Abstract
Neuroblastoma (NB) is a pediatric cancer that develops from immature nerve cells in the peripheral sympathetic nervous system. NB is remarkably heterogeneous, ranging from spontaneous regression to aggressive progression, and is characterized by widespread dissemination and relapse. Approximately half of all NB patients present with widespread metastasis at diagnosis and are classified as high-risk with a substantial likelihood of treatment failure, despite receiving aggressive multimodal therapies, including surgery, chemotherapy, radiotherapy, autologous hematopoietic stem cell transplantation, and monoclonal antibody treatment. Beyond conventional multimodal approaches, immunotherapy has emerged as an essential part of cancer treatment, by boosting the immune system to recognize and eliminate tumor cells. In this review, we provide an overview of current therapeutic strategies for NB patients and summarize recent advances in the development of next-generation NB immunotherapies, highlighting their potential to improve NB management. We further discuss future directions for therapeutic improvement, and the potential limitations and challenges associated with translating these approaches into long-term benefits for NB patients.
keywords
Introduction
Neuroblastoma (NB) is the most frequently occurring childhood malignancy of the sympathetic nervous system. NB arises from neural crest-derived progenitor cells1 and occurs with an incidence of 8.3 per million in children <15 years of age. NB is frequently detected in and around the adrenal gland, but can also be found in sympathetic ganglia, with metastatic sites in the bone marrow, lymph nodes, bones, liver, and orbital sites2. NB usually occurs sporadically, but inherited NB has a prevalence of 1-2%. Nearly 90% of NB cases occur in children <5 years of age and account for >15% of pediatric cancer-related deaths3,4. The overall survival in patients with NB is 65%, and a subset of NB can spontaneously regress without any specific treatment. However, NB patients with widespread metastasis (high-risk) have refractory disease with a grim outcome: >50% of these patients do not survive despite aggressive surgery and dose-intensified chemotherapy.
The International Neuroblastoma Risk Group (INRG) developed the Children’s Oncology Group (COG) Risk Classifier, based on the International Neuroblastoma Staging System (INSS), to better identify pretreatment risk stratification. In INSS, a postsurgical staging system that classifies disease stages according to tumor location, locoregional tumors are classified as INSS stage 1, 2A, 2B, or 3, and metastatic tumors are classified as INSS stage 4 and 4S. To create a staging system independent of the extent of surgical resection, the INRG staging system (INRGSS) was developed, in which patients with NB are classified into four categories (very low-, low-, intermediate-, or high-risk) according to the presence of several biological and clinical factors, including age at diagnosis, tumor histology, grade of tumor differentiation, DNA ploidy, MYCN amplification status, and chromosome aberrations5. To improve the accuracy of treatment planning for patients with NB, the COG Risk Classifier was recently updated to version 2, which incorporates the segmental chromosome abbreviation (SCA) score as an additional genomic biomarker6.
Genetic predisposition of NB
NB is a complex extracranial neoplasm with remarkable biological and genetic heterogeneity reflected in several chromosomal and gene alterations, as well as aberrant expression of key oncogenes that drive tumor initiation and disease progression7. In recent decades, increasing attention has been directed toward the genetic predisposition for NB. Familial NB represents a small subset of NB cases (1-2%). Approximately 85% of familial NB cases arise from heritable ALK8 and PHOX2B9 mutations. More recently, germline pathogenic variants in BARD1 have been identified to impair DNA repair and contribute to NB tumorigenesis10. Moreover, NB patients with higher germline variant burden tend to exhibit greater somatic mutational burden and poorer overall survival, thereby underscoring the interplay between inherited and acquired genetic factors in disease progression11.
Beyond these familial cases, most NB cases occur sporadically in people without an identifiable family history. Somatically acquired segmental gain of 17q, hemizygous deletion of 1p and 11q, and MYCN oncogene amplification are among the common genetic alterations associated with poor prognosis in patients with NB7. MYCN amplification, a key oncogenic event observed in approximately 20% of NB cases, is a prominent prognostic marker for high-risk disease. MYCN amplification is rarely acquired during disease progression, thus suggesting that MYCN amplification is a key driver in high-risk NB development12. MYCN promotes NB tumorigenesis primarily through the transcriptional regulation of large numbers of genes that encode essential proteins for fundamental cellular processes, including proliferation, cell growth, ribosome biogenesis, and metabolism13.
Beyond these well-characterized somatic events, genome-wide association studies have revealed numerous loci associated with NB susceptibility14,15. For example, variants in LMO1 are strongly associated with poor clinical outcomes, thereby emphasizing the potential contributions to more aggressive disease phenotypes16,17. Additional rare variants, including SNPs within DUSP12, DDX4, and HSD17B12, may increase susceptibility to NB development; these findings highlight the complex genetic landscape underlying both inherited predisposition and sporadic tumor development18.
Current treatment options for NB
Given the complex nature of NB, disease management is based on risk stratification (Figure 1). The use of precision oncology and targeted therapy in NB management is aimed at decreasing the frequency of relapse, as well as improving overall survival and quality of life among patients with NB19. The combination of surgery and multidrug chemotherapy significantly improves overall survival rates among non-high-risk NB patients. For high-risk patients, intensive multimodal treatment includes surgery, chemotherapy, radiotherapy, autologous hematopoietic stem cell transplantation (AHSCT), and anti-GD2 therapy.
Current treatment strategies for NB by risk group. Very low- and low-risk patients are treated with surgery alone. Intermediate-risk patients typically receive a combination of surgery and chemotherapy. High-risk patients undergo multimodal therapy consisting of three sequential phases (induction, consolidation, and maintenance). AHSCT: autologous hematopoietic stem cell transplantation. Figure created with BioRender.com.
Surgery is performed primarily to remove NB that is localized and has not metastasized to other parts of the body. Surgery can effectively improve survival among patients with NB, regardless of risk group20. A complete macroscopic resection or resection based on a >90% cut-off of the NB primary tumor is associated with an improved 5-year survival rate and diminished local disease progression21,22.
In addition, chemotherapy is an essential first-line treatment for NB, particularly for metastatic cases. The use of anti-cancer chemotherapy drugs increases NB cell cytotoxicity and leads to tumor eradication. Chemotherapy may be used in conjunction with, before, or after surgery, because it decreases the risk of surgery23. Use of a different chemotherapy intensity has been found to enable the intermediate-risk group to achieve an 80% overall survival rate24 and high-risk patients to achieve an improved 5-year overall survival rate25.
A low-to-moderate dose of radiotherapy is necessary for metastatic cases in which tumors cannot be removed by chemotherapy and surgery26. Because of the potential deleterious effects on normal tissues, a newer form of radiotherapy using a radioactive form of the chemical metaiodobenzylguanidine (MIBG) is incorporated into current treatment27. MIBG therapy typically results in mitigated adverse effects, because only NB cells that take up the radioactive compound are targeted, thereby sparing the surrounding normal tissues.
The first NB AHSCT was introduced in the 1990s28. Healthy stem cells are collected from the patient before high-dose chemotherapy and preserved until needed. AHSCT is usually combined with high-dose chemotherapy and radiotherapy. AHSCT helps replenish the blood cells destroyed during intense chemotherapy and consequently promotes the immune response against NB. Multiple systematic reviews have shown that event-free survival favors AHSCT over conventional chemotherapy or no further treatment29,30.
The retinoid acid (RA) signaling pathway, involving heterodimerization of the RA receptor and retinoid X receptor, is critical in early embryonic development and neurodevelopment. Impaired retinoid acid signaling is associated with NB31,32. Isotretinoin, also known as 13-cis-RA, is the only RA form used in current clinical practice. Isotretinoin is usually administered together with an anti-GD2 monoclonal antibody in NB maintenance therapy. Patients with NB receiving isotretinoin maintenance therapy have an improved 5-year event-free survival with respect to that in patients receiving no isotretinoin33.
Unfortunately, this intensive multimodal treatment might not be effective for high-risk patients with NB, <40% of whom survive more than 5 years without relapse34. Therefore, immunotherapy has also been considered an alternative approach for NB treatment by empowering the immune system to recognize and eliminate tumor cells. Monoclonal antibodies (mAbs), designed to interact with a specific target, were subsequently introduced as a targeted immunotherapy. In 2015, the Food and Drug Administration (FDA) approved the first mAb (dinutuximab) for high-risk NB treatment. Dinutuximab is a chimeric mAb (Ch14.18) that targets disialoganglioside 2 (GD2), a glycolipid with high expression on the surfaces of NB cells but limited expression in normal tissues35. The interaction of dinutuximab with GD2 on NB cells initiates a series of killing mechanisms that eliminate NB cells, including antibody-dependent cell-mediated cytotoxicity (ADCC), complement-dependent cytotoxicity (CDC), and antibody-dependent phagocytosis. Dinutuximab is always used with granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin-2 (IL-2) to promote antitumor immunity, because GM-CSF activates macrophages and neutrophils, whereas IL-2 activates natural killer cells36. Although immunotherapy with dinutuximab improves outcomes in high-risk NB patients, the adverse reactions of this treatment, such as transient neuropathic pain, fever, hypersensitivity reactions, and capillary leak syndrome, remain the greatest management challenge in clinical practice. Approximately a decade later, another anti-GD2 mAb (Hu3F8), naxitamab, was approved in the United States (US), Israel, China, Brazil, and Mexico, and other countries, for use in combination with GM-CSF in pediatric patients (>1 year of age) and adults with relapsed or refractory high-risk NB. Naxitamab induces immune-mediated antitumor cytotoxicity, primarily via CDC and ADCC37.
On the basis of the principles of mAbs, an alternative class of dual-specificity therapeutics, known as bispecific antibodies (bsAbs), was introduced. BsAbs contain two target-binding units in one antibody-based molecule (usually a tumor-associated antigen and an immune cell receptor, such as CD3 on T cells)38. Each unit can independently recognize a unique epitope. After sequential or simultaneous binding, bsAbs act as a biophysical bridge between two antigens, by enabling interactions between the tumor and immune cells, which lead to a synergistic tumor-specific immune response. BsAbs have been successfully used in the treatment of hematologic diseases and are currently in various stages of solid tumor trials39. Although NB cells decrease the expression of human leukocyte antigen (HLA), thus facilitating evasion of conventional T cell immunity, they remain susceptible to T cells activated by bsAbs, which engage polyclonal T cells in a major histocompatibility complex (MHC)-independent manner40.
Emerging immunotherapies for patients with NB
Beyond antibody-based immunotherapy, substantial efforts are focused on the development of new immunotherapeutic strategies. Designing effective NB immunotherapies is exceptionally challenging, given the complex and minimally immunogenic nature of the NB tumor microenvironment (TME). The TME is a highly organized ecosystem in which tumor cells are surrounded by diverse cellular components, including immune cells, cancer-associated fibroblasts, endothelial cells, pericytes, and other stromal elements, all of which are embedded within an altered and aberrantly vascularized extracellular matrix41. The cellular components of the TME are dynamic; they therefore co-evolve as NB progresses and often acquire immunosuppressive features that facilitate tumor immune evasion. Whereas early tissue microenvironments and immune surveillance can exert tumor-suppressive effects, NB cells can evade these constraints and progressively remodel the TME to promote disease progression and metastatic outgrowth. Masih and colleagues42 have demonstrated that the NB immune microenvironment strongly correlates with patient outcomes. Specifically, immunologically “cold” NB tumors exhibit diminished survival and poorer prognosis, whereas “hot” tumors correlate with more favorable outcomes.
The identification of actionable tumor-associated antigens (TAAs) is a fundamental and ongoing challenge in the development of effective immunotherapeutic strategies. TAAs are generally self-proteins that are aberrantly expressed during tumorigenesis or are normally restricted to specific developmental stages, and are found in minimal or negligible levels in most healthy normal tissues43. An ideal immunotherapy selectively targets these antigens to achieve tumor control while minimizing deleterious effects on normal tissues.
TAAs expressed on the plasma membrane have long been widely used as therapeutic targets. Extensive efforts have been directed toward identifying clinically relevant TAAs strongly associated with NB progression44. A limited number of NB-associated TAAs, including GD2, glypican-2 (GPC2), New York esophageal squamous cell carcinoma 1 (NY-ESO-1), B7-H3, neural cell adhesion molecule (NCAM), L1 cell adhesion molecule (L1CAM), preferentially expressed in melanoma antigen (PRAME), and paired-like homeobox 2B (PHOX2B), have been identified and incorporated in the development of multiple immunotherapeutic strategies (Table 1). Among these, GD2 is the most established and extensively studied TAA in NB68. Several oncogenic roles of GD2 have been suggested, including enhanced NB cell proliferation, migration, invasion, and metastasis through the modulation of ganglioside biosynthesis and cellular signaling pathways. In addition, GD2 can be released from the cytoplasmic membrane through extracellular vesicles, and elevated levels of circulating GD2 are significantly associated with poorer event-free survival among patients with stage M disease or with MYCN-amplified tumors. Building on extensive research involving GD2, a variety of immunotherapeutic strategies for NB have been developed that specifically target this antigen.
Tumor-associated antigens (TAAs) that can be potentially targeted by NB-specific immunotherapies
Active immunotherapies
Active immunotherapy stimulates the immune system to specifically recognize and eliminate tumor cells, primarily through vaccination. In contrast to conventional vaccines, which are developed to target infectious diseases, cancer vaccines are designed to treat existing cancer or to prevent relapse after initial treatments (Figure 2). An effective cancer vaccine coordinates humoral immunity (B cells) and cell-mediated tumor killing (CD8+ cytotoxic T cells) with CD4+ helper T cells, thus serving as a key regulator enhancing immune activation and promoting durable protection through generation of long-term immune memory. Currently, eight active clinical trials are investigating GD2-targeted vaccines for NB68. Among these GD2-targeted vaccines, a bivalent vaccine (NCT00911560) incorporates two NB-associated TAAs (GD2 and GD3). The inclusion of β-glucan into this bivalent vaccine may enhance the leukocyte-mediated killing of tumor cells. In high-risk NB patients with a history of disease progression, this NB bivalent vaccine has been found to elicit robust antibody responses, and higher anti-GD2-IgG1 titers are associated with improved survival69. Phase 2 data have further indicated that incorporation of oral β-glucan during vaccine priming increases anti-GD2 IgG1 titers among responders carrying the A/A or A/C genotypes of the dectin-1 SNP rs3901533, without additional toxicity70.
Current and emerging immunotherapies for NB. (A) A cancer vaccine stimulates B cells to produce antibodies and CD8+ cytotoxic T cells to kill tumor cells. For maximal effects, a cancer vaccine also activates CD4+ helper T cells, which enhance CD8+ T-cell responses and promote long-term T-cell memory. (B) NK cell adoptive transfer enhances the susceptibility of NB to recognition and elimination by cytotoxic T cells. (C) CAR-T cells recognize and induce direct tumor cell killing by granzyme and perforins in an MHC-independent manner. (D) CAR-NK cells mediate efficient direct lysis of NB cells after recognition. (E) CAR-N triggers tumor phagocytosis and promotes transcription of various transcription factors that facilitate the secretion of cytokines, thus leading to activation of various immune cells. CAR-N also mediates tumor lysis through neutrophil extracellular trap (NET) production and reactive oxygen species (ROS) secretion. (F) CAR-M enhances tumor phagocytosis, facilitates antigen presentation, and promotes transcription of pro-inflammatory genes, thus leading to the activation of T-cell-mediated immunity against the tumor. The release of proinflammatory elements also counteracts the immunosuppressive TME. (G) mAb binding to NB cells initiates a series of killing mechanisms, including antibody-dependent cell-mediated cytotoxicity (ADCC), complement-dependent cytotoxicity (CDC), and antibody-dependent phagocytosis (ADP) to eliminate cancer cells. (H) bsAbs act as a biophysical bridge between two antigens, thus enabling interactions between the tumor cells and immune cells that lead to a synergistic tumor-specific immune response. CAR: chimeric antigen receptor; DC: dendritic cell; M: macrophage; N: neutrophil; NK: natural killer. Figure created with BioRender.com.
Beyond the GD2-targeted vaccine, a phase 1 trial (NCT01241162) evaluated a combination chemoimmunotherapy regimen consisting of decitabine and a dendritic cell vaccine containing overlapping peptide mixtures derived from full-length NY-ESO-1, MAGE-A1, and MAGE-A371. Six of nine patients demonstrated a response after vaccination. Overall, the treatment regimen was well tolerated, and decitabine-related myelosuppression was the primary toxicity in half the patients. Moreover, Tuccitto72 has reported that vaccination with the HLA-A*0201-restricted peptide NY-ESO-1157-165(V) enhances infiltration of functionally active, NY-ESO-1-specific T cells into the TME in patients with advanced NB. Together, these studies highlight the potential of cancer vaccines to elicit antigen-specific immune responses in NB and support continued vaccine development.
However, NB is characterized by a “cold” TME, and the ability of tumor cells to establish an immunosuppressive milieu that promotes immune evasion and limits the effectiveness of cancer vaccine-induced antitumor responses. Moreover, although cancer vaccines show promise in NB management, their efficacy is largely restricted to immunocompetent patients. Unfortunately, most patients with NB undergo intensive chemotherapy, which results in profound immune cell depletion and treatment-induced immunosuppression, and significantly decreases the effectiveness of active immunotherapies. Consequently, therapies that provide immediate antitumor activity independent of the patient’s immune competence may serve as a complementary strategy in this setting.
Passive immunotherapies
Passive immunotherapies exert immediate yet transient therapeutic effects through the administration of exogenous agents, including mAbs, bsAbs, and adoptive cell transfers using genetically engineered immune cells. Unlike active immunotherapy, passive immunotherapy does not rely on the patient’s immune system to generate long-term immunity but instead facilitates a rapid antitumor effect through direct immune-mediated mechanisms. mAbs are a key form of passive immunotherapy used in the current management of NB, particularly in high-risk patients, and have demonstrated promising clinical efficacy.
Chimeric antigen receptor (CAR)-based strategies
Beyond the use of antibodies in current NB immunotherapy, CAR-T cell therapy represents a major advancement in personalized immunotherapy. Although CARs were first proposed in the late 1980s, these synthetic antigen receptors were not evaluated clinically as a cancer treatment until 200673. These synthetic receptors can recognize pre-defined surface antigens with a high degree of specificity and later effectively activate T cells in an MHC-independent manner. Because CAR-T cells do not require antigen processing and presentation during tumor eradication, they are insensitive to the tumor evasion mechanism associated with MHC loss.
CAR contains an extracellular domain for tumor antigen recognition, which frequently comprises single-chain variable fragment (scFv), a transmembrane domain, and an intracellular domain that mediates immune cell activation74 (Figure 3). The scFv portion of a CAR consists of the variable heavy (VH) and variable light (VL) chains of an antibody joined by a linker. This extracellular domain is then anchored in the transmembrane domain by an ~15-residue peptide spacer. The transmembrane domain is responsible for CAR stabilization. The intracellular domain, which is responsible for inducing immune responses after antigen recognition, comprises the T cell receptor CD3ζ signaling chain and optional tandem co-stimulation domains, such as the 4-1BB or CD28 signaling domains. Binding of ligands to CARs induces dimerization of the transmembrane domain, thus leading to phosphorylation of immunoreceptor tyrosine-based activation motifs (ITAMs) on CD3ζ75. This phosphorylation triggers a signaling cascade and upregulates the transcription and translation of various genes primarily involved in CAR-immune cell proliferation, cytokine secretion, and tumor cell death (e.g., perforin and granzyme).
Schematic diagram of different generations of CAR design. The extracellular domain contains a binding region responsible for antigen recognition and is connected to the transmembrane domain via a hinge region. Upon antigen engagement, the intracellular domain is activated, thus initiating downstream signaling. First-generation CARs consist of an scFv extracellular domain and a CD3ζ TCR signaling intracellular domain. Second-generation CARs incorporate an scFv extracellular domain, a single costimulatory domain (usually CD28, 4-1BB, and OX40), and CD3ζ. Third-generation CARs include an scFv extracellular domain, two costimulatory domains, and CD3ζ. Fourth-generation CARs contain additional genetic elements that enable the expression of proteins to modulate the TME. Fifth-generation CARs build upon second-generation design by incorporating a truncated IL-2 receptor β (IL-2Rβ) domain and a STAT3-binding YXXQ motif, thus enabling three antigen-dependent signals, including TCR signaling via CD3ζ, co-stimulatory signaling (e.g., CD28), and cytokine signaling through the JAK-STAT3 pathway. Figure created with BioRender.com.
First-generation CARs contain the CD3ζ signaling domain, whereas second-generation CARs incorporate an additional co-stimulatory domain, such as CD28, 4-1BB, and OX40, fused with CD3ζ (Figure 3). The inclusion of co-stimulatory domains enhances CAR-T persistence and functional potency. Building on this concept, third-generation CARs were developed to incorporate two or more co-stimulatory domains, thereby further augmenting activation, proliferation, and cytokine production of CAR-expressing immune cells. Hannah and colleagues76 have demonstrated that third-generation CAR-T cells exhibit enhanced phosphorylation of signaling proteins involved in cell cycle regulation, cell adhesion, and exocytosis, thus suggesting stronger intracellular signaling than second-generation CARs. Table 2 provides a comprehensive overview of CAR-based immunotherapeutic strategies for the management of NB, including the associated toxicity profiles across different immune effector platforms, whereas Table 3 summarizes the currently active CAR-based clinical trials in NB registered on ClinicalTrials.gov.
Summary of CAR-based immunotherapy strategies for NB management
Summary of currently active or recently active CAR-based clinical trials for NB, identified through ClinicalTrials.gov
Notably, CAR-T cells remain the predominant modality in clinical translation, accounting for all 26 ongoing trials, whereas alternative platforms, such as CAR-natural killer (NK), CAR-NKT, CAR-neutrophils (CAR-N), and CAR-macrophages (CAR-M), are largely confined to preclinical or early exploratory stages. Across the limited subset of trials with reported outcomes (n = 7), CAR-T therapies in NB appear to have a generally manageable safety profile with predominantly low-to-moderate adverse events, including cytokine release syndrome (CRS) and peripheral or neurologic pain77–81. However, the reporting of dose-limiting toxicities (DLTs) varies across studies, which are limited by small sample sizes and data heterogeneity77,79–84.
Clinical efficacy remains variable, and most patients experience stable disease (SD) or progressive disease (PD) and relatively low complete response (CR) rates. Therefore, although current CAR-T approaches are feasible and generally well tolerated in patients with NB, their therapeutic efficacy remains suboptimal. Collectively, these findings indicated that CAR-based immunotherapy in patients with NB demonstrates a favourable safety profile in early-phase studies and clear translational momentum. The following sections therefore examine individual CAR-engineered immune cell types in NB.
CAR-T cells
Given the success of CAR-T therapy in hematologic malignancies, CAR-T cells have emerged as a promising immunotherapeutic approach for NB, which was also the first pediatric cancer to be targeted with CAR-T cells in a clinical trial. In this context, a clinical trial of first-generation CAR-T cells incorporating the same antigen recognition domain (i.e., scFv) as dinutuximab demonstrated several clinical responses without signs of on-target, off-tumor toxicity82. In an updated report on a phase 1 clinical trial in NB patients treated with Epstein-Barr virus-specific T cells and CD3-activated first-generation GD2-CAR-T cells, 5 of 11 remained disease-free at least 10 years after their final infusion85. Complete remission was observed in 3 of 100 patients, including one individual who has remained in remission for >18 years. However, because first-generation CAR-T cells were limited by poor expansion and insufficient long-term persistence, second- and third-generation anti-GD2 CAR-T cells were developed79,81. Subsequent studies further demonstrated that targeting NB with anti-GD2 CAR-T cells is a feasible and generally safe strategy. In a recent clinical trial report, GD2-CART01 has shown potential for sustained antitumor activity and was well tolerated in patients with relapsed or refractory high-risk NB, with manageable treatment-related toxicity77. To mitigate the risk of severe toxicity, an inducible caspase-9 suicide gene was incorporated into the CAR-T design, thus enabling controlled elimination of GD2-CART01 cells if needed.
Beyond GD2, Yarmarkovich and colleagues65 have developed HLA-restricted (HLA-A*24:02 and HLA-A*23:01), PHOX2B peptide-centric (PC)-CAR-T cells exhibiting potent tumor killing and growth inhibition in both in vitro study and in vivo NB patient-derived xenografts. PC-CARs may offer an additional advantage of initiating a feed-forward loop of MHC upregulation and increasing antigen density beyond that with conventional CARs, thereby potentially enhancing tumor recognition and immune responses. This approach is currently being evaluated in a phase 1 clinical trial assessing the safety of PHOX2B PC-CAR T cells in patients with advanced, high-risk NB (NCT07007117).
Rather than targeting a single TAA, bi-specific CARs have been developed to recognize two TAAs simultaneously. A novel approach using a synthetic Notch (SynNotch) system functions as a gating mechanism, in which expression of a CAR targeting a TAA depends on a transactivating signal initiated by engagement of a second TAA86. Binding of the first antigen to the SynNotch receptor induces receptor cleavage and release of intracellular transcription factors, which translocate to the nucleus and drive transcription of the CAR gene. The newly synthesized CAR is then expressed on the cell surface. After recognition of the second TAA, T cells become fully activated and eliminate tumor cells through the release of perforin and granzymes. GD2/B7-H3 CAR-T cells developed by Moghimi et al.87 use GD2 as the gating antigen and B7-H3 as the target. The SynNotch-based approach minimizes off-tumor toxicity while enabling precise regulation of T-cell activation and effector function. In a preclinical study, GD2/B7-H3 CAR-T cells have demonstrated high specificity and efficacy in suppressing tumor growth in vitro and in metastatic xenograft mouse models. These effects might be attributable to enhanced metabolic fitness and diminished exhaustion with respect to those of conventional CAR-T cells.
In 2023, ALK/B7-H3-CAR-T cells were developed by combining an ALK-targeting CAR with anti-B7-H3 or anti-GD2 chimeric costimulatory receptors in an “AND” gate configuration to address the challenge of low-antigen density88. This co-CAR strategy enables effective recognition of single- and double-antigen targets, thereby enhancing cytokine production and proliferative responses against ALK-low NB cells and outperforming conventional second-generation ALK CAR-T cells.
Collectively, these strategies provide a strong foundation for the continued development of CAR-T therapy. CAR-T cells may offer superior tumor control to mAbs, because of their enhanced tumor binding through polyvalent CAR expression on the T cell surface; additional cytotoxicity mechanisms that promote effective tumor clearance; and persistence, which may help prevent relapses82. Despite these advantages, several limitations remain. An immunosuppressive TME can impair CAR-T cell trafficking to tumor sites, whereas off-target effects and CAR-T cell malfunction induced by an immunosuppressive TME cannot be overlooked. Therefore, other immune effector cells with intrinsic antitumor effects, including NK cells, NKT cells, neutrophils, and macrophages, are being actively investigated as alternative platforms for CAR-based immunotherapy.
CAR-NK cells
NK cells are a major effector population in immunotherapy, particularly in mAb-based treatment that facilitates tumor clearance through NK cell mediated ADCC. NB cells significantly upregulate the surface expression of MHC class I after exposure to activated NK cells, thereby increasing susceptibility to recognition and elimination by cytotoxic T cells89. However, NK cell cytopenia is a common challenge in patients with a high tumor burden and following chemotherapy. Data from multiple clinical trials have indicated that patients receiving allogeneic NK cells transfer may experience improved progression-free survival90–92. In addition, reconstituted endogenous NK cells exhibit enhanced cytotoxic activity with respect to pretherapy levels without causing additional identifiable toxicity.
Building on the intrinsic antitumor activity and favorable safety profile of NK cells, CAR-NK has emerged as an alternative in NB immunotherapy. Unlike CAR-T cells, CAR-NK cells are generally associated with milder toxicity, probably because of innate immune characteristics and a higher activation threshold that contribute to more controlled cytokine production93. In addition, CAR-NK cell therapy has gained considerable attention as a potential “off-the-shelf” approach, because these cells can be generated in advance from sources such as induced pluripotent stem cells (iPSCs), thus enabling the production of readily available therapeutic product.
The concept of CAR-NK therapy was first demonstrated with NK-92 cells, a human NK cell line94. GD2-targeted CAR-NK cells have been shown to mediate efficient lysis of NB cells in vitro and to significantly delay tumor growth, while prolonging survival in drug-resistant GD2+ NB xenograft models95. These antitumor effects are likely to be driven by direct cytotoxic activity of CAR-NK cells independently of conventional NK receptor-ligand interactions. In addition to GD2, ex vivo-expanded ROR1-targeted CAR-NK cells have demonstrated significantly enhanced cytotoxicity against ROR1+ NB cells96. When combined with N-803, an IL-15 superagonist, ROR1-CAR-NK cells have been found to further improve the survival of human ROR1+ NB xenografted NSG mouse models with respect to that observed in control groups.
CAR-NKT cells
NK and NKT cells share many phenotypic and functional features in antitumor immunity97. However, they exhibit distinct patterns of dysfunction in advanced stages of cancer. NK cells tend to become senescent, whereas NKT cells develop features of immune exhaustion with impaired cytotoxicity capability. NKT cells represent an innate-like T-cell subset that uses a semi-invariant T-cell receptor (TCR) and mediates potent antitumor activity by directly recognizing CD1d-restricted antigens and by producing cytokines, such as IL-2, that indirectly enhance NK cell cytotoxicity98.
The first-in-human phase 1 trial (NCT03294954) using ex vivo-expanded autologous NKT cells co-expressing a GD2-specific CAR with interleukin 15 (IL-15) (GD2-CAR.15) demonstrated that these cells can expand in vivo after infusion, traffic to tumor sites (including metastatic lesions), and mediate antitumor responses in heavily pre-treated patients with relapsed or refractory NB83,84. However, increased expression of BTG anti-proliferation factor 1 (BTG1), a key regulator in hyporesponsiveness in exhausted NKT and T cells, has been observed after GD2-CAR.15 NKT cell treatment. Subsequent in vitro and in vivo studies have indicated that BTG1 knockdown enhances the functional activity of GD2-CAR.15 NKT cells and promotes effective elimination of metastatic NB in mouse models. Accordingly, current efforts are focused on targeting BTG1 to improve the efficacy of GD2-CAR.15 NKT cell therapy.
CAR-neutrophils
Neutrophils, the most abundant leukocytes in the circulation, serve as key effector cells of the innate immune system against invading pathogens. Increasing evidence indicates that neutrophils also have a critical role in regulating tumor immunity, because their hyperinflammatory phenotypes can contribute to tumorigenesis and immune modulation within the TME. Neutrophils’ natural ability to infiltrate tumors makes them an attractive alternative cellular platform for immunotherapy, complementing existing CAR-lymphoid-based approaches. Recent studies have shown that human pluripotent stem cell (hPSC)-derived CAR neutrophils exhibit an antitumor phenotype and enhanced efficacy in vitro and in vivo99. In agreement with these findings, iPSC-derived GD2 CAR neutrophils (GD2-CAR-N) have demonstrated cytotoxic activity against GD2-expressing NB cells in vitro, accompanied by elevated production of IFNγ, TNFα, IL-13, IL-4, and IL-10 upon tumor engagement100. Importantly, the limited toxicity observed in in vivo studies further supports the feasibility of CAR-N as a therapeutic strategy in NB.
CAR-macrophages
In addition to direct cytotoxicity against NB, effective tumor control requires targeting the immunosuppressive TME. Among immune cells, macrophages are a dominant component of the NB TME with central roles in shaping immune suppression and tumor progression. Macrophages are key regulators of the immune system, contributing to pathogen and tumor clearance, removal of cellular debris, and modulation of inflammatory responses101. Macrophages’ natural ability to infiltrate tumors, together with the functional plasticity within the TME, makes them an attractive candidate for NB immunotherapy.
Tumor-associated macrophages (TAMs) are among the most abundant innate immune cells that constitute as much as 50% of the cell mass within the TME of many solid tumors102. In the TME, macrophages can adopt distinct polarization states in response to local cytokines and signals, broadly classified as classically activated (M1) and alternatively activated (M2). M1 macrophages typically exhibit pro-inflammatory properties and antitumor functions, thereby contributing to antibacterial and tumoricidal activity. In contrast, M2 macrophages are involved primarily in tissue remodeling and immunomodulation, and are often associated with pro-tumor characteristics, including promoting tumor growth and immune evasion. TAMs frequently display an M2-like phenotype, thereby contributing to an immunosuppressive TME supporting tumor progression103.
Building on these features, CAR-engineered macrophages (CAR-M) can leverage tumor-infiltrating, phagocytic, and TME-modulating abilities to achieve effective targeted immunotherapy. Binding of tumor antigen to CARs on the surfaces of engineered macrophages activates a cascade of antitumor mechanisms104. This interaction triggers tumor phagocytosis and promotes the activation of transcription factors, such as NF-κB, thus leading to the secretion of cytokines and chemokines, including IL-12, which in turn enhance T cell-mediated antitumor immunity. In addition, the release of proinflammatory mediators, such as interferons, can counteract the immunosuppressive TME by upregulating the MHC class I and II expression on tumor cells, thereby improving antigen presentation to T cells105. The CAR-M construct developed by Kang et al.106 incorporates an interferon gamma (IFNγ) gene to promote repolarization of macrophages from an M2 phenotype (pro-tumor) to an M1 phenotype (antitumor), thus resulting in enhanced tumoricidal activity.
The potential of CAR-M in NB treatment was first demonstrated in 2023. Specifically, hPSC-derived GD2-targeted CAR-M were found to exhibit superior cytotoxic effects against NB cells in in vitro and in vivo models107. Although studies of CAR-M in NB are limited, the fast-track FDA approval of the first-in-human phase 1 clinical trial (CT-0508) in 2021, which evaluated CAR-M therapy in patients with recurrent or metastatic HER-2 overexpressing solid tumors, further supports the feasibility of this approach in solid tumor treatment. CT-0508 has been reported to be safe and well tolerated with no dose-limiting toxicities or major organ toxicities observed108. Notably, antitumor immunity was induced, as evidenced by increasing T-cell clonality and significant expansion of newly emergent T-cell clones within the TME, along with enhanced CD8+ T-cell activation. Therefore, CAR-M might provide a promising antigen-specific adoptive cell therapy for NB, following the paradigm established by CT-0508.
Advanced CAR engineering and functional optimization
Current CAR-based NB immunotherapies are largely based on first-, second-, and third-generation CAR designs, which have progressively improved immune cell activation, persistence, and effector function through the addition of costimulatory domains and optimized intracellular signaling. However, despite these advances, their clinical efficacy in solid tumors remains limited, largely because of the highly immunosuppressive TME and insufficient sustained immune cell activity.
Given the heterogeneity and immunosuppressive nature of TME, fourth-generation CARs were engineered to incorporate genes that modulate TME, such as IL-2, IL-15, and IL-12109. For example, a fourth-generation CAR-T expressing the IL-23 p40 subunit has shown enhanced antitumor capacity in vitro, elevated granzyme B expression, diminished PD-1 levels, and robust efficacy in syngeneic solid tumor mouse models, including NB110. Beyond targeting the immunosuppressive TME, CAR constructs incorporating IL-6 and TNF-α have also been explored for autoimmune diseases, such as rheumatoid arthritis (RA), in which the CAR constructs have shown robust in vivo expansion, complete B-cell depletion, clearance of RA-related autoantibodies, and sustained clinical responses111. Fifth-generation CARs have built on second-generation designs, by incorporating a truncated IL-2 receptor β domain (IL-2Rβ) and a STAT3-binding tyrosine-X-X-glutamine (YXXQ) motif112. This truncated IL-2Rβ-YXXQ domain links JAK-STAT signaling directly to CAR engagement, thus enabling antigen-dependent cytokine signaling independently of exogenous IL-2. After antigen recognition, these CARs deliver three coordinated signals, including TCR signaling through CD3ζ (signal 1), costimulatory signaling through CD28 (signal 2), and cytokine signaling via the JAK-STAT3 pathway (signal 3). The integration of these signals enhances CAR-T cell persistence and antitumor activity while maintaining a favorable safety profile in preclinical models.
Beyond incorporating new signaling domains, fine-tuning the structural components of CARs can further improve targeting efficacy113. Antigen specificity remains a central challenge in CAR-based therapy. Beyond the selection of appropriate TAAs, factors such as antigen expression patterns, spatial localization, and functional relevance are critical determinants of therapeutic success. After epitope selection, the stability of scFv must be carefully considered, because instability can lead to receptor misfolding, segregation, tonic signaling, and immune cell exhaustion. The interaction between the TAA and the scFv is largely governed by binding affinity, which directly influences CAR activation, expansion, survival, and persistence. Although high-affinity scFvs may enhance antigen sensitivity and induce robust T cell activation, excessively high affinity may increase the risk of on-target/off-tumor toxicity. In contrast, moderate-affinity scFvs often provide a more favorable balance between therapeutic efficacy and safety, and therefore lead to improved clinical outcomes114.
The hinge region, which connects the extracellular and intracellular domains, also has a critical role in providing structural flexibility and stability. Hinge length influences CAR function by modulating antigen accessibility, signal transduction, and receptor conformation. Hinge length can also influence immune exhaustion and signaling dynamics115. Therefore, optimization of hinge length and design with antigen selection and epitope targeting is essential for maximizing CAR efficacy.
Importantly, most CAR designs have been largely optimized in T cells and are often extrapolated to other immune cell types. However, such direct translation may be suboptimal, because the functional properties and intracellular signaling pathways differ substantially among immune cell populations. Therefore, cell type-specific optimization of CAR constructs is necessary to achieve optimal therapeutic performance. In this context, phage display technologies provide a powerful platform for identifying scFv binders with optimized affinity and specificity to target antigens, facilitating the development of next-generation CAR therapies tailored to diverse immune cell populations.
The need for combination therapies to improve NB outcomes
Despite the demonstrated efficiency of these immunotherapies across multiple diseases, including NB, accumulating evidence suggests that single-agent approaches are often insufficient to achieve durable clinical responses, because of tumor immune evasion. Tumor evasion is highly dynamic and complex, involving various inhibitor pathways that suppress antitumor immunity116. These mechanisms include the production of immunosuppressive cytokines, regulation of immune checkpoints, such as PD-1/PD-L1 and CTLA-4, impairment of antigen presentation and recognition, and establishment of an immunosuppressive TME by creating physical and biochemical barriers that hinder immune cell infiltration and activation. Consequently, targeting a single immune axis carries a high risk of incomplete immune activation and the development of adaptive resistance.
The limitations of single-agent immunotherapies are exemplified by the clinically established anti-GD2 mAb (i.e., dinutuximab), to which approximately 40% of patients do not respond or develop resistance due to tumor evasion117. Anti-GD2 mAb treatment can induce a phenotypic shift in tumor cells from an adrenergic to a mesenchymal state through PRC2 pathway-mediated downregulation of ST8SIA1, thus resulting in decreased GD2 expression and diminished treatment responsiveness118. Notably, inhibition of EZH2 has been shown to reverse this mesenchymal transition, by restoring adrenergic-like phenotypes through transcriptional, epigenetic, and chromatin remodeling. This reprogramming leads to re-expression of ST8SIA1, increased GD2 expression, and renewed sensitivity to anti-GD2 mAb.
Beyond tumor-intrinsic resistance mechanisms, NB-derived small extracellular vesicles contribute to therapeutic failure by suppressing NK cell maturation and infiltration into the TME, thereby limiting dinutuximab-induced NK cell-mediated ADCC119. However, combination therapy with dinutuximab and tipifarnib, an inhibitor of small extracellular vesicle secretion, enhances antitumor efficacy, inhibits tumor growth, and prevents the development of an immunosuppressive TME, thus highlighting the importance of simultaneous targeting of tumor cells and the immune microenvironment.
Evidence from other solid tumors further supports the need for combination immunotherapies to overcome tumor plasticity, immunosuppression, and therapeutic resistance. For example, combined GD2-targeted CAR-T cells and FLASH radiotherapy significantly enhance antitumor efficacy in medulloblastoma, a malignant pediatric brain tumor120. FLASH radiotherapy modulates lipid metabolism within the TME, thereby reprogramming TAM toward a pro-inflammatory M1 state and reversing TAM-mediated immunosuppression. This TME remodeling facilitates CAR-T cell infiltration and activation and significantly decreases tumor burden.
Beyond radiotherapy, combinations of immunotherapy and chemotherapy have also demonstrated improved antitumor effects over immunotherapy alone. CAR-M therapy targeting c-Met, an abundantly expressed TAA, enhances tumor phagocytosis and cytotoxic activity in both in vitro and in vivo models of pancreatic cancer by promoting a pro-inflammatory immune state121. Notably, this approach exhibits synergistic effects when combined with chemotherapy, thus further enhancing therapeutic efficacy.
Beyond combining immunotherapy with conventional treatments, dual targeting within immunotherapy offers another potential strategy. Emerging NB immunotherapies, such as GD2/B7-H3 CAR T and ALK/GD2 CAR T, use the Boolean logic gates design (“AND”, “OR”, and “NOT”) to target multiple TAAs (Figure 4). This logic-gated approach is aimed at decreasing tumor antigen escape while minimizing on-target, off-tumor toxicity. Beyond logic-gated co-CAR design, dual CAR-immune effector strategies, in which distinct CAR-engineered immune cells independently target different TAAs, may also be considered. By leveraging complementary antitumor mechanisms, such approaches have the potential to improve tumor coverage, increase therapeutic durability, and reduce resistance associated with single-target or single-cell-type therapies.
Boolean logic gate strategies in CAR-based immunotherapies. (A) AND gate: CAR-immune cells are activated only when both antigens are simultaneously present, thus enabling highly specific tumor recognition and decreasing on-target, off-tumor toxicity. (B) OR gate: CAR-immune cells are activated upon recognition of either antigen, improving tumor coverage and helping overcome antigen heterogeneity and escape. (C) NOT gate: CAR-immune cell activation is suppressed by the presence of a designated antigen associated with an inhibitory signaling domain, thereby protecting normal tissues that express this marker.
Collectively, these findings underscore the limitations of single-agent immunotherapies and provide a strong rationale for combination strategies that concurrently target both tumor plasticity and immunosuppressive TME. Future immunotherapy development in NB should prioritize mechanism-driven combination approaches, rather than further optimization of single-agent immunotherapies, to achieve durable therapeutic benefit.
Future perspectives
NB remains a clinically challenging disease, particularly in high-risk patients and in the relapse setting, in which conventional therapies often fail to achieve ideal disease control. Immunotherapy has shown promise by enabling antigen-specific targeting of tumor cells. CAR-based approaches offer unique advantages, including enhanced persistence, adaptability, and the potential for long-term improved disease management. The expansion of CAR-based platforms from T cells to other immune cells, such as NK, NKT, neutrophils, and macrophages, highlights the importance of engaging diverse immune effectors to address the complex biology of NB.
Despite advances in CAR engineering, tumor heterogeneity, antigen escape, and the immunosuppressive TME continue to limit the efficacy of single-agent therapies. Future treatment strategies should emphasize mechanism-driven combination approaches, including multi-antigen targeting, logic-gated CAR designs, and the integration of CAR-based therapies with radiotherapy, chemotherapy, or immune-modulatory agents.
Advances in CAR delivery are equally important for clinical feasibility and scalability122. CAR constructs can be introduced into immune cells through physical methods (e.g., electroporation and mechanoporation), chemical methods [e.g., lipid nanoparticles (LNPs) and poly β-amino esters], and biological methods (e.g., lentivirus, retrovirus, and adenovirus). Conventional ex vivo approaches are labor-intensive, costly, and time-consuming, and they require immune cell collection from patients or donors, in vitro engineering, and reinfusion after lymphodepletion. In contrast, CAR-engineered immune cells derived from iPSCs provide a scalable alternative enabling the generation of large quantities of functional cells without repeated donor collection100.
Despite the demonstrated effectiveness of these strategies, conventional ex vivo manufacturing remains a bottleneck. Consequently, rapid, scalable, off-the-shelf in vivo engineering strategies are needed. Among emerging delivery systems, LNP-based platforms have gained substantial attention after successful clinical application in an mRNA COVID-19 vaccine. LNPs offer a flexible, transient, and potentially repeatable approach that bypasses ex vivo cell manipulation. Precise targeting of LNPs to a specific immune cell population is essential to minimize off-target delivery and unintended interactions with tumor or non-target cells. Cargo selection is also critical; i.e., CAR-encoding DNA enables long-term expression but carries a risk of genome integration, whereas CAR-encoding RNA enables transient expression with minimal risk of insertional mutagenesis but may require repeated dosing122. Emerging RNA platforms, such as self-amplifying RNA and circular RNA, have demonstrated prolonged gene expression in a vaccine setting and may be adapted for CAR delivery123.
Key challenges remain, including achieving immune cell-specific targeting, improving delivery efficiency, ensuring durable yet controllable CAR expression, and minimizing systemic immune activation. Addressing these issues will be essential for translating CAR-based therapies into clinically feasible, off-the-shelf treatments.
Conclusions
Overall, next-generation immunotherapies for NB require a multifaceted approach that prioritizes immune cell-specific CAR design, advanced delivery platforms, and rational integration with complementary therapies to achieve durable, safe, and broadly accessible treatment for patients.
Conflict of interest statement
No potential conflicts of interest are disclosed.
Author contributions
Conceptualization: Ke-En Tan.
Investigation: Ke-En Tan, Kok Siong Yeo.
Writing of original draft: Ke-En Tan.
Writing-review and editing: Ke-En Tan, Kok Siong Yeo, Yat-Yuen Lim, Shizhen Zhu.
Supervision: Yat-Yuen Lim, Shizhen Zhu.
- Received March 3, 2026.
- Accepted May 20, 2026.
- Copyright: © 2026, The Authors
This work is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License.
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.
- 46.
- 47.
- 48.
- 49.
- 50.
- 51.
- 52.
- 53.
- 54.
- 55.
- 56.
- 57.
- 58.
- 59.
- 60.
- 61.
- 62.
- 63.
- 64.
- 65.↵
- 66.
- 67.
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.
- 79.↵
- 80.
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.↵
- 122.↵
- 123.↵
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.