Donor lymphocyte infusion as immune modulation after pediatric allogeneic transplantation

Abstract

Relapse remains a leading cause of failure after allogeneic hematopoietic cell transplantation for pediatric hematologic malignancies. Donor lymphocyte infusion (DLI) offers a uniquely accessible and scalable form of nonengineered adoptive cellular therapy. Although historically viewed as salvage therapy, emerging data support a paradigm shift toward using DLI as strategic immune modulation rather than rescue. This review synthesizes the biological rationale and clinical evidence for DLI in children, integrating the adult data when necessary to define translatable principles. Across platforms, outcomes rely on timing, disease burden, and the immune context. We highlight the superiority of prophylactic and preemptive DLI over therapeutic use and review contemporary structured approaches relying on treatment initiation at low and safe doses and escalation based on strict tolerance and monitoring criteria. Special consideration is given to haploidentical transplantation and mechanisms of immune escape such as HLA loss that limit T-cell–based interventions. Furthermore, we examine emerging combinations, including DLI with blinatumomab in B-cell acute lymphoblastic leukemia and hypomethylating agents in myeloid malignancies. Finally, we discuss T-cell subset manipulation to mitigate graft-versus-host disease while preserving the graft-versus-leukemia effects. All these data support a modern framework for pediatric DLI that prioritizes early, biology-informed interventions and rational combinations, underscoring the urgent need for pediatric-specific prospective studies.

Introduction and biological basis

Donor lymphocyte infusion (DLI) is the most established form of nonengineered adoptive cellular therapy following allogeneic hematopoietic cell transplantation (allo-HCT). As one of the most tested and cost-effective cellular therapies available to prevent or treat relapse, DLI consists of the administration of unmanipulated lymphocytes collected from the donor, typically via leukapheresis.^1–5 Its therapeutic intent is to augment graft-versus-leukemia (GVL) effects through donor T cells, which is achieved by normalizing the T-cell repertoire, improving the coordination between T-cell and B-cell immunity, and reversing local T-cell exhaustion among preexisting CD8⁺ T cells infiltrating the bone marrow.^6,7 Recent studies also indicated that the bone marrow microenvironment differs between DLI responders and nonresponders in acute myeloid leukemia (AML), with the presence of clonally expanded ZNF683⁺ CD8⁺ cytotoxic T lymphocytes showing in vitro specificity for patient-matched AML.⁸ However, the clinical utility of DLI is constrained by the competing risk of graft-versus-host disease (GVHD).⁹ A consistent paradigm has emerged across pediatric and adult studies; the efficacy and safety of DLI are determined primarily by timing, disease burden, and immune context at the time of infusion, rather than by the procedure itself (Figure 1).

Figure 1. Posttransplant DLI strategies in pediatrics.

AML, acute myeloid leukemia; B-ALL, B-cell acute lymphoblastic leukemia; DLI, donor lymphocyte infusion; FLT3, Fms-like tyrosine kinase 3; GVHD, graft-versus-host disease; HCT, hematopoietic cell transplantation.

The shift toward prophylactic and preemptive strategies

The benefit of DLI has been historically most pronounced in chronic myeloid leukemia and to a lesser extent in other myeloid malignancies,^10,11 whereas its efficacy in acute lymphoblastic leukemia (ALL) remains limited, but not negligible.^12,13 Therapeutic DLI administered for overt hematologic relapse is the least effective application because rapid leukemia kinetics frequently outpace the delayed expansion of donor T-cell–mediated GVL, resulting in modest response rates and inferior survival.^14–21 Outcomes are consistently superior when DLI is deployed earlier: preemptive DLI triggered by minimal residual disease (MRD) positivity or declining donor chimerism and prophylactic DLI administered to high-risk patients in remission, are both associated with improved disease control.^{9,14,15,17,21–23} Structured prophylactic DLI can recapitulate the protective effects of spontaneous GVL without requiring clinically overt GVHD, initiating treatment at a low, safe dose and escalating based on strict tolerance and monitoring criteria. Pediatric series (Table 1) have demonstrated effective disease control with acceptably low GVHD rates following delayed, structured DLI, supporting the concept that GVL can be more precisely and safely titrated than spontaneous GVHD, provided timing and dose escalation are carefully controlled.^{16,19,20,24–31}

Table 1.

Pediatric donor lymphocyte infusion studies (2020–2026)

Study	Strategy	n	Malignancies	Transplant platform	DLI timing	Other agents	Key outcomes	GVHD
Walther et al20	DLI	34	ALL, AML, JMML, CML, MDS, and solid tumors	MUD/haplo	Pro-DLI/pre-DLI/tDLI	Various	OS: pro-DLI, 100%; pre-DLI, 96%; and tDLI, 80%	GI-II aGVHD, 6%
Suryaprakash et al19	DLI	144	ALL, AML, lymphoma, and other leukemia	MSD/MUD/haplo	Pre-DLI	No	ORR mixed chimerism: MSD, 45%; MUD, 33%; and haplo, 43%	aGVHD DLI <30 d, 38% vs >30 d, 15%; and cGVHD DLI <30 d, 15% vs >30 d, 3%
Wei et al31	DLI	101	ALL and AML	haplo	Pro-DLI	No	5-y OS, 88%; EFS, 79%	Post-DLI GVHD: GI-II, 41%; GIII-IV, 5%,
Qi et al30	DLI	54	High-risk myeloid neoplasms	haplo	Pro-DLI	No	DFS, 79%	GIII-IV aGVHD, 17%; moderate/severe cGVHD, 8%
Hou et al16	DLI	23	ALL, AML, JMML, CML, and lymphoma	MSD/MUD/MMUD	Pro-DLI/pre-DLI/tDLI	No	EFS pro-DLI/pre-DLI > tDLI	GIII-IV aGVHD, 39%; moderate/severe cGVHD, 30%
Chandar et al27	DLI	17	Acute leukemia	MSD/MUD/haplo	Pre-DLI	±Lenalidomide	RFS, 63%	aGVHD, 77%; cGVHD, 38%; severe acute/chronic rare
Huschart et al29	DLI	10	High-risk AML	MSD/MUD/haplo	Pro-DLI >120 d	AZA	DFS, 90%; median F/U, 17 mo	No severe aGVHD and mild cGVHD
Booth et al26	DLI	17	AML	MSD/MUD/MMUD/haplo	Pro-DLI >120 d	AZA	2-y OS and LFS, 88%	No increase in GVHD
Bica et al25	DLI	16	AML	MUD/MSD/haplo	Pro-DLI/pre-DLI/tDLI	AZA	OS: pro-DLI/pre-DLI, 80%; tDLI, 45.4%	aGVHD, 12% (only in tDLI) and no cGVHD
Blagov et al80	CD45RA depletion	131	AML, ALL, and other malignancies	αβTCR-depleted haplo/MUD	Pro-DLI	No	Sustained T-cell diversity/viral specific immunity	GIII-IV aGVHD, 2%; cGVHD, 7%
Dunaikina et al84	CD45RA depletion	76	AML, ALL, and other malignancies	αβTCR-depleted haplo/MUD/MSD	Pro-DLI/pre-DLI	No	Improved CMV specific immunity	GIII-IV aGVHD, 8%; cGVHD, 6%
Gasior Kabat et al85	CD45RA depletion	61	ALL, AML, MDS, and lymphoma	MSD/MUD/MMUD/haplo	Pro-DLI/pre-DLI/tDLI	No	2-y OS, 64%; DFS, 63%; NRM, 19%; and reduced CMV viral load	De novo GVHD, 4% and 10% of preexisting GVHD improved
Cheuk et al89	CD62L depletion	6	Not specified	αβTCR-depleted haplo	Pro-DLI day 0	No	Preservation of EM and TEMRA cells and early CD3 recovery	GII-IV aGVHD, 0%; cGVHD, 0%

aGVHD, acute GVHD; cGVHD, chronic GVHD; CML, chronic myeloid leukemia; CMV, cytomegalovirus; DFS, disease-free survival; EFS, event-free survival; F/U, follow-up; GI-II, grade I-II; GIII-IV, grade III-IV; haplo, haploidentical; JMML, juvenile myelomonocytic leukemia; LFS, leukemia-free survival; MMUD, mismatched unrelated donor; MSD, matched sibling donor; MUD, matched unrelated donor; NRM, nonrelapse mortality; ORR, overall response rate; pre-DLI, preemptive DLI; pro-DLI, prophylactic DLI; RFS, relapse-free survival; TCR, T-cell receptor; tDLI, therapeutic DLI.

Timing from transplantation has emerged as the dominant determinant of GVHD risk following DLI. Large pediatric data sets demonstrate a marked reduction in both acute and chronic GVHD when DLI is delayed beyond the early posttransplant period, with the risk declining progressively as immune reconstitution stabilizes.¹⁹ Scheduled prophylactic or preemptive DLI at 4 to 6 months post-HCT has been associated with high rates of chimerism conversion and low clinically significant GVHD, even in haploidentical settings, illustrating that transplant platform and immune reconstitution status profoundly shape DLI outcomes.^17,19,30,32

Early pediatric series highlighted the hazards of unstructured dose escalation, with higher CD3 ⁺ cell doses associated with increased rates of moderate-to-severe chronic GVHD.¹⁶ Contemporary practice has shifted toward lower starting doses with cautious escalation, particularly in mismatched or haploidentical settings,^19,28 an approach now codified in the European Society for Blood and Marrow Transplantation (EBMT) best-practice recommendations, which stratify dosing by intent (prophylactic, preemptive, and therapeutic), donor type, and time from transplantation, with conservative initial doses as low as 1 × 10⁵ CD3⁺ cells per kg in early prophylactic settings. Real-world validation of these recommendations has demonstrated improved response rates, improved overall survival (OS), and reduced nonrelapse mortality (NRM),^33,34 with pediatric studies employing similar risk-adapted strategies reporting lower GVHD rates than historical controls.^19,20

DLI considerations in haplo-HCT

EBMT consensus statements addressing haploidentical HCT (haplo-HCT) emphasize the need for particularly cautious DLI strategies in this setting.^3,35 Recommended practices include lower starting doses, predefined stopping rules, and reassessment after each infusion. Preemptive DLI should be discontinued once MRD negativity is achieved or significant GVHD develops, whereas the treatment of mixed chimerism should be reassessed after each infusion, with consideration of stopping once donor chimerism exceeds 90%.³⁵ Despite the increasing technical refinement of DLI strategies, relapse after haplo-HCT may arise through biological immune escape mechanisms that fundamentally limit the efficacy of donor T-cell–based immunotherapy. One of the most clinically consequential mechanisms is the genomic loss of the recipient-specific mismatched HLA haplotype within leukemic cells, most commonly mediated by the copy-neutral loss of heterozygosity encompassing the mismatched HLA region. This phenomenon renders leukemic blasts effectively invisible to donor-derived T cells and therefore intrinsically resistant to DLI, while preserving full susceptibility to GVHD toxicity. Recent pediatric data now establish HLA loss as a frequent and clinically relevant mechanism of relapse after haplo-HCT, occurring in ~25% of relapses in a large single-institution cohort of children with acute leukemias.³⁶ Importantly, HLA-loss relapse was enriched among patients who received transplantation with active disease or advanced remission status and was associated with poor long-term survival, with no benefit observed from DLI or donor T-cell–based interventions. These findings mirror the adult experience but critically demonstrate that HLA loss is not a rare or late phenomenon in pediatric haplo-HCT, underscoring the necessity of systematic evaluation at relapse. In this context, the administration of DLI in the absence of HLA characterization risks exposing patients to GVHD without any realistic prospect of GVL benefit, emphasizing that biological escape must be excluded before the escalation of T-cell–based immunotherapy. Therefore, routine HLA-loss testing is essential to identify patients who would benefit from alternative approaches, such as natural killer (NK) cell–based therapies, in which immune recognition also relies on the interactions of leukemic cells with activating receptors such as NKG2D or NKp46.³⁷

Across multiple studies, both acute and chronic DLI-associated GVHD appear to be reduced in posttransplant cyclophosphamide (PT-CY)–based platforms compared with historical non–PT-CY cohorts. In pediatric haplo-HCT, prophylactic DLI administered after PT-CY–based conditioning was shown to be feasible and well tolerated, with favorable survival outcomes and a relatively low incidence of GVHD.³⁰ Similar safety and feasibility have also been observed with early prophylactic DLI administered after T-cell–depleted haplo-HCT.³⁸ However, improved safety has not consistently translated into improved efficacy. Although dose-escalated haploidentical DLI following PT-CY can induce remission in a subset of patients, OS after haploidentical DLI was found to be inferior to that observed with HLA-identical donor DLI in relapsed myeloid malignancies.^23,39 Moreover, the efficiency of therapeutic DLI in the PT-CY setting was particularly limited when administered without serotherapy-containing conditioning, with extremely low complete response rates reported in selected relapsed or refractory cohorts.⁴⁰

Collectively, despite substantial variability in DLI application strategies and heterogeneity in outcomes, DLI remains a cornerstone of posttransplant immunotherapy.³⁴ Both pediatric and adult data support a unified paradigm in which DLI is the most effective and safest when deployed early in immune modulation rather than salvage therapy. Prophylactic or preemptive DLI, administered after immune stabilization with conservative dosing and careful escalation, can enhance GVL effects while limiting GVHD, even in mismatched or haploidentical settings. Although most recommendations are largely adult-derived, their close alignment with contemporary pediatric outcomes supports the use of standardized protocols and cautious application in children while underscoring the need for pediatric-specific prospective studies.

DLI combined with blinatumomab in B-ALL

Biological rationale for combining blinatumomab with DLI

Relapse of B-cell ALL (B-ALL) after allo-HCT remains a major therapeutic challenge because B-ALL is less consistently sensitive to DLI alone, with durable responses largely confined to settings of low disease burden.¹² This limitation has driven interest in combination strategies that enhance GVL while maintaining acceptable GVHD risk. Blinatumomab, a CD19-directed bispecific T-cell engager, has emerged as the most biologically coherent and clinically mature partner for DLI in CD19⁺ malignancies. By simultaneously binding CD19 on leukemic blasts and CD3 on T cells, blinatumomab redirects polyclonal T-cell cytotoxicity toward malignant B cells. In the posttransplant setting, DLI restores or augments the donor T-cell pool, whereas blinatumomab provides targeted antigen engagement, amplifying antileukemic immunity without relying on nonspecific alloreactivity.⁴¹ Critically, blinatumomab-driven T-cell activation is largely target-dependent, requiring the presence of CD19-expressing cells, a property that explains why the combination does not uniformly precipitate severe GVHD despite robust T-cell activation and that underscores the necessity of confirming CD19 expression before initiating therapy.

Clinical experience: feasibility and safety

Across retrospective cohorts, prospective studies, and case series, blinatumomab combined with DLI has proven feasible and generally well tolerated after allo-HCT, with toxicities largely mirroring those expected from each agent individually, including cytopenias, infections, neurotoxicity, and GVHD, without a consistent signal of prohibitive synergistic toxicity.^42–49 These observations span both matched and haploidentical platforms, although careful patient selection and immunosuppression management have been consistently emphasized.

The intent and timing of the combination critically determine outcomes. In the therapeutic setting, blinatumomab serves primarily as a cytoreductive agent with DLI providing immunologic reinforcement; however, responses remain inconsistent and rarely durable in high-burden or extramedullary disease, reinforcing the principle that immune-based strategies perform suboptimally against high tumor burden.^42,44,50 Preemptive use, triggered by MRD positivity or early molecular relapse, represents a more biologically favorable window, with blinatumomab achieving MRD clearance followed by DLI as immunologic consolidation.⁴⁵ A prophylactic strategy is conceptually attractive in pediatric patients at extremely high-risk with adverse biology, although robust pediatric-specific data remain limited.⁵¹ In the haploidentical setting, feasibility has been demonstrated but the therapeutic window is narrow, reinforcing the importance of structured protocols over empirical escalation.⁴⁶

Two broad sequencing paradigms have emerged: (1) concurrent administration, which maximizes donor T-cell availability during active CD19-directed engagement and may benefit patients with impaired immune reconstitution and (2) sequential administration, in which blinatumomab first reduces disease burden before DLI consolidation, leveraging DLI’s known efficacy in minimal disease states.⁴⁴ Neither approach has proven universally superior, and selection should be individualized based on immune competence, disease kinetics, and transplant platform.

Given the inconsistent responses in overt relapse and the lifelong morbidity of chronic GVHD in children, current evidence most strongly supports early, preemptive deployment at low disease burden.⁴² Future pediatric-focused studies are needed to define optimal timing, sequencing, and patient selection.

DLI plus blinatumomab vs CAR T-cell therapy

CD19-directed chimeric antigen receptor (CAR) T-cell therapy has demonstrated high remission rates in relapsed pediatric B-ALL and offers potent HLA-independent antileukemic activity even at high disease burden.^52–57 However, its use is constrained by manufacturing time, cost, access, toxicity (including cytokine release syndrome, neurotoxicity, and prolonged B-cell aplasia), and risk of relapse through antigen loss or limited persistence.^53,56 DLI combined with blinatumomab represents a readily available, rapidly deployable, donor-based alternative that can be integrated into existing care pathways and repeated or modulated over time and that critically preserves future treatment options including CAR T-cell therapy or secondary transplantation.⁵⁸ Although its cytoreductive potency is generally lower in overt relapse, it is particularly well suited to preemptive or low-burden settings, in which immune modulation rather than rapid cytoreduction is the primary goal. These strategies should be viewed as complementary within a broader posttransplant immunotherapy framework, with selection guided by disease burden, relapse timing, donor availability, institutional resources, and long-term toxicity considerations, all of which carry particular weight in the pediatric context.

DLI combined with other agents in myeloid malignancies

Biological rationale for combining DLI with other agents

Relapse of AML or myelodysplastic syndromes (MDS) after allo-HCT remains associated with poor outcomes, particularly in pediatric patients with high-risk disease biology. Although DLI is more effective in myeloid malignancies than in lymphoid diseases, durable responses are still largely confined to settings of low disease burden.⁵⁹ These limitations have driven the development of combination strategies for pairing DLI with agents that provide modest cytoreduction, enhance leukemic immunogenicity, or modulate donor immune responses, with the aim of strengthening GVL while minimizing GVHD.

HMAs combined with DLI

Among the available partners, hypomethylating agents (HMAs), particularly azacitidine (AZA), represent the most mature and pediatric-relevant agents combined with DLI in AML/MDS. HMAs offer limited cytoreduction but may increase leukemic antigen expression, alter the tumor microenvironment, and facilitate immune-mediated clearance, providing a biological rationale for synergy with DLI. A key pediatric study demonstrated that scheduled AZA combined with prophylactic DLI is feasible and well tolerated in children, with encouraging relapse control in an exceptionally high-risk population. Importantly, this strategy was implemented in a preemptive/prophylactic setting, reinforcing the principle that immune-based interventions are most effective when applied before overt hematologic relapse (Table 1).^26,29 Very recently, Bica et al similarly reported that the combination of AZA and DLI was both effective and safe in pediatric patients with AML, leading to a durable remission and OS of 80% when DLI was used prophylactically or preemptively, without observing any chronic GVHD or NRM. Although outcomes when DLI was used therapeutically were less favorable, the incidence of GVHD was low, suggesting that the concurrent administration of AZA and DLI is feasible.²⁵ Studies in adults have reinforced and extended these findings. Multiple retrospective and prospective series using AZA with preemptive or prophylactic DLI in AML/MDS reported improved relapse-free survival compared with historical controls, with acceptable GVHD rates when DLI was delivered in a dose-escalated and risk-adapted manner.^60–63 Across studies, the benefit of HMA plus DLI appears to be driven less by cytotoxicity and more by immune sensitization, positioning AZA as an enabling agent rather than a definitive salvage therapy.

Although the available studies are heterogeneous, several consistent principles emerge. Disease burden at the time of intervention remains the dominant determinant of outcome regardless of the partner agent, and sequential approaches, such as cytoreduction followed by DLI-based immune consolidation, outperform concurrent escalation in the setting of advanced relapse. Toxicity is driven primarily by the transplant platform and DLI dosing rather than by HMAs themselves.^60–66 For children, AZA combined with DLI may be sufficient for molecular or low-level relapse, and until prospective pediatric data are available, this combination should remain the backbone of immune-modulatory therapy in this setting.

Venetoclax and DLI

AML or MDS relapse after allo-HCT remains highly resistant to therapy, particularly when it occurs early or with high disease burden. Although HMAs combined with DLI represent a tolerable immunomodulatory strategy, incomplete responses have prompted interest in adding venetoclax to enhance cytoreduction while preserving GVL activity. In adult cohorts, venetoclax-based regimens achieved response rates exceeding those historically observed with DLI alone.^67,68 A multicenter retrospective study of venetoclax combined with DLI for early posttransplant AML relapse reported objective responses, including complete remissions, in approximately half of the patients, even among those relapsing within 6 months of HCT; notably, responders were more likely to develop GVHD, which correlated with improved survival, suggesting that venetoclax may augment immune-mediated alloreactivity beyond its cytoreductive role.⁶⁹ Systematic reviews further support the feasibility of this combination without a clear increase in nonrelapse mortality, although prolonged cytopenias and frequent infections often necessitate dose reductions or treatment discontinuation.^67,70 Pediatric-specific data are lacking, and venetoclax-based intensification should therefore be reserved selectively for overt or rapidly progressive disease rather than applied routinely.

FLT3 inhibitors and DLI

Preclinical evidence has demonstrated that the Fms-like tyrosine kinase 3 (FLT3) inhibitor sorafenib augmented interleukin-15 expression and synergized with T cells via reduced ATF4 expression.⁷¹ Moreover, gilteritinib administration after allo-HCT was found to reduce the expression of inhibitory receptors TIGIT and programmed cell death protein 1 on donor CD8⁺ T cells and to enhance GVL against FLT3-ITD⁺ leukemia without affecting GVHD.⁷² In a retrospective study of post–allo-HCT relapsed AML, the combination of sorafenib, DLI, and chemotherapy prolonged complete remission, OS, and progression-free survival, with sorafenib independently contributing to improved OS and without either agent increasing GVHD incidence or mortality.⁶⁶ A subsequent single-center report demonstrated promising outcomes with preemptive escalating DLI doses combined with sorafenib,⁷³ and a more recent study in FLT3 wild-type AML relapsing after allo-HCT found that sorafenib added to DLI improved 2-year OS and event-free survival without increasing GVHD risk.⁷⁴ Although no FLT3 inhibitors have been US Food and Drug Administration–approved for pediatric use, they have shown promise in pediatric patients with FLT3-ITD–positive AML, a subgroup that has been associated with inferior outcomes.⁷⁵ All these data suggest that combining FLT3 inhibitors with DLI may represent a feasible and biologically rational strategy for pediatric FLT3-mutated AML in the posttransplant setting.

MRD status critically informs the application of posttransplant DLI in myeloid malignancies: preemptive DLI is triggered by emerging MRD positivity, whereas prophylactic DLI is reserved for patients at high-risk to prevent relapse. The detection of actionable mutations such as FLT3-ITD further refines this framework, supporting the integration of DLI, either as prophylactic monotherapy or in combination with targeted agents in the preemptive setting, alongside mutation-directed therapies. Although such combinatorial approaches remain largely investigational in the pediatric context, they represent a biologically rational and clinically promising direction for posttransplant immune modulation in high-risk myeloid disease.

T-cell subset–modified DLI

Despite advances in DLI dosing and combination strategies, GVHD remains a major limitation of conventional DLI. The selective depletion of T-cell subsets most strongly associated with alloreactivity, while preserving those responsible for GVL, has driven the development of several subset-modified DLI platforms designed to improve safety without sacrificing antileukemic potency.

CD8-depleted DLI

CD8-depleted DLI represented an early subset-modification strategy, grounded in the observation that CD8⁺ T cells, particularly naïve subsets, are major mediators of alloreactive tissue injury, whereas effector memory (EM) CD4⁺ T cells preserve antileukemic activity with lower GVHD risk.⁷⁶ Although promising results were reported,^77–79 CD8-depleted DLI has not been widely adopted due to logistical complexity, residual GVHD risk, and the emergence of more refined alternatives. Nevertheless, these foundational studies established the critical principle that selective T-cell manipulation can reshape the therapeutic index of DLI, informing all subsequent strategies aimed at uncoupling GVL from GVHD.

CD45RA-depleted DLI

CD45RA-depleted DLI is rooted in the observation that CD45RA⁺ naïve T cells are the principal mediators of alloreactivity and GVHD, whereas CD45RA⁻ memory T cells retain pathogen-specific immunity and GVL activity with substantially lower GVHD potential.^80,81 A phase 1 dose-escalation study of prophylactic CD45RA-depleted DLI demonstrated no dose-limiting grade III to IV acute GVHD at doses up to 1 × 10⁷ CD3⁺ cells per kg, with only isolated grade II acute GVHD and low chronic GVHD rates,⁸² and a subsequent study reported ~90% viral clearance rates in HCT recipients with COVID-19.⁸³ Although both studies were conducted in adults, they established the safety foundation for pediatric application.

Pediatric data now constitute a substantial component of this literature. A randomized trial in children undergoing αβ T-cell–depleted haplo-HCT demonstrated that CD45RA-depleted DLI did not increase grade II to IV acute GVHD, preserved high survival rates, and significantly enhanced cytomegalovirus-specific immune reconstitution (Table 1).⁸⁴ Gasior Kabat et al similarly reported low de novo GVHD rates and reduced cytomegalovirus viral load with high-dose memory CD45RA⁻/RO⁺ DLI,⁸⁵ and additional pediatric cohort studies confirmed feasibility, with extremely low GVHD incidence and durable T-cell receptor repertoire reconstitution,^86–88 findings of particular relevance given that infectious morbidity and delayed immune reconstitution are leading drivers of posttransplant mortality in children. CD45RA depletion using the Clin-iMACS platform reliably achieves near-complete naïve T-cell removal while enriching for memory CD4⁺ T cells without abrogating antiviral or antileukemic function,⁸¹ establishing naïve T-cell–depleted DLI as the most clinically translatable modified DLI platform currently available for pediatric prophylactic and preemptive applications.

CD62L-depleted DLI

A recent pediatric pilot study introduced CD62L-depleted DLI as a next-generation refinement of naïve T-cell–depleted strategies (Table 1).⁸⁹ Conventional T cells can be classified into CD45RA⁺CD62L⁺ naïve, CD45RA⁻CD62L⁺ central memory (CM), CD45RA⁻CD62L⁻ effector memory (EM), and CD45RA⁺CD62L⁻ terminally differentiated EM re-expressing CD45RA (TEMRA) subsets with progressively decreasing GVHD potential.⁸⁹ Although CD45RA depletion removes naïve T cells, it retains CD62L⁺ CM T cells, which can still drive alloresponses, and eliminates TEMRA subsets that contribute importantly to early antiviral protection. CD62L depletion addresses both limitations by removing naïve and CM T cells while preserving EM and TEMRA subsets, theoretically minimizing GVHD risk while maintaining antiviral and antileukemic function. In this first-in-human pediatric study following T-cell receptor (TCR) αβ–depleted haplo-HCT, CD62L-depleted DLI was feasible, achieved efficient depletion of target subsets, and was associated with minimal acute GVHD, no chronic GVHD, and early recovery of EM and TEMRA compartments consistent with the intended biological design.⁸⁹

The evolution from unmanipulated to subset-modified DLI, spanning CD8-, CD45RA-, and CD62L-depletion strategies, underscores a central principle: the therapeutic index of DLI depends less on dose escalation than on immune composition, timing, and biological context. Pediatric haplo-HCT stands to benefit most from this shift toward rationally engineered products, and future prospective studies integrating immune monitoring with biologically informed design will be essential to safely uncouple GVL from GVHD and maximize disease control in children.

Limitations of the current evidence base

The evidence supporting DLI in pediatric allo-HCT is subject to several important limitations that warrant explicit acknowledgment. First, the field lacks prospective randomized controlled trials in pediatric populations. The overwhelming majority of available data derive from retrospective, single-center, or small-cohort studies, many of which conflate pediatric and adult patients or extrapolate findings from adult trials without adequate adjustment for the biological and immunologic differences inherent to pediatric leukemia. The mutational landscape, immune ontogeny, and disease biology of childhood leukemia differ meaningfully from adult AML and ALL, and these distinctions have not been systematically accounted for across most DLI studies.

Second, the absence of standardized definitions for prophylactic, preemptive, and therapeutic DLI across institutions and trials substantially limits cross-study comparisons and the generation of reproducible, evidence-based clinical guidelines. Similarly, DLI dosing strategies, escalation schedules, and thresholds for intervention remain highly heterogeneous, with no consensus on optimal T-cell doses across different transplant platforms, disease contexts, or HLA-matching configurations.

Third, MRD assessment methodologies vary considerably across centers, encompassing multiparameter flow cytometry, polymerase chain reaction–based approaches, and next-generation sequencing, leading to inconsistent definitions of MRD positivity and divergent criteria for triggering preemptive intervention. The lack of standardized HLA-loss testing protocols further compounds this because relapse driven by immune escape may be mischaracterized or missed in centers where routine posttransplant HLA surveillance is not embedded in clinical practice.

Finally, although emerging cellular therapies, including CAR T cells and cytokine-induced memory-like NK cells, have demonstrated substantial efficacy in pediatric lymphoid and myeloid malignancies, respectively, and in some settings, have outperformed conventional DLI, the comparative effectiveness of these approaches has not been evaluated in prospective head-to-head trials.^90–92 Despite these limitations, DLI retains a unique and durable position in transplant medicine: it offers broad immune reconstitution, immediate availability from the original donor, no requirement for specialized good manufacturing practices manufacturing, and a capacity for rational combination with targeted agents, HMAs, checkpoint modulators, and subset-modified cellular products. These attributes sustain its clinical relevance as both a standalone strategy and as a platform for next-generation immunotherapeutic combinations.

Conclusions

DLI remains one of the most biologically compelling and clinically accessible posttransplant immunotherapies in pediatric allo-HCT. Its efficacy is determined not by dose alone, but by the interplay of timing, disease burden, immune reconstitution, donor-recipient HLA configuration, and cellular composition. Across studies, earlier structured deployment, particularly in prophylactic and preemptive settings, consistently yields superior outcomes compared with salvage use in overt relapse, in which disease burden and immune exhaustion conspire against durable responses.

In the past decade, a meaningful evolution has been witnessed from empirical unmanipulated infusions to biologically informed strategies integrating MRD surveillance, mutation-guided decision-making, T-cell subset engineering, and rational pharmacologic combinations. Blinatumomab has expanded the reach of DLI in B-ALL by redirecting donor T cells with antigen specificity while attenuating bystander alloreactivity. HMAs have created synergistic immunologic and cytoreductive contexts in myeloid disease that amplify GVL while maintaining acceptable toxicity. T-cell subset–modified DLI platforms have advanced the uncoupling of GVL from GVHD through mechanistically distinct but complementary approaches.

Critical gaps nonetheless remain. The field urgently requires prospective pediatric trials designed with harmonized MRD end points, standardized HLA-loss testing, and consensus dosing frameworks. Multicenter collaboration through the Pediatric Blood and Marrow Transplant Consortium and EBMT Pediatric Diseases Working Party will be essential to generate the sample sizes and biological correlates needed to define optimal DLI strategies across pediatric disease subtypes. Future trials should embed biological risk stratification, longitudinal immune monitoring, and prespecified criteria for DLI initiation and escalation, moving the field beyond retrospective inference toward evidence-based individualized immunotherapy.

Ultimately, DLI’s enduring value lies not only in its immediate therapeutic utility but in its role as a conceptual model for harnessing the GVL effect. The principles established through decades of DLI research, immune surveillance, clonal T-cell expansion, leukemic escape mechanisms, and the feasibility of separating GVL from GVHD continue to inform every subsequent generation of adoptive cellular therapy. Appropriately refined and prospectively evaluated, DLI remains an indispensable pillar of posttransplant immune modulation for children with high-risk hematologic malignancies.