Transition to a mesenchymal state in neuroblastoma may be characterized by a high expression of GD2 and by the acquisition of immune escape from NK cells

Abstract

Background: Neuroblastoma (NB) is characterized by both adrenergic (ADRN) and undifferentiated mesenchymal (MES) subsets. The ganglioside sialic acid-containing glycosphingolipid (GD2) is widely overexpressed on tumors of neuroectodermal origin promoting malignant phenotypes. MES cells are greatly enriched in post-therapy and relapsing tumors and are characterized by decreased expression of GD2. This event may cause failure of GD2-based immunotherapy. NK cells represent a key innate cell subset able to efficiently kill tumors. However, the tumor microenvironment (TME) that includes tumor cells and tumor-associated (TA) cells could inhibit their effector function.

Methods: We studied eight NB primary cultures that, in comparison with commercial cell lines, more faithfully reflect the tumor cell characteristics. We studied four primary NB-MES cell cultures and two pairs of MES/ADRN (691 and 717) primary cultures, derived from the same patient. In particular, in the six human NB primary cultures, we assessed their phenotype, the expression of GD2, and the enzymes that control its expression, as well as their interactions with NK cells, using flow cytometry, RT-qPCR, and cytotoxicity assays.

Results: We identified mature (CD105⁺/CD133^-) and undifferentiated (CD133⁺/CD105^-) NB subsets that express high levels of the MES transcripts WWTR1 and SIX4. In addition, undifferentiated MES cells display a strong resistance to NK-mediated killing. On the contrary, mature NB-MES cells display an intermediate resistance to NK-mediated killing and exhibit some immunomodulatory capacities on NK cells but do not inhibit their cytolytic activity. Notably, independent from their undifferentiated or mature phenotype, NB-MES cells express GD2 that can be further upregulated in undifferentiated NB-MES cells upon co-culture with NK cells, leading to the generation of mature mesenchymal GD2^bright neuroblasts. Concerning 691 and 717, they show high levels of GD2 and resistance to NK cell-mediated killing that can be overcome by the administration of dinutuximab beta, the anti-GD2 monoclonal antibody applied in the clinic.

Conclusions: NB is a heterogeneous tumor representing a further hurdle in NB immunotherapy. However, different from what was reported with NB commercial cells and independent of their MES/ADRN phenotype, the expression of GD2 and its displayed sensitivity to anti-GD2 mAb ADCC indicated the possible effectiveness of anti-GD2 immunotherapy.

Keywords: NK cells; ganglioside sialic acid-containing glycosphingolipid-2; immunotherapy; neuroblastoma; primary tumor cells.

Figure 1

Ganglioside sialic acid-containing glycosphingolipid 2 (GD2) expression in neuroblastoma (NB) primary mesenchymal and adrenergic cultures. (A) Flow cytometry analysis of the surface expression of the mesenchymal markers CD133 and CD105 in NB primary cultures WU-1 and WU-2. Representative histograms of three independent experiments. (B) RT-PCR analysis of the NB mesenchymal transcripts WWTR1, SIX4, and ZFLP36LI and NB adrenergic markers ASCL1, DBH, DLK1, and GATA2 in NB primary cultures WU-1 and WU-2. mRNA levels were normalized with ACTB and were expressed as fold increase relative to the untreated controls. (C) Flow cytometry analysis of the surface expression of GD2 in mesenchymal NB primary cultures (WU-1, WU-2, 691-MES, and 717-MES) and in NB adrenergic primary cultures (691-ADRN and 717-ADRN). Representative histograms of three independent experiments. (D) RT-PCR analysis of the sialyl-transferase GD3 synthase (GD3S, ST8SIA1) transcript (left panel) and GD2 synthase (GD2S, B4GALNT1) transcript (right panel) in NB primary mesenchymal and adrenergic cell cultures. The adrenergic cell line SH-SY5Y and its GD2 silenced counterpart SH-SY5Y/GD2⁻ were employed as controls of expression. Results were expressed as means ± SEM (n = 3).

Figure 2

Natural killer (NK) cells mediate cytotoxic activity against NB cells. (A) Susceptibility of CMFDA-labeled NB primary cultures to allogenic IL-2-activated NK cell-mediated lysis at different effector:target ratios (E:T ratio). CMFDA-labeled K-562 and SK-N-AS cells were used as positive and negative lysis controls, respectively. Data were expressed as mean ± SEM (n = 4) of the percentage of cell lysis (propidium iodide-positive cells, PI⁺ cells). Statistical analysis was performed using two-way ANOVA with Tukey’s multiple comparisons test. *p ≤ 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not statistically significant. (B) Surface expression of the degranulation marker CD107a on NK cells after co-culture with the NB mesenchymal/adrenergic pairs 691 and 717 in the presence or absence of the anti-GD2 mAb (αGD2) at 1:1 E:T ratio by flow cytometric analysis. NK cells cultured without target cells (NK only and NK only + αGD2) or with K-562 (K-562 and K-562 + αGD2) were used as internal controls. Data were expressed as the percentage of CD107a-positive cells (mean ± SEM; n = 4). Statistical analysis was performed using the Mann–Whitney test. *p ≤ 0.05; ns, not statistically significant.

Figure 3

Interactions between primary mesenchymal and adrenergic cell cultures with NK cells. (A) Evaluation of the cytotoxic activity of freshly isolated PB NK cells against CMFDA-labeled K-562 target cells at different E:T ratios. NK cytotoxicity assay was performed after 4 days of co-culture with NB primary mesenchymal cells WU-1 and WU-2. NK cells cultured for 4 days in an IL-2-supplied medium were used as control (Ctrl NK). Data were expressed as mean ± SEM (n = 4) of the percentage of K-562 cell lysis (PI⁺ cells). Statistical analysis was performed using two-way ANOVA with Tukey’s multiple comparisons test. *p ≤ 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not statistically significant. (B) Flow cytometric analysis of the NK cell phenotype for NKGD2, NCRs (NKp46, NKp30, NKp44), CD56, and CD69 on NK cells after co-culture with WU-1 and WU-2 NB primary cells. Freshly isolated PB NK cells were co-cultured for 4 days either in the presence of WU-1 and WU-2 primary cells cell lines or alone (Ctrl NK only) in an IL-2-supplied medium. Fold change mean fluorescence intensity (MFI) was expressed as mean ± SD (n = 3). *p < 0.05 sample co-culture with NK vs. Ctrl NK only. (C) Phenotype of WU-2 primary NB cells, after 4 days of co-culture, with freshly isolated PB NK cells by flow cytometry analysis. The MFI ratio of the indicated markers was expressed as mean ± SD (n = 3). *p < 0.05 sample co-culture with t0 vs. t4 days. (D) Percentage of CMFDA-labeled K-562 cell lysis in cytotoxicity assays using freshly isolated PB NK cells after 4 days of co-culture with NB cells 691 and 717. NK cells cultured alone were used as controls (Ctrl NK). Values are expressed as mean ± SEM (n = 3). Statistical analysis was performed using two-way ANOVA with Tukey’s multiple comparisons test. *p ≤ 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not statistically significant.