Osteosarcoma (OS) is the most common primary bone tumor in children and adolescents, with an incidence of 5 cases per million in children and adolescents up to 19 years [11]. It is a malignant neoplasm of osteoid-secreting mesenchymal stem cells which typically arises in the metaphysis of long bones. The standard of care includes surgical resection and chemotherapy with methotrexate, cisplatin, and doxorubicin [11]. While cure rates for nonmetastatic osteosarcoma have improved to greater than 60%, cure rates for patients with metastatic disease are as low as 19% [29, 30]. The pleomorphism of osteosarcoma cells presents a challenge regarding chemotherapy resistance, as 15–20% of patients with metastatic OS are resistant to first-line chemotherapeutics [31]. This pleomorphism has also made engineering a CAR-T antigen that precisely targets OS difficult.

The diganglioside GD2 has emerged as a target for cancer immunotherapy due to its relatively restricted expression to solid tumors such as sarcomas [32]. It was found to be present on 55% of primary sarcomas, with the highest expression among OS and alveolar rhabdomyosarcoma [32]. While GD2-directed CAR-T cells demonstrated potent in vitro and in vivo activity against OS, tumor-derived G-CSF limited the effectiveness of GD2-directed CAR-T cells and contributed to the immunosuppressive TME by driving the expansion of myeloid-derived suppressor cells (MDSCs) [32]. Due to the immunosuppressive effects of G-CSF and its ubiquity in the OS TME, its inhibition remains a potential target for the augmentation of CAR-T therapy. A study by Sterner et al. demonstrated that the neutralization of G-CSF with Lenzilumab enhanced CD-19 directed CAR-T cell proliferation in leukemia and led to a reduction in cytokine release syndrome (CRS), which they hypothesized was due to G-CSF’s role in leukemia cell proliferation [33].

In an abstract, Ramakrishna et al. reported the initial results of a phase I trial of GD2-CARTs in children and young adults with GD2-positive solid tumors, including OS and neuroblastoma (NCT02107963). Thirteen patients were infused, with two cases of grade-1 cytokine release syndrome and no neurological or dose-limiting toxicity [34]. On day 28 after infusion, 10 of 13 patients had stable disease and three had progressive disease, with all patients eventually progressing [34]. Following these results, they analyzed the immune profiles of the patients in order to determine factors influencing adequate CAR-T cell expansion, concluding that proliferation of the GD2-CARTs was correlated with a larger baseline population of naïve and central memory T cells [34].

Human epidermal growth factor receptor 2 (HER2) is expressed on a wide range of cancers, including lung, ovary, prostate, brain, and sarcomas [35]. Most sarcomas, however, express HER2 at levels which are too low to allow for effective targeting by HER-directed monoclonal antibodies (mAbs) such as trastuzumab. Ahmed et al. reported the results of a phase I/II clinical trial of HER2-CARTs in patients with HER2-positive tumors (NCT00902044). Nineteen patients had been enrolled, 16 of whom had osteosarcoma. The HER2-CARTs persisted for 6 weeks in seven of the nine evaluable patients [36]. Three patients had their tumors removed, and one showed greater than 90% tumor necrosis [36].

Activated leukocyte cell adhesion molecule (ALCAM, CD166) is a membrane glycoprotein that, by binding to CD6, mediates interactions between adjacent leukocytes [37]. In a study by Wang et al., the expression of ALCAM on four OS cell lines ranged from 36.9 to 96.7% [37]. After establishing the high levels of ALCAM in OS cell lines, Wang et al. constructed an orthotopic OS model using immunodeficient mice and tibially implanted OS cells. When anti-ALCAM CAR-T cells were transfused into these mice, there was a significant reduction in tumor weight compared to control [37].

B7-H3 is a cell surface glycoprotein and another promising target antigen for CAR-T cells. While its precise physiological function has not been elucidated, it is believed to function as an immune checkpoint molecule. It is highly expressed among cell surfaces of pediatric solid tumors, including neuroblastoma, rhabdomyosarcoma, Wilms tumor, Ewing sarcoma, osteosarcoma, and diffuse midline glioma, but has limited expression on healthy tissue [38,39,40,41]. In addition to being expressed on cancer cells, B7-H3 is also upregulated on component cells of the TME, such as dendritic cells (DCs), tumor-associated macrophages (TAMs), myeloid-derived suppressor cells (MDSCs), neutrophils, and tumor-associated endothelial cells. The immunosuppressive functions of B7-H3 interfere with IFN-γ release and NK cell cytotoxicity, reducing the efficacy of CAR-T treatment [39]. Consequently, CAR-T cells targeting B7-H3 may have the dual function of eliminating tumor cells and peripheral cells contributing to the TME [38, 42]. Upregulation of B7-H3 has also been demonstrated on B7-H3-directed CAR-T cells when co-cultured with antigen-positive solid tumor cells, indicating that CAR-T cell fratricide may be a potential limitation of this therapy [43].

Zhang et al. tested anti-B7-H3 CAR-T cells both in vitro and in vivo in a murine model. They first confirmed high B7-H3 expression of 73.3% among 60 pathological OS sections [44]. In vitro, their CAR-T cells recognized and killed B7-H3-positive OS cells. In vivo, these CAR-T cells showed excellent anti-tumor activity in both the high- and low-dose group [44].

Talbot et al. investigated B7-H3 CAR-T cells in a metastatic OS model in which mice were implanted with an OS cell line (LM7) with high B7-H3 expression and monitored for the development of metastases [45]. B7-H3 CARTs demonstrated excellent antitumor activity in vivo and prevented the development of pulmonary metastases, which was monitored using firefly luciferase (ffLuc) expressed by the OS cells [45]. In a subsequent study, Talbot et al. harnessed chemokines secreted by OS cells to more precisely direct B7-H3 CAR-T cells to their target tumor. They identified IL-8 and CXCL16 as most highly expressed chemokines on OS samples and modified anti-B7-H3 CAR-T cells with CXCR2 and CXCR6, these chemokines’ cognate receptors. These modifications enhanced CAR-T cell homing towards OS cells both in vitro and in vivo, including in a metastatic murine OS model [19].

Adeshakin et al. constructed anti-B7-H3 CAR-T cells with a knockout of the reg1 gene, which produces a potent negative immune regulatory molecule known as regnase-1, and tested them in a murine model of OS [46]. Deletion of reg1 improved the function of these CAR-T cells, which were able to promote a more effective proinflammatory landscape compared to nonmodified CAR-T cells [47]. Second-generation B7-H3 CAR-T cells developed by Majzner et al. demonstrated promising activity against OS, ES, and medulloblastoma in an orthotopic murine model [48].

Pinto et al. reported results from the STRIvE-02 clinical trial of B7-H3 CAR-T cells in pediatric patients (NCT04483778). While three patients with OS were enrolled, the study was open to all B7-H3 positive tumors, and patients with Ewing Sarcoma, neuroblastoma, and rhabdomyosarcoma were also included [49]. Nine patients in total were treated at two different dose levels, with three experiencing stable disease and six experiencing progressive disease [49]. A clinical response was noted in one patient after being treated with a second infusion at dose level two [49].

Other tumor antigens are also highly expressed on OS cell lines, such as ALPL-1, an isoform of alkaline phosphatase, and IL-11RA, a receptor for IL-11 [50, 51]. CAR-T cells directed towards these antigens demonstrated preferential targeting of tumor cells without significant systemic toxicity in in vivo models of orthotopic primary and metastatic OS [50, 51].

Xin Huang et al. investigated insulin-like growth factor receptor (IGF1R) and tyrosine kinase-like orphan receptor 1 (ROR1) as potential CAR-T targets and determined that both were highly expressed among OS, ES, and RMS cell lines [52]. Transfusion of these CARTs into NSG mice intravenously inoculated with OS cells led to reduced tumor growth and prolonged survival [51].

Shin et al. investigated the feasibility of programmed cell death ligands 1 and 2 (PD-L1, PD-L2) as CAR-T targets by assessing their expression on OS, NB, GBM, and RMS cell lines. They found that PD-L1 and PD-L2 expression were upregulated in OS after stimulation with IFN-γ and TNF-α, while their expression remained low on other cell lines [53]. They then tested PD-1 CAR-T cells against an OS cell line that expresses moderate levels of PD-L1 and observed that the CAR-T cells displayed cytotoxic activity in a dose-dependent manner in vitro [53].

Combating the hostile tumor microenvironment remains an obstacle in the treatment of OS, and fourth-generation “armored” CAR-T cells attempt to overcome this by co-expressing or secreting cytokines with antitumor activity. Hui et al. developed an NKG2D-directed CAR-T construct based on previous studies’ success using this ligand against OS cell lines in vitro and in murine models and engineered it to co-express IL-7 and CXCR5, a chemokine receptor whose ligand is overexpressed in OS [22, 54, 55]. CXCR5/IL-7 CAR-T cells demonstrated superior cytokine production, decreased exhaustion, and superior cytotoxicity against OS in mouse models compared to wild-type CAR-T cells [56]. An IL-18-armored anti-NKG2D CAR-T engineered by Breman et al. showed high levels of IL-18 secretion upon antigen challenge and superior antitumor activity compared to non-armored anti-NKGD2 CAR-T cells in a murine model [23].

Ewing sarcoma (ES) is the second most common primary bone malignancy in the pediatric population after OS. The cornerstone of treatment of ES is chemotherapy induction followed by radiation, surgery, or both. While the 5-year survival rate of localized ES is 75 to 80%, patients with regional or distant metastases have a poorer 5-year survival rate of around 30% [57]. Similar to OS, the heterogeneity of tumor-associated antigens in ES has been a major focus of CAR-T cell development against this disease.

Like neuroblastoma, ES belongs to a group of tumors known as primitive neuroectodermal tumors (PNETs). Since other PNETs had previously demonstrated aberrant expression of GD2, Kailayangiri et al. investigated its expression in ES and demonstrated GD2 expression in 100% of ES cell lines and cell cultures [58]. They engineered GD2-specific CAR-T cells and found that anti-GD2-CAR-T cells lysed ES cells both in vitro and in OS xenografts [58].

Charan et al. posited that combining anti-GD2 CAR-T cells with immunotherapy could increase their ability to target and kill ES cells [59]. Hepatocyte growth factor (HGF) appears to be crucial in the formation of a hostile TME by facilitating cross-talk between malignant cells and adjacent stroma [60]. Charan et al. determined that HGF expression was upregulated in ES samples via RT-qPCR and then administered both anti-GD2 CAR-T cells and anti-GD2 CAR-T cells with an adjuvant antibody directed at HGF (AMG102). Superior tumor regression was demonstrated with the anti-GD2 CAR-T cells plus AMG102 compared to the control arm, which did not effectively control the growth of orthotopically implanted primary and metastatic ES [59].

Increased angiogenesis is a hallmark of tumor growth and allows for unchecked vascular proliferation and recruitment of nutrients necessary for tumor survival. The expression of vascular endothelial growth factor receptor 2 (VEGFR2) on ES cells and tumor-associated endothelial cells has been implicated in ES proliferation and correlates with a poor prognosis [61,62,63]. Englisch et al. investigated the efficacy of VEGFR2-directed CAR-T cells and found that they specifically lysed VEGFR-expressing ES cells in a murine model [64].

Overexpression of the ephrin (EPH) family of receptors is also believed to play a role in the development and progression of several types of cancer, including ES. Funasaka et al. are currently enrolling patients for the CARTiEr study, a single-center, single-arm, phase I study of anti-EPHB4 CAR-T in patients with ES or other EPHB4-positive tumors [65]. Hsu et al. tested EphA2-directed CAR T cells in vitro and in vivo based on this target’s high expression in ES and OS and low expression in normal bone [66, 67]. These CAR-T cells displayed potent antitumor activity in vitro and eliminated ES and OS tumors in a murine model [68]. Additionally, they led to the eradication of OS which had metastasized to murine lungs and livers [68].

Most studies to date have focused on engineering antibodies and CAR-T cells against tumor surface antigens. There have been early successes; however, utilizing proteins secreted by tumors as targets of CAR-T cells. One such cancer-secreted protein is oncofetal tenascin C (TNC). Expression of aberrant isoforms of TNC has been found in several adult tumors [69, 70]. Wickman et al. investigated whether TNC could serve as a target for CAR-T cells in pediatric cancers and found that it was expressed in diffuse intrinsic pontine glioma (DIPG), OS, rhabdomyosarcoma (RMS), and ES [24]. They generated CAR-T cells directed at TNC, which lysed TNC-positive tumor cells in vitro but had poor efficacy in vivo. To combat the hostile TME, they armored their anti-TNC CAR-T construct with a constitutively active IL-18 receptor (IL-18R). IL-18 had previously been demonstrated to enhance the antitumor activity of both CAR-T cells and endogenous lymphocytes [71]. Accordingly, the anti-TNC CAR-T cells expressing a constitutively active IL-18R displayed superior antitumor effector function in vitro and prolonged mouse survival in a murine OS xenograft compared to non-armored anti-TNC CAR-T cells [24].

The G-protein coupled receptor 64 (GPR64) is physiologically expressed only in epididymal tissue; however, it has been noted to be specifically expressed in ES and some other sarcomas [72]. In a conference abstract, Schirmer et al. reported the development of GPR64-directed CAR-T cells which controlled tumor growth in mice bearing ES xenografts [73].

Neuroblastoma (NB) accounts for 8–10% of pediatric cancers and is the most common extracranial solid tumor in children [74, 75]. Outcomes in NB patients have improved over the past three decades, with a 5-year survival increasing from 52 to 74% [75]. Similar to other solid tumors, the prognosis remains poor for patients with high-risk metastatic disease. Treatment of NB is stratified based on risk, ranging from monitoring or surgery for low-risk cases to high-dose chemotherapy, resection, and stem cell transplantation for high-risk patients with metastatic disease [76]. Treatment of metastatic NB remains one of the primary challenges in the management of this disease, with a 2-year survival rate of approximately 46% [77].

Like pediatric sarcomas, GD2 is highly expressed in most NB cell lines and is, accordingly, a promising target for immunotherapy. It was the first cancer antigen targeted using CAR-T cell therapy for NB due to the previous success of denutuximab, an anti-GD2 monoclonal antibody. The second anti-GD2 mAb for NB, naxitamab, improved on but did not eliminate the side effects of denutuximab, most notably peripheral nerve damage [78].

Early studies involving GD2-directed CAR-T cells demonstrated adequate safety but poor clinical response and CAR-T cell persistence. Pule et al. generated GD2-directed CAR-T cells which persisted in vivo for greater than 6 weeks [79]. In a phase 1 clinical trial from 2004 to 2009, children with NB were treated with first-generation CAR-T cells targeting GD2 (NCT00085930). In the first set of results from this trial, Louis et al. reported improved persistence of GD2-directed CAR-T cells up to 192 weeks and demonstrated that even low levels of circulating CAR-T cells were associated with a sustained antitumor response [80]. At an 18-year update in 2025, three of 11 patients with active disease achieved a complete antitumor response, which was sustained in two patients. Of the eight patients with a history of NB but no active disease at the time of infusion, five were disease free at their last follow-up. The overall survival rate was 31.6% at 15 years [81].

Che-Hsing Li et al. reported the long-term outcomes of a clinical trial that ran from 2004 to 2009 of first-generation CAR-T cells targeting GD2 in children with neuroblastoma (NCT00085930). Three of 11 patients with active disease at infusion achieved a complete response. The response was sustained in two patients, one for 8 years and one for more than 18 [81]. Of the eight patients with no disease at the time of CAR-T infusion, five remain disease-free 10 to 15 years after follow-up. Despite these CAR-T cells being first-generation constructs without co-stimulatory domains, they were able to control relapsed and refractory NB in several patients.

A phase 1 and 2 clinical trial of GD2-directed CAR-T cells in pediatric patients with NB or other GD2-positive tumors began in 2018 at the Bambino Gesù Hospital and Research Institute in Rome and remains active (NCT03373097). Del Bufalo et al. reported the first results from this trial in 2023. Twenty-seven children with relapsed or refractory NB received GD2-directecd CAR-T cells. These cells proliferated and remained detectable in blood samples in 26 of the 27 children enrolled up to 30 months after the initial infusion [82]. Seventeen of the 27 children had a positive response, with 9 achieving a complete response and an overall 3-year survival of 60%. Cytokine release syndrome (CRS) occurred in 74% of patients but was mild in all but one. In an updated report of this trial, the total number of children treated with these GD2-CAR-T cells had increased to 54. The overall response rate was 66%, with a complete remission rate of 40% at 6 months and 5-year overall survival rate of 42.7% [83]. CAR-T cells were detected in samples after 12 months in 64% of patients. Four children treated experienced neurotoxicity which was effectively mitigated with the activation of a caspase-9 suicide gene in the CAR-T cells induced by treatment with rimiducid [83]. A follow-up investigation of these results determined that polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) impaired the efficacy of the GD2-CARTs in these patients [84]. These cells were found to increase in peripheral blood after treatment with GD2-CAR-T cells and downregulate genes involved in cell activation, inflammatory response, and cytokine secretion. The levels of circulating PMN-MDSCs correlated inversely with the levels of GD2-CAR-T cells in these patients, suggesting that a strategy to limit PMN-MDSC expansion could improve overall CAR-T proliferation [84].

Quintarelli et al. presented a case series of five children with refractory NB who were treated with GD2-CAR-T cells. After treatment, two complete responses were achieved and one was maintained [85]. In addition, one patient had a partial response and one achieved disease stability [85].

Heczey et al. developed a third-generation GD2-directed CAR-T cell after observing poor expansion and long-term persistence with a first-generation construct [86]. They administered these third-generation CAR-T cells in three cohorts: one with the CAR-T cells alone, one with CAR-T cells plus lymphodepletion with cyclophosphamide and fludarabine (Cy/Flu), and one with CAR-T cells, lymphodepletion, and a programmed death-1 (PD-1) inhibitor (NCT01822652). They determined that lymphodepletion prior to CAR-T infusion improved expansion but that the addition of PDL-1 inhibition did not provide an additional benefit [86].

In an abstract, Straathof et al. reported the results of a phase 1 clinical trial in which children with relapsed or refractory NB were treated with escalating doses of GD2-CAR-T cells and lymphodepletion (NCT02761915). No patients had objective clinical response 28 days after infusion; however, six patients receiving a higher dose of CAR-T cells experienced grade 2 to 3 CRS [87]. Three patients did demonstrate regression of soft tissue and bone marrow disease, suggesting that targeting GD2 with CAR-T cells in patients with NB is safe but may require modification to improve CAR-T cell persistence [87].

A GD2-directed CAR-T cell was constructed by Lihua Yu et al. and tested in children with refractory or recurrent NB (NCT02765243). Ten patients were included, with six experiencing stable disease (SD) at 6 months and four with SD at 1 year [77]. Four patients remained alive three to 4 years after infusion. The median overall survival time was 25 months, and the median progression-free survival time was 8 months [77].

Based on the promising results for GD2-directed CAR-T cells in pediatric patients with NB, several fourth-generation armored CAR-T cells have been tested, with more currently in development. Chen et al. investigated the benefits of armoring GD2-CAR-T cells with IL-15 to improve persistence and antitumor function. They constructed both GD2-CAR-T cells and GD2-CAR-T cells with IL-15 incorporated into the CAR cassette and investigated their efficacies in vitro and in vivo in a murine model. The IL-15-armored CAR-T cells demonstrated superior antitumor activity and survival compared to the non-armored GD2-CAR-T cells. Additionally, they observed that the IL-15 secreted by the GD2-CAR-T constructed produced cells with “memory” and “stem cell-like” phenotypes and was independent of antigen encounter [12].

Fischer et al. construc