Immune reset: what do we learn from bone marrow transplantation Authors: Huidong Guo, Xiao-Jun Huang Published: 2026 ABSTRACT AbstractBone marrow transplantation (BMT) has traditionally been viewed as a curative strategy for hematologic malignancies through replacement of the diseased hematopoietic system. However, accumulating evidence indicates that BMT represents a far more profound biological process: an enforced collapse and subsequent reconstruction of the hematopoietic and immune system. This process provides a unique human model to investigate how immune homeostasis can be reprogrammed beyond physiological conditions. Importantly, post-transplant immunity does not revert to a pre-existing state but instead converges on a newly defined equilibrium shaped by conditioning-induced injury, donor-recipient immune interactions, stromal regeneration, and therapeutic interventions. In this review, we conceptualize BMT as an immune reset and synthesize emerging insights into the mechanisms that govern tolerance reinstallation and immune stability after transplantation. By reframing BMT as a model of immune reset rather than simple immune reconstitution, we highlight general principles of immune homeostasis restoration that extend beyond transplantation and may inform therapeutic strategies for autoimmunity, immune aging, and immune-mediated diseases. Introduction The immune system maintains homeostasis through tightly regulated processes that balance immune responsiveness and tolerance within the host. Under physiological conditions, this equilibrium is established early in life during immune development and is maintained with remarkable stability thereafter. Bone marrow transplantation (BMT) fundamentally disrupts this paradigm by forcibly eliminating the recipient immune system and initiating de novo immune reconstruction in adulthood. First performed in 1956 and applied routinely since the 1970 s, allogeneic BMT (allo-BMT) has become a curative therapy for hematologic malignancies and disorders of hematopoiesis [1], 2]. Advances in transplant strategies have progressively overcome classical immunological barriers. Notably, human leukocyte antigen-haploidentical (HLA-haploidentical) BMT now achieves clinical outcomes comparable to those of human leukocyte antigen-matched (HLA-matched) transplantation [3], 4]. This success underscores the capacity of BMT to fundamentally reprogram immune tolerance and positions immune reset as a promising paradigm for restoring immune homeostasis in pathological conditions. Historically, immune reconstitution following BMT described primarily in quantitative terms, focusing on the kinetics of leukocyte recovery. However, conditioning regimens administered before hematopoietic stem and progenitor cell (HSPC) infusion do more than facilitate donor engraftment, they eradicate not only host immune cells, but also immune memory, tolerance networks, and stromal support structures [5], 6]. Consequently, the post-transplant immune system does not simply “recover”; it is fundamentally rebuilt. During this process of immune reassembly, outcomes diverge. Some patients develop immune dysregulation leading to transplant-related complications such as graft-versus-host disease (GVHD). In contrast, some recipients successfully re-establish immune homeostasis, discontinue immunosuppressive therapy, and maintain long-term disease-free survival. This stands in sharp contrast to solid organ transplantation, where lifelong immunosuppression is generally required to prevent rejection. These observations raise fundamental questions: How is immune homeostasis re-established after BMT What mechanisms underlie durable tolerance in this setting To what extent can our understanding of immune restoration after BMT be translated into therapeutic strategies for other conditions marked by immune dysregulation, including autoimmune diseases, immune aging, and solid organ transplantation Despite the profound immunologic reorganization that occurs after transplantation, immune homeostasis restoration has rarely been examined as a distinct biological process. Moreover, biomarkers capable of capturing the qualitative transformation of immune regulation after transplantation remain poorly defined. Here, we propose that BMT should be conceptualized as a model of immune reset: a biological reboot in which the pre-existing immune system is eradicated and subsequently re-established under profoundly non-physiological conditions. Although the concept of “immune reset” was first introduced in the context of hematopoietic stem cell transplantation (HSCT) in 1997, it is now more widely discussed in the treatment of autoimmune diseases [7]. Indeed, therapeutic strategies for autoimmune diseases have gradually shifted from conventional immune modulation toward immune reset. Immune modulation generally refers to the non-selective suppression of pathogenic immune cells and aberrantly activated immune responses, often resulting in temporary immunodeficiency. In contrast, immune reset encompasses two key processes aimed at achieving sustained remission: depletion of pathogenic immune subsets and restoration of immune homeostasis. Thus, immune reset represents more than a simple replacement of immune cell origin; it involves systemic reprogramming of the immune repertoire, activation thresholds, metabolic states, and epigenetic landscapes. Clinically, immune reset in both BMT and autoimmune diseases is characterized by sustained disease remission with minimal or no maintenance therapy [4], 7]. Therefore, systematic investigation of the immune reset process after BMT may provide important cellular and molecular insights into the restoration of immune homeostasis across a broad range of immune imbalance-associated diseases. The immune reset process following BMT is fundamentally distinct from physiological immune development, which unfolds gradually during embryogenesis and early life within an intact and evolutionarily optimized microenvironment. Notably, much of our current understanding of immune tolerance has been derived from studies of this steady, developmental context. In contrast, immune reconstitution after BMT occurs in a severely perturbed host shaped by cytotoxic conditioning, inflammation, tissue injury, and alloimmune pressure. Precisely because of this extreme disruption, BMT uniquely exposes core regulatory principles governing immune tolerance and stability that remain obscured during steady-state or naturally developing immunity. Conditioning: programming the depth and direction of immune reset Conditioning regimens represent the initiating event that drives immune collapse prior to BMT. Irrespective of their specific agents or intensity, conditioning strategies are designed around three fundamental principles: (1) eradication of residual malignant or host hematopoietic cells, (2) creation of ecological space for donor stem cell engraftment, and (3) suppression of host immunity to prevent graft rejection [6]. Clinically, conditioning is categorized as myeloablative conditioning (MAC), reduced-intensity conditioning (RIC), or non-myeloablative (NMA) [8]. Myeloablative, or “high-dose” regimens, typically combines alkylating agents with or without total body irradiation (TBI) and results in complete ablation of bone marrow (BM) hematopoiesis, precluding autologous recovery. In contrast, NMA regimens induce minimal cytopenia and do not require stem cell support, whereas RIC regimens fall between these extremes, causing prolonged cytopenia while still necessitating hematopoietic stem cell infusion [9]. Malignant BMT: myeloablation and microenvironmental remodeling In hematologic malignancies, conditioning has traditionally prioritized maximal malignant cell eradication, and MAC has long been regarded as the standard approach. However, the impact of MAC extends far beyond the elimination of hematopoietic cells. Non-hematopoietic stromal compartments are profoundly affected: BM niches, thymic epithelial cells, and the structural framework of secondary lymphoid organs sustain lasting damage, leading to simultaneous impairment of central and peripheral tolerance as well as tissue-specific immune education. This widespread disruption creates a biological vacuum in which immune homeostasis must be reconstructed de novo. Unlike neonatal immune development, post-BMT immune reconstitution occurs within an inflamed, damaged, and often aging host environment, conditions that fundamentally shape the trajectory of immune reset. At the cellular level, chemotherapy ablates rapidly proliferating hematopoietic progenitors and forces normally quiescent hematopoietic stem cells (HSCs) into proliferation and differentiation [10]. Sinusoidal vessels and stromal populations, including perivascular and osteolineage cells, are significantly depleted, while BM adipocytes expand, reflecting an adipogenic skewing of mesenchymal stromal cells [11], [12], [13]. Single-cell transcriptomic analyses following 5-fluorouracil-induced (5-FU-induced) myeloablation reveal coordinated molecular reprogramming across niche compartments [14]. Leptin receptor (LEPR+) stromal cells expand adipocyte-primed populations and upregulate adipogenic pathways, while osteolineage-associated genes are broadly suppressed. Vascular (VE-Cad+) and perivascular (LEPR+) cells exhibit parallel transcriptional shifts, indicating system-wide remodeling [14]. Simultaneously, regenerative programs are activated, characterized by increased niche cell proliferation and altered expression of hematopoietic regulators, including upregulation of angiopoietin, interleukin-7 (IL-7), bone morphogenetic protein 4 (BMP4), and wnt5a, alongside downregulation of notch ligands (Delta-like ligand 1 [Dll1], Delta-like ligand 4 [Dll4]), sele, and wnt4, thereby reshaping wnt and notch signaling dynamics [14]. Importantly, the dismantling of host hematopoietic and immune compartments by conditioning is not merely destructive, it may also actively facilitate immune reset. Conditioning profoundly remodels the BM microenvironment, which in malignant states may be corrupted to support disease progression [15]. Our recent work demonstrates that chemotherapy can correct acute myeloid leukemia-altered BM endothelial progenitor cells and restore a microenvironment capable of supporting normal hematopoiesis in patients achieving complete remission [16]. Thus, MAC should be viewed not only as cytoreduction, but as a profound ecological reset of the BM microenvironment that critically determines subsequent hematopoietic and immune reconstitution. Non-malignant BMT: balancing immunoablation and toxicity As BMT indications expand beyond malignancy, conditioning strategies require recalibration. In non-malignant disorders, such as aplastic anemia, primary immunodeficiencies, hemoglobinopathies, and autoimmune diseases, the objective is durable correction of hematopoietic or immune dysfunction rather than tumor eradication. Full myeloablation is often unnecessary and may be harmful, particularly in patients with DNA repair defects or pre-existing organ dysfunction. Moreover, unlike malignant transplantation, there is no therapeutic benefit from graft-versus-tumor effects, and thus no justification for tolerating significant GVHD. Minimizing acute and chronic GVHD while ensuring stable donor chimerism is therefore central to transplant optimization in this setting. Reduced-intensity and non-myeloablative regimens, commonly fludarabine-based and combined with immune depleting agents such as anti-thymocyte globulin (ATG) or alemtuzumab, exemplify this approach. These strategies preferentially target host-reactive or pathogenic immune compartments rather than indiscriminately ablating hematopoiesis. In autoimmune and immune dysregulation disorders, clinical benefit often derives from immune reprogramming rather than complete hematopoietic replacement. Accordingly, strategies that preferentially target pathogenic T and natural killer cells [5], which are commonly employed in autologous transplantation settings, further emphasize the primacy of immunoablation over myeloablation. However, attenuated conditioning increases the risk of graft failure, particularly in chemotherapy-naïve or sensitized recipients [17], [18], [19], [20]. This is especially relevant in hemoglobinopathies and certain primary immune regulatory disorders. Strategies such as prolonged post-transplant immunosuppression or pre-transplant immunomodulation can improve engraftment but carry increased risks of infection and immunosuppression-related complications. Treosulfan-based conditioning has emerged as an important alternative in non-malignant transplantation, offering adequate myeloablation with a more favorable toxicity profile compared with conventional busulfan-based regimens. Large retrospective analyses in children with non-malignant disorders demonstrate high engraftment rates and low regimen-related toxicity [21], although some studies suggest higher rates of mixed chimerism or graft failure in specific diseases [22]. Prospective comparative trials remain needed to define optimal exposure and long-term immune reconstitution outcomes. Collectively, conditioning in non-malignant BMT should be conceptualized not in terms of maximal cytotoxicity, but as calibration of immune reset: sufficient to dismantle pathological immunity and enable tolerance re-establishment, yet restrained enough to preserve organ function and microenvironmental integrity. Refining conditioning to target immune reset Given the toxicities and treatment-related mortality associated with conventional conditioning, biologically targeted conditioning approaches, including antibody-based and immune-targeted approaches, have been developed to achieve selective depletion of hematopoietic stem and progenitor cells (HSPCs) and pathogenic immune subsets while sparing non-hematopoietic organs. A central focus of antibody-based conditioning has been targeting CD117 (c-Kit), the receptor for stem cell factor (SCF), which is critical for hematopoietic stem cell (HSC) survival and niche retention. Monoclonal antibodies such as JSP191 have demonstrated effective HSPC depletion in preclinical models and early-phase clinical trials, permitting donor engraftment without the systemic toxicities associated with alkylating agents or irradiation [6], 23], 24]. Preclinical studies further showed that CD117 targeting can deplete not only normal HSCs but also myelodysplastic syndromes-derived (MDS-derived) aberrant HSCs, while promoting engraftment of healthy donor HSCs, suggesting its dual potential for both niche opening and elimination of pathological hematopoietic clones [23]. In addition, combined targeting of CD117 and CD47 achieved more than 99 % HSC depletion, indicating that blockade of phagocytic inhibitory signaling can further enhance antibody-mediated HSPC clearance [25]. Antibody-drug conjugates (ADCs) further enhance specificity by linking cytotoxic payloads to lineage-restricted antibodies. CD45−saporin and CD117−saporin ADCs efficiently depleted host HSPCs and support durable multilineage donor engraftment after HSCT [26]. Compared with cyclophosphamide-based regimens, ADC-treated mice exhibited markedly reduced toxicity, including the absence of chemotherapy-induced hepatic inflammation [26]. In CD117−ADC studies, a single dose achieved more than 99 % depletion of host HSCs and induced high-level donor myeloid and bone marrow HSC chimerism [27]. In hemophilia A and non-human primate hemoglobinopathy models, CD117−ADC also promoted donor engraftment and showed potential to reduce long-term toxicities compared with conventional busulfan-based conditioning [28], 29]. By contrast, CD45−ADC targets a pan-hematopoietic antigen and therefore mediates broader hematopoietic and immune compartment depletion. As a single-agent conditioning approach, CD45−ADC supported high-level multilineage engraftment and enabled donor chimerism and immune reconstitution in severe combined immunodeficiency/recombination-activating gene-deficient (SCID/RAG-deficient) and Fanconi anemia models [30], [31], [32]. Moreover, combining CD45- and CD117−targeted ADCs with the Janus kinase (JAK) 1/2 inhibitor baricitinib achieved up to 99 % donor myeloid chimerism, even across fully major histocompatibility complex-mismatched (MHC-mismatched) barriers, while maintaining stable multilineage engraftment [33]. Importantly, this strategy effectively balanced GVHD and graft-versus-leukemia (GVL) effects, improving both safety and immunologic precision in allo-HSCT [33]. Recent clinical trials demonstrated that CD117−targeted approaches combined with nonmyeloablative conditioning constitute a feasible and safe BMT conditioning strategy in older adults with high-risk acute myeloid leukemia (AML) and MDS [34]. In patients with Fanconi anemia, briquilimab-based conditioning has been incorporated into a TBI- and busulfan-free haploidentical transplantation platform, further supporting its value as a reduced-toxicity conditioning strategy for populations highly vulnerable to genotoxic injury [35]. Beyond HSPC targeting, immune-modulating monoclonal antibodies such as alemtuzumab (anti-CD52), rituximab (anti-CD20), and vedolizumab (anti-α4β7 integrin) have been incorporated into conditioning regimens to mitigate GVHD and reshape post-transplant immune dynamics. However, careful titration is required to preserve GVL effects and avoiding excessive infection risk. Emerging cellular approaches further extend biologically targeted conditioning. Preclinical studies demonstrate that CD117−directed chimeric antigen receptor T (CAR-T) cells showed effective killing of CD117+ AML cells and healthy HSPCs in vitro and in vivo [36]. However, because CD117 is expressed on both normal HSPCs and a subset of AML cells, persistent CD117 CAR-T cells may damage subsequently infused donor HSPCs. Therefore, safety switches, controllable elimination systems, and careful optimization of the interval between CAR-T administration and HSC infusion represent critical issues for clinical translation. Notably, recent clinical observations suggest that CD7 CAR-T cell therapy may enable direct transition to allogeneic transplantation without traditional cytoreductive conditioning [37], indicating the potential of immune-selective preparative regimens. Similarly, donor-derived CAR-T cells have been explored as reduced-intensity conditioning in relapsed/refractory acute lymphoblastic leukemia (R/R ALL), and CD19/CD20/CD22 CAR-T cells have been incorporated into allo-HSCT conditioning for R/R B-cell non-hodgkin lymphoma (B-NHL), collectively suggesting that CAR-T therapy is evolving from a purely bridging cytoreductive treatment toward an immune-depleting module within transplant conditioning [38], 39]. To better summarize conditioning-refinement strategies investigated in preclinical animal studies and clinical trials, we have summarized the relevant studies in Tables 1 and 2. Table 1:Preclinical studies of targeted conditioning strategies.CategoryTargetModelDiseaseKey findingsRef.mAbCD117MouseMDSAnti-human CD117 antibody depleted both normal HSCs and MDS HSCs and promoted engraftment of normal donor HSCs[23]mAbCD117, CD47MouseNoneAchieved >99 % depletion of HSCs[25]mAbCD117, anti-CD4, anti-CD8, anti-CD40L, anti-CD47, and anti-CD122MouseNoneDonor granulocyte chimerism reached 52 % at 8 weeks; the regimen supported MHC-mismatched transplantation[40]ADCCD117Rhesus macaqueHemoglobinopathyCD117−ADC promoted donor engraftment and showed greater potential than conventional busulfan-based conditioning for fertility preservation[29]ADCCD117MouseNoneAchieved >99 % depletion of host HSCs, 98 % donor myeloid chimerism, and >99 % donor chimerism among bone marrow HSCs[27]ADCCD117MouseHemophilia AEffectively depleted HSCs; donor myeloid chimerism reached 90.6 % at 4 weeks[28]ADCCD45/CD117MouseFanconi anemiaSignificantly depleted HSCs and achieved superior engraftment compared with cytarabine-based conditioning[26]ADCCD45MouseNoneAs a single-agent conditioning regimen, CD45−ADC enabled high-level multilineage engraftment in murine allo-HSCT models[30]ADCCD45MouseNoneCD45−ADC served as a single-agent regimen capable of fully depleting human HSCs and supporting engraftment of gene-modified autologous and allogeneic HSCs[41]ADCCD45MouseSCID/RAG deficiencyConditioning with CD45−ADC alone enabled donor chimerism and immune reconstitution in Rag1-deficient mice, even after haploidentical transplantation[31]ADCCD45MouseFanconi anemiaA single dose of CD45−ADC induced stable multilineage donor chimerism across three Fanconi anemia mouse models with lower toxicity and significantly reduced GVHD-related mortality compared with lethal irradiation[32]ADCCD45MouseSickle cell disease modelAchieved 99 % depletion of host HSCs and approximately 90 % donor chimerism among BM HSCs at 4 months, with lower toxicity than TBI. In immunocompetent mice, anti-c-Kit alone was insufficient, whereas combined CD47 blockade enhanced phagocytic clearance and enabled donor HSC engraftment without chemotherapy or irradiation[42]ADCCD45MouseSCID/RAG deficiencyAchieved 91.7–95.2 % depletion of host HSCs and 32.04–100 % donor HSC chimerism[43]ADC+ mAbCD117 ADC + anti-CD4/anti-CD8/anti-CD40L mAbsMouseNoneHigh-level and durable donor chimerism was achieved in 14 of 15 mice, accompanied by induction of skin allograft tolerance[44]ADC+ JAK inhibitorCD45/CD117 ADC + JAK1/2 inhibitor (baricitinib)MouseNoneSignificantly depleted HSCs, achieved 99 % donor myeloid chimerism, and supported MHC-mismatched transplantation[33]ADC+ JAK inhibitorCD45/CD117 ADC + JAK1/2 inhibitor (ruxolitinib)MouseNoneAnti-CD117 combined with CD47 blockade and JAK inhibition enabled antibody-based conditioning without the need for cytotoxic payloads[45]CAR-TCD117 CAR-TMouseNoneDonor chimerism among LT-HSCs reached 27 %, while donor chimerism among peripheral blood granulocytes, B cells, and T cells reached 20–30 %[46]CAR-TCD117 CAR-TMouseNoneDepleted 98 % of CD117+ cells and reduced total bone marrow cellularity[36]MDS, myelodysplastic syndromes; HSCs, hematopoietic stem cells; MHC, mismatched major histocompatibility complex-mismatched; ADC, antibody-drug conjugate; HSCT, hematopoietic stem cell transplantation; SCID/RAG, severe combined immunodeficiency/recombination-activating gene; GVHD, graft-versus-host disease; BM, bone marrow; TBI, total body irradiation; JAK, Janus kinase; CAR-T, chimeric antigen receptor T. Preclinical studies of targeted conditioning strategies. MDS, myelodysplastic syndromes; HSCs, hematopoietic stem cells; MHC, mismatched major histocompatibility complex-mismatched; ADC, antibody-drug conjugate; HSCT, hematopoietic stem cell transplantation; SCID/RAG, severe combined immunodeficiency/recombination-activating gene; GVHD, graft-versus-host disease; BM, bone marrow; TBI, total body irradiation; JAK, Janus kinase; CAR-T, chimeric antigen receptor T. Table 2:Clinical translational studies of targeted conditioning and CAR-T–Mediated bridging to allo-HSCT for immune reset.Strategy categoryTargetDiseaseStudy phase/typeRegistration no.Pre-transplant conditioning regimenTransplant typeFunctional positioningKey findingsRef.CD117 antibody-targeted conditioningBriquilimab/JSP191Fanconi anemiaPhase 1b/1/2NCT04784052TBI- and busulfan-free regimen; combined with immunosuppressive agents including ATG, fludarabine (Flu), cyclophosphamide (Cy), and rituximabHaplo-HSCT with a TCRαβ+ T-cell/CD19+ B-cell-depleted graftHSPC niche clearance; reduced-toxicity conditioningDemonstrated the feasibility of TBI- and busulfan-free conditioning in patients with Fanconi anemia[35]CD117 antibody-targeted conditioningJSP191/briquilimabAML/MDSPhase 1a/1bNCT04429191JSP191 + low-dose TBI + fludarabineallo-HSCTHSPC niche clearance; reduced myeloablative toxicityRepresents early clinical translation in older or high-risk patients with AML/MDS[34]CD7 CAR-T bridging to allo-HSCTCD7 CAR-TR/R T-ALL/T-LBLSingle-center, prospective, single-arm clinical studyChiCTR2000040641Reduced-intensity conditioning: A Total body irradiation (TBI)/fludarabine-based regimen or a busulfan/fludarabine-based regimenSame-donor allo-HSCTBridging cytoreduction plus functional immune depletionTransplantation after CD7 CAR-T therapy was feasible and may improve DFS[47]CD7 CAR-T all-in-one strategyCD7 CAR-TR/R CD7+ leukemia/LymphomaPhase 1NCT04599556/NCT04538599Fludarabine/cyclophosphamide/etoposide before CAR-T therapy; no additional pharmacologic myeloablative conditioning before HSCT and no conventional GVHD prophylaxisHaplo-HSCTFunctional conditioning/all-in-one strategyProvides one of the closest clinical examples of CAR-T therapy replacing part of the conventional conditioning function[37]Donor-derived CAR-T incorporated into conditioningCD19/CD20/CD22 CAR-TR/R B-NHLSingle-center, prospective, single-arm clinical studyChiCTR2000040665T/Bu/Flu-based conditioning; CAR-T cells infused on days −9 and −8allo-HSCTImmune cytoreductive module within conditioningPatients achieved CAR-T expansion, engraftment, and donor chimerism[39]HSPC, hematopoietic stem and progenitor cell; ATG, anti-thymocyte globulin; HSCT, hematopoietic stem cell transplantation; CAR-T, chimeric antigen receptor T; AML, acute myeloid leukemia; MDS, myelodysplastic syndromes; GVHD, graft-versus-host disease; B-NHL, B-cell non-hodgkin lymphoma. Clinical translational studies of targeted conditioning and CAR-T–Mediated bridging to allo-HSCT for immune reset. HSPC, hematopoietic stem and progenitor cell; ATG, anti-thymocyte globulin; HSCT, hematopoietic stem cell transplantation; CAR-T, chimeric antigen receptor T; AML, acute myeloid leukemia; MDS, myelodysplastic syndromes; GVHD, graft-versus-host disease; B-NHL, B-cell non-hodgkin lymphoma. Together, these strategies represent a paradigm shift from indiscriminate cytotoxic ablation toward precision immune depletion. By selectively dismantling hematopoietic and immune compartments while sparing systemic tissues, these approaches aim to minimize toxicity, broaden transplant eligibility, and more precisely coordinate immune elimination with durable immune reset. Importantly, future conditioning strategies may need to move beyond simply “clearing the room” toward actively “programming” the immune rebuilding process. In addition to targeted depletion of pathogenic immune and hematopoietic cells, preservation and modulation of the stromal and thymic microenvironment may play critical roles in directing subsequent immune reconstitution. Incorporating approaches such as thymic and stromal protection, inflammatory signal modulation, damage-associated molecular pattern (DAMP) antagonism, and cytokine-guided immune regeneration may help establish a microenvironment that favors central tolerance, balanced immune recovery, and suppression of GVHD. Continued refinement of antigen selection, combinatorial strategies, and microenvironment-directed interventions will be essential to fully realize the next generation of conditioning platforms that support not only immune reset, but also immune reprogramming. Immune reset: from reconstitution to homeostasis Under physiological conditions, immune homeostasis reflects a balance refined over years of antigen exposure and regulatory adaptation. Following BMT, this historical context is erased. The newly developing immune system must establish tolerance and competence de novo, but within a profoundly altered host environment. Accordingly, immune homeostasis after BMT should not be viewed as a simple return to baseline. Rather, it represents the establishment of a new equilibrium shaped by donor immune repertoires, recipient tissues, inflammatory cues, and therapeutic interventions (Figure 1). This “reset homeostasis” may differ fundamentally from both donor and recipient pre-transplant immune states. Immune reconstitution following allo-BMT. Immune reconstitution after allo-BMT proceeds through two complementary pathways: (1) a peripheral pathway, in which donor-derived mature immune cells undergo homeostatic expansion; and (2) a central pathway, in which donor-derived hematopoietic stem and progenitor cells engraft within the recipient microenvironment and undergo positive and negative selection in the thymus and bone marrow, thereby redefining recipient tissue antigens from “non-self” to “self” and generating de novo T and B cells. These two processes are tightly interconnected. Alloreactive cells generated during peripheral reconstitution can induce GVHD, leading to thymic injury and delayed central reconstitution. Conversely, peripherally reconstituted immune cells recover rapidly after transplantation and play essential roles in early immune surveillance and elimination of infections and residual tumor cells. Therefore, selectively modulating peripherally reconstituted T cells is required to maintain peripheral immune homeostasis. In parallel, peripheral hematopoietic cells, particularly DCs, can migrate into the thymus and participate in antigen presentation and central tolerance establishment. However, re-establishment of central tolerance after BMT differs substantially from physiological immune development and remains incompletely understood. Because donor-derived immune cells must ultimately recognize recipient antigens as “self” during central reconstitution, restoration of thymic function and promotion of thymus-dependent immune regeneration are likely key mechanisms underlying long-term immune homeostasis after BMT. Looking forward, immune reconstitution after BMT should not be viewed merely as numerical recovery of immune cells, but rather as a process of “immune reset” – a coordinated orchestration of peripheral and central immune reconstitution that ultimately restores immune homeostasis. Allo-BMT, allo-bone marrow transplantation; GVHD, graft-versus-host disease; DCs: Dendritic cells. Previous studies have largely emphasized the numerical recovery of immune cell subsets after BMT. However, the extent to which quantitative reconstitution translates into functional immune homeostasis remains poorly defined. In the following sections, we extend beyond numerical reconstitution to integrate mechanistic insights into the key components that govern immune reprogramming after transplantation. T cells: the cornerstone of immune homeostasis remodeling As the central effectors of adaptive immunity, T cells are pivotal regulators of immune homeostasis under steady-state conditions and during immune reassembly following BMT. Their roles in immune tolerance and post-transplant immune reconstitution have been extensively discussed elsewhere [4], 48]. Notably, modern haploidentical BMT strategies, including the granulocyte colony-stimulating factor/anti-thymocyte globulin-based (G-CSF/ATG-based) Beijing protocol and post-transplant cyclophosphamide (PTCy)-based regimens, have achieved robust clinical success across HLA barriers [3]. A striking feature of these approaches is that many recipients are able to eventually discontinue long-term immunosuppressive therapy after allogeneic BMT, providing compelling evidence for durable immune reset that extends beyond physiological immune development. Despite these advances, the immunological features of the re-established homeostatic state after BMT remain poorly defined. In a recent study, we characterized immune cell subset distribution in patients who successfully restored immune homeostasis following transplantation. Our findings revealed that HLA-matched BMT and haploidentical BMT exhibit distinct patterns of immune homeostasis remodeling. After BMT, both CD4 and CD8 T cells can reconstitute through peripheral expansion of donor-derived mature T cells or through de novo generation via the thymus-dependent central pathway. In HLA-matched BMT recipients, CD4 and CD8 effector T cells preferentially reconstitute through peripheral expansion, whereas in haploidentical BMT recipients, effector T cells were more prominently regenerated through the thymus-dependent central pathway and accompanied by higher levels of thymic output [49]. Peripherally reconstituted T cells are readily activated by DAMPs and pathogen-associated molecular patterns (PAMPs) induced by conditioning regimens, as well as by alloimmune stimulation. However, these cells recover rapidly and play essential roles in early immune surveillance and the elimination of infections and residual tumor cells. Therefore, selectively modulation of peripherally reconstituted T cells to limit alloimmune responses while preserving protective immunity is essential for maintaining peripheral immune homeostasis. Importantly, our findings suggest that restoration of thymic function and promotion of thymus-dependent T-cell regeneration may be key mechanisms enabling successful immune re-establishment across HLA barriers in haploidentical BMT. However, the thymus is particularly susceptible to injury from cytoreductive conditioning [50], 51]. Even sublethal irradiation induces acute damage to thymic epithelial cells (TECs), with medullary TECs, especially autoimmune regulator (AIRE+) medullary thymic epithelial cells (mTECs), showing persistent functional impairment despite apparent recovery of overall thymic cellularity [51], 52]. Such dysfunction can persist for weeks to months after transplantation [51], 52]. Additionally, AIRE+ mTECs are also could be injured by alloreactive donor T cells during thymic GVHD, leading to reduced mTEC numbers and diminished expression of self-antigen [53], 54]. Quantitative or functional defects in Aire+ mTECs compromise central tolerance, permitting the escape of newly generated autoreactive T cells into the periphery [48], 55]. These cells subsequently drive secondary tissue damage and contribute to the development of chronic GVHD (cGVHD) [56], particularly in haploidentical BMT settings. Furthermore, our recent study further demonstrates that major histocompatibility complex (MHC) disparity itself impedes thymus-dependent T cell reconstitution after BMT, and inhibiting the transforming growth factor-β1 (TGF-β1) and low-density lipoprotein receptor-related protein 6 (LRP6) pathways partially rescued the differentiation and maturation of thymic T cells in the haploidentical BMT group [57]. Despite its central role, how thymus-dependent central tolerance is re-established following BMT remains unclear. Mechanistic studies to date have largely focused on how key clinical interventions, such as G-CSF, ATG and cyclophosphamide, induce peripheral T cell tolerance [58], [59], [60]. In contrast, their effects on thymus-dependent central tolerance remain poorly defined. Evidence suggests that G-CSF can protect the thymus from radiation-induced injury [61], 62], and could facilitate central T cell reconstitution after BMT [63]. The success of immune reset achieved by contemporary allo-BMT protocols thus implies that these regimens may also promote thymic repair and the re-establishment of central tolerance, providing opportunities to extend these principles to other contexts of immune injury. In this regard, allo-BMT serves not only as a therapeutic intervention but also as a unique platform for dissecting the regulatory mechanisms that govern immune reset beyond normal physiological development. Our recent work provides insight into this process. By mapping the immune cellular landscape of allo-BMT recipients who achieved stable immune homeostasis, we identified a previously unrecognized subset of CD8+ T cells, termed CD8 regulatory precursors (CD8 Trp) [49]. These cells exhibit a distinct transcriptomic profile that is clearly separable from conventional CD8+ T cell subsets and partially overlaps with that of CD4 regulatory T cells (Tregs). CD8 Trps were selectively expanded in recipients who successfully restored immune homeostasis after allo-BMT, with particularly pronounced expansion in haploidentical transplant recipients. Further phenotypic characterization identified CD8 Trp cells as CD8+HLA-DR+CD27+ [64]. This population also exists under physiological conditions, accounting for approximately 5 % of total CD8+ T cells in healthy individuals. Functionally, CD8 Trp cells exhibit regulatory precursor-like properties and can exert immunoregulatory effects by promoting forkhead box protein P3 (FOXP3) expression in both CD4+ Tregs and themselves [64]. These findings suggest that BMT-mediated immune reset may selectively engages immune cell subsets that appear functionally redundant under steady-state conditions but become critical for restoring immune homeostasis following profound immune disruption. Further studies are needed to investigate whether the CD8 Trp signature is conserved across both homeostatic and pathogenic conditions. Such investigations may not only identify novel regulators of immune homeostasis, but also facilitate the development of new therapeutic strategies for autoimmune diseases and immune homeostasis restoration in solid organ transplantation. B cells: orchestrators of durable immune homeostasis restoration Once regarded primarily as “antibody factories,” B cells are now recognized as versatile regulators of immune response, functioning as antigen-presenting cells, cytokine producers, and architects of lymphoid organization. These diverse roles position B cells as key determinants of durable immune reset rather than passive downstream effectors. Similar to T cells, B cells undergo tightly regulated positive and negative selection during their development in the BM to establish discrimination between self and non-self [65]. Successfully selected B cells subsequently migrate to peripheral lymphoid organs, including the spleen and lymph nodes, where they complete maturation and acquire the capacity for activation and high-affinity antibody production [65]. Failure of central or peripheral B cell tolerance permits the persistence of autoreactive clones and the generation of pathogenic autoantibodies, contributing to autoimmune diseases such as systemic lupus erythematosus (SLE) and rheumatoid arthritis. In the context of BMT, the re-establishment of B cell homeostasis also can be understood through both central and peripheral regulatory mechanism. Central B cell reconstitution derives from donor HSPC, which must undergo de novo selection processes to establish tolerance to recipient antigens. However, conditioning-induced injury profoundly disrupts the BM niche, impairing the generation and maturation of B cell progenitors [66]. Within the BM, stroma components, including osteoblasts, endothelial cells, and mesenchymal stromal cells (MSCs), provide essential cytokines and adhesion signals required for B cell development [67]. High-dose irradiation and busulfan-based conditioning severely damage osteoblastic niches and vascular endothelium, leading to loss of key B cell supportive factors such as IL-7 and CXCL12 (SDF-1) [68], 69]. As a result, B cell reconstitution after transplantation is markedly delayed, often requiring 6–12 months, largely because the damaged BM niche cannot support the transition from pro-B to pre-B cells [70]. Notably, selective injury to the osteoblastic niche preferentially impairs B cell lineages, a phenomenon describes as “bone marrow GVHD” when exacerbated by donor-derived immune responses but initially triggered by conditioning-induced damage [66]. Meanwhile, profound lymphopenia following transplantation results in markedly elevated levels of B cell activating factor (BAFF). In the absence of a sufficient reconstituted B cell compartment to buffer BAFF, elevated BAFF lowers the survival threshold for B cells, allowing autoreactive B cell clones that would normally be deleted to persist and expand [71]. These pathogenic B cells further exacerbate immune dysregulation by promoting T cell activation and producing pro-inflammatory cytokines such as interleukin-6 (IL-6), thereby sustaining germinal center (GC) reactions that drive chronic inflammation, tissue fibrosis, and organ damage [71]. As a result, defects in B cell regeneration after BMT contribute to the pathogenesis of cGVHD, whose clinical and immunological features closely resemble those of autoimmune disorders. In this setting, B cells transition from protective regulators to key drivers of a self-perpetuating cycle of immune activation and tissue injury [72], 73]. Beyond their intrinsic reconstitution, B cells may also contribute to T cell central tolerance after BMT. Under steady state conditions, thymic B cells can present self-antigens and promoting central CD4+ T cell tolerance [74], [75], [76]. Following transplantation, donor-derived B cells may enter the recipient thymus and participate in T cell regeneration. However, the precise mechanisms and regulatory roles of B cells in T cell central reconstitution remain poorly defined. Given the profoundly altered microenvironment during B cell regeneration, dysregulated B cell responses emerge as a central driver of cGVHD. Accordingly, B cell depletion has become an important therapeutic strategy in both autoimmune disease and cGVHD [71], 77]. Anti-CD20 therapies, such as rituximab, can ameliorate clinical manifestations, although clinical benefit often correlates poorly with reductions in circulating autoantibodies. Increasing evidence suggests that efficacy instead derives from disruption of antigen presentation, inflammatory cytokine networks, and pathogenic B-T cell interactions rather than from elimination of antibody-secreting plasma cells alone [71], 77]. Importantly, the pattern of B cell reconstitution after depletion critically influences outcome: repopulation dominated by transitional and naïve B cells is associated with immune tolerance, whereas early emergence of activated or plasmablast-like populations predicts relapse and persistent inflammation [77]. These insights have driven the development of emerging strategies such as CD19−directed CAR-T cells and bispecific antibodies, which aim to eradicate the entire B cell repertoire, including cells in protected niches, thereby enabling a deeper and more sustained immune reset [78], 79]. Dendric cells: arbiters of immunity and tolerance Dendritic cells (DCs) are rare, heterogeneous, BM-derived professional antigen-presenting cells (APCs) that occupy a central position at the interface of innate and adaptive immunity. They originate from HSCs in the BM, arise from both myeloid and lymphoid progenitors, and are widely distributed in the blood, secondary lymphoid organs, and peripheral tissues [80]. In the steady state, DCs typically reside in an immature or semimature state, characterized by efficient antigen uptake and processing but low expression of costimulatory molecules. In this context, DCs continuously sample self-antigens and promote peripheral tolerance by inducing T cell anergy, clonal deletion, or the differentiation of Tregs. Upon sensing pathogen-associated or damage-associated molecular patterns through pattern recognition receptors (PRRs), such as Toll-like receptors, DCs undergo functional maturation, marked by upregulation of MHC and costimulatory molecules, and acquisition of potent T cell priming capacity, thereby enabling robust effector T cell responses [81]. DCs comprise several major subsets, including conventional DCs (cDCs), which are further divided into cDC1 subsets specialized in antigen cross-presentation and CD8+ T cell activation, and cDC2 subsets that preferentially support CD4+ T cell response. Additional subsets include plasmacytoid DCs (pDCs), monocyte-derived DCs (moDCs), and Langerhans cells (LCs). The development and maintenance of these DC populations are tightly regulated by hematopoietic growth factors, with fms-like tyrosine kinase 3 ligand (Flt3L) playing a central role, alongside context-dependent contributions from GM-CSF and M-SCF [82]. Following BMT, DCs play critical roles in both central and peripheral immune homeostasis restoration. Conditioning regimens profoundly perturb DC number, origin, and functional state, resulting in the dynamic coexistence of residual host DCs and newly generated donor-derived DCs. Both populations are capable of presenting alloantigens and shaping post-transplant T cell responses. Previous studies have demonstrated that either host or donor DCs alone are sufficient to induced aGVHD, as host APCs can directly present alloantigen-MHC complexes, whereas donor APCs can indirectly present host antigens [83], [84], [85]. Despite this pro-inflammatory potential, the functional heterogeneity of DC subsets also enables the induction of immune tolerance. Immature DCs residing in the bone marrow, as well as pharmacologically induced tolerogenic DCs (tolDCs), have been shown to suppress alloreactivity and promote tolerance in experimental models of organ transplantation and BMT [81], [86], [87], [88]. Accordingly, tolDCs-based therapies have emerged as a promising strategy for inducing immune reset in autoimmune disease, with multiple ongoing clinical investigations [81]. Beyond their role in peripheral tolerance, DCs also contribute to the establishment of central tolerance [89]. In the setting of organ transplantation, the infusion of donor bone marrow into appropriately conditioned recipients can induce durable donor-specific tolerance [90]. This process is largely attributed to donor antigen presentation by DCs within the recipient thymus, which promotes negative selection of donor-reactive T cells and prolongs graft survival [91], [92], [93]. However, the clinical application of combined organ transplantation and BMT remains limited by the toxicity associated with cytoreductive conditioning regimens. In this context, tolDC-based approaches may offer a less toxic alternative for delivering donor-derived antigenic signals to the immune system while preserving overall immune competence and promoting long-term tolerance. Conclusions and perspective With the rapid advancement of clinical and cellular technologies, resetting the human immune system beyond its original developmental window has become an increasingly tangible reality. Yet, most immunological research has focused on how the immune system develops and functions under steady-state conditions, while the mechanisms governing immune reset, and the criteria defining its success, remain poorly understood. BMT uniquely exposes the immune system to a controlled collapse followed by de novo reconstruction, providing an unparalleled opportunity to investigate how immune homeostasis is rebuilt after catastrophic disruption. By reframing BMT as a model of immune reset rather than simple immune reconstitution, we uncover fundamental principles that are often concealed in physiological immunity, including re-establishment of central and peripheral tolerance, dependence of immune recovery on stromal regeneration, and dynamic rebalancing of regulatory networks across hematopoietic lineages. These insights support a therapeutic shift from broad immunosuppression toward rational immune recalibration and ultimately immune reprogramming. A major future direction will be the establishment of integrated functional biomarkers that can more accurately define successful immune reset beyond conventional quantitative immune recovery. Potential biomarkers may include T cell receptor (TCR) repertoire diversity, epigenetic signatures associated with stable Treg identity, thymic-output-associated naïve T-cell states, and network-based indices reflecting the balance between pro- and anti-inflammatory immune programs. These multidimensional biomarkers may help define not only whether immune recovery occurs, but also the quality, stability, and “depth” of immune reset. From a therapeutic perspective, next-generation immune reset protocols will likely require sequential and coordinated interventions rather than single-step immune depletion. Early-stage strategies may focus on promoting central tolerance and stromal regeneration through thymic-protective approaches, tolerogenic dendritic cells, cytokine-guided regeneration, or microenvironmental modulation. Subsequent interventions may then fine-tune peripheral immune homeostasis through approaches such as Treg-based therapies, selective cytokine blockade, or targeted modulation of pathogenic immune subsets. Such staged strategies may more effectively support durable immune tolerance while preserving protective immunity. Extending immune reset strategies beyond malignant diseases to autoimmunity and solid organ transplantation presents introduces additional challenges. A central question is how to achieve a sufficiently deep immune reset capable of eliminating pathogenic or autoreactive immune clones without the toxicity of full myeloablation. Therefore, defining and controllably titrating the “depth of reset” may become a critical goal for future immune reset therapies. Ultimately, translating these principles into less toxic, more precise, and programmable interventions may enable durable restoration of immune homeostasis across a broad range of immune-mediated diseases.