Antibody is a kind of protective protein secreted by plasma cells under the stimulation of antigen. The function of antibodies in acute myeloid leukemia depends on the following aspects: First, after the antibody binds to the surface antigen of tumor cells, it uses its Fc domain to recruit immune effector molecules, thereby triggering natural killer (NK) cell-mediated antibody-dependent cytotoxicity (ADCC), antibody-dependent phagocytosis or complement-dependent cytotoxicity; secondly, through receptor-mediated endocytosis of antibody-drug conjugates (ADCs) and radionuclide conjugates, it can effectively deliver toxic carriers to kill leukemia cells. The third is to use bispecific or multispecific antibodies to bind these cells to target leukemia cells, thereby enhancing the anti-leukemia effect of T cells or NK cells.

Since most leukemia cells are associated with the positive expression of specific antigens (such as CD123, CD33, CD96 and CLL-1) (72), targeted immunotherapy, including monoclonal antibodies (73) and bispecific antibodies, is currently in the clinical evaluation stage. These therapies aim to target multiple antigens expressed on AML blast cells, of which CD33 and CD123 are the most common targeted antigens (74). The bispecific T cell conjugative molecule (BiTE) is composed of a single variable fragment (Fv). Bispecific T cell conjugator (BiTE) targeting CD33 can activate and amplify T cells in autologous clinical samples of patients with AML, and mediate the lysis of primary AML cells and normal myeloid cells in vitro in a dose-dependent manner (75–77). The use of anti-CD19 BiTE triggered CD33-independent activation, resulting in CD33 expression on a small number of T cells. This phenomenon is related to the self-killing of T cells, but has little effect on the function of cytotoxic T cells associated with AML (78). The single-chain Fv trisomics (sctb) based on CD16 single-chain Fv fragment (scFv) targeting CD123 exhibits significant anti-leukemia activity, but its effect is not as good as the dual-targeting sctb targeting both CD33 and CD123 (79). 7370 anti-FLT3 bispecific IgG can activate T cells in patients with acute myeloid leukemia, thereby inducing cytotoxicity to autologous blasts. The results of in vivo experiments showed that the bispecific antibody could also effectively guide human T cells to target AML cell lines and produce killing effects in vivo. This effect depends on the expression of FLT3 antigen on the surface of AML cells, and is not affected by FLT3 mutation status. Both double-targeted single-chain Fv trisomy (sctb) and single-targeted sctb use isolated monocytes (MNC) as effector cells, which can effectively induce antibody-dependent cytotoxicity (ADCC) of two different acute myeloid leukemia (AML) -derived CD33 and CD123 double-positive cell lines at low concentrations (80). In addition, a bispecific single-chain Fv (bsscFv) was formed by binding another scFv that is specific for low-affinity Fcγ receptor III (CD16). This bsscFv can effectively mediate the cleavage of AML-derived leukemia cell lines in a specific concentration of ADCC reaction (81). Bispecific antibodies produced by chemical coupling of anti-CD3 and anti-CD13 Fab ‘ fragments enhance the cytotoxicity of peripheral blood mononuclear cells (PBMC) to CD13-positive AML cells stimulated by IL2 or IL7 (82). At the same time, the clinical potential of monoclonal antibodies (mAbs) against CD123 in the treatment of AML has been supported by relevant evidence (83). BI836858 is an antibody against CD33 with an Fc domain. Decitabine can increase the expression of NKG2D ligands on the surface of leukemia cells, thereby enhancing BI836858-mediated antibody-dependent cell-mediated cytotoxicity (ADCC) (84). Evorpacept (ALX148) is mainly composed of a modified SIRPαD1 domain targeting CD47 and can bind to the inactivated human IgG1 fragment (Fc) (85). At present, the molecule is being studied in an ongoing phase I/II clinical trial (NCT04755244) to evaluate its efficacy in combination with the BCL2 inhibitor venetox and AZA, especially for patients with relapsed/refractory acute myeloid leukemia (R/R AML) who have not been treated or are not suitable for standard induction chemotherapy (86). At the same time, in vivo experimental results showed that under the action of activated human T cells, AMG 330 targeting CD33 could inhibit the growth of xenografts of humanized mouse subcutaneous AML cell lines, thereby significantly improving the survival rate (76, 77). In addition, treatment with anti-CD28 activation antibody not only enhanced the killing effect of AMG 330 on human AML cell lines, but also increased the cytotoxicity of primary AML samples from patients with refractory leukemia (87). Lintuzumab (SGN-33) is a monoclonal antibody (mAb) against CD33. Studies have shown that lintuzumab may have an anti-leukemia effect, which can induce remission after cytoreductive surgery with low-dose cytarabine in untreated AML patients (88). Studies on AML animal models have shown that MGD024 combined with cytarabine or venetox can almost completely eliminate tumor cells. Therefore, a drug dose escalation study was proposed to evaluate the safety of MGD024 in refractory/recurrent hematological malignancies (including AML and BPDCN).

A few years ago, researchers developed an antibody-drug conjugate (ADC) treatment strategy, which showed good therapeutic effect and ideal side effects (89). ADC has gradually become an emerging chemotherapeutic drug for the treatment of cancers including AML (90). IMGN632 is an antibody-drug conjugate (ADC) targeting CD123, which uses a new IGN payload (91). In addition, compared with X-ADC, IMGN632 showed cytotoxic effects on AML samples at doses that had no adverse effects on normal myeloid progenitor cells (92). Clinical studies have shown that IMGN632 and venetoclax (a BCL-2 inhibitor for patients with acute myeloid leukemia) combined with azacitidine showed a consistent anti-leukemia synergistic effect, and the synergistic anti-leukemia effect between IMGN632 and venetoclax was verified in leukemia cell lines and xenotransplantation (PDX) derived from patients with acute myeloid leukemia. Mylotarg is a kind of ADC, which is used to treat newly diagnosed or refractory CD33 positive AML. Compared with Mylotarg, ADC targeting AML-specific antigens is expected to provide a more effective treatment for AML, and has a wider range of indications (93). SGN-CD123A is an antibody-drug conjugate (ADC) based on humanized CD123 antibody, which can effectively induce the apoptosis of leukemia cells expressing CD123. The drug shows a significant effect on promoting the apoptosis of leukemia cells in myeloid leukemia cell lines, primary AML blasts and patient-derived xenograft models (94). Antibody therapy has made significant progress in the clinical research of acute myeloid leukemia, but it still needs to be continuously optimized and improved to provide more effective and safe treatment for patients.

Immune checkpoint inhibitors (ICIs) relieve the inhibition of tumor cells on the immune system by interfering with the signal transmission mechanism between tumor cells and immune cells, thereby enhancing the immune system ‘s ability to attack tumor cells. Specifically, immune checkpoint inhibitors can effectively kill tumor cells by blocking the inhibitory molecules on the surface of T cells and transforming T cells from an ‘ exhausted ‘ state to an ‘ activated ‘ state, which has been proved to be a promising therapeutic option (95). But ICIs are not very promising drugs in AML treatment, and can be consider only as an addition to the treatment, do not work in monotherapy. In the current study, PD-1 and CTLA-4 are the two most active checkpoint receptors in the current study. They play a key role in different stages of anti-tumor immune response. At the same time, inhibition of CTLA-4 and PD-1 is the most widely used two immune checkpoint blocking strategies in clinical practice.

PD-1 inhibitors commonly used in clinical research for the treatment of AML include drugs such as ipidilizumab, nivolumab, pembrolizumab, durvalumab and atezolizumab (96). The results showed that in patients with AML, the complete remission rate (CR) of anti-PD-1 antibody nivolumab combined with azacitidine was 18%, and the hematological improvement rate was 15% (97). Studies have shown that the combination of azacitidine, nivolumab and ipilimumab (an anti-CTLA-4 antibody) has achieved complete remission (CR) or complete remission with incomplete hematological recovery (CRi) in 3% of patients with AML (98). The application of nivolumab combined with conventional induction chemotherapy (such as idarubicin plus cytarabine) in newly diagnosed AML patients is feasible (99). The efficacy of Pembrolizumab combined with decitabine or azacitidine in patients with R/R AML was similar to that of azacitidine combined with nivolumab. Tiragolumab is an anti-TIGIT antibody that can improve the prognosis of lung cancer patients when combined with atezolizumab. Anti-TIGIT antibody reshapes the tumor microenvironment by enhancing the blocking effect of PD-L1 on bone marrow cells and Treg cells, thereby improving the prognosis of tumor patients. The study found that tumor patients with higher baseline levels of macrophages and regulatory T cells in tumors had a better prognosis when treated with atezolizumab combined with tiragolumab, while atezolizumab alone did not have this effect. This indicates that TIGIT checkpoint inhibitors can play a role by reshaping the immunosuppressive tumor microenvironment (100). And the study reported the safety of azacitidine combined with duralizumab in patients with MDS and AML.

CTLA-4 (CD152) transmits immunosuppressive signals to terminate the immune response by interacting with CD80 and CD86 ligands. In the study of the treatment of melanoma patients, we found that anti-CTLA-4 antibody ipilimumab can effectively increase the proportion of Teff/Tregs, enhance the activity of NK cells, and restore the function of T effector cells, thereby significantly prolonging the survival of patients (101). Anti-CTLA-4 can enhance the function of AML-specific T cells, which is manifested in increasing its frequency, cytotoxicity and IFN-γ secretion (102). When ipilimumab monotherapy was used in patients undergoing hematopoietic stem cell transplantation (HCT) for recurrent hematological malignancies, a complete remission (CR) rate of 23% and a partial remission (PR) rate of 9% were observed (103). A phase I clinical trial for recurrent or refractory AML and myelodysplastic syndrome (MDS) is underway to evaluate the efficacy of ipilimumab in combination with decitabine (DAC) in patients who have received or have not received allogeneic hematopoietic stem cell transplantation (HCT). At present, the trial has not yet begun to recruit subjects (clinical trial registration number: NCT2890329).

The concept of chimeric antigen receptor (CAR) T cell therapy was first proposed in 1993 (104). This therapy has achieved remarkable results in the treatment of lymphatic and hematological malignancies by targeting unique targets such as CD19, CD22 and BCMA (105). CAR-T cells are genetically engineered from autologous peripheral blood T cells and allogeneic CAR-T cells. These cells have specific extracellular antigen recognition domains, which are usually composed of single-chain variable fragments of monoclonal antibodies and are connected to the intracellular signal transduction domain (106). The advantages of chimeric antigen receptor (CAR) T cell therapy are reflected in the following aspects: (a) CAR-T cells are activated only when recognizing specific targets (107); (b) CAR-modified cells can be effectively eliminated after disease eradication by using non-persistent cell types or safe switching mechanisms. The idea of safety switch is to trigger it when a high grade of CRS occurs to save the patient’s life; if the patient is safe without a serious toxicity, keeping CAR T cells in the body provides the treatment for many more years after disease eradication; the CAR T cells not active sitting in the body, ready to attack the cancer cells again when needed. “The safety switch” works like “an insurance policy” for the patient, in case of the relapse of the disease, high T cells activation followed by a high grade of CRS. (c) Genome editing of hematopoietic stem cells (HSCs) to generate hematopoietic systems that are resistant to CD33 (destruction of normal bone marrow) (108).

At present, a variety of strategies have been developed to improve the safety and reliability of CAR-T cell therapy in AML treatment ( Figure 4 ) ( Table 1 ). These strategies include using inducible caspase-9 to control the death of CAR-T cells and adjusting the affinity of CAR-T cells to only target cells with high expression levels (109). A large number of studies have shown that anti-CD123 CAR-T cells exhibit significant anti-leukemia activity before clinical practice. However, some studies have also raised concerns about the toxicity of hematopoietic stem cells (HSCs) (110). A variety of anti-CD33 CAR-T cell constructs have shown significant preclinical therapeutic effects on primary acute myeloid leukemia (AML) cells in vitro and in humanized animal models, and their toxic effects on leukemia cells and non-leukemia myeloid cells have also been observed (111). At present, in a clinical trial report of anti-CD33 CAR-T cell therapy, the treatment of a patient with refractory acute myeloid leukemia (AML) was recorded in detail. After receiving autologous anti-CD33 CAR-T cell infusion, the patient not only developed cytokine release syndrome (CRS), but also observed a temporary decrease in the original cells in the bone marrow (BM) (112). Studies have shown that human peripheral blood T lymphocytes transduced with CD19 CAR can completely eliminate lymphoma and leukemia in immunodeficient mice (113). CAR-T cells targeting CD19 in combination with T cell inhibitors can destroy pathological B cells and regulate T cell response, thereby inducing remission of refractory antisynthase syndrome (114). In a phase I clinical trial (ClinicalTrials.gov identifier:NCT02842138) for patients with B-cell lymphoma,11 patients received CD19-BBz (86) CAR-T cell therapy. Each patient received a dose of 2×10⁸ to 4×10⁸ cells of CD19-BBz (86) CAR-T cells, of which 6 patients achieved complete remission (115). A study showed that 23 of 27 adult and pediatric ALL patients (including 11 patients with extramedullary diseases) achieved complete remission after the first CD19 CAR-T treatment (116). CAR-T cell therapy has been reported to show great potential in improving the treatment of relapsed or refractory B-cell malignancies (117–119). According to relevant reports, a bispecific split CAR (BissCAR) T cell targeting CD13 and TIM-3 has a significant effect in clearing patient-derived acute myeloid leukemia (AML). At the same time, in mouse and patient-derived xenograft models, this treatment method is less toxic to normal hematopoietic stem cells (HSC), myeloid cells, and healthy organ systems (120). CD123 CAR-T cells have shown significant anti-leukemia effects in vitro on leukemia cell lines and primary patient leukemia cells, as well as in vivo using leukemia mouse models. However, it is unclear whether CD123 CAR-T cells affect the normal hematopoietic function of bone marrow (121, 122). A study selected ADGRE2 and CLEC12A as the target antigens by analyzing samples from patients with resistant AML and doing a series of tests in mouse models beginning in 2018 (123). They showed that CAR T cells designed for these antigens were safe and effective against human AML cells implanted in mice. Based on these results, the clinical trial was approved to go forward.CRISPR-Cas9 technology has been proved to be feasible for gene ablation of CD33 antigen in human HSPCs, and this technology shows multi-lineage hematopoietic recovery ability in in vivo model system (124). Based on this result, we will initiate a clinical trial that will combine allogeneic HSCT using transgenic CD33-negative HSCs and CD33-targeted CAR-T cell therapy (125). In addition to CD33 and CD123, targets such as folate receptor β (126), FLT3 (127), NKG2D ligand (128) and CD70 (129) are also being tested for the development of CAR-T cells.

The latest research provides strong evidence for the effectiveness of the combination therapy of anti-FLT3 CAR-T cells and crenolanib. It has been found that pretreatment with crenolanib can significantly increase the expression level of FLT3 on the cell surface, thereby enhancing the ability of T cells to recognize AML cell lines. This effect has been observed in both in vitro experiments and animal models (130). The efficacy and safety of CAR-T therapy in AML are not convincing enough, which is mainly due to the lack of ideal therapeutic targets, a series of complex factors related to AML microenvironment, and CAR-T cell exhaustion (131). At present, the most promising candidate targets for CAR-T treatment of AML include CD33, CD123 and CLL1.A large number of literatures have reviewed these targets and other possible targets in detail (132). The resistance of AML to CAR-T therapy involves a variety of mechanisms, including antigen escape, inhibition of tumor immune microenvironment, CAR-T cell dysfunction, and tumor heterogeneity. In response to these challenges, CAR-T cells that simultaneously target multiple antigens can be developed to reduce the risk of antigen escape; the persistence and efficacy of CAR-T cells are enhanced by genetic modification or combined use of immune checkpoint inhibitors. Although CAR-T therapy has a significant effect, it may also cause side effects such as cytokine release syndrome (CRS) and immune effector cell-related neurotoxicity syndrome (ICANS).

The cancer vaccines currently studied mainly include Peptide vaccines and dendritic cell (DC) vaccines (133) ( Table 2 ). Peptide vaccines mainly target leukemia-associated antigens including nephroblastoma 1 (WT1), protease 3 (PR3), hyaluronic acid-mediated motor receptor (RHAMM), and mucin 1 (MUC1) (134). Mutations in the WT1 gene may cause abnormalities in cell growth and differentiation, which play an important role in the occurrence of leukemia (135). At the same time, it has been reported that the overexpression of WT1-specific T cells is significantly associated with a variety of hematological diseases (136). In AML patients, humoral immune response and cytotoxic response to WT1 protein have been observed (137). The above findings provide a theoretical basis for the application of WT1 protein in cancer vaccine therapy. The multivalent WT1 peptide vaccine galinpepimut-S (GPS) is considered to induce a specific immune response and is associated with the 5-year survival rate of AML patients participating in this clinical trial (138). CV-501 is a different HLA class II restricted peptide vaccine based on WT1, which has been studied in patients with acute myeloid leukemia (AML). However, the test results of this vaccine did not show any significant immune response (139). High-dose RHAMM-R3 peptide vaccine can effectively stimulate the body ‘s immune response and has a good effect in the treatment of hematological malignancies (140). Studies have explored the effects of combined use of PADRE, adjuvants, and WT1 or PR3 vaccines in AML patients. The results showed that the combination failed to produce clinical or immune responses (141).

Dendritic cells may be derived from different progenitor cells, but in the study of vaccine therapy, the most widely discussed sources are acute myeloid leukemia dendritic cells (AML-DC) (142) and monocyte-derived dendritic cells (mo-DC) (143). Most DC-based vaccine therapies are carried out in the state of minimal residual disease (MRD) to prevent the recurrence of leukemia. Leukemia-derived dendritic cell-based vaccines have shown good tolerance and a low probability of adverse reactions (144). Mo-DC can effectively load complete apoptotic leukemia cells, leukemia cell lysates or RNA/mRNA derived from leukemia cells, thereby further enhancing its therapeutic effect (145, 146). WT1 mRNA-electroporated DCs can improve the overall survival rate of AML patients with high risk of recurrence and further promote vaccine-induced WT1-specific CD8 T cell response (147). The efficacy of allogeneic dendritic cell vaccine DCP-001 in AML patients was studied. The results showed that DCP-001 had good safety and feasibility, and could induce cellular and humoral immune responses (148).

Telomerase is considered to be an important target in the study of cancer vaccine therapy. This is because the expression of telomerase is more in AML patients, and it plays a key role in maintaining the microenvironment of leukemia stem cells, especially in the context of high-risk cytogenetics. Studies have shown that telomerase (human telomerase reverse transcriptase, hTERT) and lysosomal-associated membrane protein (LAMP) can significantly enhance immunoreactivity by encoding mRNA, and this signal specifically targets lysosomes (149).

Exosome-based vaccines are a new approach to cancer treatment. The first study on tumor suppression using exosome-based vaccines was reported in 1998 (150). Some studies have evaluated the therapeutic effect of a novel interferon-modified exosome vaccine on prostate cancer (151). The inoculation of exosome vaccine not only reduces the expression level of vascular endothelial growth factor receptor 2, but also effectively reduces the ability of tumor metastasis (152). A study has shown that trastuzumab emtansine (T-DM1) can specifically bind to exosomes derived from HER2-positive cells, while some studies have explored the potential of exosomes as HER2-positive tumor vaccines by targeting and activating CD4 + and CD8 + T cells, and promoting long-term immune response through CTL memory cells (153, 154). Studies have found that exosomes derived from human adipose-derived mesenchymal stem cells (haMSC) can effectively induce apoptosis of ovarian cancer cells by blocking the cell cycle, up-regulating the molecular levels of BAX, CASP9 and CASP3, and down-regulating the anti-apoptotic protein BCL2 (155). A number of clinical studies have confirmed that vaccine therapy has shown significant efficacy in improving patient survival and reducing the risk of recurrence (156, 157). Leukemia-derived exosomes can induce the polarization of regulatory T cells and macrophages, and these exosomes promote the formation of tumor microenvironment in bone marrow, and activate HF-1α, AKT, VEGF, c-Myc, IL-8 and cyclin D1 signaling pathways by transporting tyrosine kinase receptor MET. AML-derived exosomes can transform the bone marrow microenvironment into conditions conducive to the development of leukemia by regulating multiple molecules (158). In addition, circulating exosomes produced by AML can also carry immunosuppressive substances, thereby inhibiting the body ‘s anti-tumor immune response (159). Exosomes can efficiently deliver drugs and antigens and accurately target malignant tumor cells in the blood system. At present, there are few reports on the application of MSC-derived exosome vaccines in acute myeloid leukemia.