Adoptive Cellular Therapy for Pediatric Solid Tumors: Beyond Chimeric Antigen Receptor-T Cell Therapy

Abstract

Children and adolescents with high-risk (metastatic and relapsed) solid tumors have poor outcomes despite intensive multimodal therapy, and there is a pressing need for novel therapeutic strategies. Adoptive cellular therapy (ACT) has demonstrated activity in multiple adult cancer types, and opportunity exists to expand the use of this therapy in children. Employment of immunotherapy in the pediatric population has realized only modest overall clinical trial results, with success thus far restricted mainly to antibody-based therapies and chimeric antigen receptor T-cell therapies for lymphoid malignancy. As we improve our understanding of the orchestrated cellular and molecular mechanisms involved in ACT, this will provide biologic insight and improved ACT strategies for pediatric malignancies. This review focuses on ACT strategies outside of chimeric antigen receptor T-cell therapy, including completed and ongoing clinical trials, and highlights promising preclinical data in tumor-infiltrating lymphocytes that enhance the clinical efficacy of ACT for high-risk pediatric solid tumors.

Solid tumors account for approximately 25% of all pediatric cancers, when excluding tumors of the central nervous system.¹ The most common types of pediatric malignant solid tumors (pMST) include neuroblastoma, sarcomas, and Wilms tumor.^2,3 Some subsets of pMST (e.g., favorable histology Wilms tumor and germ cell tumors) demonstrate chemotherapy sensitivity and high curative potential. However, most patients with high-risk pMST (defined here as metastatic and relapsed pMST) have unsatisfactory outcomes despite combinatorial intervention with chemotherapy, surgical resection, and/or radiation that carry high treatment-related toxicity.^4-9

Progress in immunotherapeutic approaches to cancer treatment, including augmenting or even replacing cytotoxic chemotherapy, continues to accelerate in both adult and pediatric malignancies. Despite these advancements, most pediatric cancers including pMST do not yet incorporate immunotherapy in standard-of-care practices. The purpose of this review is to highlight successes and challenges in immunotherapy for pediatric cancer with a focus on pMST, as well as present new opportunities for adoptive cellular therapy (ACT) outside of chimeric antigen receptor T-cell (CAR-T) therapy in this patient population.

IMMUNOTHERAPY SUCCESSES AND CHALLENGES IN pMST

Monoclonal antibody therapy is the most widely used immunotherapy in pediatric oncology to date. Their successes in pediatric leukemia and lymphoma were thoroughly reviewed in a recent publication.¹⁰ In regard to pMST, extensive investigation of antibody therapy targeting disialoganglioside (GD2) expressed by neuroectodermal cancers has improved survival rates and become a frontline immunotherapy in high-risk neuroblastoma.^11-13 Thorough reviews of anti-GD2 in neuroblastoma have been published.¹⁴ Briefly, treatment consists of administration of a GD2-specific monoclonal antibody (dinutuximab, naxitamab, etc.), often in combination with isotretinoin, interleukin 2 (IL-2) and/or granulocyte-macrophage colony-stimulating factor.^12,15 Unfortunately, use of GD2-specific antibodies in other pMST such as osteosarcoma and soft tissue sarcoma has shown limited responses accompanied with multiple toxicities including severe leg pain, vision dysfunction, motor weakness, and loss of vibratory sensation.^13,16-18 One death, possibly attributable to dinutuximab given on a 2-day schedule plus cytokine therapy, was reported on a phase II study in relapsed osteosarcoma.¹⁹ The effectiveness of this therapy relies on host immune-cell–mediated antibody-dependent cellular cytotoxicity (ADCC), especially natural killer (NK) cells. One factor that may influence the potency of ADCC and clinical outcomes is genetic polymorphisms of the Fc-γ-receptor expressed on the immune cells and killer cell immunoglobulin-like receptor (KIR) and KIR ligand expressed by cancer cells. One cohort of neuroblastoma patients treated with anti-GD2 therapy showed those with high-affinity Fc-γ receptor and the activating KIR 2DS2 had improved event-free survival and increased ADCC.²⁰

Immune checkpoint inhibition (ICI), including monoclonal antibody therapy targeting programmed cell death protein 1 (PD1), programmed death-ligand 1, and cytotoxic T-lymphocyte–associated protein 4, is now widely used in many adult solid tumors. One factor strongly correlating with patient response to ICI is tumor mutational burden (TMB; the sum of mutations in a tumor), which correlates with the presence of neoantigens.²¹ Tumors with high TMB have increased prevalence of suppressed tumor-specific T cells, leading to higher potential for response to ICI.²² Unfortunately, the TMB in pMST is generally quite low, and predicted neoantigens are present at low levels relative to common adult malignancies such as cutaneous melanoma and non–small cell lung cancer.^5,23,24 With the lack of neoantigens, minimal tumor-specific T-cell activation has led to largely disappointing clinical trials of single-agent ICI in pMST, including Children's Oncology Group phase I/II trial ADVL1412 with nivolumab alone and in combination with ipilimumab, KEYNOTE-051 with pembrolizumab, and adult trials that extended down to childhood ages (e.g., SARC028 with pembrolizumab in bone sarcomas, NCT03141684 with atezolizumab in alveolar soft-part sarcoma.^25-28 With the exception of alveolar soft-part sarcoma, very few or no objective responses were seen in pediatric patients with solid tumors. Leading groups such as ACCELERATE have suggested that ICI agents be studied in combination with other immunotherapies in pediatric histologies.²⁹ Some combination strategies are already in clinical trials for pMST, including combining anti-PD1 with the DNA methylation inhibitor 5-azacitidine for relapsed osteosarcoma (NCT03628209), a neuroblastoma cancer vaccine plus anti-PD1 (NCT04239040), the multityrosine kinase inhibitor regorafenib with anti-PD1 for osteosarcoma (SARC038 - NCT04803877), and the multityrosine kinase inhibitor sunitinib with anti-PD1 and standard chemotherapy for bone and soft tissue sarcomas (NCT03277924).

ACT APPROACHES IN pMST

Adoptive cell therapies come in 2 main varieties: those that redirect immune cell specificity (e.g., CAR T cells and T-cell receptor [TCR]–engineered T cells, CAR NK cells) and those that seek to augment the function of preexisting endogenous T-cell responses by ex vivo manipulation or boosting after adoptive transfer. We will first discuss advances in therapies that seek to redirect immune cell specificity. Chimeric antigen receptor T-cell therapy represents an ACT strategy with clinical success in hematologic malignancies, including B-cell acute lymphoblastic leukemia, chronic lymphocytic leukemia, and large B-cell lymphoma.^30-32 Chimeric antigen receptor T-cell therapy involves collection and expansion of circulating lymphocytes via apheresis with subsequent engineering of T cells via viral transduction to express a CAR that enables targeting of a specific tumor-associated antigen (TAA). Because tumor cells often downregulate their human leukocyte antigen (HLA) as an immune escape mechanism, this engineered T-cell approach is a valuable strategy as it bypasses the need for tumor antigen cross-presentation on HLA to T cells and instead enables direct tumor recognition by CAR-redirected T cells.³³ The technique also allows the generation of large numbers of T cells with identical antigen specificity mixing both CD4⁺ and CD8⁺ T cells. Chimeric antigen receptor T-cell therapy does carry risk of severe toxicities including cytokine-release syndrome and neurotoxicity.³⁴ Most CAR-T clinical trial successes in pediatrics have applied to lymphoid malignancy, specifically B-cell acute lymphoblastic leukemia.³⁵

Chimeric antigen receptor T-cell therapy trials conducted in pMST, including targeting of GD2 and human epidermal growth factor receptor 2, have thus far demonstrated limited efficacy.^36,37 A key obstacle to expanded utilization of CAR-T in pMST involves the identification of candidate TAAs, as most pediatric cancer types have limited known TAAs. Even with identification of novel TAAs, the heterogeneity of expression within a single tumor and potential for antigen loss and immune escape may dramatically limit the efficacy of single-antigen targeted CAR T-cell therapy.

In regard to TCR-engineered T cells, groups have cloned the α and β chains of the TCR, which can then be expressed by genetic engineering of peripheral T cells to enable infusion of large numbers of tumor-specific T cells that target the antigen presented in a major histocompatibility complex-dependent setting setting.⁷ Genetically modified TCR therapy has been tested in synovial sarcoma (SS).^38,39 Seventy percent of SSs express the cancer testis antigen NY-ESO-1, which is considered an ideal target because of its immunogenicity and lack of expression in nonmalignant tissues.⁴⁰ Unfortunately, objective clinical responses have been hampered by insufficient induction of immune response and tumor escape from immune recognition; thus, more recent efforts have focused on overcoming these barriers.⁴⁰ A pilot trial testing genetically modified autologous T cells targeting NY-ESO-1 in advanced sarcoma or melanoma patients was conducted in combination with concurrent dendritic cell (DC) vaccination with or without cytotoxic T-lymphocyte–associated protein 4 targeting antibody, ipilimumab.⁴¹ Although this trial included only patients older than 16 years, the diagnoses enrolled have overlap with the pediatric population. Of 6 sarcoma patients receiving ACT with DC vaccination (3 SSs, 1 osteosarcoma, 1 liposarcoma, and 1 malignant peripheral nerve sheath tumor), 4 had transient tumor regression on PET/CT. Two SS patients achieved a partial response (PR) on day 90 of the study. Rapid expansion of NY-ESO-1–specific T cells in peripheral blood was evidenced by cell levels peaking within 2 weeks of being infused. Tumor infiltration of NY-ESO-1–specific T cells was demonstrated using sequencing of tumor biopsies, and T cells were shown over time to shift away from a memory phenotype toward terminally differentiated effector T cells. Results demonstrated that engineered T cells expressing transgenic TCR infused in conjunction with pulsed DC is both feasible and capable of providing some antitumor activity.⁴¹

Other trials attempting to enhance the clinical activity of NY-ESO-1–targeting T cells have been reported.^38,42 Using autologous T cells whose TCRs were engineered for increased affinity for HLA-A2–restricted NY-ESO-1/LAGE1a–derived peptide, investigators examined safety and efficacy in targeting metastatic SS. Among 12 patients, 6 (50%) had objective response of median duration 30.9 weeks. Engineered T cells persisted in peripheral blood for 6 months in the responsive patients. Following ex vivo expansion, most T cells were characterized as having effector memory phenotype, whereas most engineered T cells that persisted in responders were characterized as central memory and stem-cell memory subsets, sustaining polyfunctionality and evading exhaustion. Results indicated the engineered T cells underwent prolonged regeneration to provide multiclonal, antigen-specific effector cells, affording a sustained antitumor response.³⁸ Another trial utilizing T cells with affinity-enhanced TCR to target advanced-stage SS expressing NY-ESO-1–specific HLA-A*02–restricted peptide SLLMWITQC (NY-ESO-1 SPEAR T cells) included 42 patients.⁴² Efficacy was demonstrated with 15 objective responses (1 with complete response [CR], 14 with PR). Abrupt expansion of engineered T-cell postinfusion correlated with high tumor expression of NY-ESO-1.⁴² This approach is being further expanded tweaking the NY-ESO-1–targeted TCR-engineered T cells either to express CD8-α to improve stabilization of the TCR signaling or to express a dominant negative transforming growth factor β (TGF-β) receptor to reduce the suppressive signals from TGF-β in the tumor microenvironment TME; NCT03967223).

PREVIOUS NON–CAR-T ACT TRIALS IN pMST

The rest of this review will focus on use of ACT approaches that do not redirect immune cell specificity, with special focus on the therapeutic potential of tumor-infiltrating lymphocyte (TIL) ACT. Most non–CAR-T ACT trials involving pMST have been combined pediatric/adult trials or trials for adolescents and young adults with peripheral blood serving as the immune cell source. In an attempt to address the poor prognosis of metastatic Ewing sarcoma (ES) and metastatic alveolar rhabdomyosarcoma, a clinical trial aimed at consolidating remission in 52 metastatic and recurrent patients was undertaken in children and young adults.⁴³ Before chemotherapy, blood was collected with apheresis for isolation of immune cells. After standard multimodal therapy, 30 patients continued on the trial and were divided into 3 cohorts. All cohorts received autologous T cells in conjunction with a DC vaccine pulsed with ES or alveolar rhabdomyosarcoma translocation breakpoint peptides and a control E7 peptide from human papillomavirus. Interleukin 2 was administered in moderate-dose, low-dose, or none to cohorts 1, 2, and 3, respectively. All patients apheresed had a 31% 5-year survival, whereas those receiving ACT had a 43% 5-year survival. Treatment was well tolerated with minimal toxicity, and patients analyzed had favorable influenza immune responses implying functional immune capacity following their up-front cytotoxic chemotherapy treatment. Cohort 1 (moderate-dose IL-2) seemed to have improved immune responses. Investigators concluded that despite the profound lymphopenia and alteration of T-cell subsets known to occur with sarcoma chemotherapy, this did not preclude use of postchemotherapy ACT.⁴³

Patients with recurrent or refractory Epstein-Barr virus (EBV)–positive nasopharyngeal carcinoma have been studied in a phase I/II trial enrolling children and adults with autologous, EBV-specific cytotoxic T lymphocytes (EBV-CTLs).⁴⁴ This approach utilizes known TAAs (EBVantigens) with ACT attempting to infuse large numbers of tumor-specific T cells expanded and activated ex vivo and outside the tumor environment. Of 23 total patients, 8 patients were in remission, and 15 had active disease at the time of cell infusion. No significant toxicity was observed. For the 8 patients treated in remission, 5 remained disease-free after ACT infusion for a median of 42 months (range, 25–82 months). Of the 15 patients treated with relapsed or metastatic disease, 33% had a CR, with 15.4% achieving a PR. Outcome was not statistically influenced by EBV antigen specificity of the infused CTLs or by postinfusion expansion. Both progression-free survival and overall survival were lower in patients with metastatic disease compared with locoregional disease in this study. Investigators concluded that infusion of EBV-CTLs for treatment relapsed/refractory EBV-positive nasopharyngeal carcinoma is safe and provides potential long-term benefit, especially for patients with locoregional disease.⁴⁴

DIFFERENCES BETWEEN ADULT AND PEDIATRIC SOLID TUMORS: CHALLENGES AND OPPORTUNITIES

Two significant challenges exist for ACT approaches, which do not redirect immune cell specificity in pMST compared with adult solid tumors: (1) differences between adult and pediatric adaptive immune systems and (2) the histologic origin of tumors within the respective age groups. In the adaptive immune system, children have higher levels of regulatory T cells, and their CD4 T cells are biased toward assisting T_H2 responses, whereas DC populations are skewed toward a higher plasmacytoid DC–to–myeloid DC ratio.⁴⁵ The immune system gradually matures through adolescence into a more balanced T_H1:T_H2 system capable of mounting antipathogen and non–self-tissue responses in adulthood.⁴⁶ Before this process, the early landscape of infancy is geared toward preventing autoimmune activation at a cost of decreased ability to mount strong T-cell–mediated responses against pathogens and malignant tissue.^45,46

Regarding tissue of origin, pediatric cancers generally arise from less-differentiated mesenchymal cell lineage, whereas adult cancers typically derive from more differentiated cells that form components of individual organs.^47,48 For example, adult tumors arising from known environmental factors such as UV and cigarette smoking involved in melanoma and lung cancer, respectively, can harbor hundreds of nonsynonymous mutations associated with tumorigenesis.⁴⁹ Furthermore, DNA repair defects that allow cellular division to occur in cells with potentially mutagenic DNA damage over lengthy periods can have thousands of mutations.⁵⁰ In contrast, pediatric tumors have an average of 9.6 mutations.⁴⁷ This lower TMB in pMST also leads to lower immunogenicity and less lymphocytic infiltration.⁵¹ Thus, when the source of the ACT is the patient's endogenous T-cell repertoire, there may be fewer tumor-specific T cells present initially. The TME of solid tumors provides additional hurdles to effective ACT in children.^52-54 Similar to adult tumors, pediatric tumors maintain an immunosuppressive microenvironment including infiltration with tumor-associated macrophages and myeloid-derived suppressor cells as well as metabolic perturbations that create a TME hostile to proliferation of cytotoxic lymphocytes.^55-59

POTENTIAL FOR TILS IN pMST

While research continues to improve upon antibody-based immunotherapy, CAR-T, and TCR therapy, TIL-based ACT is also an attractive candidate therapy for treatment of refractory and advanced-stage solid tumors. Tumor-infiltrating lymphocyte–based ACT involves the ex vivo expansion of tumor-derived T cells from a patient and infusing the expanded cell product back to the patient.^7,60 The rationale for this approach is that TILs should recognize the patient's TAA but are suppressed within the TME. By expanding the T cells away from the immunosuppressive TME, they should have greater capacity to mount antitumor immune responses upon infusion back into the patient.⁶⁰ Unlike toxicities with CAR-T therapy, off-target adverse effects are minimized because TAAs are less likely to be coexpressed on nonmalignant tissue.⁶¹ Another important difference between TIL-based ACT compared with gene-modified T-cell therapy is the polyclonal nature of the infusion product, which enables a diverse T-cell response to match the heterogeneity seen in solid tumors, potentially decreasing the risk of antigen loss and immune escape.

Most trials reported have focused on adult metastatic melanoma with a 40% to 50% durable objective response rate and a 10% to 20% CR rate.^61-65 Preclinical and early clinical TIL trials targeting adult solid tumors illustrate the potential of this modality.^60,65-70 According to ClinicalTrials.gov, there more than 100 actively recruiting TIL-based ACT trials for solid tumors enrolling patients 18 years or older. In contrast, the number of actively recruiting TIL-based ACT trials enrolling patients 18 years or younger is only 3 (NCT03645928, NCT03449108, NCT01955460), and these trials focus on diseases that, with the exception of sarcoma, are rare in pediatrics.

INVESTIGATION OF TILS IN pMST: PRECLINICAL AND EARLY CLINICAL

In neuroblastoma, TILs have been successfully expanded from freshly resected tumors with identified CD8- and CD4-positive T cells (both central memory and effector), as well as a significant number of NK T and γδT cells.⁷¹ When isolated TILs were nonreactive against neuroblastoma cell lines, they were then transduced with a CAR-T product targeting GD2, which showed in vitro cytotoxicity. MYCN nonamplified tumors have been found to have significantly more cytotoxic TIL signatures than MYCN amplified.⁷² Another group established 3 allogeneic neuroblastoma cell lines from different metastatic deposits in the same patient, and TILs seemed to preferentially kill neuroblastoma cells in vitro.⁷³ Tumor-associated lymphocytes obtained from neuroblastoma also seem to have an activated phenotype, especially in the CD8⁺ T cells.⁷⁴

Wilms tumor samples have demonstrated a higher concentration of CD8⁺ T cells in the tumor as compared with the peritumoral area.⁷⁵ Tumor-infiltrating lymphocytes have also been derived from Wilms tumors that demonstrated cytotoxic activity.⁷⁶ Another study found that Wilms tumor is infiltrated by immune cells both before and after chemotherapy with elevated NK cells and CD4/CD8⁺ T cells, which displayed an activated phenotype (HLA-DR, PD1, and CD57 expression increased).⁷⁷ Germ cell tumors have also been assessed for TILs preclinically: seminomas and germinomas were found to be highly infiltrated by T cells, whereas embryonal carcinomas, choriocarcinomas, and yolk sac tumors demonstrated high infiltration of T cells combined with an immunosuppressive microenvironment.⁷⁸

For osteosarcoma, T-cell infiltration may correlate with survival; preclinical studies in dogs suggest that there are decreased CD8⁺ TILs during tumor progression versus regression.⁷⁹ In a study isolating TILs from 27 patients with pediatric bone tumors (osteosarcoma, ES, giant cell tumor, chondrosarcoma), it was found that only TILs extracted from osteosarcomas were cytotoxic.⁸⁰ In a study of osteosarcoma of the jaw, 21 samples were analyzed for TILs and positive for CD8⁺ T cells in approximately half of cases and CD4⁺ T cells in a third of cases.⁸¹ Preclinical data also seem to suggest that exposure to chemotherapy may increase TILs in osteosarcoma. A study of the TARGET database using 27 matched osteosarcoma patient samples assessed changes in the immune environment after chemotherapy and found that neoadjuvant chemotherapy was associated with increased infiltration of all T cells, and specifically CD8⁺ T cells, into the tumors, with decrease in myeloid-derived suppressive cells.⁸² Additional work has found that genomic instability of osteosarcoma leads to neoantigen generation, potentially making it more immunogenic or “hot”; in this study, increased TILs were found in metastatic disease compared with the primary tumor, hinting a use for T-cell–based immunotherapy in metastatic osteosarcoma.⁸³ Two active trials using TIL therapy along with lymphodepleting chemotherapy and post-TIL IL-2 are currently enrolling patients as young as 16 years old (NCT03449108, NCT04052334).

Although primarily focused on adults, recent preclinical work has demonstrated the ability to harvest and expand TILs with tumor-specific function in soft tissue sarcoma samples.⁸⁴ Successful TIL expansion on this study included diagnoses that may affect children and adolescents, including SS, desmoplastic small round cell tumor, and malignant peripheral nerve sheath tumor. Recently, we also demonstrated the ability to harvest and expand TILs from 14 of 18 pMST samples taken from numerous diagnoses including Wilms tumor, neuroblastoma, soft tissue sarcoma, and bone sarcoma, providing preliminary data to support a clinical trial of TIL-based ACT in pMST.⁸⁵ The work in the adult population supports the currently accruing phase I clinical trial of TIL-based ACT in adolescents and young adults with soft tissue sarcoma (NCT04052334).

OBSTACLES AND FUTURE DIRECTIONS

To optimize the therapeutic potential of TILs in pMST, several obstacles will need to be addressed likely through a combination of ex vivo manipulation of TILs and adjuvant therapeutics.⁸⁶ First, lack of T-cell recognition of tumor may occur through paucity of tumor-specific neoantigens or through tumor downregulation of major histocompatibility complex molecule or neoantigen expression.⁸⁷ This may be especially true for pMST that have relatively low TMB. Second, the TME may inhibit optimal TIL activity through multiple mechanisms. Immunosuppressive cells (myeloid-derived suppressor cells, regulatory T cells) may secrete immunosuppressive cytokines (e.g., IL-10, TGF-β) inhibiting TIL function.^88,89 The TME may upregulate checkpoint molecules to induce T-cell exhaustion. The coadministration of ICIs alongside TILs is currently being studied in multiple adult solid tumor types to address this. The TME may also impose physical or chemical barriers to the trafficking of TILs to the tumor.⁹⁰ One strategy to modulate the TME and enhance TIL function includes insertion of a synthetic CD8α:MyD88 coreceptor into TILs during ex vivo expansion conferring improved trafficking to the tumor and resistance to signals inducing T-cell exhaustion.⁹¹

CONCLUSIONS

Prior immunotherapy investigation in pediatric cancer has resulted in clinical efficacy in only a small subset of diagnoses, primarily hematologic malignancies. While CAR-T is a transformational therapy for lymphoid malignancy, its success thus far has not translated to pMST. While genetically engineered TCR products can address the paucity of known TAA in pMST and has demonstrated clinical activity in SS, this approach limits the scope of targets within heterogeneous tumors. With the possibility of disease recurrence due to nonimmunogenic tumor cell subclones, TIL-based ACT offers an approach that can limit the refractory response to therapy because of the polyclonal nature of the infusion product. It is therefore an opportune time to fully investigate the utility of TIL-ACT in pMST and also potentially exploit the advantages of targeting multiple TAAs within heterogeneous tumors.

Hurdles presently exist for pMST TIL-based ACT, and differences between adult and pediatric immune systems will require modifications to achieve greater efficacy. Yet, its potential to address unmet needs among refractory, high-tumor-burden and metastatic pMST patients is far too great to overlook. The low presence of mutational burden in the pMST landscape illustrates a need to isolate tumor-reactive T cells from TILs ex vivo away from the suppressive environment within the patient. These isolated T cells can then be reintroduced to initiate an antitumor response likely more reflective of the adult immune system. Another benefit of this strategy is that tumor-reactive T cells expanded in this manner should help circumvent the noted off-target effects of CAR-T ACT. Preclinical work in pMST demonstrates that the hurdle of isolating TILs and expanding them ex vivo to clinically relevant numbers in pMST is certainly surmountable and should undergo further preclinical evaluation as well as clinical investigation in early-phase trials.