified Kelly cells, endogenous CCL2 levels were significantly lower compared to MYCN non-amplified SH-SY5Y cells. Treatment with 5 μM RA increased CCL2 release in both cell lines, but reduced N-Myc levels and cell numbers in Kelly cells. Over-expression of MYCN enhanced viability in SH-SY5Y cells, but did not affect RA-induced CCL2 release, while supplementation of CCL2 in Kelly cells did not prevent RA-mediated growth reduction. Impaired N-Myc or CCL2 signaling reduced the survival of all RA-treated cells and inhibition of N-Myc also decreased CCL2 levels. However, attenuated survival signaling was not generally associated with reduced levels of N-Myc or CCL2. Co-application of RA and the growth factor receptor inhibitors cediranib or crizotinib decreased N-Myc levels only in Kelly cells, while CCL2 release was dependent on the cell type and stimulus. Conclusion: CCL2 and N-Myc promote the viability of RA-treated cells, although the levels of these mediators were not consistently correlated with cellular outcomes, especially during apoptotic signaling.
Neuroblastoma is the most common extracranial solid tumor in childhood and accounts for approximately 15% of pediatric cancer-related mortality (1, 2). Dysregulation of differentiation and cell death in sympathoadrenal neural crest cells during embryonic development promotes tumorigenesis in the adrenal medulla or sympathetic ganglia along the sympathetic chains, which can lead to metastasis in advanced stages (1, 2). The amplification of the oncogene MYCN, a hallmark of high-risk neuroblastoma, was identified in 20% of neuroblastoma cases (3) and is usually associated with poor outcome (3). Inversely correlated with MYCN expression levels is the secretion of the chemokine CCL2, which is thought to be particularly important for antitumor immunity in neuroblastoma (4, 5). However, when CCL2 is released by cancer cells, different responses have been observed. On one hand, it can stimulate host anti-tumor activity (4, 5). On the other hand, CCL2 promotes cancer cell proliferation as well as angiogenesis and is involved in the formation of an immunosuppressive tumor environment (6, 7), thereby facilitating cancer progression and metastasis (8).
As a morphogen, retinoic acid (RA) is an important factor for processes, such as differentiation, growth inhibition or apoptosis (9). For these antineoplastic properties, RA can therapeutically be used in acute promyelocytic leukemia (APL) or high-risk neuroblastoma (3, 10). Especially, when proliferation is MYCN-driven, treatment with RA reduces cell growth and promotes differentiation (11). However, various mechanisms can lead to the formation of RA resistant tumor cells (12) limiting its clinical use. Whether the induction of physiological RA target genes like CCL2 (13, 14) might impair the therapeutic response is still elusive.
Therefore, the aim of the present study was to examine the RA-mediated regulation of N-Myc and CCL2 and their function in two neuroblastoma cell lines, which differed in MYCN copy numbers. Interestingly, both mediators supported cellular viability, although the regulation of their levels was cell type-specific and could not consistently be linked to the initiation of cell death or survival signaling.
Materials and Methods
Cell culture. Human Kelly and SH-SY5Y neuroblastoma cells were obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany; RRID: CVCL_2092 and CVCL_0019). Kelly cells were cultured on 10-cm plates in RPMI 1640 medium (Thermo Fisher Scientific, Darmstadt, Germany) supplemented with 10% v/v fetal bovine serum (Bio&Sell, Nürnberg, Germany) and 1% penicillin/strepotmycin solution (Thermo Fisher Scientific) at 37°C and 5% CO2. SH-SY5Y cells were kept under the same conditions in Dulbecco’s modified Eagle’s medium (Thermo Fisher Scientific) supplemented with 20% fetal bovine serum (Bio&Sell) and 1% penicillin/strepotmycin solution (Thermo Fisher Scientific). Cells were mycoplasma-free as monthly tested with the Venor®GeM qOneStep kit (Minerva Biolabs, Berlin, Germany). For the experiments, both cell lines were plated at low density (5×103 cells per well on 96-well plates; 1.2-1.8×106 cells per 10-cm plate) and incubated with the following substances: all-trans retinoic acid (5 μM, Tocris, Wiesbaden, Germany), CCL2 neutralizing antibody (R&D Systems/Bio-Techne, Wiesbaden, Germany), C021 (1 μM or 5 μM, Absource Diagnostics, Munich, Germany), cediranib (5 μM, Absource Diagnostics), crizotinib (1 μM, Absource Diagnostics), JQ1 (5 μM, Absource Diagnostics) or RS504393 (5 μM or 10 μM, Absource Diagnostics). The substances were applied to the cells for the indicated duration. Cells termed as controls were routinely incubated with a comparable amount of DMSO if used as a solvent. To examine whether constant exposure to elevated CCL2 levels in cell culture supernatants affected Kelly cell viability, the cells were plated on 6-well plates (0.1×106 cells per well) and treated with different CCL2 concentrations (1 ng/ml, 10 ng/ml, or 50 ng/ml, R&D Systems/Bio-Techne) for five weeks. Kelly cells were provided with fresh medium containing different CCL2 concentrations every 2-3 days. Once a week, they were detached from the plates, centrifuged (1,000×g for 10 min at room temperature) and plated on new 6-well plates (0.1×106 cells per well). After 5 weeks, the cells were detached from the plates, centrifuged, plated on 96-well plates (5×103 cells per well) and cultured in medium containing different concentrations of CCL2 and 5 μM RA for three days, before ATP levels were determined. If not indicated otherwise, chemicals were purchased from Sigma-Aldrich (Taufkirchen, Germany) or Carl Roth (Karlsruhe, Germany).
MYCN plasmid construction and transfection. MYCN was amplified from the ORF clone (NM_005378.5; OHu27902, GenScript, Rijswijk, the Netherlands) with the gene-specific primers 5′-GGACTCAGAT CTCGAATGCCGAGCTGCTCCACG-3′ and 5′-CGCGGTACCGTC GACGCAAGTCCGAGCGTGTTC-3′ at 67°C annealing temperature using the CloneAmp PCR Premix (Takara, Saint-Germain-en-Laye, France). The enhanced green fluorescent protein (EGFP) expression vector pEGFP-N3 (Takara) was digested using the restriction enzymes XhoI and PstI (both New England Biolabs, Frankfurt am Main, Germany) according to the manufacturer’s recommendations. The MYCN amplicon was inserted into the vector as a 1,395 bp fragment using the In-Fusion HD Cloning Kit (Takara) including cloning enhancement following the manufacturer’s protocol. Successful insertion was validated using Sanger sequencing. Transfection of the MYCN-EGFP construct was performed using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s instructions. Cells were plated in 96-well plates (5×103 cells per well) in 100 μl growth medium without antibiotics 48 h before transfection. For each well, 0.2 μg plasmid DNA and 0.5 μl Lipofectamine 2000 were used. Individual stable clones expressing the MYCN-EGFP construct or the pEGFP control vector were selected by adding 500 μg/ml G418 (Thermo Fisher Scientific) to the growth medium. Transfected clones were kept in the presence of 500 μg/ml G418 to maintain the selective pressure. Comparable results were obtained with three different clones per construct and three independent experiments for each clone.
siRNA transfection. Cells were plated in 6-well plates (8×105 cells per well) in 2 ml or 96-well plates (4×103 cells per well) in 100 μl growth medium without antibiotics 72 h before transfection. Two validated siRNAs against CCL2 (s12566, s12567, Thermo Fisher Scientific) and one negative control (Silencer Select Negative Control 1, 4390843, Thermo Fisher Scientific) were used. Transfections were performed with Lipofectamine™ RNAiMAX (Thermo Fisher Scientific) according to the manufacturer’s instructions. Briefly, Opti-MEM® was used to dilute the siRNAs for a final concentration of 50 nM and Lipofectamine™ RNAiMAX (0.2 μl Lipofectamine/100 μl or 5 μl Lipofectamine/2 ml), before both preparations were mixed. After an incubation of 15 min, the siRNA Lipofectamine complexes were added to the cells. The effects induced by specific siRNAs were determined after two days, as the amount of CCL2 was reproducibly reduced at that timepoint.
Determination of the cell number, viability and caspase activity. Cells were grown and treated with the different substances on a 96-well-plate (5×103 cells per well). To determine the amount of metabolically active cells, 50 μl of the ATP-detecting CellTiter-Glo® reagent (Promega, Mannheim, Germany) were added to each well. The contents were mixed for 2 min, and the plate was incubated for 10 min at room temperature. The plate was placed into an Infinite M200 reader (Tecan, Crailsheim, Germany), and luminescence was recorded. To determine the number of viable cells after three days of RA treatment, trypan blue exclusion assays were performed. The cells were detached from the plates, centrifuged (1,000×g for 10 min at room temperature) and resuspended in 1 ml PBS. Next, 20 μl of cell suspension were added to 20 μl trypan blue staining solution and the living cells were counted. The activation of caspase 3 or 7 were analysed by performing Caspase-Glo® 3/7 assays (Promega). Directly before use, Caspase-Glo® 3/7 substrate was dissolved in the assay buffer as recommended by the manufacturer. The cells were grown and stimulated on a 96-well-plate (5×103 cells per well) and 100 μl Caspase-Glo® 3/7 reagent was added to each well. The contents were briefly mixed, and the plate was incubated for 30 min at room temperature. Luminescence was recorded using an Infinite M200 reader (Tecan).
Whole cell extracts. Whole cell lysates were generated from subconfluent cells. Before harvesting, cells were washed with PBS. The cells were resuspended in denaturing lysis buffer [20 mM Tris, pH 7.4; 2% w/v sodium dodecyl sulfate (SDS); 1% w/v phosphatase inhibitor and 1% w/v protease inhibitor (both from Roche Applied Sciences)], incubated at 95°C for 5 min, briefly sonicated, and centrifuged to remove insoluble material (15,000×g for 15 min; 4°C).
Western blots. Twenty μg of protein extract were separated on 10-15% SDS-polyacrylamide gels and transferred to nitrocellulose transfer membranes (LI-COR Biosciences, Bad Homburg, Germany). The membranes were blocked with Odyssey Blocking Buffer (TBS) (LI-COR Biosciences) and incubated with the primary antibodies according to the manufacturer’s recommendations overnight at 4°C. After four washing steps with TBST (TBS plus 0.1% v/v Tween-20), the membranes were incubated with the corresponding IRDye 800CW or IRDye 680RD secondary antibody (LI-COR Biosciences) for 1 h. After four more washing steps with TBST, the membranes were kept in TBS before scanning with the AutoScan function of an Odyssey CLx reader (LI-COR Biosciences). To normalize protein activation or cleavage, two-colour detection was used. Densitometrical and statistical analysis was performed with the Image Studio™ Lite Software (LI-COR Biosciences). Antibodies against the following targets were used: GAPDH (sc-47724, Santa Cruz Biotechnology, Heidelberg, Germany, 1:1,000), N-Myc (sc-53993, Santa Cruz Biotechnology, 1:1,000).
ELISA. To measure CCL2 levels in cell culture supernatants, Quantikine® solid phase ELISAs (R&D Systems/Bio-Techne) were used according to the manufacturer’s instructions. Cells were grown and stimulated on a 96-well-plate (5×103 cells per well). After adding standards or samples to each well, the ELISA plates were incubated at room temperature followed by washing (three times) and an incubation with CCL2 (1 h) conjugate at room temperature. For the detection of CCL2, substrate solution was added to each well after washing. When the stop solution was added, the absorbance was measured within 30 min at 450 nm (reference wavelength 540 nm) with an Infinite M200 reader (Tecan).
RNA extraction and quantitative PCR. Total RNA was isolated using the E.Z.N.A Total RNA Kit II (Omega Bio-Tek, Norcross, GA, USA) according to the manufacturer’s instructions. For mRNA analyses, 1 μg of total RNA was reversely transcribed using random hexamer primers and the Transcriptor High Fidelity cDNA Synthesis Kit (Roche Applied Sciences) on the GeneAmp PCR System 9700 device (Applied Biosystems/Thermo Fisher Scientific) according to the manufacturer’s recommendations. qRT-PCRs of MYCN (Hs00232074_m1) and CCL2 (Hs00234140_m1) mRNA were performed in triplicates using TBP (TATA-box binding protein, Hs00427620_m1) as an internal control, with Universal Master Mix II, without UNG (Thermo Fisher Scientific) on the QuantStudio 7 device (Thermo Fisher Scientific) under default cycling conditions. Data were analyzed using the ΔΔCt-method as previously described (15).
Statistical analyses. All data were obtained from independent sets of experiments. The numbers of the repetitions (n) are included in the figure legends of the respective experiments, whereby biological replicate is equivalent to one repetition. For assays performed in 96-well plates, generally two technical replicates were used for each sample. All quantitative data represent the mean±standard deviation of these independent experimental replications. Statistical analyses were performed with GraphPad prism software (Version 8 for Windows, La Jolla, CA, USA) using one-way ANOVA with Bonferroni post hoc test.
Results
Gene expression and protein levels of N-Myc and CCL2 in Kelly and SH-SY5Y cells. First, the gene expression of the protooncogene MYCN was analyzed in Kelly and SH-SY5Y cells using RT-qPCR. Kelly cells are known to be MYCN-amplified and accordingly displayed a 1,700-fold higher expression than SH-SY5Y cells (relative expression of 158 vs. 0.093, p<0.001, Figure 1A). Similarly, western blots revealed 23-fold higher protein levels of N-Myc in Kelly compared to SH-SY5Y cells (p<0.001).
mRNA expression and protein levels of N-Myc and CCL2 in Kelly or SH-SY5Y cells. (A) Expression of MYCN was analyzed in Kelly and SH-SY5Y cells using RT-qPCR (n=4). (B) Protein levels of N-Myc were determined using western blotting (n=4). (C) Expression of CCL2 mRNA was analyzed in Kelly and SH-SY5Y cells using RT-qPCR (n=4). (D) Protein levels of CCL2 were determined in cell culture supernatants of treatment-naïve Kelly and SH-SY5Y cells using ELISA (n=4). ***p<0.001 for expression or protein levels of CCL2 or N-Myc in Kelly vs. SH-SY5Y cells.
For CCL2, qRT-PCR showed a 770-fold higher expression in SH-SY5Y compared to Kelly cells (relative expression of 10.7 vs. 0.014, p<0.001, Figure 1C). Likewise, the amount of CCL2 in cell culture supernatants was 1,300-fold enhanced in SH-SY5Y cells compared to Kelly cells (12,800 pg/ml vs. 9.97 pg/ml, p<0.001, Figure 1D).
Retinoic acid-mediated regulation of the cell number, CCL2 levels and MYCN levels. In previous studies, appropriate RA concentrations and incubation times had been determined for the analysis of different RA-mediated cellular effects (16, 17). To be able to associate the regulation of CCL2 or N-Myc with a putative function in RA signaling, RA-induced cellular reactions such as the cell number, as well as the intracellular ATP content were analyzed first. When Kelly and SH-SY5Y cells were treated with 5 μM RA for 3 days, the number of RA-treated Kelly cells were significantly reduced by 27% compared to control cells (p<0.001), while the number of RA-incubated SH-SY5Y cells was significantly enhanced (128%, p<0.001). Correspondingly, ATP levels were examined. The content of intracellular ATP was decreased in Kelly cells to 69% (p<0.001) compared to control cells after incubation with RA for three days. In contrast, ATP levels in RA-treated SH-SY5Y cells were 1.22-fold higher than in control cells (p<0.001) (Figure 2A).
Effects of 3-day RA treatment (5 μM) on the cell number, CCL2 release, MYCN expression and protein levels of N-Myc in Kelly and SH-SY5Y cells. (A) Effect on the cell number (n=4) or ATP levels (n=6). (B) Release of CCL2 determined using ELISA (n=4). (C) Expression of MYCN analyzed using RT-qPCR (n=4). (D) Protein levels of N-Myc determined using western blots (n=4). ***p<0.001, **p<0.01, for control vs. RA-treated cells. Control, treatment-naïve cells; RA: retinoic acid.
When CCL2 release was determined in RA-treated Kelly and SH-SY5Y cells, a similar regulation was observed in both cell lines. Incubation with RA for three days increased CCL2 levels in Kelly and SH-SY5Y cell culture supernatants 1.64-fold (p<0.001) and 2.8-fold (p<0.001), respectively, compared to control cells (Figure 2B).
Next, the effects of RA treatment on N-Myc levels were examined in both cell lines. After incubation with 5 μM RA for three days, MYCN mRNA as well as protein levels of N-Myc were analyzed. In Kelly cells, the gene expression of MYCN was significantly reduced to 60% (p<0.001) by RA treatment compared to control cells (Figure 2C). Similarly, protein levels were decreased in RA-treated cells by 53% (p<0.001) (Figure 2B). In SH-SY5Y cells, neither MYCN gene expression (Figure 2C) nor the levels of N-Myc protein were altered in response to RA (Figure 2D).
Taken together, RA attenuated N-Myc levels and the number of Kelly cells, while the number of SH-SY5Y cells increased. Interestingly, RA induced CCL2 release in both cell lines.
Cellular effects of N-Myc signaling in RA-treated neuroblastoma cells. So far, a RA-mediated reduction of N-Myc levels was associated with a decreased viability in Kelly cells, while no RA-induced effects on MYCN could be observed in MYCN non-amplified SH-SY5Y cells. For the analysis of MYCN-mediated effects, two lines of experiments were conducted: (i) enhancing MYCN expression in MYCN non-amplified SH-SY5Y cells and (ii) inhibition of N-Myc signaling in both cell lines. When SH-SY5Y cells were stably transfected with an MYCN-encoding plasmid, the amount of N-Myc protein was significantly increased (3.8-fold, p<0.001; vs. vector-transfected cells, Figure 3A). In addition, transfection with MYCN significantly enhanced ATP levels to 156% compared to vector controls (p<0.001) suggesting an increased number of viable cells. When the cells were treated with 5 μM RA for 3 days, cell viability of vector-transfected cells was increased to 114.7% (p<0.05 vs. untreated vector controls). In MYCN-over-expressing cells, ATP levels were elevated to 133% (p<0.001; compared to untreated vector controls) after incubation with RA for three days. Thereby, the intracellular ATP content of RA-treated MYCN-transfected cells was substantially lower than that observed in untreated MYCN-transfected cells (156%; p<0.001) (Figure 3B).
The role of N-Myc in RA-treated Kelly and SH-SY5Y cells. (A) Protein levels of N-Myc were determined in vector- or MYCN-transfected SH-SY5Y cells using western blots (n=3). (B) Effect of MYCN transfection on ATP levels in SH-SY5Y cells with or without a 3-day incubation with 5 μM RA (n=6). (C) ATP levels in Kelly cells or SH-SY5Y cells after a 3-day incubation with 5 μM RA alone or in combination with 5 μM JQ1 (n=4). (D) CCL2 expression (by RT-qPCR) or the release of CCL2 (by ELISA) in MYCN-transfected SH-SY5Y cells after a 3-day treatment with 5 μM RA (n=6). (E) CCL2 release (by ELISA) in Kelly or SH-SY5Y cells after a 2-day incubation with 5 μM RA, 5 μM JQ1 or a combination of both (n=4). ***p<0.001, **p<0.01, *p<0.05, for untreated (vector) control cells vs. MYCN-transfected cells, RA-treated cells or cells incubated with a combination of RA and JQ1; •••p<0.001, for untreated MYCN-transfected cells vs. RA-treated cells MYCN-transfected cells; ###p<0.001, for RA-treated cells vs. cells incubated with RA and JQ1. RA: Retinoic acid.
We also examined whether N-Myc inhibition affected cell viability. Both cell lines were co-incubated with 5 μM RA and 5 μM JQ1, which attenuates MYCN transcription (18). After three days, the viability of Kelly cells only treated with RA was significantly reduced to 72.5% (p<0.001; vs. untreated controls). Co-application of JQ1 and RA further decreased ATP levels to 20.5% (p<0.001) compared to control cells (Figure 3C). In SH-SY5Y cells, viability was increased after incubation with RA alone (127%, p<0.001; vs. control cells) and reduced by co-stimulation together with JQ1 to 42.6% (p<0.001) compared to untreated controls (Figure 3C). Thus, N-Myc sustained cell viability, but over-expression of MYCN did not abrogate the response to RA.
Another important aspect was to examine if modified N-Myc signaling affected the release of CCL2. In untreated Kelly and SH-SY5Y cells the levels of N-Myc and CCL2 were inversely correlated (Figure 1). In MYCN-over-expressing SH-SY5Y cells, neither expression nor release of CCL2 were altered compared to vector controls (Figure 3D). Incubation with 5 μM RA for three days increased mRNA and protein levels of CCL2 in vector-transfected cells 3.9-fold (p<0.001; vs. untreated vector controls) and 2.3-fold (p<0.01; vs. untreated vector controls), respectively. In MYCN-transfected cells, application of RA considerably elevated expression (4.3-fold; vs. untreated vector controls) and release of CCL2 (5.3-fold; vs. untreated vector controls), which was also significantly higher than in untreated MYCN-transfected cells (p<0.001) (Figure 3D).
To further examine the connection between N-Myc and CCL2, the release of CCL2 was determined after the cells were treated with RA or a combination of RA and JQ1. As cellular viability was considerably impaired after a 3-day co-incubation with RA and JQ1 (Figure 3C), which might also have affected signal transduction events, we examined the release of CCL2 after two days. In Kelly cells, the levels of CCL2 were increased (1.41-fold, p<0.01; vs. untreated controls), while co-treatment with RA and JQ1 reduced CCL2 levels (0.29-fold, p<0.001) compared to untreated controls (Figure 3E). In SH-SY5Y cells, incubation with RA for two days, enhanced CCL2 release (2.67-fold, p<0.001; vs. untreated controls). Incubation with RA and JQ1 significantly reduced CCL2 levels (0.25-fold, p<0.001) compared to control cells (Figure 3E). In summary, the levels of N-Myc and CCL2 are not inversely correlated, when analyzed in MYCN-over-expressing SH-SY5Y cells or after N-Myc inhibition.
Cellular effects of CCL2 signaling in RA-treated neuroblastoma cells. Next, it was examined if targeting CCL2 signal transduction affected cellular viability. Again, two different approaches were chosen: first, increasing CCL2 levels in Kelly cells, which displayed a lower endogenous CCL2 release compared to SH-SY5Y cells. Second, the effects of attenuated CCL2 signaling were analyzed in both cell lines. For supplementing CCL2 levels, Kelly cells were continuously incubated with different CCL2 concentrations (1 ng/ml, 10 ng/ml or 50 ng/ml) for five weeks, before 5 μM RA was co-applied for three days. When ATP levels were determined, the viability of CCL2-incubated cells was reduced compared to control cells (1 ng/ml: 77%, 10 ng/ml: 73%, 50 ng/ml: 67%; all p<0.001), but to a similar extent as RA treatment of Kelly cells, which had not been cultivated with additional CCL2 (Figure 4A).
The role of CCL2 in RA-induced effects in Kelly and SH-SY5Y cells. (A) ATP levels in Kelly cells after continuous incubation with CCL2 (1 ng/ml, 10 ng/ml or 50 ng/ml) for five weeks and an additional 3-day treatment with 5 μM RA (n=6). (B) CCL2 levels analyzed using ELISA after transfection with either a negative control siRNA (neg.) or specific CCL2 siRNAs (siRNA1 or siRNA2) for two days in Kelly or SH-SY5Y cells (n=3). (C) ATP levels in siRNA-transfected Kelly or SH-SY5Y cells (negative control siRNA or two specific CCL2 siRNAs) treated with or without 5 μM RA determined after 2 days (n=3). (D) ATP levels in Kelly or SH-SY5Y cells after a 2-day treatment with 5 μM RA or the combination or RA and a CCL2-neutralizing antibody (neutr. ab; 5 μg/ml or 10 μg/ml) (n=3). (E) ATP levels in Kelly or SH-SY5Y cells after a 3-day treatment with 5 μM RA or a combination with C021 (1 μM or 5 μM) and RA (n=4). ***p<0.001, **p<0.01, *p<0.05, for control vs. cells incubated with RA, a combination of RA and C021, a CCL2-neutralizing antibody or transfected with a specific CCL2 siRNA; ###p<0.001, ##p<0.01, for RA-treated cells vs. cells transfected with a specific CCL2 siRNAs or incubated with RA and C021 or a CCL2-neutralizing antibody; •••p<0.001, ••p<0.01, for cells transfected with CCL2-specific siRNAs vs. transfected cells additionally incubated with RA. siRNA1: siRNA 12566; siRNA2: siRNA 12567; siRNA1/2: combination of siRNAs 12566 and 12567; C0: C021; neg.: negative control; RA: retinoic acid.
To reduce CCL2 signaling, siRNA experiments were performed. Both cell lines were transfected with control siRNA or two different validated CCL2 siRNAs (s12566, s12567). When CCL2 levels were determined 48 hours after transfection, CCL2 release was significantly (p<0.001) reduced in Kelly and SH-SY5Y cells for both CCL2 mRNA targeting siRNAs as well as their combination compared to cells transfected with negative control siRNAs (Figure 4B). Additionally, ATP levels were measured in transfected cells. Interestingly, down-regulation of CCL2 did not affect the viability of untreated Kelly cells. When transfected cells were additionally incubated with 5 μM RA for two days, ATP levels in Kelly cells were significantly reduced compared to the respective untreated controls (79% vs. negative control cells; 73.9% for RA-treated siRNA1/2 vs. untreated siRNA1/2). However, ATP levels of all RA-incubated Kelly cells were similar irrespective of CCL2 levels (Figure 4C). In SH-SY5Y cells, transfection of CCL2-specific siRNAs significantly increased viability compared to negative controls (120%, p<0.01 for siRNA1/2). Incubation with RA enhanced ATP levels in negative control cells (120%, p<0.01; vs. untreated negative control cells). In contrast, RA treatment significantly reduced viability of transfected SH-SY5Y cells when compared to RA-incubated negative controls or the respective untreated specific siRNA control (99.9% for RA-treated siRNA1/2 vs. 120% of untreated siRNA1/2-transfected cells) (Figure 4C).
Similar experiments were performed by incubating Kelly and SH-SY5Y cells with RA or the combination with RA and a neutralizing antibody (5 μg/ml or 10 μg/ml) against CCL2 for two days. When viability was determined in Kelly cells, ATP levels were significantly reduced compared to untreated controls in RA-treated cells (84%, p<0.01) as well as in cells incubated with RA and the CCL2-neutralizing antibody (5 μg/ml: 79%, p<0.001; 10 μg/ml: 69%, p<0.001). The intracellular ATP content of Kelly cells, which were treated with RA and 10 μg/ml of the CCL2-neutralizing antibody, was even lower than that observed in RA-treated cells (p<0.05) (Figure 4D). In SH-SY5Y cells, treatment with RA increased ATP levels (122%, p<0.01, vs. untreated controls). After co-application of RA and 10 μg/ml of the CCL2-neutralizing antibody, the cellular viability was significantly lower than that of RA-treated cells (91% vs. 122%; p<0.001) (Figure 4D).
Finally, inhibition of CCL2 signaling in RA-treated Kelly or SH-SY5Y cells was analyzed by application of different CCL2 receptor inhibitors. RS504393 was chosen to attenuate CCL2 binding to the receptor CCR2 and C021 to reduce CCL2 activity via the receptor CCR4. While co-incubation with RS504393 and RA had no effects on the viability of Kelly or SH-SY5Y cells (data not shown), co-treatment with 1 μM or 5 μM C021 and 5 μM RA for three days significantly decreased ATP levels compared to control or RA-treated cells. When Kelly cells were stimulated with RA, cellular viability was reduced to 77.7% (p<0.05, vs. untreated controls). Simultaneous application of 1 μM or 5 μM C021 and RA attenuated the number of viable cells by 41.3% (p<0.001 vs. control) or 96.3% (p<0.001 vs. control and RA-treated cells), respectively (Figure 4E).
When SH-SY5Y cells were incubated with RA for three days, cell viability was increased (129%, p<0.001, vs. untreated controls). While co-application of RA and 1 μM C021 did not affect ATP levels compared to controls, simultaneous incubation with RA and 5 μM C021 decreased cell viability to 26.6% (p<0.001, vs. untreated controls and RA-treated cells). Thus, inhibition of CCR4 by C021 reduced the number of viable cells compared to RA-treated cells after three days (p<0.001) (Figure 4E). Therefore, CCL2 contributes to cellular viability in Kelly and SH-SY5Y cells, although its levels could not consistently be associated with improved or decreased cellular viability in both cell lines.
CCL2 levels in response to apoptotic stimuli in RA-treated neuroblastoma cells. So far, the results of our experiments suggested that basal CCL2 levels or N-Myc activity might be beneficial for the viability of RA-treated cells, whereby the amount or the regulation of each mediator is cell type-specific. Therefore, we examined whether the inhibition of growth signaling led to a consistent regulation of both molecules. To attenuate survival signaling, the VEGFR2 inhibitor cediranib or the c-Met/ALK inhibitor crizotinib were used. When Kelly cells were treated with 5 μM RA or a combination of RA and 5 μM cediranib for three days, ATP levels were decreased to 44.6% (p<0.001, vs. untreated controls and RA-treated cells) (Figure 5A). Co-applied RA and 1 μM crizotinib reduced Kelly cell viability to 26.3% (p<0.001, vs. untreated controls and RA-treated cells). In SH-SY5Y cells, concomitant incubation with RA and cediranib or crizotinib attenuated ATP levels by 75.7% or 19.2%, which was markedly lower than ATP levels in unstimulated cells (p<0.001) as well as in RA-treated cells (p<0.001) (Figure 5A). It was also investigated whether apoptosis was induced in response to co-treatment with RA and cediranib or crizotinib by quantifying the activity of caspase 3/7. As caspase activation usually occurs before cell loss can be quantified, the experiments were conducted after two days. In Kelly cells, incubation with RA reduced caspase activity to 84% (p<0.05), while co-stimulation with RA and cediranib or crizotinib enhanced caspase cleavage (1.89-fold or 2.12-fold, respectively, p<0.001; vs. untreated controls and RA-treated cells). In SH-SY5Y cells, treatment with RA did not affect caspase activity, but co-application of RA and cediranib or crizotinib enhanced it (4.67-fold or 1.34-fold, respectively, p<0.001; vs. untreated controls and RA-treated cells) (Figure 5B).
Effects of co-application of RA and cediranib or crizotinib on cell viability and CCL2 release in Kelly and SH-SY5Y cells. (A) ATP levels in Kelly or SH-SY5Y cells after a 3-day incubation with 5 μM RA or a combination of RA and 5 μM cediranib (cedi) or 1 μM crizotinib (crizo) (n=5). (B) Activation of caspase 3/7 (Caspase-Glo® 3/7 assays) in Kelly or SH-SY5Y cells after a 2-day treatment with RA or a combination of RA and cediranib or crizotinib (n=4). (C) Release of CCL2 (by ELISA) in Kelly or SH-SY5Y cells after a 2-day incubation with RA or a combination of RA and cediranib or crizotinib (n=4). (D) Protein levels of N-Myc in Kelly cells determined in response to cediranib, crizotinib or a combination of either inhibitor with RA using western blots after two days (n=5). ***p<0.001, **p<0.01, for control vs. cells treated with RA or a combination of RA and cediranib or crizotinib; ###p<0.001, for RA-treated cells vs. cells incubated with RA and cediranib or crizotinib. RA: Retinoic acid.
Next, it was examined if co-stimulation with RA and cediranib or crizotinib affected CCL2 release. Again, signaling events involved in the initiation of cell death or survival precede reduced viability as detected after three days (Figure 5A). Therefore, ELISAs were performed after incubation with RA or the combination of RA and cediranib or crizotinib for two days (Figure 5C). In Kelly cells, incubation with RA enhanced CCL2 release (1.8-fold, p<0.05). Co-application with cediranib and RA substantially increased CCL2 levels (10.5-fold, p<0.001), which was significantly higher than the levels of CCL2 in control or RA-treated cells (p<0.001). The simultaneous incubation with RA and crizotinib, however, significantly reduced CCL2 release to 45.3% (p<0.001). When SH-SY5Y cells were incubated with RA for two days, CCL2 release was enhanced (2.5-fold, p<0.001). Co-treatment with cediranib and RA, in contrast, decreased CCL2 levels to 35.8% (p<0.001 vs. control or RA-treated cells), while the concurrent incubation with RA and crizotinib increased CCL2 release (1.51-fold, p<0.001).
Finally, we analyzed N-Myc levels after co-stimulation with RA and cediranib or crizotinib after 2 days. In Kelly cells, stimulation with crizotinib or cediranib decreased the N-Myc signal to 75% or 74%, respectively (p<0.01; vs. untreated cells). Co-incubation with RA and crizotinib reduced N-Myc levels to 34% (p<0.001; vs. untreated cells) and to 23% (p<0.001; vs. untreated cells) in response to RA and cediranib (Figure 5D). The amount of N-Myc protein was not affected in SH-SY5Y cells (data not shown). Therefore, CCL2 or N-Myc levels were not generally decreased, when the inhibitors were co-applied with RA, but regulated in a cell type-specific and stimulus-dependent manner.
Discussion
RA induces differentiation in a variety of neuroblastoma cell lines (19) and even attenuates MYCN-driven proliferation (20-22). At the same time, RA-mediated growth inhibition can be impaired by several resistance mechanisms (12), but it is still unclear if the induction of physiological targets like CCL2 (13) might contribute to treatment failure. When Kelly or SH-SY5Y cells were incubated with RA, the levels of CCL2 were enhanced in both cell lines, which was accompanied by a decrease in the amount of N-Myc or the cell number only in Kelly cells. The inhibition of CCL2 or N-Myc signaling generally reduced viability of RA-treated cells, but particularly the regulation of CCL2 levels was dependent on the cell type and the stimulus and could not be consistently linked to cell survival or death.
When the levels of N-Myc or CCL2 were examined in cells not exposed to RA, MYCN-amplified Kelly cells exhibited lower endogenous CCL2 release compared to MYCN non-amplified SH-SY5Y cells. Inversely correlated MYCN and CCL2 levels were observed before in neuroblastoma cells (4, 5). Even direct repression of CCL2 transcription by over-expressed MYCN has been shown (5), but was not observed in our study, which might be due to different transfection systems, since lentiviral-based gene expression usually yields higher efficiency rates (5) compared to lipofection as used in our experiments.
When Kelly or SH-SY5Y cells were incubated with RA, N-Myc levels and the number of cells were decreased in Kelly cells, while both parameters remained similar to untreated controls in SH-SY5Y cells. Exposure to RA induced differentiation, growth arrest or apoptosis in several neuroblastoma cell lines (19), which can be preceded by down-regulation of N-Myc (20). Especially, when proliferation is MYCN-driven, treatment with RA reduces cell growth and promotes differentiation (11). Acyclic RA in combination with the β1-integrin-activating peptide TNIIIA2 has recently been described as an additional therapeutic option to reduce N-Myc levels and malignant properties in neuroblastoma cells (23). Similarly, MYCN knockdown attenuates proliferation and even promotes apoptosis in neuroblastoma cells (24, 25).
As for SH-SY5Y cells, it has been noted before that they were less responsive towards RA (17, 26). Only higher RA concentrations stopped proliferation, but did not induce cell death (26). Furthermore, N-Myc levels were not reduced by RA treatment in SH-SY5Y cells, which is characteristic of neuroblastoma cells, when MYCN is not used as an enhancer for its own expression (11).
In both cell lines, RA treatment significantly increased CCL2 levels, which had not been examined so far in neuroblastoma cells, but in APL cells, where CCL2 has been identified as a target gene of RA (13, 14). Still, the overall effect of RA on CCL2 regulation depends on the activation or inhibition of specific intracellular pathways, which is usually context-dependent. When, for instance, NFkB signaling is inhibited by the application of RA, CCL2 release is attenuated, as observed in melanoma cells or diabetic nephropathy (27, 28).
We also examined whether N-Myc or CCL2 contributed to the viability of RA-treated cells. As described before, MYCN drives proliferation and promotes malignancy in neuroblastoma (29-32) and its down-regulation by RA can be associated with differentiation, growth arrest or even apoptosis (20). When RA was co-applied with the bromodomain and extra-terminal domain inhibitor JQ1, cell viability was reduced in both cell lines, which has been observed for Kelly and SH-SY5Y cells as well as for other neuroblastoma cell lines (33). Even in acute myeloid leukemia, when RA signaling is impaired, the combination with JQ1 synergistically inhibited proliferation and induced differentiation or apoptosis by depleting c-Myc or hTERT (34).
The effects of differentially regulated CCL2 levels are less clear. The reduction of CCL2 release by transfection of siRNAs did not affect cellular viability in Kelly cells and only attenuated the RA-mediated increase of the number of SH-SY5Y cells. It has been shown that treatment with an anti-CCL2 antibody did not inhibit neuroblastoma proliferation, but migration towards an increasing CCL2 gradient (8), which also suggests that the reduction of CCL2 levels is not sufficient to inhibit proliferation or even survival. In contrast, blocking the CCL2 receptor CCR4 attenuated cell viability of RA-treated Kelly and SH-SY5Y cells, while targeting CCR2 by RS504393 had no effects. In a variety of tumor diseases, e.g., colorectal cancer, ovarian cancer, lung cancer, renal cell carcinoma or prostate cancer cells, CCL2 contributes to survival or growth by activating different survival pathways (35-38). Often, CCL2 enhances tumorigenic properties by binding to CCR2 (36, 39) or uses both receptors to induce its full range of cellular functions (39). However, the CCL2/CCR4 axis can be selectively activated to promote cancer metastasis (40). Thus, with regard to the increased CCL2 levels in RA-treated Kelly and SH-SY5Y cells, it is conceivable that elevated CCL2 release can facilitate the emergence of RA-resistant cells, especially when it acts in concert with growth factors like VEGF (6).
Co-treatment with RA and cediranib or crizotinib induced apoptosis in Kelly and SH-SY5Y cells. So far, the concomitant use of RA and cediranib or crizotinib has not been tested before. Cediranib has effectively suppressed growth in neuroblastoma xenographs as a single substance (41) or when co-applied with rapamycin (42). Similarly, crizotinib has been used in monotherapy or in combination with chemotherapeutic substances in neuroblastoma cells, xenographs or patients (43-45), which led to reduced proliferation and even clinical improvement (43, 44), albeit not in relapsed/refractory neuroblastoma patients (46). Especially in MYCN-amplified neuroblastoma, crizotinib could decrease proliferation by interfering with MYCN transcription (47), which is in line with the findings of the present study. The effect of cediranib on N-Myc levels, which is similar to crizotinib in Kelly or SH-SY5Y cells, has not been examined before.
Likewise, it had not been known, if cediranib or crizotinib affected RA-mediated release of CCL2. In our experiments, differential effects were observed. Although the combination of either compound and RA increased apoptosis, a cell-type specific regulation of CCL2 was observed. In Kelly cells, CCL2 levels were increased after co-treatment with RA and cediranib but reduced after the simultaneous application of RA and crizotinib. An opposite regulation was detected in SH-SY5Y cells. These findings indicate that CCL2 is not decisive for the induction of cell survival or death, but the regulation of its release rather mirrors the activity of the signaling pathways responsible for the cellular outcome.
Conclusion
In summary, despite its potential to decrease cell growth by reducing N-Myc levels, RA might facilitate resistance development by increasing CCL2 release in neuroblastoma cells, although the levels of N-Myc or CCL2 are not generally indicative of the cellular outcome, especially when examined during apoptotic signaling.
Acknowledgements
The Authors thank Irina Naujoks and Annika Muetze for their excellent technical assistance.
Footnotes
Funding
This work was supported by an intramural grant of the Medical Faculty of Kiel University (Kiel, Germany).
Conflicts of Interest
The Authors declare no conflicts of interest in relation to this study.
Authors’ Contributions
NR, NSP, RSI, and VT conducted experiments. MK and HB conducted experiments and reviewed the manuscript. IC reviewed and revised the manuscript. VW designed and supervised the study, designed and conducted experiments and wrote the manuscript.
- Received September 29, 2024.
- Revision received October 8, 2024.
- Accepted October 9, 2024.
- Copyright © 2025 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
