Acute Myeloid Leukemia (AML) is a type of blood cancer that affects the cells of the myeloid lineage (1). It is the most common type of blood cancer with poor survival due to the aggressiveness of the disease and high relapse rates (2, 3). To date, limited therapeutic options are available to treat AML (1, 3). Chemotherapy is the primary treatment given to AML patients. Depending on the risk of the disease, some patients may undergo Hematopoietic Stem Cell Transplantation (HSCT) (2, 3).

In the past few years, novel treatment options have been explored to improve the survival of patients suffering from AML. Targeted gene therapies, cytokine therapy, immunotherapy, etc., are some of the upcoming novel therapeutic options. Of these, immunotherapy is a promising strategy that explores the use of the patient’s immune system to kill leukemic cells (2–4).

Natural Killer cells play an essential role in innate immunity, and their anti-leukemic properties have been explored in the past few years (5). This anti-leukemic activity is driven by receptors on their surface, which transduces either activating or inhibitory signals. Most of these receptors belong to Killer Immunoglobulin-like receptors (KIRs), natural cytotoxicity receptors (NCRs) or NKG2 (Natural Killer Group 2) receptors. A balance between the activating and inhibitory receptors drives NK cell cytotoxicity. KIRs, NCRs and NKG2 can either activate NK cells by signaling through their immune receptor tyrosine-based activating motifs (ITAMs) or inhibit NK cells by signaling through their immune receptor tyrosine-based inhibitory motifs (ITIMs). These receptors need to recognize their cognate ligands to achieve their cytotoxic potential.

The KIR genes are present on chromosome 19q13.4. 16 KIR genes encode for the KIR receptors, namely KIR 2DL1, 2DL2, 2DL3, 2DL4, 2DL5, 2DS1, 2DS2, 2DS3, 2DS4, 2DS5, 3DL1, 3DL2, 3DL3, 3DS1 and two pseudogenes 2DP1, 3DP1 (5, 6). Inhibitory KIRs have a long cytoplasmic tail and hence are denoted as ‘L,’ whereas activating receptors have a short cytoplasmic tail and are denoted as ‘S.’Two KIR haplotypes, A and B, are determined by an individual’s KIR gene content. The A haplotype is inhibitory since it has more inhibitory KIR genes (KIR3DL1, KIR3DL2, KIR3DL3, KIR2DL1, KIR2DL3, KIR2DL4) and one activating KIR gene (KIR2DS4). On the other hand, the B haplotype is said to be an activating haplotype since it has more activating KIR genes (KIR2DS1, KIR2DS2, KIR2DS3, KIR2DS5, KIR3DS1) in addition to the inhibitory KIR genes (KIR3DL2, KIR3DL3, KIR2DL1, KIR2DL2, KIR2DL4, KIR2DL5A, KIR2DL5B) (7). The B haplotype is associated with lower relapse rates as it carries more activating KIR genes (8). KIR gene can further be divided into centromeric (Cen) and telomeric (Tel) regions to calculate the KIR B content score which can range from 0-2 B-content motifs for each of the centromeric and telomeric regions (i.e. AA, AB, or BB) (9). The haplogroup, which does not carry any of the B-content motifs is designated as CenAA/Tel-AA, whereas haplogroups that have scores ranging from 1 to 4, are designated as combinations of the following haplotypes CenAB/CenBB/TelAB/TelBB. The ligands for KIRs are the HLA Class I molecules (Bw4, C1, and C2). The inhibitory KIR2DL1 and activating KIR2DS1 recognize the C2 group alleles (possessing Lysine at position 80), the inhibitory KIR2DL2/3 and activating KIR2DS2/3 identify the C1 group alleles (having Asparagine at position 80), KIR3DL1 recognizes the HLA B alleles (with exceptions to B*13:01 and B*13:02) in the Bw4 group and KIR3DL2 identifies the HLA A*03/A*11 alleles (10). In the event of an inhibitory KIR receptor-ligand match, the NK cell activity is inhibited, leading to self-tolerance, whereas, in the case of an activating KIR receptor-ligand match, the NK cell activity is enhanced, leading to an increase in graft versus leukemia effect in the event of a transplant. Downregulation of the HLA ligands is an immune escape mechanism often used by tumor cells (11).

Among the NCRs, the activating NCRs are NKp30, NKp44, and NKp46, the ligands for which include various viral hemagglutinins (12–14), human CMV pp65 protein (15), and ligands such as B7-H6 and BAG6 which are expressed on the tumor cell surface. The NKP46 gene is very close to the Killer immunoglobulin-like receptor family on the human chromosome 19 (19q13.42) (16, 17). NKp44 and NKp30, on the other hand, are located in the MHC Class III region of Chromosome 6 (18, 19). NCRs such as NKp30 and NKp46, which were initially thought to be expressed only on resting NK cells (19– 20), has more recently been reported to be present on some T cells (Vd1 þ T cells for NKp30 and αβ T cells and δγ T cells for NKp44) and NK like cells (21, 22) where they perform similar functions as in NK cells. NKp44, on the other hand, is expressed only in activated NK cells. NCRs bind to their corresponding ligands, increasing cytotoxic activity and cytokine release by NK cells (23).

There are 7 members in the NKG2 family namely NKG2A, NKG2B, NKG2C, NKG2D, NKG2E, NKG2F and NKG2H. These transmembrane glycoproteins are encoded by genes in the NK complex present on chromosome 12. Of these 7 members, NKG2A, NKG2B (mRNA splice variant of NKG2A), NKG2C, NKG2E and NKG2H (mRNA splice variant of NKG2E) form heterodimers with CD94. NKG2F is expressed intracellularly (24).

NKG2A is an inhibitory receptor (25) CD56^bright NK cell subsets show a higher expression of NKG2A than CD56^dim NK cells (26). CD94/NKG2A recognizes its ligand, HLA-E, a non-classical HLA molecule and transmits inhibitory signals via ITIMs present in the cytoplasmic tails thus inhibiting NK cell activity (24–26). NKG2A is overexpressed on the NK cells of AML patients which is associated with poor remission (26, 27). NKG2C is an activating receptor that is expressed on NK cells during the later stages of maturation (25). It binds to its ligand HLA-E with lower affinity than NKG2A. NKG2C interacts noncovalently to DAP12 containing ITAMs. NKG2D is an activating receptor expressed as a homodimer on NK cells. It binds to its ligand MICA/B and ULBP (also known as NKG2D-L) (28). NKG2D interacts with 2 dimers of DAP10 which recruits phosphatidylinositol-3 kinase (PI3 kinase) thus activating further signaling and NKG2D associated cytotoxic response. Tumor cells tend to shed NKG2D thus facilitating tumor escape (24, 25). NKG2D is underexpressed on NK cells of AML patients (27). NKG2D releases signals to activate the immune system by binding to the NKG2D-L present on AML cells and lyses the leukemic cells. However, in an event of downregulation of NKG2D-L tumor cells evade surveillance facilitating tumor progression (28). This NKG2D-NKG2DL axis is involved in clearing tumor cells in the early phases of cancer development. There are various mechanisms by which tumor cells escape NK-mediated cytotoxicity ( Figure 1 ). These could be due to numerical defects in functional NK cells, defective expression of KIR and NCR receptors on NK cells, and defective maturation of NK cells, among others. This tumor evasion, if left unchecked, leads to the progression of the disease.

One factor that correlates with AML disease progression is the anti-leukemic activity exhibited by the NK cells, which also depends on the number of functionally active NK cells present. The number of NK cells is lowest during active disease, i.e., at the time of diagnosis and in the event of disease relapse. In contrast, it increases when the patient is in remission (29, 30). Reconstitution of higher NK cells post HSCT has been associated with better 2 year OS, lower rates of CMV reactivation (31), lower 2 year relapse risk (32), lower non relapse mortality, higher progression free survival (33). Conversely, low NK cell counts post 60 days of HSCT was associated with higher relapse risk (34).

There are two distinct Natural killer cell groups based on CD56 cell surface expression: 90% belong to the CD56^dim CD16^high group and are cytotoxic, and 10% belong to the CD56^bright CD16^dim/neg group and exhibit a more immunomodulatory role by releasing different cytokines (44). The maturation process of NK cells starts from CD34⁺ hematopoietic progenitor cells in the bone marrow, wherein cytokines like IL-15 and c-kit ligand, and flt-3 ligand help them reach their full potential from an NK cell progenitor (CD34^dim/CD117^-/NKG2A^-) to immature CD56⁺NK cells (CD34^-/CD117⁺/CD56^+/-) precursor intermediate and finally a functional CD56^bright NK cell (CD34^-/CD117^+/-/CD94⁺). They attain full functional maturity by expressing the NKG2A or KIRs (CD34^-/CD117^-/CD94⁺/CD16⁺/KIR⁺) (45). Another marker of importance in the identification of mature NK cells is CD57. Cells expressing CD57 have high cytotoxic potential with low proliferation capacity. NK cells are the first immune cells to reconstitute post-transplant and are the primary cells to keep leukemia in check immediately post-HSCT. The normal maturation process is essential in conferring NK cells the ability to recognize and eliminate cancer cells. Mundy-Bosse et al. (46) demonstrated that the immature NK phenotype is upregulated in murine AML models, whereas the intermediate phenotype is almost absent. Chretien et al. (47) in their study divided the AML patients into three groups based on their NK maturation profile: hypo-maturation group with CD56^bright/dim KIRs^- CD57^-; intermediate group with CD56^dim KIRs^-/+ CD57^-/+ and hyper-maturation group with CD56^dim KIRs⁺ CD57⁺ profile. Their study reported that the patients with the hypo-maturation profile had inferior outcomes, such as three-year OS and RFS. This study highlighted that the maturation status of NK cells plays a vital role in predicting long-term outcomes. Post-transplant infections such as cytomegalovirus (CMV) have also been implicated in influencing the reconstitution of mature NK cells post HSCT resulting in lower rate of relapse. Cichocki et al. reported that mature NK cells (CD56^dimCD57⁺NKG2C⁺) predominantly expand in post HSCT recipients after CMV reactivation leading to better DFS (48). Armin et al. (49) showed that recipients with CMV reactivation had a higher frequency of KIR⁺NKG2A^- NK cells than in CMV seropositive recipients without reactivation and seronegative recipients. Hassan et al. (50) observed a rapid and sustained increase in NK cell numbers in patients with CMV reactivation, with an increase in the proportion of NKG2C expressing NK cells. Mundy-Bosse et al. (46) further showed that AML patients with defective NK cell maturation carrying lower levels of the transcription factors required for terminal NK differentiation (TBET and EOMES) was associated with elevated levels of miR-29b. These are important in controlling final NK cell differentiation. Moreover, lower T-bet expression leads to lower perforin in mature NK cells, which results in NK cells with reduced cytotoxic potential.

Immune checkpoint molecules are essential in maintaining the immune response by protecting self-cells from lysis. Immune checkpoints like PD-1, TIM-3, and TIGIT regulate NK cell activity (51). PD-1 is expressed in mature NK cells when stimulated by infected cells or tumor cells. NK cells expressing PD-1 have reduced proliferative potential and lower cytotoxic and cytokine-producing capabilities. Blocking the interaction of PD-1 with its ligand PDL-1 by using specific antibodies activates NK cytolytic activity. PDL-1 expression in AML patients has been reported with lower immune recognition (52). Goltz et al. (53) showed that reduced PDL-1 expression in AML cells correlated with better outcomes, such as a lower risk of relapse and prolonged OS. Another checkpoint TIGIT is expressed on NK cells. The ligand for TIGIT is CD112, CD113 and CD155 same as that for DNAM-1. Sanchez-Correa et al. (54) showed low expression of DNAM-1 in AML patients, which could favor the binding of TIGIT to its ligands, sending inhibitory signals and leading to tumor evasion. Blocking TIGIT promotes NK cell-mediated cytotoxicity, as shown in cancer mouse models by Zhang et al. (55).

The above studies highlight the role of different NK cell defects in AML disease susceptibility and progression ( Figure 1 ). Many of these studies formed the basis on which various targets for cell-based therapies were devised for AML.

The tumor microenvironment (TME) is diverse, made up of stromal cells, immune cells, extracellular matrix and secreted factors (56). Interaction of AML blasts with the TME can be responsible for AML disease development, relapse, progression and resistance to therapy. Mesenchymal stem cells (MSCs) modulate the development and differentiation of hematopoietic stem cells, in the case of AML, the bone marrow MSCs can contribute to tumor progression by providing anti-apoptotic signals to AML blasts (57, 58). Regulatory T cell (Tregs) subsets are responsible for maintaining the peripheral homeostasis and central tolerance by suppressing T- helper (T_h) cell proliferation. A few studies have reported increased numbers of Tregs population in AML patients (59, 60), and is associated with unfavorable treatment outcomes (61). Depletion of these tumor associated Tregs results in better cytotoxic T (T_c) cell therapy (62). At the time of diagnosis NK cells in the AML patient are defective due to factors such as downregulation of activating receptors, upregulation of inhibitory receptors and lower cytotoxic and higher immature NK subsets (40, 47, 63). Paczulla et al. (64) showed that downregulation of the ligand for the NKG2D receptor (NKG2DL) via the PARP1 enables tumor evasion in AML blasts (64). Moreover, susceptibility of AML cells to NK cells is affected by the direct interactions between AML cells and mesenchymal stomal cells in the tumor microenvironment (65). Most common reasons for inhibition of NK activity in the TME are hypoxia, low glucose concentrations, cytokines, tumor cell derives factors such as IL6, IL10, TGF-Beta, prostaglandin E₂ (PGE₂). Myeloid derived suppressor cells (MDSCs) have an immunosuppressive role and expansion of MDSCs in AML patients leads to an immunosuppressive TME with a possible role in tumor progression (66–68). The expansion of MDSCs can take place through different pathways such as MUC1 oncogene expression through c-myc expression. Another group of cells in the TME is the Tumor associated macrophages (TAMs) which promote tumor progression. Al-Mataray et al. (2016) showed that AML promotes infiltration of TAMs (69). Further, AML blasts also secrete various immunosuppressive cytokine and chemokines such as IL-10, TGF B, IL-35 into the tumor microenvironment giving it an immune-inhibiting property (70–72).

Epigenetic modifications such as DNA methylation, histone modifications, Chromatin remodeling, control normal gene expression. Such epigenetic changes can also influence NK cell development, differentiation and functionality. DNA methylation involves transfer of a methyl group to 5’ position of cytosine molecule in CpG sites leading to normal gene expression. DNA methyltransferases (DNMTs) control these DNA methylation patterns. Nucleoside analogues of cytosine are integrated into the DNA and help in covalent bond formation with the DNMT, leading to DNA degradation and inhibition of DNA methylation (73, 74). Hypermethylation of the CpG islands can result in gene silencing. In case of tumor suppressor genes, hypermethylation promotes tumorigenesis (75, 76). Hypomethylating agents (HMAs) such as DNMT inhibitors and 5-aza-2’deoxycytidine (decitabine) have been used for at least a decade to treat high risk AML patients (77). Post translational modifications (PTMs) at the histone N-terminal tails, such as acetylation, methylation, phosphorylation, deamination etc. weakens the packed chromatin structure favoring the binding of transcription factors and normal genetic regulation. However, enzymes such as histone deacetylases (HDACs), demethylases, dephosphorylase reverse this function by removing the acetyl group from the histone tails leading to more compact packing of the DNA around the histones resulting in poor or no binding of transcriptional factors and hence affecting the gene transcription (78, 79). Such modifications of these epigenetic processes favor malignancy and tumor progression.

Treatment for AML varies based on the stage of the disease and the type of mutation involved. Stem cell transplantation remains the most suitable treatment for AML patients to reduce the risk of relapse and mortality. However, the risk of relapse depends on several other factors, such as the blast clearance status in bone marrow, the presence of FLT3 mutations and cytogenetics, and molecular mutations, among others (80, 81).

Conventional treatment strategies for AML include chemotherapy, targeted therapy, and radiation treatment, with hematopoietic stem cell transplantation (HSCT) as the final treatment option. The most common chemotherapy regimens include standard protocol of 7 days of cytarabine and three days of anthracycline (82), fludarabine–Ara-C–granulocyte colony-stimulating factor–idarubicin (FLAG-IDA) for relapsed childhood AML (83), or similar kind of induction, followed by consolidation chemotherapy and hematopoietic stem cell transplantation in patients who are at a high risk of relapse. Although, these methods have resulted in complete remission in 60-80% of elderly patients the downside to this approach is induction-related mortality due to poor tolerance to these drugs (82, 84, 85).

HiDAC consolidation doses on days 1, 3, and 5 for younger patients not undergoing HSCT is a standard norm (86–88). Consolidation along with an added targeted drug results in more prolonged relapse-free survival (RFS) (89). Most AML patients achieve complete remission (CR) with intensive chemotherapy alone. However, many of those patients would relapse. In such a situation, hematopoietic stem cell transplantation (HSCT) becomes the treatment of choice.

In AML, the leukemic cells interact with the bone marrow microenvironment and secrete a cytokine called KIT ligand, which favors tumor progression by interfering with normal hematopoeisis (90–92). There have been studies to investigate the role of CXCR4/CXCL12 inhibitors, TGF-Beta neutralizing antibodies and use of monoclonal antibodies to block IL-6 (93–96). Hypoxia is another feature of the TME which favors the growth and proliferation of AML cells. Hypoxia downregulates the NK cell ligands like NKp30, NKp44, NKp46, NKG2D, perforin and granzymes. The NKG2D can be restored by administration of IL-2 (97). Hypoxia-activated prodrugs such as Evofosfamide and PR-104 have been shown to inhibit hypoxia associated resistance to therapy in AML (98–101). Glucose is a pre-requisite for generating ATP and NADPH in order for NK cells to function correctly, however, the TME is glucose deprived which impairs NK cell glycolytic activity. Some metabolic regulators like GLUTS have also been shown to regulate glucose influx (102). Recombinant cytokines such as Il2, Il-12, Il-18, IL-15, IL-21, IFN-gamma, GM-CSF have been shown to promote, mediate and enhance NK cell expansion and cytolytic potential (103–106). Although this therapeutic option of targeting different metabolites in the TME is fascinating, in AML the microenvironment is so complex that targeting a single metabolite to get the desired outcomes is a challenge.

Epigenetics is regulated by molecules such as DNA and histone methyltransferases (HMAs) and histone deacetylases (HDACs) (107). DNMT inhibitors dectabine and 5-aza-2’deoxycytidine (DAC) are known HMAs that have been approved for clinical treatment of AML and MDS. DAC is phosphorylated to decitabine triphosphate which then incorporates into the DNA and inhibits DNA synthesis causing a cytotoxic effect, whereas, 5-AZA gets integrated into the RNA structure and interferes with protein formation (108–110). There have been a few clinical trials investigating HMAs for treatment of AML. Issa et al. (111) have reported good efficacy and tolerance with lower doses of these drugs. In the first trials doses of 1500 to 2500 mg/m² were used to stop DNA synthesis and cause cytotoxicity (112). Cashen et al. (113) in their Phase II clinical trial of decitabine treatment in elderly AML patients showed complete remission (CR) in 24% and overall response rate (ORR) in 25% with a median survival of 14 months. Lubbert et al. (114) treated their AML patients with decitabine + ATRA and observed ORR in 26%, with CR in 13.2% and PR in 12.8% with a median survival time of 5.5 months and 1 year OS of 28%. HDACs regulate cell apoptosis and proliferation by deacetylating lysine residues on proteins. HDAC inhibitors on the other hand inhibit this deacetylation process and help in controlling tumor progression. Some of the HDAC agents currently used are Valproic acid, Vorinostat and Entinostat among others. Vorinostat has a hydroxamic acid component which when binds to the zinc pocket of HDAC 1,2,3 and 6 results in reversible inhibition of HDACs and apoptosis of cancer cells (115).Valproic acid targets the AML1/ETO complex and induces apoptosis of AML cells (116). Entinostat helps in promoting loss of leukemia and longtime survival s proven in animal models (117, 118). Although by themselves HMA and HDAC inhibitors have shown moderate results combination therapies involving these molecules with other therapies may be promising.

Factors such as viral infections, oxidative damage, DNA damage increase the expression of metalloproteinases on tumor cells. Metalloproteinases have been suggested to be involved in the cleavage of NKG2D-L in AML (119–124). Raneros et al. (125) showed that azacitidine a DNA methyltransferase (DNMT) inhibitor reduces the release of soluble NKG2D-L in AML cells. Poly ADP ribose polymerase 1 (PARP1) inhibits the expression of NKG2D-L proteins in leukemic stem cells (LSCs) helping them evade surveillance (64). In case of DNA damage, the ATM/ATR pathway activates the PARP1 (126, 127), which in turn upregulates the expression of NKG2D-L on the surface of leukemic stem cells in AML (64). Nanbaksh et al. (128) reported that lower expression of c-Myc results in decreased expression of NKG2D-L in AML. Moreover, HDAC inhibitors and demethylating agents can also induce NKG2D-L in tumor cells making them susceptible to NK cells lysis (129–132). A few ongoing clinical trials are focused on targeting the NKG2D/NKG2DL axis in AML. CAR-NKG2D’s have been manufactured by introducing NKG2D into a CAR-T that will target the NKG2D-L in relapsed/refractory (r/r) AML patients (133–138). Baumeister SH et al. (134) confirmed the safety of such a CAR-T in a recent phase 1 clinical trial. These studies provide a new option for the treatment of AML through NK cells by targeting the NKG2D-NKG2D-L axis.

HSCT has been designated as the penultimate therapy for adult AML patients in CR1. It can reduce the risk of relapse by more than 60% compared to intensive chemotherapy alone by virtue of its potent graft-versus leukemia (GvL) effect (139). The advance of strategies, such as the use of post-transplant cyclophosphamide, has led to an increase in the number of suitable haplomatched donors for patients who do not find a fully matched HLA donor (140). However, even with changes in conditioning regimens (myeloablative to reduced intensity) and use of graft versus host disease (GvHD) prophylaxis such as cyclosporine, tacrolimus, mycophenolate mofetil, the success of HSCT is marred by post-transplant complications such as relapse, GvHD, and TRM. Hence, newer treatment options with fewer complications are sought. In recent years the role of KIR receptor-HLA ligand mismatch between the patient and donor favoring HSCT outcomes has gained much attention. Table 1 highlights some of the published data on patient-donor KIR genotype, KIR ligand match/mismatch and KIR receptor ligand match on transplant outcomes. Most of these studies show the role of KIR B genotype in reduced risk of leukemia relapse and improved disease-free survival, lower GvHD, lower risk of CMV reactivation (9, 141–143). On the other hand, Bultitude et al. (144) showed higher non relapse mortality (NRM) in AML patients undergoing T cell depleted HLA-matched, unrelated donor HSCT where the donor carried the centromeric KIR B haplotype. Bao et al. (145) reported a higher rate of relapse in patients transplanted with graft from KIR A/A donors, whereas Mansouri et al. (146) reported similar results when the patient had a KIR A/A haplotype. Some reports suggest no role of KIR mismatch on transplant outcomes (147–149), whereas higher TRM, reduced OS and DFS has been reported by Kroger et al. (150), DeSantis et al. (151), in the presence of KIR mismatch transplants. On the other hand, favorable outcomes such as lower aGvHD (152, 153), lower risk of relapse (154–157) and better OS and DFS (158) has been reported by other studies. These contradictory results across studies could be due to the difference in study designs such as heterogenous population, HLA match/mismatch, donor being related/unrelated to the patient and the conditioning regimen used. These studies suggested some insights into the role of KIR and NK cells as the basis for immunotherapies.

Recently the focus has shifted to more and more adoptive cell therapies for treating leukemias. Chimeric Antigen receptor (CAR)-T cell therapy was the first cell therapy to receive FDA approval after showing success in patients with aggressive B cell malignancies. The idea that T cells could be used to target specific antigens was brought forward by Eshar et al. in 1989 (169). The use of CAR-T by engineering the patient’s self-cells to target CD19 which is an important target in ALL has been evaluated in various studies (170–172). CAR T cells secrete anti-tumor cytokines, perforin, and granzymes into the tumor microenvironment (TME) (173, 174). Generation of CAR-T involves high costs and takes a long time from generation to implementation as the patient’s self-cells are engineered. Various studies have evaluated the role of CAR-Ts targeted towards different AML antigens.

The cell surface expression of proteins differs on cancer cells compared to healthy cells. Cancer cells may over-express or under-express a surface protein and cause aberrant expression of proteins (175). CD33, CD45, CD123, and CD244, are some of the surface proteins ubiquitously expressed on AML blast cells (176). These molecules are promising targets for immunotherapy, and currently, there are many ongoing trials focused on developing monoclonal antibodies (mAbs) against these molecules to achieve better outcomes in AML patients. mAbs can be either naked antibodies that function through NK mediated ADCC, such as lintuzumab (anti-CD33) or mAbs conjugated to toxins (e.g., gemtuzumab ozogamicin – anti-CD33), or conjugated to radioactive particles (176).

Natural Killer (NK) cell immunotherapy is a novel immunotherapeutic treatment option. In AML patients, the NK cell function is suppressed, thus enabling the cancer cells’ to evade immune surveillance. NK cell immunotherapy prevents the suppression of NK cells so that they can carry out target lysis. Many clinical trials for various NK cell-based immunotherapies like adoptive NK cell transfer, chimeric antigen receptor (CAR) NK cells, cytokine-induced memory-like (CIML) NK cells, etc., are currently being carried out (2, 4).

NK cells lyse their target firstly by forming an immunological synapse with them, followed by the release of cytolytic granules and cytokines. The second mode of lysis is antibody-mediated cell cytotoxicity (ADCC), wherein NK cells recognize antigen-coated target cells by their FCgammaIIIA (CD16) receptor and trigger ADCC and cytokine production (177, 178). In 2012, Romee et al. (179) demonstrated memory like function of NK cells by showing that if NK cells are preactivated with IL12, IL15, and IL18 followed by 1-3 weeks’ rest, they were able to generate enhanced IFN-gamma production when exposed to cytokines or K562 leukemia cell line. These hypotheses led to the investigation of different cell sources that could be used to generate NK cells in adoptive cellular therapies. Some of the cell sources are cord blood, peripheral blood mononuclear cells (PBMC), cell lines such as NK-92, hematopoietic stem and progenitor cells, and induced pluripotent stem cells (179–188). There are advantages and disadvantages associated with every NK cell source used. Generation of NK cells from donor PBMCs is the safest and most favorable option as they carry various NK cell activating markers such as CD16, NKG2D, NKp44, and NKp46 which strongly recognize and kill non-self or tumor cells; however, the percentage of NK cells in peripheral blood is too low, and expansion of NK cells derived from PBMCs is costly and time-consuming (189–191). NK cells expanded from NK-92 cell lines, which are an immortalized NK lymphoma cell line, raises safety concerns and irradiation of products derived from these cell lines requires irradiation before infusion. NK-92 cell lines also lack various activating KIRs and Cd16 (FCRIII), which may limit their killing potential even through ADCC. The advantages of using NK-92 cell line-derived NK cells are the speed at which they multiply without the need for feeder cells and their ability to secrete higher amounts of perforin granzymes and cytolytic cytokines (192–195). Induced pluripotent stem cells are a good source for NK cells as they have a higher expansion rate; however, these, too, express lower CD16 levels, which could reduce their cytotoxic capacity, although this could be rectified by genetic engineering (196). Not only various sources of NK cells but also different NK cell-based immunotherapies such as adoptive transfer, CAR-NKs, bi-specific and tri-specific killer engagers (BiKEs and TriKEs) are recently being investigated in clinical trials to understand their role as a curative option for AML ( Figure 2 ). Table 2 lists the various completed clinical trials of NK cells in AML.