Pediatric solid tumors account for approximately 50% of pediatric cancers and heterogeneous diseases [1]. The survival rates of patients with childhood cancers have improved dramatically over the decades and 80% of patients achieve long-term remission after treatment. However, treatment outcomes for pediatric patients with relapsed/refractory (R/R) solid tumors are still poor. Furthermore, the most common treatment for pediatric cancer is non-specific and aggressive cytotoxic chemotherapy, which is related with excessive toxicities, long-term toxicities and decreased quality of life. Novel therapeutic strategies are desperately needed to improve survival rates and to reduce toxicities. Cancer immunotherapy is a type of treatment that works by leveraging the patient’s own immune system. Although immunotherapy has been proven to be effective in some pediatric hematologic malignancies and solid tumors, especially neuroblastoma (NBL), there is still a long way to go before it can be effectively used to treat pediatric solid tumors. This review demonstrated the immunological background of why adult-like approaches have limitations in pediatric tumors. It also summarizes various aspects of immunotherapy currently under investigation and ongoing efforts for pediatric solid tumors, focusing on monoclonal and bispecific antibodies, checkpoint inhibition, chimeric antigen receptor T cell (CAR-T) therapy, and cancer vaccine therapy.

1) Tumor microenvironment

There are some similarities and differences between pediatric and adult cancers from an immunological perspective. Commonly, tumors are composed of infiltrating stromal cells and immune cells in addition to malignant cells. In pediatric cancer, tumor-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs) are two of the most prominent infiltrating cells [2]. On the other hand, the tumor active lymphocytes are insufficient, hence the notion of the “cold tumor” [3]. TAMs and MDSCs originate from myeloid precursors in the bone marrow and are major contributors to the tumor microenvironment (TME) [4]. They have various roles that affect and regulate tumorigenesis, vasculogenesis, tumor cell growth, extracellular matrix deposition/remodeling and response to therapy [5,6]. MDSCs produce immunosuppressive factors such as interleukin-10 (IL-10), prostaglandin E2, and TGF-β [7]. CD4+FOXP3+ regulatory T cells (Tregs) and MDSCs inhibit cytotoxic T lymphocyte (CTL) proliferation and activation [8]. The chronic cytokine and chemokine production in TME create a contradictory inflamed and immune-suppressive environment in both adult and pediatric cancers [2]. T cells are exposed to chronic tumor antigen stimulation, making them exhausted and ineffective [9,10]. In order for immunotherapy to succeed, it is necessary to overcome this TME.

2) Paucity of neo-antigens

Most pediatric cancers originate from embryonic cells rather than from epithelial cells. Most pediatric solid tumors have either one or just a few driver mutations [11, 12]. Additionally, compared to adult cancer, pediatric cancer has fewer genotoxic environmental stressors and less cell division, so there are fewer background mutations [2]. Instead, 8-10% of childhood cancer patients have germline mutations that predispose to cancer [13]. Chromosome rearrangements that can activate proto-oncogenes or inactivate tumor suppressor genes are more common in pediatric cancers [14]. Thus, pediatric tumors have significantly lower mutation burdens and express less neo-antigens, which cause decreased response to immunological targeting [15]. However, developing new neo-antigens in pediatric cancer will provide more potential therapeutic targets [16]. When analyzing the mutation spectrum of NBL at diagnosis compared to that at relapse, it was discovered that the overall mutation burden was not high, but increased significantly at relapse [17]. These results show that mutation analysis in recurrent solid tumors can help identify treatment targets.

Immunotherapy is functionally classified into two treatment categories: direct utilization of the immune system and modulation of the immune system.

1) Direct utilization of the immune system

Monoclonal antibody (mAb) therapy is utilized by targeting tumor-associated antigens (TAAs). Target cancer cells are removed through Fcg receptor-mediated antibody-dependent cell-mediated cytotoxicity (ADCC) or complement-dependent cytotoxicity [18,19]. The most impressive mAb in childhood cancer is the one targeting the GD2 diganglioside, which is universally expressed in NBL. In addition to NBL, GD2 is also expressed in melanoma, soft tissue sarcoma, osteosarcoma, desmoplastic small round cell tumor (DSRCT), small cell lung cancer, and various other embryonal tumors [20].

Currently, high-risk NBL treatment is composed of induction chemotherapy and surgery, consolidation therapy with high-dose chemotherapy and autologous stem cell transplantation (ASCT) and radiotherapy, and post-consolidation therapy to treat minimal residual disease. Dinutuximab (ch 14.18) and naxitamab (humanized 3F8), chimeric anti-GD2 antibodies have improved the survival of high-risk NBL patients. Yu AL, et al. showed that combination therapy with isotretinoin, ch14.18, granulocyte-macrophage colony stimulating factor (GM-CSF), and IL-2 significantly improves event free survival and overall survival than isotretinoin alone, which was the standard treatment after ASCT at the time [21]. Patients enrolled in HR-NBL/SIOPEN trials who have completed induction chemotherapy, ASCT, and radiotherapy, were randomly assigned either to the dinutuximab-β group or the dinutuximab-β with IL-2 group. It was found that the addition of IL-2 did not improve the survival rate, but increased the risk of side effects such as hypersensitivity, capillary leak, and fever [22]. We may guess that ADCC via natural killer (NK) cells is not the major contributor of anti-GD2 mAbs, given that IL-2 augments NK cell activity [23,24]. The common adverse effects of anti-GD2 Abs are neuropathic pain, fever, hypersensitivity and on-target off-tumor adverse effects.

Anti-GD2 Ab has been mainly effective in minimal residual disease of NBL. In R/R NBL, a combination treatment of chemotherapy and anti-GD2 therapy showed the possibility of improving the survival rate. In a phase 2 study of patients with R/R NBL, patients were randomly assigned to either the temsiroimus or the dinutuximab group with irinotecan and temozolomide. The object response rate was significantly higher in the dinutuximab group [25].

There are other mAb targets such as insulin growth factor 1 receptor (IGF-1R), human epidermal growth factor receptor 2 (HER2) oncogene and B7-H3, a surface immunomodulatory glycoprotein B7 homolog 3 protein. Enoblituzumab, B7-H3-targeting antibody is underway in phase 1 trials for adult refractory solid tumors (NCT 02628535).

The bi-specific T-cell engagers (BiTE) comprise two binding domains; one is a single-chain variable fragment that binds to the tumor and the other fragment engages an activating receptor on T cells [26]. BiTE technology activates a T cell response by binding to CD3 on T cells. Blinatumomab, a CD19/CD3 BiTE, showed efficacy and safety in R/R pediatric acute lymphoblastic leukemia (ALL) patients [27]. Although no approved BiTEs are yet available in pediatric solid tumors, research on the development of BiTE is ongoing. Previously, anti-GD2 murine 5F11-single-chain variable fragment (scFv) and humanized anti-CD3 OKT3-scFv BiTE were developed for patients with NBL [28]. Cheng M, et al substituted the 5F11-scFv into the higher affinity humanized 3F8-scFv, showing increased tumor cell killing in vitro [29]. A phase I/II clinical trial using anti-GD2 BiTE is ongoing (NCT 02173093). A multicenter phase 1 study of solitomab, a bispecific epithelial cell adhesion molecule (EpCAM)/CD3 T-cell engager, was performed in adult refractory solid tumors. Dose limiting toxicities prevented dose escalation to potentially therapeutic levels [30].

Normal T-cell functions are controlled with T-cell receptor (TCR), co-stimulatory and inhibitory signals. The immune system distinguishes cancer cells from normal cells and activates cytotoxic T cells to lead the apoptosis of the cancer cells. This process requires the regulation of the TCR and co-stimulatory and inhibitory signal, which is called the immune checkpoint [31]. However, cancers often avoid immune surveillance by immunosuppressive tumor environments and immune tolerance [32]. The development of immune checkpoint inhibitors (ICIs) was a breakthrough in immune-oncology. Immunity checkpoint blockage removes inhibitory signals of T-cell activation that make T cells surmount regulatory mechanisms and fight against cancers [33,34].

The most frequently studied molecules in relation to immune checkpoint blockade are cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) and programmed cell death receptor (PD-1) and its ligands PD-Ls. CTLA-4 is expressed on the surface of activated T cells and dampens T-cell activation. CTLA-4 (providing inhibitory co-stimulatory signal) and CD28 (providing positive co-stimulatory signal) share ligands (CD80 and CD86) with each other. However, because CTLA-4 has higher avidity and affinity than CD28, it becomes possible to competitively attenuate T-cell activation [33,35]. PD-1 mainly regulates effector T-cell activity within tissues and tumors and maintains peripheral tolerance [36]. PD-1 is more broadly expressed than CTLA-4. CLTA-4 is mainly present in T cells, but PD-1 is present in B cells and NK cells as well. CTLA-4 and PD-1 are highly expressed on Tregs. Tregs may promote the proliferation by the ligand binding [33]. PD-1 modulates T-cell activation through binding with PD-L1 and PD-L2 [37,38]. PD-L1 is known to be upregulated in many cancers. PD-Ls inhibit cytokine production and anti-tumor lymphocytes in the TME. The molecular mechanisms of checkpoint inhibitors are shown in Fig. 1.

Figure 1. Schematic diagrams of the mechanisms of action of CTLA-4, PD-1 and PD-L1 blockade with or without chimeric antigen receptor T cell (CAR-T) therapy [36,58]. (A) T-cell activation, attenuation by normal inhibitory signals, and negative regulation using anti-CTLA-4, anti-PD-1 or anti-PD-L1 antibody therapy. (B) CAR-mediated activation, attenuation by tumor and checkpoint mediated inhibitory signals. Checkpoint receptor binding and checkpoint ligand blocking antibodies may synergize with chimeric antigen receptor therapy. APC, antigen presenting cell; CTLA-4, cytotoxic T-lymphocyte-associated antigen-4; PD-1, programmed cell death receptor; PD-L1, programmed death-ligand 1; TAA, tumor-associated antigen; TCR, T-cell receptor.

Immunohistochemistry is commonly used to check PD-L1 upregulation of the tumor cells. A study was conducted on PD-L1 expression in pediatric cancers, and 39 (9%) of the 451 evaluable tumors were identified to show PD-L1 expression in at least 1% of tumor cells. The highest incidence of PD-L1 expressing tumors included Burkitt lymphoma (80%; 8/10 tumors), glioblastoma multiforme (GBM, 36%; 5/14 tumors), and NBL (14%; 17/118 tumors) [39].

There are several studies that evaluate the safety and efficacy of ICIs in pediatric populations. In the first phase I/II pediatric trial of nivolumab, a PD-1 inhibitor was evaluated in patients aged 1-18 years with R/R solid tumors and lymphomas [40]. Of the total 85 patients, 63 (74.1%) had solid tumors (22 NBL, 13 osteosarcoma, 12 rhabdomyosarcoma, 11 Ewing sarcoma, 2 epithelioid sarcoma, 2 other sarcomas). The recommended phase 2 dose for pediatric patients was 3 mg/kg every 14 days of nivolumab. There were no objective responses in solid tumors, while 30% of Hodgkin lymphomas and 10% of non-Hodgkin lymphomas showed complete or partial response (PR). Stable disease was the best response in 33% of sarcoma and 50% of NBL patients. The common immune-related adverse events were hepatic toxicity and pleural or pericardial effusions. PD-L1-positivity was reported in 7-50% of solid tumors, 88% of non-Hodgkin lymphoma and 100% of Hodgkin lymphoma cohorts. Another PD-1 inhibitor trial in a pediatric population was the pembrolizumab study for advanced melanoma; or PD-L1-positive, advanced, R/R solid tumor, or lymphoma [41]. A total of 863 patients were enrolled. The recommended phase 2 dose for pediatric patients was 2 mg/kg every 3 weeks of pembrolizumab. Hodgkin lymphoma showed the highest objective response of 60%, while 5.8% (8 of 136) solid tumor and other lymphoma patients had PR. The diagnoses of patients who had PR were adrenocortical carcinoma, mesothelioma, malignant ganglioglioma, epithelioid sarcoma, lymphoepithelial carcinoma and malignant rhabdoid tumor. Except for melanoma, most tumor samples showed PD-L1 expression. Nevertheless, the low response rate is presumed to be due to the low mutation burden of pediatric cancer, low expression of major histocompatibility complex (MHC), the immaturity of the immune system in young children, and so on [42-44].

① Biomarkers: Several biomarkers are being presented to predict the response to ICIs. These include the expression of PD-L1 and PD-L2 on tumor cells, RNA expression signatures, lymphocytic infiltrates, and tumor mutation burden (TMB) [45-49]. Researchers discovered a small group in colorectal cancer showing highly activated Th1- and CTL-rich environments. Through the genetic analysis, it was discovered that the group has the loss of DNA mismatch repair activity, which is called microsatellite instability (MSI) [50]. MSI-high cancer means there is a high number of mutations within microsatellites. MSI-high tumors in this environment may survive due to a dramatic overexpression of immune checkpoint-related proteins such as CTLA-4, PD-1 and lymphocyte-activation gene 3 (LAG-3) [47]. They also have a higher Foxp3+ Tregs and PD-L1 expression. Colorectal cancer and other solid tumors with mismatch repair deficiency showed a greater response to pembrolizumab [51]. Le DT, et al. prospectively evaluated the efficacy of PD-1 blockade in different subtypes of mismatch repair-deficient cancers. A total of 86 patients were enrolled, the object response rate was 53%, complete response rate was 21%, and 11 patients were in CR two years after the treatment [52]. These results signify that patients with a mismatch repair deficiency, regardless of the tumor type, can get benefit from PD-1 blockade.

TMB, another important biomarker, is defined by the total number of mutations per coding area of a tumor genome. Hypermutation is defined as >10 mut/Mb [15]. MSI-high tumors have a high mutation burden, therefore potentially more neo-antigens. Palles C, et al identified specific heterozygous POLE or POLD1 germline variants in multiple-adenoma and/or early-onset colorectal cancer cause replication errors and coupled repair of base pair-level mutations, thus contain high mutation burden without MSI [53]. Rizvi NA, et al. revealed that a high mutation burden in non-small cell lung cancer was related with improved survival when treated with pembrolizumab [49]. Noskova H, et al. reported that TMB along with tumor progression was increased in pediatric solid tumor samples [54]. Clinical trials are ongoing to clarify the safety and effectiveness of PD-1 blockade in pediatric cancer patients with MSI-high and high tumor mutation burden tumors. Efforts to discover various biomarkers in pediatric cancers will be continued. Early-phase trials involving ICIs are presented in Table 1.

Table 1 . Clinical trials involving immune checkpoint inhibitors monotherapy or combination therapy for pediatric solid tumors.

Intervention	Description	Tumor	Age	Phase	ClinicalTrials.gov Identifier	Status
PD-1 inhibitor	Nivolumab	R/R hypermutated malignancies	12 months-18 years	Pilot	NCT02992964	Recruiting
		bMMRD positive patients
	Pembrolizumab	Advanced melanoma	6 months-18 years	I/II	NCT02332668	Recruiting
		Advanced, R/R PD-L1-positive malignant solid tumor or other lymphoma
		R/R classical HL
		Advanced, R/R MSI-H solid tumors
	Pembrolizumab	R/R DIPG, NB-HGG, ependymoma, MBL or hypermutated brain tumors	1-29 years	I	NCT02359565	Recruiting
	Pembrolizumab	R/R hepatocellular carcinoma	Up to 30 years	II	NCT04134559	Recruiting
	Pembrolizumab, Decitabine, hypofractionated index site radiation	R/R solid tumors (excluding primary CNS tumors)	12 months-40 years	Pilot	NCT03445858	Recruiting
		Lymphoma in second or greater relapse or with refractory disease
	Nivolumab, Entinostat	R/R solid and CNS tumors with high mutational load, PD-L1 mRNA expression or MYC/N amplification	6-21 years	I/II	NCT03838042	Recruiting
PD-1 inhibitor+CTLA-4 inhibitor	Nivolumab, Ipilimumab	Metastatic melanoma, R/R ES, HL, NBL, OSA, RMS, other solid tumors	12 months-30 years	I/II	NCT02304458	Active, not recruiting
	Nivolumab, Ipilimumab	DIPG, HGG, recurrent or progressive MBL, ependymoma or other CNS high-grade tumors	6 months-21 years	Ib/II	NCT03130959	Active, not recruiting
PD-L1 inhibitor	Avelumab	Solid malignant tumors (including CNS tumors) or lymphoma for which no standard therapy is available	Up to 18 years	I/II	NCT03451825	Recruiting
	Durvalumab	R/R solid tumors, lymphoma, CNS tumors except germ cell tumors (both CNS and non-CNS) and DIPG	1-17 years	I	NCT02793466	Recruiting
PD-L1 inhibitor+CTLA-4 inhibitor	Durvalumab, Tremlimumab	Solid malignant tumors (except primary CNS tumors)	Up to 18 years	I/II	NCT03837899	Recruiting

bMMRD, biallelic mismatch repair deficiency; CNS, central nervous system; DIPG, diffuse intrinsic pontine glioma; ES, Ewing sarcoma; HL, hodgkin lymphoma; MBL, medulloblastoma; MSI-H, microsatellite-instability-high; NBL, neuroblastoma; NB-HGG, non-brainstem high-grade gliomas; PD-L1, programmed death-ligand 1; RMS, rhabdomyosarcoma; R/R, recurrent/refractory..

2) Modulation of the immune system

CAR-T usually contain a single-chain variable fragment from a monoclonal antibody that can identify TAA; a transmembrane hinge region; and a signaling domain such as CD28, CD3z, or 4-1BB [55]. MHC independent TAA recognition induces tumor-directed cytotoxicity. As CAR-T has demonstrated its effectiveness in pediatric ALL [56], CAR-T studies are now being conducted in pediatric solid tumors. Incorporation of synthetic immunotherapy such as CAR-T is an option to overcome limitation of ICIs in pediatric cancers lack of tumor reactive effector lymphocytes [57].

① GD-2: Clinical trials of GD2-CAR-T in NBL were the first experiments where CAR-T cells have been tested in pediatric solid tumors [58]. In a phase 1 clinical trial of first-generation GD2-CAR-T in refractory NBL, 3 (27%) of 11 patients achieved complete remission. In addition, two of them have sustained remission for more than 5 years [59]. In 2017, a study regarding third-generation GD2-CAR-T incorporating the CD28 and the OX40 (CD134) costimulatory domains was reported. In this study, patients received CAR-T alone, CAR-T plus cyclophosphamide and fludarabine, or CAR-T plus cyclophosphamide, fludarabine and pembrolizumab. They demonstrated the third-generation GD2-CAR-T was safe and antitumor responses were modest. Lymphodepletion by cyclophosphamide and fludarabine helped to increase GD2-CAR-T expansion, but PD-1 inhibitor did not further increase GD2-CAR-T expansion [60].

② HER2: HER2 is a member of the epidermal growth factor receptor family with tyrosine kinase activity. HER2/neu is highly expressed in medulloblastoma, osteosarcoma, Wilms tumor and GBM [61]. Mineo JF, et al. reported that humanized anti-HER2/neu antibody have the ability to induce apoptosis and cellular-dependent cytotoxicity of HER2/neu-expressing GBM cell lines [62]. The first study evaluating HER2-CAR-T cells in pediatric and young adult sarcoma including osteosarcoma, primitive neuroectodermal tumor and DSRCT was reported in 2015. This phase I/II study showed persistence of HER2-CAR-T cells and tolerable safety. Three osteosarcoma and one DSRCT patients showed stable disease for 3 to 14 months [63]. HER2-CAR-modified autologous virus-specific T cells (HER2-CAR VSTs) were evaluated in adult and pediatric relapsed or progressive GBM patients. In this phase I study, HER2-CAR VSTs were safe and showed some clinical benefits. Maximal response was PR in a 17-year-old male with right thalamic GBM [64]. A preclinical model of HER2-CAR-T showed potential efficacy in osteosarcoma [65].

③ IL13Ra2: Interleukin-13 receptor alpha 2 (IL13Ra2), a glioma-associated antigen, is associated with a reduced rate of survival of glioma patients [66]. One adult relapsed GBM patient was enrolled in IL13Ra2-CAR-T cell therapy. After CAR-T cell intracranial infusions, complete tumor regression was observed and the response continued for 7.5 months [67].

④ B7-H3: B7-H3 is one of the checkpoint molecule highly expressed on pediatric sarcomas and brain tumor cells. CAR-T targeting B7-H3 represented potent preclinical activity in various pediatric solid tumors and brain tumors [68].

Recently, it was reported that CAR-T can be exhausted by immunosuppressive cells, including MDSC cells or Treg cells [69,70]. Researches to overcome T-cell exhaustion and restore immune memories are being investigated. Combination therapy with ICIs and CAR-T may synergize the CAR-T activity (Fig. 1) [57]. Cytokine release syndrome, CAR-T-related encephalopathy syndrome and sustained on-target off-tumor effects are also considered. Early-phase trials using second and third-generation CAR-T are ongoing (Table 2).

Table 2 . Clinical trials involving chimeric antigen receptor T-cells for pediatric solid tumors.

Intervention	Description	Tumor	Age	Phase	ClinicalTrials.gov Identifier	Status
HER2	HER2-specific CAR T cell	HER2 positive R/R pediatric CNS tumors	1-26 years	I	NCT03500991	Recruiting
EGFR	EGFR806-specific CAR T cell	EGFR positive R/R pediatric CNS tumors	1-26 years	I	NCT03638167	Recruiting
	Second generation 4-1BBx EGFR806-EGFRt and a second generation 41BBx CD19-HER2tG	EGFR positive R/R solid tumors in children and young adults	1-26 years	I	NCT03618381	Recruiting
B7-H3	SCRI-CARB7H3(s); B7H3-specific CAR T cell	DIPG/diffuse midline glioma or R/R pediatric CNS tumors	1-26 years	I	NCT04185038	Recruiting
IL13Ra2	IL13Ra2-specific, hinge-optimized, 41BB-costimulatory CAR/truncated CD19-expressing autologous T lymphocytes	Grade III or IV glioma	12-75 years	I	NCT02208362	Recruiting
GD2	GD2-CAR-T01	High risk or R/R neuroblastoma	12 months-18 years	I/II	NCT03373097	Recruiting
	Third generation GD2 CAR T cell, co-expression of IL15 and inducible caspase9	R/R neuroblastoma	18 months-18 years	I	NCT03721068	Recruiting
GPC3	GPC3-CAR-T	GPC3-positive solid tumors (currently only liver tumors)	1-21 years	I	NCT02932956	Recruiting

CAR, chimeric antigen receptor; CNS, central nervous system; DIPG, diffuse intrinsic pontine glioma; EGFR, epidermal growth factor receptor; EGFRt, truncated EGFR; GPC3, glypican-3; HER2, human epidermal growth factor receptor 2; HER2tG, truncated HER2 extracellular protein; IL13Ra2, interleukin-13 receptor alpha 2; R/R, recurrent/refractory..

Dendritic cells (DCs) are the most powerful antigen presenting cells and are typically used as anti-cancer vaccines. They play a role as a bridge between the adaptive and the innate immune responses and allow to generate immune responses against tumors [71-73]. A phase I trial combining decitabine/dendritic cell vaccine for children aged 2.5-15 years with relapsed NBL, Ewing sarcoma, osteosarcoma and rhabdomyosarcoma was conducted. The DC vaccines were tolerated, and six of the 10 patients had a T cell response; one of the 10 patients had a complete tumor response [74,75].

Cancer treatment is entering a new era through immunotherapy that stimulates the patient’s own immune system. As with adult cancer, innovative immunotherapy is also being tried in pediatric solid tumors. However, there are still very few patients with pediatric solid tumors who can benefit from immunotherapy. Pediatric solid tumor is less likely to respond to ICIs because it has a lower mutation burden than adult cancer. Several ICIs are in clinical trials for testing the potentiality for MSI-high and high mutation burden pediatric solid tumors. Furthermore, to overcome the limitations of monotherapy, studies are also being conducted on combination therapy such as checkpoint inhibitors with CAR-T. Further research is needed on the TME of children and adolescents who may be different from that of adults but are not yet known. Immunotherapy in pediatric solid tumors is still in the early stages, but it is encouraging that various attempts are being made. The future treatment is expected to optimize the selection of therapeutic strategies, and ultimately improve patient outcome and quality of life.

The author has no conflict of interest to declare.