High mobility group box 1 (HMGB1): a pivotal regulator of hematopoietic malignancies

The HMGB1 gene is located on chromosome 13q12 and includes five exons and four introns. The TATA box promoter of the HMGB1 gene contains binding sites for several transcription factors, such as activator protein 1 (AP1), and a silencing element [4]. Human HMGB1 protein is a highly conserved nuclear protein consisting of 215 amino acids with a molecular weight of approximately 30 kD. Structurally, HMGB1 is divided into three functional regions (Fig. 1): A-box (9-79 aa), B-box (89-162 aa), and acidic C-terminus (186-215 aa). The A-box and B-box are composed of 80–90 amino acid residues, with similar amino acid repeats and nonspecific DNA binding sites; the B-box is a functional structural region that causes an inflammatory response [5]. However, the A-box has a certain antagonistic effect on the B-box [6]; the acidic C-terminus containing aspartic acid and glutamic acid is mainly involved in regulating the binding affinity between HMGB1 and DNA, and mediates gene transcription and chromosome derotation [7]. The N-terminus of HMGB1 (6–12 aa) contributes to heparin-binding activity. After binding to HMGB1, heparin impacts the spatial conformation of HMGB1, reduces the affinity of HMGB1 for its receptor, and inhibits its proinflammatory activity [8, 9]. The B-box domain has two crucial binding sites for Toll-like receptor 4 (TLR4) and receptor for advanced glycation end products (RAGE), which regulate the release of proinflammatory cytokines. The RAGE binding site of HMGB1 is located between amino acid residues 150 and 183, and the 20 amino acids of the TLR4 binding site (89-–108 aa) are the minimal sequence necessary to induce cytokine activity [10, 11]. Although HMGB1 is an evolutionarily conserved multifunctional protein, the biological function of HMGB1 depends on its modifications, cellular location, redox state, and binding partners.

The HMGB1 protein shuttles between the nucleus and cytoplasm because it contains two nuclear localization sequences (NLSs) and two putative nuclear export signals (NESs). HMGB1 interacts with the nuclear receptor chromosome-region maintenance-1 (CRM-1), which is a nuclear transport receptor involved in the export of leucine-rich NES proteins and is then released from the nucleus into the cytoplasm [12]. The conserved lysine residues in the NLSs are sensitive to acetylation and can activate nuclear exclusion and HMGB1 translocation [13‐15]. In addition to acetylation, HMGB1 is regulated by extensive posttranslational modifications (PTMs) including methylation, phosphorylation, ADP-ribosylation, glycosylation, and ubiquitination. These PTMs redirect it toward secretion and modulate its interactions with DNA and other proteins [16]. Then, the oxidation of extracellular HMGB1 determines its bioactivity in mediating inflammation and innate immune responses.

There are two mechanisms for releasing HMGB1 into the extracellular environment: passive release and active release (Fig. 2). In response to infection and injury, HMGB1 can be actively secreted from activated immune cells or passively released from damaged or necrotic cells and transferred outside the cell [37, 38]. Active release of HMGB1 from macrophages or monocytes requires a proinflammatory stimulus that could cause an immune response. Active HMGB1 release promotes neutrophil recruitment and macrophage release of proinflammatory cytokines, such as TNF-α and interleukin-6 (IL-6) and dendritic cell (DC) activation [39]. HMGB1 can be passively secreted from the nuclei of necrotic cells and damaged cells and then triggers inflammatory responses by functioning as necrotic cell death markers [36].

Once released from the cells, HMGB1 binds to cell-surface receptors, inducing a reaction as a prototypical DAMP. Classic HMGB1 receptors include RAGE, TLRs (TLR2, TLR4, and TLR9), CXCR4, and T cell immunoglobulin mucin-3 (TIM-3) [40, 41].

HMGB1 can regulate HSC multipotency and self-renewal at the transcriptional level (Fig. 3). In conjunction with FOS, TCFEC, and SFPI1, HMGB1 confers a clear repopulation advantage to HSCs via a non-cell-autonomous phenomenon [70]. A recent study demonstrated that HMGB1^−/− mouse embryonic fibroblasts (MEFs) showed slight telomere shortening but significantly decreased telomerase activity and DNA damage [71]. This indicates that HMGB1 may modulate chromosomal stability of HSCs by altering the functional chromatin structure of telomeres. HMGB1 can also bind to p53 DNA and stimulate DNA linearization, which increases p53 activity [72]. Moreover, the HMGB1 A-box has strong p53 binding activity based on crosslinking chemical and biophysical measurements [73]. HMGB1 regulates not only the transcriptional activity of p53 but also the subcellular localization and phosphorylation of p53.

HMGB1 plays an important role in the mobilization of HSPCs, thus regulating BM microenvironment formation. Altmann S et al. found that HMGB1 was broadly expressed in canine hematopoietic cells and directly induced the proliferation of peripheral blood mononuclear cells (PBMCs) [74]. Furthermore, mobilization of HSPCs is mainly the result of a sterile inflammatory response to mobilizing stimuli in the BM microenvironment. In the initiation stage of the mobilization process, HMGB1, which binds to mannan-binding lectin (MBL), regulates the mobilization of HSPCs into peripheral blood (PB) [75].

HMGB1 participates in granulocyte colony-stimulating factor (G-CSF)-induced mobilization of HSCs from the BM into the systemic circulation [76]. Additionally, a clonogenic assay for CFU-granulocyte-monocytes indicated that HMGB1 was required to prevent HSC exhaustion and maintain immune/hematopoietic homeostasis. HMGB1 is linked to substance P (SP) and neurokinin-A (NK-A) to protect the most primitive hematopoietic cells and ensure hematopoietic homeostasis. Mechanistically, HMGB1 negatively regulates hematopoietic stimulation, while SP, a hematopoietic stimulator, decreases HMGB1 expression. Furthermore, NK-A can negatively regulate SP-mediated hematopoietic stimulation [77‐79]. The dysfunction of HMGB1 may promote the occurrence and development of hematological malignancies by interfering with the hematopoietic function of the BM.

The BM is a soft viscous tissue that occupies cavities within the bone [80]. The BM microenvironment is a dynamic network composed of growth factors, cytokines, and stromal cells, which provides a supportive environment for the occurrence and development of hematopoietic malignancies [81]. As a cytokine, HMGB1 can bind to RAGE and TLR4 to activate proinflammatory signaling pathways, such as the NF-κB pathway, and sustain the inflammatory BM microenvironment by inducing cytokine release and recruiting leukocytes. Subsequently, the inflammatory BM microenvironment can accelerate neoplastic transformation and support tumor growth, invasion, and metastases. Infiltrating leukocytes and cancer cells have the ability to secrete HMGB1 in response to hypoxia, injury, inflammatory stimuli, or environmental factors. This loop promotes inflammatory responses and the development of an inflammatory BM microenvironment.

Myeloid-derived suppressor cells (MDSCs) are newly identified immature myeloid cells with immunosuppressive activity. During tumor microenvironment (TME), MDSCs suppress the host anti-tumor immune response through inhibition of T cell proliferation, cytokine secretion, and recruitment of regulatory T cells in hematological malignancies. In all hematological malignancies, several strategies to target MDSCs could improve immune therapies via multiple mechanisms, such as hampering MDSCs function, promoting MDSCs maturation, and depleting MDSCs [82‐84]. HMGB1 can facilitate MDSCs differentiation in BM and inhibit the activation of antigen-driven CD4⁺ and CD8⁺ T cells. HMGB1 also increases MDSC-mediated IL-10 production, enhances crosstalk between MDSCs and macrophages, and promotes MDSCs to downregulate the expression of the T cell-homing receptor L-selectin [85]. Circulating complement C1q can stimulate leukocyte-associated Ig-like receptor-1 (LAIR-1) and maintain monocyte quiescence [86]. Very high levels of HMGB1 induce proinflammatory M1-like macrophage differentiation, and high levels of HMGB1 synergize with C1q via RAGE and LAIR-1 to induce the differentiation of monocytes to anti-inflammatory M2-like macrophages [87]. HMGB1 could be released into the BM microenvironment by DCs as a potential immunomodulatory factor to bind with RAGE on the T cell surface and mediate the interaction between DCs and T cells, which is involved in the occurrence and development of hematological malignancies [88]. A study also showed that HMGB1 enhances the maturation and accumulation of DCs by promoting CCR5 and CXCR3 production and inducing potent T cell cytotoxicity [89].

Mesenchymal stem cells (MSCs) play a “double-edged sword” role in hematological malignancies. Studies indicate that MSCs appear to influence pathways that can suppress both proliferation and apoptosis [90]. MSCs protect T cell acute lymphoblastic leukemia (T-ALL) cells from drug-induced apoptosis though mitochondria transfer mechanism, which eventually leads to chemotherapy resistance [91]. Tumor-associated MSCs are essential components of the TME and also associated with a protumorigenic effect by enhancing tumor cell stemness. HMGB1 also regulates MSCs to promote the inflammatory BM microenvironment formation. HMGB1 acts as a chemoattractant to MSCs. Substantial evidences have revealed that HMGB1 significantly upregulates epidermal growth factor receptor (EGFR) and activates the Ras/MAPK pathway to regulate the differentiation of MSCs [92]. These results demonstrate that HMGB1 induces MSCs to secrete multiple cytokines, which are predominantly associated with the development of an inflammatory BM microenvironment. Furthermore, HMGB1 in the inflammatory BM microenvironment can promote the senescence of MSCs via the TLR2/4 and NF-κB signaling pathways, and inhibition of HMGB1 by ethyl pyruvate (EP) can improve lupus nephritis and reverse senescence-associated secretory phenotype (SASP) development [93, 94]. These findings suggest that nuclear HMGB1 can redistribute or relocalize to the extracellular environment in senescent cells. Moreover, senescent fibroblasts secrete oxidized HMGB1, which stimulates cytokine secretion through TLR4 signaling, inducing p53-dependent cellular senescence. Therefore, the alarmin HMGB1 has been considered a central mediator of senescent phenotypes [95]. Interestingly, a recent study found that metformin, a widely used drug for type 2 diabetes, can block HMGB1 translocation and inhibit catabolic production and cell senescence in stem cells (SCs) [96]. Cellular senescence is considered a tumor-suppressive mechanism that permanently arrests cells that are at risk for malignant transformation, can secrete SASP into the BM microenvironment, and transform senescent fibroblasts into proinflammatory cells that have the ability to promote tumor progression [97, 98].

Myelodysplastic syndrome (MDS) is a heterogeneous group of clonal disorders that is characterized by abnormal differentiation of HSCs, ineffective hematopoietic function of BM, and the risk of conversion to acute myeloid leukemia (AML). The inflammatory BM microenvironment is involved in the development and progression of MDS by inducing the apoptotic death of BM progenitor cells. Charoonpatrapong et al. found that DCs released HMGB1 as a potent immunomodulatory cytokine into the BM microenvironment. HMGB1 binds to RAGE on the surface of T cells to mediate the interaction between DCs and T cells [88]. In addition, Velegraki et al. revealed that TLR4 was overexpressed in the BM mononuclear cells of MDS patients compared with those of the control group. TLR4 inhibitors can also inhibit the production of proinflammatory cytokines released by monocytes in patients. Moreover, a study has illustrated that TLR4-dependent inflammatory cytokines not only increase cell apoptosis but also impair the cell clearance capacity of macrophages under the influence of the endogenous ligand HMGB1 [52]. Recently, Angel Y.F. and Kam et al. identified HMGB1 as a previously undescribed target that modulated the innate immune system in MDS. This group used combined siRNAs and the small molecule inhibitor sivelestat to study the loss of function of HMGB1 compared to standard chemotherapy. In MDS cells, sivelestat, a neutrophil elastase inhibitor, increases the expression of PUMA and DNA double-strand breaks and activates caspase-3, which indicates that sivelestat can downregulate HMGB1 and suppress the TLR and NF-κB pathways to promote apoptosis in the BM [99]. The reduction in HMGB1 levels is sufficient to impair MDS cell self-renewal and promote apoptotic cell death. Inhibitors of HMGB1 signaling can provide a first-in-class therapeutic option for patients with MDS and can be used as monotherapy or in combination with chemotherapies to improve the sensitization of MDS cells.

In acute promyelocytic leukemia (APL), the direct molecular target of all-trans-retinoic acid (ATRA) in human myeloid cells is the PML-RARα oncoprotein that mediates differentiation [100]. It has been proved that upregulated endogenous HMGB1 promoted autophagy and induced NB4 cell differentiation via ubiquitin-binding adaptor protein p62/SQSTM1-mediated degradation of PML-RARα oncoprotein [101]. However, long-term exposure of ATRA and arsenic trioxide (ATO) results in hyper-inflammation and development of the differentiation syndrome (DS) [102]. HMGB1 promoted ATRA/ATO-induced DS by enhancing inflammation through the MEK/ERK signaling pathway [103]. Under the induction of specific chemical reagents, murine erythroleukemia (MEL) cells release HMGBl and promote self-differentiation. However, HMGB1 also mediates the differentiation of MEL cells through pathways other than HMGB1-RAGE [104‐107]. During the progression of AML, HMGB1 is secreted to induce TNF-α production and subsequent secretion of IL-1β, which stimulates endothelial cells to release stem cell factor (SCF), which can further promote the proliferation of AML cells. HMGB1 is also dependent on the immune receptor Tim-3 to induce angiogenic VEGF secretion and participate in tumor angiogenesis [69]. Additionally, Liu et al. confirmed that miR-34a suppressed the expression of HMGB1 by directly binding with its 3′-untranslated region (UTR). Overexpression of miR-34a can dramatically reverse apoptosis inhibition by downregulating the expression of HMGB1 in AML [108].

HMGB1 not only is highly expressed and directly induces autophagy in AML cells but also indirectly promotes autophagy to result in therapeutic resistance by enhancing the effect of Beclin-1/PI3KC3 and Atg5-Atg12-Atg16. Reducing the expression of HMGB1 by miR-34a, miR-181b and miR-142-3p enhances the drug sensitivity of AML cells by inhibiting autophagy; moreover, miR-142-3p directly targets HMGB1 to not only represses autophagy but also reduces P-gp to enhance the drug sensitivity of AML cells [108‐110]. HMGB1 also increases the expression of monocyte chemoattractant protein 1 (MCP1) and myeloid cell leukemia 1 (Mcl1) to promote the migration of human leukemia monocytic THP1 cells, which is inhibited by glycyrrhizin (GL) [111]. It has been reported that HMGB1 knockdown reduces the expression of RAGE, which is developed from a cell adhesion molecule family and acts as an adhesion molecule in mammalian cells [112]. Extracellular HMGB1 is not only an important DAMP that is released by cells upon necrosis but also a regulatory factor to prevent AML cell necroptosis. When Z-VAD-fmk inhibits caspase activity, etoposide induces necroptosis by triggering cIAP1/2 depletion [113]. However, extracellular HMGB1 prevents this necroptosis. Interestingly, HMGB1 enhances cell viability and regulates necroptosis through the NF-κB pathway rather than preventing cIAP1/2 degradation [114].

In chronic myeloid leukemia (CML), HMGB1 knockdown can arrest the cell cycle at the G1 phase and inhibit cell proliferation by downregulating the expression of cyclooxygenase-2 (COX-2) [115]. COX-2 activates the Akt/survivin- and Akt/ID3 signaling pathways, which are related to promoting proliferation [116]. Second-generation tyrosine kinase inhibitors (TKIs), such as dasatinib, can reactivate the induction of apoptotic cell death in patients with imatinib-refractory CML and Philadelphia chromosome-positive ALL (Ph⁺ALL). However, HMGB1-mediated necroptosis gives rise to dasatinib-induced cardiotoxicity, which reduces the clinical applications of dasatinib, indicating that targeting HMGB1 may be a viable strategy to prevent dasatinib-induced cardiotoxicity [117]. A tetrahydrobenzimidazole derivative TMQ0153 has a strong pro-oxidant effect against imatinib-sensitive and imatinib-resistant CML cells. TMQ0153 treatment significantly stimulates the release of HMGB1, leading to immunogenic cell death (ICD), which is a form of chemotherapy-induced tumor cell death [118]. Yang et al. found that cytoplasmic HMGB1 reduced the sensitivity of CML cells to death induced by anticancer drugs by upregulating the autophagy pathway. HMGB1 overexpression increases the transcriptional activity of JNK, ERK, and Beclin-1 [119]. Chen et al. found that HMGB1 knockdown promoted the apoptosis of K562 cells by increasing Bax protein and reducing Bcl-2 protein [115]. Conversely, HMGB1 overexpression inhibits ADM-induced apoptosis in K562 cells by regulating Bcl-2 protein levels and the activity of caspase-3/9 [120]. Moreover, knockdown of HMGB1 significantly inhibits the adhesion of K562 cells [115].

Vincristine, corticosteroids, and l-asparaginase in conjunction with intrathecal therapy can completely alleviate 95% of ALL in patients [121]. Compared with the healthy control and ALL patients with complete remission, serum HMGB1 level of ALL patients was increased, but there was no significant difference in HMGB1 level between the healthy control group and ALL complete remission group, indicating that serum HMGB1 is a useful biomarker to evaluate the prognosis of childhood ALL. Moreover, HMGB1 may be associated with the stages of hemocyte differentiation and maturation. HMGB1 is released from ALL cells and promotes inflammation by stimulating leukemic cells to secrete TNF-α through a MAPK-dependent mechanism [122]. This finding indicates that HMGB1 expression is positively correlated with the clinical status of ALL patients. Moreover, the Ulk1-Atg13-FIP200 complex, which is upstream of HMGB1-Beclin1 and PI3KC3-Beclin1 complexes, promotes HMGB1 trafficking and consequently upregulates autophagy. Therefore, targeting the transformation of autophagic complexes or HMGB1 translocation may inhibit autophagy, and thus reverse ALL drug resistance [123]. Unlike common miRNA effects with negative correlations, inhibition of miR-181a expression induces a decrease in HMGB1 protein in T- and B-ALL cells. This suggests that dysregulation of HMGB1, perhaps due to miR-181a dysregulation, promotes leukemogenesis [124]. Anthracycline can induce a tumor-specific immune response through HMGB1 release in the late stage, and play a role in enhancing the antigen expression of dead tumor cells to DCs through the TLR4 receptor in ALL cells [125].

The expression level of HMGB1 in many primary lymphomas is higher than the average level in normal lymph nodes, and HMGB1 is only detected in lymphoma cells. There is a correlation between HMGB1 expression and classification [126]. Chronic lymphocytic leukemia (CLL) is the most common subtype of non-Hodgkin lymphoma (NHL) and is mainly characterized by mature small lymphocytes invading PB and lymphoid tissues such as BM, lymph nodes, and the spleen. Jia et al. found that the plasma HMGB1 levels in CLL patients were significantly higher than those of the healthy control group, and the HMGB1 concentration was related to the absolute lymphocyte count. Furthermore, CLL cells passively release HMGB1 through the HMGB1-RAGE/TLR9 pathway and differentiate CD14⁺ monocytes from CLL cells into nurse-like cells (NLCs), thus regulating the microenvironment. The high number of NLCs is related to the short survival time of CLL patients [127]. Cutaneous T cell lymphoma (CTCL), the second most common extranodal NHL, is characterized by clonal accumulation of postthymic T cells residing in the skin and represents a group of diseases such as mycosis fungoides (MF) and Sézary syndrome (SS) [128, 129]. Senda et al. demonstrated that HMGB1 expression in sera is increased in CTCL patients and correlates with serum levels of soluble IL-2 receptor, lactate dehydrogenase, thymus and activation-regulated chemokines, and the number of eosinophils in PB. It was also found that the level of HMGB1 mRNA in CTCL-injured skin was significantly increased and positively correlated with IL-4, IL-10, IL-19, and angiogenin mRNA levels [130]. It has been reported that IL-4, IL-10, and IL-19 are associated with Th2 polarization [131‐133]. These results suggest that enhanced HMGB1 expression may contribute to the progression of CTCL through Th2 polarization and promotion of angiogenesis [130]. Notably, Fredholm S. et al. proved that 72% of CTCL patients had pY-STAT3-positive malignant T cells, and staining for eosinophils and the trafficking factor HMGB1 was also positive, which supports HMGB1 as a possible therapeutic target [134].

To evaluate the significance of HMGB1 in patients with T cell lymphoma, a study found that the expression of HMGB1 in 120 cases of T cell lymphoma was significantly higher than that in 40 cases of reactive lymphoid hyperplasia. Furthermore, the positivity rate of HMGB1 was used as an indicator for diagnosing T cell lymphoma in patients with lymph node biopsy. The specificity of this finding was 63.7%, which was significantly associated with malignancy and clinical stage but not gender, age, or tumor location. Elevated expression of HMGB1 may be a potential diagnostic marker for the development and progression of T cell lymphoma [135]. Zhao T et al. demonstrated that rituximab-induced inhibition of STAT3 activity led to an increase in HMGB1 release and a decrease in IL-10 secretion, triggering immune responses and greatly improving the clinical outcome of patients with diffuse large B cell lymphoma (DLBCL), suggesting that indirectly affecting the immune system rather than directly killing cells led to the elimination of DLBCL [136]. Conversely, HMGB1 stimulates DLBCL cell proliferation by activating the Src/ERK pathway, which is inhibited by EP, causing an accumulation of p27 and cell cycle arrest in the G1 to S phase transition. It has been suggested that EP-mediated blockade of the HMGB1-mediated signaling pathway can effectively inhibit the occurrence of DLBCL and disease progression [137]. Moreover, in their studies, HMGB1 plays a dual role in DLBCL as an inflammatory factor that promotes tumorigenesis and as a cytokine that induces immune responses, which further indicates that HMGB1 has a potential application in the pathogenesis and treatment of DLBCL [138].

In anaplastic large-cell lymphomas (ALCLs), Dejean et al. found that HMGB1 could activate the MMP-9, PAR-2, and NF-κB pathways to induce the release of IL-8, which bound to CXCR1 and CXCR2 on the surface of ALK-positive lymphoid cells to promote the proliferation and metastasis of lymphoid cells. After treatment with the HMGB1 inhibitor glycyrrhiza, the invasion and metastatic abilities of lymphoma cells were significantly decreased [139]. Adult T cell leukemia (ATL) patients have high plasma HMGB1 levels compared with normal controls [140]. It has been reported that high plasma HMGB1 levels in patients with ATL are caused by infection with human T cell lymphotropic virus type I (HTLV-I) [141]. In addition, HMGB1 mRNA is abundantly expressed in HTLV-I-infected T cell lines. The HTLV-I oncoprotein Tax enhances the expression of the HMGB1 gene at the transcriptional level by interacting with C/EBP and inducing extracellular release of HMGB1 by T cells. These results suggest that HMGB1 is a potential biomarker and a therapeutic target for ATL [140, 142].

In MM, high expression of HMGB1 is negatively associated with the 3-year survival of MM patients, which may be involved in promoting MM drug resistance. HMGB1 could participate in DNA damage repair and autophagy. In contrast, when HMGB1 is downregulated, the sensitivity of MM cells to dexamethasone (Dex) is enhanced by activating the mTOR pathway to inhibit autophagy and induce apoptosis [143]. Similarly, Gao et al. found that the expression of the lncRNA MALAT-1 and HMGB1 was dramatically increased in patients with untreated MM, while MALAT-1 expression and HMGB1 protein levels in patients with complete remission were significantly decreased. Furthermore, MALAT-1 increases the expression of HMGB1 at the posttranslational level by inducing HMGB1 ubiquitination in MM cells, thereby promoting autophagy and inhibiting apoptosis [29]. In addition, Roy M. et al. revealed that the expression of HMGB1 increased in MM bortezomib-resistant cells, and bortezomib combined with lycorine efficiently resensitized resistant cells to bortezomib. Mechanistically, the proteasomal degradation of the HMGB1 by lycorine inactivates the MEK-ERK pathway, inhibiting Bcl-2 dissociation from Beclin-1 and consequently suppressing autophagy [30]. Therefore, HMGB1 is an important target for MM patients to increase chemotherapy drug sensitivity.

Interestingly, HMGB1 can also participate in other pathological processes in MM. Similar to DAMPs emitted by apoptotic MM cells, HMGB1 fosters an immunogenic microenvironment to promote antitumor immunity. Recently, studies showed that chemotherapeutic agents, such as melphalan and docosahexaenoic acid (DHA), promoted the release of HMGB1 by ICD, leading to an immune response [144, 145]. Moreover, HMGB1 can act as a thrombosis-related biomarker in patients with MM. After Mel-P, bortezomib, and lenalidomide therapies, the plasma concentrations of HMGB1 were reduced in association with the risk of thrombosis [146, 147]. CXCR4 plays an important role in proliferation, invasion, dissemination, and drug resistance in MM [148]. This indicates that its ligand HMGB1 could regulate MM physiological processes. Because of its pivotal role in the progression of MM, HMGB1 is considered one of the most important potential targets for inhibiting tumor growth, metastasis, and drug resistance and optimizing current anti-MM treatment strategies (Table 1).

Hematopoietic stem cell transplantation (HSCT) is an intensive therapy to treat hematologic malignancies, but graft-versus-host disease (GVHD) is a frequent severe inflammatory complication that is associated with poor outcomes [149]. Yujiri et al. found increased serum levels of HMGB1 in patients who developed acute GVHD (aGVHD) after HSCT, which indicates that HMGB1 may be a useful indicator of GVHD [150]. Additionally, enhanced HMGB1 is reported to promote STAT3 expression in CD4⁺ T cells via modulation of its DNA methylation, subsequently inhibiting Tregs and promoting the Th17 response during GVHD [151]. It has been demonstrated that genetic variations in cytokine genes can modulate immune reactions after HSCT. An inherited variation in HMGB1 is associated with outcomes after allogeneic HSCT (allo-HSCT) [152]. Thus, HMGB1 is likely to play an important role in the development of GVHD, known as the graft-versus-tumor (GVT) effect, and possibly engraftment because of its central role in the activation of APCs and tissue regeneration. Moreover, the compound cyclopentylamino carboxymethylthiazolylindole (NecroX)-7 could protect mice against lethal GVHD by reciprocal regulation of regulatory T/Th1 cells, attenuating systemic HMGB1 accumulation and inhibiting the HMGB1-mediated inflammatory response [153]. Cyclophosphamide (CY) in combination with either ablative doses of total body irradiation (TBI) or the oral alkylating agent busulfan (Bu) is the most common conditioning regimen for allo-HSCT. However, TBI and CY can mobilize HMGB1 to the PB, and increased levels of HMGB1 correlate with increased PAI-1 after allo-HSCT, inducing transplantation-associated coagulopathy (TAC) conditions such as veno-occlusive disease (VOD) [154]. Recombinant human soluble thrombomodulin (rhTM) is used to treat disseminated intravascular coagulation (DIC) caused by aGVHD and significantly decreases HMGB1 [155]. Extracorporeal photopheresis (ECP) depends on infusion of UVA-irradiated and 8 methoxy-psoralen (PUVA)-treated leukocytes and is an effective treatment measure for GVHD. In vitro PUVA treatment induces the expression of HMGB1 in dying T cells, especially upon T cell activation, leading to their phagocytosis by macrophages and DCs [156]. In AML patients who received allo-HSCT, a higher γδ T cell count, which is an important early source of TNF-α and IFN-γ, predicted a better prognosis. A recent study found that PD-1⁺TIM-3⁺ Vδ2 T cells, PD-L1, and HMGB1 were significantly higher in AML patients than in healthy controls, suggesting that PD-1 alone is insufficient to indicate functional impairment, and Vδ2 T cells may require anti-TIM-3 inhibition for functional revival [157].

Acquired chemoresistance is a major obstacle in the clinical treatment of hematological malignancies. Many studies have demonstrated that chemotherapy agents including docetaxel, doxorubicin (DNR), cisplatin, etoposide, and methotrexate induce HMGB1 upregulation and promote cytosolic HMGB1 translocation [158‐160]. Moreover, DNR, vincristine (VCR), etoposide (VP-16), cytosine arabinoside (Ara-C), adriamycin (ADM), and ATO can increase HMGB1 expression and promote chemoresistance in hematological malignancies [22, 44]. Downregulating HMGB1 inhibits autophagy and enhances bortezomib activity in MM [30]. HMGB1 is becoming a recognized therapeutic target for chemotherapy resistance (Fig. 4) [161].

To date, several strategies have been proposed to directly or indirectly inhibit HMGB1 expression, release, and activity to treat hematopoietic malignancies (Table 2).