Expansion of tumor-infiltrating and marrow-infiltrating lymphocytes from pediatric malignant solid tumors

Abstract

Introduction:

The expansion of tumor-infiltrating lymphocytes (TIL) for adoptive cellular therapy is under investigation in many solid tumors of adulthood. Marrow-infiltrating lymphocytes (MIL) have demonstrated antitumor reactivity preclinically. Successful expansion of TIL/MIL has not been reported across pediatric solid tumor histologies. The objective of this study was to demonstrate successful expansion of TIL from pediatric solid tumors for translation in an adoptive cell therapy (ACT) treatment strategy.

Methods:

A prospective study of TIL/MIL expansion was performed on solid tumors of pediatric patients undergoing standard-of-care procedures. TIL/MIL expansions were performed in the presence of high-dose interleukin 2. To demonstrate a full-scale expansion to clinically-relevant cell doses for TIL therapy, initial TIL culture was followed by a rapid expansion protocol for select patients. Expanded specimens were analyzed for phenotype by flow cytometry and for anti-tumor reactivity by the interferon-gamma release assay.

Results:

Eighteen tumor samples were obtained. Initial TIL cultures were successfully generated from 14/18 samples (77.7%). A median of 5.52 × 107 (range: 2.5 × 106–3.23 × 108) cells were produced from initial cultures, with 46.9% expressing a CD3 phenotype (46.9%). Eight samples underwent rapid expansion, demonstrating a median 458-fold expansion and a CD3 phenotype of 98%. Initial MIL cultures were successfully generated from five samples, with a predominantly CD3 phenotype (45.2%). Sufficient tumor tissue was only available for seven TIL samples to be tested for reactivity; none demonstrated responsiveness to autologous tumor.

Conclusions:

TIL and MIL expansion from pediatric solid tumors was successful, including the full-scale expansion process. This data supports translation to an ACT-TIL treatment strategy in the pediatric population and thus a Phase I trial of ACT-TIL in pediatric high-risk solid tumors is planned.

Introduction

Pediatric extra-cranial malignant solid tumors (pMST) comprise 30%–40% of pediatric cancers [1,2]. Despite intensive multimodality therapeutic approaches, pMST carry a heavy burden of morbidity and mortality, especially for patients with high-risk features such as metastatic disease at diagnosis and relapsed/refractory disease [3–7]. Novel therapies are required to improve survival outcomes and decrease toxicity burdens for pMST.

Adoptive cellular therapy (ACT) is based upon the expansion or modification of lymphocytes with potential anti-tumor activity (most commonly T cells) ex vivo followed by infusion into a patient to generate anti-tumor effects [8]. The use of T cell-based ACT for pMST patients is currently under clinical investigation. Encouraging clinical responses have been reported using GD-2 directed Chimeric Antigen Receptor T cells (CAR-T) for neuroblastoma and cancer-testis antigen targeting modified T cell therapy for synovial sarcoma, however, responses in other histologies and with other ACT products have been modest [9–13]. While continued progress in CAR-T cells and other genetically-modified, antigen-specific T cells for pMST is critical, alternative ACT approaches should be considered. ACT with tumor-infiltrating lymphocytes (TIL) involves obtaining tissue from a patient’s tumor, a massive ex vivo expansion of T cells from the tumor, and reinfusion of the T cells back into the patient. This approach was recently FDA-approved based on complete and durable responses in metastatic melanoma not otherwise possible with conventional therapy or emerging investigational agents [14,15]. Advantages of TIL include a polyclonal population of T cells, which are expanded away from the immunosuppressive tumor microenvironment, and the lack of genetic modification which presents specific patient risks. While TIL therapy is most heavily studied in melanoma, reliable processes for TIL expansion from several adult solid tumor histologies have been reported recent years, including investigations by our group in bladder cancer, penile cancer, and soft tissue sarcoma [16–18]. Due in part to continued improvements in the ex vivo expansion process, ACT-TIL is in clinical trials for many adult solid tumors outside of melanoma, with recent durable responses seen in anti-PD-1 resistant metastatic lung cancer [19].

The bone marrow serves as a niche for memory T cell maintenance. Preclinical data show that tumor-specific memory CD8+ T cells reside in the bone marrow of cancer patients and that these T cells (here referred to as marrow-infiltrating lymphocytes, MIL) may induce regression of established xenograft tumors in SCID mice [20–22]. A clinical trial in multiple myeloma patients demonstrated that adoptive transfer of MIL is feasible as another potential source of tumor-specific T cells for ACT, and responses to therapy were seen [23].

Full-scale TIL expansion to clinically relevant doses for ACT is a multistage process involving generation of a primary TIL culture to dissociate TIL from tumor cells and establish initial growth. This is followed by a rapid expansion protocol (REP) to achieve a number of TIL (typically >5 × 109) suitable for ACT-TIL therapy. Despite extensive investigation of TIL expansion in many adult solid tumors, there has been minimal investigation of TIL expansion in pMST, likely due unsuccessful early attempts at expansion [24,25]. In fact, with the exception of a study restricted to neuroblastoma, successful TIL expansion in pMST has not previously been demonstrated. Therefore, the primary objective of this study was to demonstrate the feasibility of TIL expansion from pMST using the contemporary expansion processes studied extensively in adult solid tumors. Secondary objectives included defining the lymphocyte subset composition of these TIL, testing the tumor-specific reactivity of pMST-generated TIL, and determining the feasibility of MIL expansion in pMST.

Materials and Methods

Patient population

This study was approved by the Johns Hopkins All Children’s Hospital Institutional Review Board (IRB00193453). Patients or families provided consent for this study as appropriate based on age. A pMST was defined as noncentral nervous system solid tumor with known local invasiveness and/or metastatic potential including but not limited to sarcomas, Wilms tumor, neuroblastoma, hepatoblastoma, and germ-cell tumors. Lymphomas, histiocytic diseases, and carcinomas were excluded. For all other diagnoses, eligibility was determined by the principle investigator (J.M.). Additional eligibility criteria included age ≤21 years and a planned standard-of-care surgical biopsy or resection of a primary tumor, tumor-involved lymph node, or metastatic site. Needle biopsies were not eligible. Samples of pMST obtained from central nervous system metastases were allowed. Samples from patients receiving neoadjuvant chemotherapy and/or radiation prior to their surgical procedure were allowed. Bone tumor resections were allowed if there was a representative extraosseous soft tissue tumor component or an accessible intraosseous site with likely viable tumor.

Sample acquisition

Immediately following tissue acquisition and under sterile conditions, a fresh tumor sample (≥1 cm3) was allocated in RPMI media by the pathologist after it was determined there was sufficient tissue available for the patient’s clinical care. Bone marrow specimens were obtained through bone marrow aspiration as part of standard-of-care procedure or, in the case of resected bone tumors, reamings were obtained from the marrow cavity prior to endoprosthetic reconstruction. An amount of 10–20 mL bone marrow was collected in either a cell preparation tube (CPT tube, BD Biosciences, Franklin Lakes, NJ) for aspirates or in heparinized phosphate buffered solution for orthopedic surgeries. All specimens were deidentified and labeled with a unique study ID. A portion or all of the specimen was processed for TIL culture, and the remaining specimen was used for tumor digest in reactivity assays, when feasible. Hematoxylin and eosin-stained (H&E) slides representative of allocated tumor samples with the lowest and highest TIL growth were reviewed by an institutional pathologist (H.M.) for determination of tumor content and evaluation of the pattern of immune cell infiltration.

Primary TIL and MIL expansion (preREP)

Tumors were prepared using techniques previously described by our group [15,26]. Briefly, tumor fragments were minced into pieces ~1mm3 in size and explanted into at least 12 wells (1 fragment/well) of a 24-well plate in culture media containing interleukin 2 (IL-2) (6000 IU/mL). One-half of the media containing IL-2 was taken out and replenished with fresh culture media supplemented IL-2 (6000 IU/mL) every 3 days to maintain the proliferation of the TIL. When TIL reached 80% confluent within the well, TIL was split 1:2 into a new well of a 24-well plate and replenished with fresh media (6000 IU/ml IL-2). After 4 weeks of initial culture, the final TIL product was collected, counted, and used for phenotypic and reactivity assays.

Bone marrow specimens were processed by Ficoll density gradient centrifugation and white blood cells were further treated with red blood cell lysis buffer for red blood cell removal. Cells were assessed for viability using trypan blue exclusion and resuspended in culture media supplemented with IL-2 (6000 IU/ml) at 5 × 10⁶ live cells/mL. A total of 1 × 10⁶ live cells were seeded along with culture media to a final volume of 2 mL in a single well of a 24-well tissue culture plate. MIL proliferation was scored every 3 days and one-half of culture media was taken out from top without disturbing cells and replenished with fresh media 6000 IU/ml IL-2 every 3 days to maintain MIL proliferation for 4 weeks. When wells reached 80% cell confluency, MILs were split 1:2 and replenished with fresh culture media supplemented with IL-2 (6000 IU/mL).

TIL/MIL phenotype

To determine the phenotype of TIL/MIL, approximately 2.5 × 105 cells were stained for Zombie NIR and surface markers CD3, CD4, CD8, and CD56. Single-cell acquisition was performed using a BD Celesta flow cytometer and FACSDiva (Becton, Dickinson and Company, Franklin Lakes, NJ). The data was further analyzed using FlowJo software (Becton, Dickinson, and Company). The phenotype of the TIL/MIL product was reported as a percentage of the live cell population.

Rapid expansion of TIL (REP)

To demonstrate the feasibility of the complete production process, initial TIL cultures were expanded using a REP developed for prior melanoma clinical trials [15,26]. TIL was loaded into a T25 flask with media containing IL-2 (3000 IU/mL), OKT-3 (30 ng/mL), and irradiated donor lymphocytes in a TIL:feeder ratio of 1:200. Media was added on days 4 and 7, and cells were transfer to a T75 flask on day 10 and to a T150 flask on day 12. Cells were counted on day 14 using trypan blue exclusion.

Tumor reactivity

TIL samples postREP were co-cultured with autologous tumor cells at a 1:1 ratio (1 × 10⁵ cells each) for 24 hours. The supernatant was collected, and interferon-gamma (IFN-γ) concentration was measured using an enzyme-linked immunosorbent assay. Anti-CD3/CD28 stimulation was performed as a positive control. Comparisons were made between IFN-γ secretion in presence or absence to autologous tumor cells. Anticlass I major histocompatibility complex blocking antibody (W6/32 clone; Becton, Dickinson, and Company) was also used to determine CD8+ T-cell–specific mediated tumor response. TIL cultures were deemed to have tumor-specific reactivity if >100 pg/mL IFN-γ was present in the supernatant and there was a statistically significant difference between the autologous tumor and anticlass I major histocompatibility complex blocking condition.

Statistical considerations

Statistical analysis for data sets comparing ex vivo preREP and postREP expansions among TIL and MIL were carried out using two-tailed unpaired t-test. TIL phenotypic comparison between preREP and postREP TIL was statistically analyzed using Ordinary Two-Way ANOVA with šidák multiple comparison test. Significant statistical difference is defined as P ≤ .5 and is represented with asterisks on top of comparison brackets; * = P ≤ .05; ** = P ≤ .01; *** = P ≤ .001; **** = P ≤ .0001.

Results

Patient data

Twenty samples were obtained from 2019 to 2021. Two samples were later deemed ineligible when pathology did not confirm pMST, leaving 18 samples from 17 unique patients for analysis (one patient was enrolled during two different surgeries, resulting in two unique samples). Five tumor samples had concurrent bone marrow collected. The median age was 12 years (range 1–20 years), and neoadjuvant treatment (defined as within 90 days of tumor sampling) was administered in eight patients (44.4%; chemotherapy, n = 5 and immunotherapy, n = 3). The prior immunotherapies included pembrolizumab (n = 2) and dinutuximab with sargramostim (n = 1). Patient and tumor characteristics are listed in Table 1.

Table 1.

Patient and tumor characteristics.

Sample	Dx	Stage, riskc	Age (years)	Sex	Disease status	Tumor sampling site	Prior therapyb	MIL analyzed?
1	NBL/GNBL	L1, low	3	M	Initial treatment	Organ/soft tissue, primary site	N	Yes
2	WT	2, standard	2	M	Initial treatment	Organ/soft tissue, primary site	N	No
3	WT	4, higher	1	M	Initial treatment	Organ/soft tissue, primary site	N	No
4	OS	4	15	M	Relapse	Organ/soft tissue, metastatic site	C	No
5	WT	4, higher	7	M	Initial treatment	Organ/soft tissue, lymph node	N	No
6	NBL/GNBL	M, high	7	M	Relapse	CNS, metastatic site	I	No
7	IMT	1	14	F	Initial treatment	Organ/soft tissue, primary site	N	No
8	ML	1	16	F	Initial treatment	Organ/soft tissue, primary site	N	No
9	OS	4	14	F	Initial treatment	Bone tumor, primary site	C	Yes
10a	ASPS	4	20	M	Relapse	CNS, metastatic site	I	No
11	OS	4	13	M	Relapse	Organ/soft tissue, metastatic site	C	No
12a	ASPS	4	20	M	Relapse	CNS, metastatic site	I	No
13	NBL/GNBL	M, high	7	M	Initial treatment	Organ/soft tissue, primary site	N	Yes
14	SS	2	13	M	Initial treatment	Organ/soft tissue, primary Site	N	No
15	ERMS	3, intermediate	3	F	Initial treatment	Organ/soft tissue, primary site	N	No
16	OS	2	14	M	Initial treatment	Bone tumor, primary site	C	Yes
17	WT	3, standard	4	M	Initial treatment	Organ/soft tissue, primary site	N	No
18	OS	2	12	M	Initial treatment	Bone tumor, primary site	C	Yes

OS, osteosarcoma; WT, Wilms tumor; NBL/GNBL, neuroblastoma or ganglioneuroblastoma; IMT, inflammatory myofibroblastic tumor; ASPS, alveolar soft part sarcoma; SS, synovial sarcoma; ERMS, embryonal rhabdomyosarcoma; ML, myxoid liposarcoma; M, male; F, female; N, no prior therapy; C, prior chemotherapy; I, prior immunotherapy; MIL, marrow-infiltrating lymphocytes.

^aDenotes tumor samples from the same patient at different timepoints.

^bDenotes therapy within 90 days of tumor sampling.

^cTumor stage and risk stratification (included for tumor types with complex risk stratification systems beyond stage) are based on National Comprehensive Cancer Center guidelines (www.nccn.org) or, for rhabdomyosarcoma, Haduong et al, Pediatric Blood and Cancer 2022.

Lymphocytic infiltrate within pMST

H&E slides from tumors with no TIL yield from initial culture (ganglioneurobastoma, alveolar soft part sarcoma, and an osteosarcoma lung metastasis) demonstrated either rare or no apparent lymphocytes by H&E (Figure 1). In contrast, two of three tumors with high yield of TIL from initial culture (inflammatory myofibroblastic tumor and embryonal rhabdomyosarcoma) demonstrated interspersed conspicuous lymphocytic infiltrates. The third tumor with high yield, an osteosarcoma primary resection, showed rare lymphocytes by H&E but interspersed lymphocytes highlighted by CD45 immunohistochemistry (Figure 1).

Fig. 1.

H&E photomicrographs of nongrowing (upper panels) and heavily growing (lower panels) TIL samples from tumor fragment culture. Upper left: GNBL, showing edge of less mature intermixed and stroma rich maturing tumor with no apparent lymphocytes. Upper middle: ASPS CNS metastasis, showing a highly cellular confluent epithelioid cells with no apparent lymphocytes. Upper right: OS metastasis in lung, showing cellular sheets of epithelioid osteogenic sarcoma cells with focal discernible osteoid matrix production and isolated stromal lymphoid cells (center) and adjacent lung (upper right). Lower left: IMT, showing sheets of myofibroblastic cells with vesicular nuclei, prominent nucleoli amid characteristic interstitial plasma cell and other lymphoreticular cell infiltrate (arrows). Lower middle: ERMS, showing embryonal cells with rhabdomyoblastic differentiation in myxoid and alternating cellular areas with isolated intercellular lymphoid cells (arrows). Lower right: OS post-therapy primary tumor, (insert with CD45 IHC) showing postchemotherapy osteogenic sarcoma, residual (10% viable) large neoplastic cells amid dense collagenous tissue and native woven bone spicule right lower corner. H&E, hematoxylin and eosin; IHC, immunohistochemistry; GNBL, ganglioneuroblastoma; ASPS, alveolar soft part sarcoma; CNS, central nervous system; OS, osteosarcoma; IMT, inflammatory myofibroblastic tumor; ERMS, embryonal rhabdomyosarcoma; Pre-REP, prior to rapid expansion.

Primary TIL and MIL culture

An average of 27 (range 12–96) tumor fragment wells were used in initial culture. Lymphocytes were successfully cultured from 14 of 18 initial cultures (77.7%). The unsuccessful cultures were a ganglioneuroblastoma, the two samples from a single patient with alveolar soft part sarcoma, and a relapsed osteosarcoma lung metastasis. The median number of TIL grown from successful cultures was 5.52 × 107 (range: 2.5 × 106–3.23 × 108, Figure 2A). Successful cultures were obtained from multiple tumor types and sites of acquisition (Figure 2B). Successful establishment of TIL culture occurred from therapy-naïve (9/10, 90%) and pretreated (chemotherapy and immunotherapy) samples (5/8, 62.5%). There was high variability in initial culture yield regardless of age, diagnosis, and site of acquisition (Figure 2B), and a comparison of yield from patients 12 and under versus 13 years and older showed no obvious differences in culture yield. (Figure 2C). Samples demonstrated high percent viability after initial culture (mean 84% viability, Figure 2D). Eleven of the 14 successful cultures underwent flow cytometry analysis; the average phenotypes of the lymphocytes from primary TIL cultures included CD3+ (51.53%), CD4+ (24.39%), CD8+(27.97%) and CD56+ (22.64%; Figure 2E).

Fig. 2.

Expansion and phenotypic characterization of lymphocytes from tumors and bone marrow of pediatric malignant solid tumors. (A) TIL and MIL expansion from initial culture (PreREP). (B) PreREP TIL expansion by age, diagnosis, and tumor acquisition site. (C) PreREP TIL expansion comparison among preteenager (12 and under) and teenager (13 and older) patients. (D) PreREP TIL and MIL cell viability. (E) Pre- and post-REP TIL Phenotype comparison. (F) Comparison of TIL and MIL preREP TIL among paired samples. (G) Pre- and post-REP MIL Phenotype comparison. (H) Fold change expansion of TIL and MIL PostREP, (I) cell viability obtained of TIL and MIL PostREP, (J) interferon-gamma concentration in supernatant of three (3) representative PreREP TIL samples when co-cultured with autologous tumor. (K) Co-culture assay positive control (anti-CD3/CD28 stimulation) for interferon-gamma release. PreREP, initial culture before rapid expansion; REP, rapid expansion protocol; PostREP, following rapid expansion; TIL, tumor-infiltrating lymphocytes; MIL, marrow infiltrating lymphocytes; OS, osteosarcoma; WT, Wilms tumor; NB, neuroblastoma/ganglioneuroblastoma; IMT, inflammatory myofibroblastic tumor; ASPS, alveolar soft part sarcoma; SS, synovial sarcoma; ERMS, embryonal rhabdomyosarcoma; ML, myxoid liposarcoma. Statistical significance was determined by two-way ANOVA; **P < .01***P < .001.

Five patients had marrow collected for MIL analysis (identified in Table 1). Primary MIL culture was successful from all five samples including all three samples originating from primary bone tumor resections. From these preREP samples, the median number of MIL was 9.55 × 106 (range 8.00 × 105–1.00 × 107, Figure 2A) and the cultured cells were highly viable (mean 88% viability, Figure 2D). Paired comparison of TIL/MIL samples showed high variability of intra-patient culture yield between specimen types (Figure 2F). The average phenotypes of the MIL cultures were: CD3+ (40.1%), CD4+ (11.3%), CD8+ (34.1%), and CD56+ (18.3%; Figure 2G).

Rapid expansion of TIL/MIL cultures

To replicate the full-scale expansion of TIL to clinically-relevant cell doses for TIL therapy, REP was attempted on eight primary TIL cultures. Five of the eight (62.5%) of samples produced REP samples with greater than 200-fold expansion. The average fold expansion during REP for all attempted specimens was 718.6 with a median of 458.6 (Figure 2H). Three MIL samples also underwent attempted REP, demonstrating mean 257- and median 241.7-fold-change (Figure 2H). Five post-REP TIL samples underwent flow cytometry analysis; the samples tested were highly viable (mean 88% viable, Figure 2I), the average phenotype of post-REP TIL were CD3+ (98%), CD4+ (72.5%), CD8+ (21.9%) and CD56+ (1.7%); a highly enriched CD3+ population, and a significant reduction in CD56 CD4+ percentage from the preREP samples (Figure 2E). One of the rapidly expanded MIL samples were phenotypically characterized and showed high percentage viability (96%, Figure 2I) a reduction of CD56+ cells, and an increase percentage of CD3+ cells and CD4+ cells as compared to preREP MIL samples (Figure 2G).

Tumor reactivity

Due to the small residual available tumor volumes available after allocation of tissue to clinical care, IFN-γ release assay for tumor-specific reactivity was feasible for 7 of 18 samples (representative data shown in Figure 2K, J). Anti-CD3/CD28 activation stimulated IFN-γ release showing high degree of TIL function, however, tumor-specific reactivity was not demonstrated.

Discussion

This report represents the first description of the feasibility of TIL to cell volumes relevant to ACT across multiple pMST histologies. We recapitulated various aspects of the multistage manufacturing process (primary TIL culture through full-scale REP) which are currently used in adult solid tumors and produced greater than 200-fold expansion in five of eight attempts at full-scale REP. TIL were grown from tumors prior and following chemotherapy and antibody-based immunotherapy. We successfully expanded TIL from both primitive embryonal tumors and sarcomas in a variety of anatomic locations. Most samples were from primary tumors or organ metastases and not from involved lymph nodes, as is common practice in melanoma. As pMST less frequently involve lymph nodes, using tumor tissue from the primary tumor and metastatic sites presents a more viable strategy for TIL expansion in pMST.

MIL expansion was also successful from five bone marrow specimens, including three from osteosarcoma resections. Memory T cells in the bone marrow have been shown to be capable of reactivation to cytotoxic effector in patients with breast cancer when exposed to antigen-presenting cells, with regression of tumor models in xenotransplanted NOD/SCID mice [20–22]. Studies in multiple myeloma have demonstrated that MIL exist in an immune-responsive microenvironment and demonstrate greater polyclonal antigen specificity than lymphocytes from peripheral blood. Further investigation into the immunogenic properties of MIL in pMST should be performed.

Despite investigation of ACT-TIL as cancer therapy for over 40 years, studies of TIL expansion from pMST have been rare. Two Wilms tumor samples and one neuroblastoma sample were included in an early study of TIL in urologic malignancies, with expansion of TIL over 100-fold achieved only from one Wilms tumor sample [24]. In another early study, only 9 of 37 pMST diagnoses were able to generate a 10-fold expansion of TIL [25]. Key differences between these studies and ours include IL-2 dosing for primary culture (10–1000 IU/mL IL-2 on prior studies versus 6000 IU IL-2 on our study) and the addition REP after initial TIL culture on our study. An investigation of TIL expansion from 30 neuroblastoma samples demonstrated effective expansion from 23 of 30 samples (76.7%), with almost 300 mean fold increase through REP [27]. While this study used a similar multistage expansion process, there are several key differences with our study. First, TIL expansion was demonstrated across a diverse range of pediatric solid tumor histologies in our study. The IL-2 culturing strategies also differed between the two studies, with higher doses of IL-2 used in primary culture and REP on our study. Finally, the initial cultures on the neuroblastoma study were generated using human T-cell activating CD3/CD28 beads to facilitate expansion, a strategy more likely to favor expansion of non-tumor specific T cells, whereas our study demonstrated successful TIL expansion using IL-2 alone, a strategy more clinically relevant because it is identical to widely-accepted procedures for TIL expansion in adult clinical trials [28].

Preclinical investigation has characterized immune cell infiltration and its relationship with clinical outcomes for many pMST. A study of pMST found significant T cell infiltration in approximately one-third of high-risk tumors [29]. Variable T cell infiltration was noted in a landscape study of 14 pMST histologies, and CD8 infiltration was noted to correlate with a favorable prognosis in osteosarcoma [30]. In a study isolating TIL from 27 patients with pediatric bone tumors extracted from osteosarcomas demonstrated higher cytotoxicity compared to other bone tumor histologies [31]. In high-risk neuroblastoma, tumors have shown clinically relevant cytotoxic TIL signatures and clonal expansion of T-cell receptors, most notably in non-MYC-amplified samples [32]. Tumor-associated lymphocytes obtained from neuroblastoma also appear to have an activated phenotype, especially in the CD8+ T-cells [33]. In germ-cell tumors, seminomas and germinomas show substantial infiltration with T-cells while embryonal carcinomas, choriocarcinomas, and yolk sac tumors show T-cell infiltration within the context of an immunosuppressive microenvironment [34]. Studies in Wilms tumor show tumor infiltration by immune cells both before and after chemotherapy with elevated NK cells and CD4/CD8+ T-cells displaying an activated phenotype [35]. TIL has also been derived from Wilms tumors that demonstrated cytotoxic activity [36]. These findings suggest TIL is present in a significant portion of pMST, and the correlation of T cell infiltration with prognosis in certain histologies suggest anticancer effects of TIL that may be exploited through expansion.

Our study has several limitations, some of which were inherent in performing a noninterventional study in the pediatric population. First, because of we used leftover samples following standard-of-care surgeries and biopsies, our tumor samples were quite small, on average yielding only 27 tumor fragments, which limited initial TIL culture size. This was due to the needs to preserve proper pathologic assessment of the tumor as well as competing tissue allocation for other high-priority studies. We anticipate that larger tumor allocation will be required when enrolling patients onto a clinical trial for ACT-TIL [23]. Strategies to increase sample size in the clinical setting may include using large tumor resections for TIL acquisition and also acquiring a larger allocation of tissue when, in the setting of a trial, there is potential for direct patient benefit. Second, also due to the small size of samples, tumor-specific reactivity testing by IFN-γ release was only performed in seven samples, with no samples reactive. It is possible that the true rate of reactivity would have been higher if IFN-γ release was tested in a higher proportion of samples. It is also possible that reactivity was not captured in the IFN-γ release assay due to the high percentage of CD4+ T cells in our TIL expansion post-REP. CD4+ T cell reactivity would not be captured in the IFN-γ release assay since the tumors may not express major histocompatibility class II molecules. While IFN-γ reactivity is often used for selection of TIL cultures to undergo REP, allowing enrichment of tumor-reactive TIL in the final cellular product, the lack of reactive TIL from primary cultures have typically not precluded production of a final TIL product on clinical trials. In fact, clinical responses have been documented in patients receiving TIL therapy with CD4+-prominent TIL fractions as well as negative tumor-specific reactivity assays [19,37].

In summary, this study successfully demonstrates that TIL expansion is feasible from pMST. A prospective biobanking study and Phase I trial of ACT-TIL for patients with relapsed pMST without curative therapy options is now planned (NCT06047977). Patients enrolling on the prospective biobanking study will have tumor resected by standard-of-care surgery submitted for primary TIL culture. This will allow additional investigation of pMST TIL phenotype including exhaustion markers, memory states, and further evaluation of TIL tumor reactivity outside of standard IFN-γ ELISA, such as intracellular cytokine staining flow cytometry. The multistage TIL expansion process will then be leveraged by cryopreserving primary preREP TIL until a patient meets eligibility for treatment on the Phase I study, and at that time TIL will enter REP and be available for therapy in a short time frame. Additional correlative testing will include TIL clonality assessment prior to infusion and from patient blood samples at serial time points after treatment to assess clonal TIL persistence. After demonstrating the safety and feasibility of this treatment in children, future efforts will be required to determine the optimal diagnostic target for TIL therapy and how to integrate TIL into combination therapy approaches in pediatrics.

Abbreviations:

ACT

Adoptive Cellular Therapy

CAR-T

chimeric antigen receptor T cells

CNS

central nervous system

H&E

hematoxylin and eosin

IFN-γ

interferon-gamma

MIL

Marrow Infiltrating Lymphocytes

pMST

pediatric malignant solid tumor

IHC

immunohistochemistry

TIL

tumor-infiltrating lymphocytes

IL-2

interleukin 2

REP

rapid expansion protocol

Glossary

Term	Definition
Tumor-infiltrating lymphocytes	Lymphocytes found within a tumor
Rapid expansion protocol	A protocol to rapidly generate large numbers of lymphocytes
Adoptive cellular therapy	A transfer of cells into a patient
Marrow-infiltrating lymphocytes	Lymphocytes residing in the bone marrow

Data Availability Statement

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.