Conventional therapies for treating acute myeloid leukemia (AML) involve induction chemotherapy, molecular-targeted therapies, hypomethylating agents, and allogeneic hematopoietic stem cell transplantation (allo-HSCT). Induction chemotherapy typically includes the administration of cytarabine and daunorubicin on a “7 + 3” treatment schedule, which can achieve cure rates of up to 70 % following consolidation therapy in patients with favorable genetic risk profiles [1]. However, some AML cells survive chemotherapy, often leading to disease relapse and thereby reducing survival rates [1]. More recently, molecular-targeted therapies (e.g. venetoclax, FLT3 inhibitors, and IDH inhibitors) and hypomethylating agents (e.g. azacitidine and decitabine) have been used in combination to increase treatment efficacy and target AML stem cells to induce more robust remission rates [2, 3, 4, 5]. Nevertheless, allo-HSCT remains the standard-of-care for high-risk AML by eliminating residual AML cells and reducing the likelihood of relapse through the graft-versus-leukemia (GvL) effect, mediated by donor-derived immune cells [1]. Unfortunately, not all patients can tolerate the allo-HSCT process [1], but its proven efficacy against AML underscores the potential of adoptive cellular therapy—a form of immunotherapy where donor- or patient-derived immune cells are infused into patients.

Human CD3+CD4−CD8− double-negative T cells (DNTs) are mature, unconventional T lymphocytes that comprise ∼1–5 % of peripheral T cells, lacking invariant natural killer-cell and mucosal-associated invariant T-cell markers [6,7]. DNTs have been discovered and explored in several different immunological contexts. Initially, DNTs were observed to induce immune tolerance and suppress T-cell effector function [8] in which ex vivo expanded DNTs could delay the onset of xenogeneic graft-versus-host disease (GvHD) [9]. Pathogenic DNTs have also been implicated in various autoimmune conditions, mediating chronic inflammatory environments and tissue damage [10,11]. However, the inflammatory and cytotoxic effects of DNTs may prove useful against malignancies. We were the first to demonstrate the potent anti-leukemic function of autologous DNTs, expanded from patients with AML [12]. Notably, not all autologous DNTs could expand ex vivo or effectively kill AML blasts, potentially due to T-cell exhaustion and prior treatment effects in patients [12]. DNT expansion from healthy donors provides an alternative and more uniform source for therapeutic DNTs.

Without any genetic modifications, we developed a protocol to expand DNTs ex vivo from healthy donors for up to 21 days using anti-CD3, IL-2, and PI3Kδ inhibitor idelalisib [6,13,14]. Expanded DNTs are heterogeneous populations of Vδ2+, Vδ1+, and αβTCR+ cells with predominantly unique αβTCR clonotypes that strongly interact with each other [15]. Vδ2– DNT subsets increase the cytotoxic capacity of the Vδ2+ DNT subset, exhibit cellular persistence upon chronic stimulation, suppress allogeneic T-cell function which may lower the risk of host rejection, and contribute to the overall anti-tumor efficacy of the DNT product [15].

DNTs kill solid and liquid tumor cells using various mechanisms depending on the type of cancer [6,7,16,17]. Against AML, DNTs can effectively target ∼75 % of primary AML blasts tested, including samples from patients with relapsed/refractory AML [6,13]. Upon AML interaction, DNTs secrete high levels of IFNγ and TNFα, which induce the upregulation of cytotoxic ligands on AML cells, including NKG2D/DNAM-1 ligands and intercellular adhesion molecule-1 (ICAM-1) [6,16] (Figure 1). The corresponding receptors on DNTs, NKG2D/DNAM-1 and lymphocyte function-associated antigen-1 (LFA-1), facilitate DNT cytolytic processes in a perforin/granzyme-dependent and TCR-independent manner, reducing the risk of GvHD [6,16,18,19] (Figure 1). The initial trigger of cytokine release by DNTs is partially explained by CD64 expression on AML cells (Figure 1). CD64 levels on AML cells positively correlate with the degree of DNT-mediated cytotoxicity, where its absence abrogates DNT killing and reduces TNFα secretion [16,20]. Interestingly, while DNTs engage DNT-susceptible AML cells, nearby DNT-resistant AML targets become sensitized by DNT-derived TNFα, making them more susceptible to DNT killing [16]. Given the heterogeneous nature of AML [21], this mechanism supports the favorable use of DNT therapy in patients with a mixture of DNT-susceptible and -resistant AML populations. Moreover, our recently published study demonstrates that SOCS1 in AML suppresses cytokine-mediated JAK1 activation, disrupting ICAM-1 upregulation and increasing AML resistance to DNT and conventional T-cell effector functions [22] (Figure 1). Pharmacological agents that induce NKG2D ligands [23], upregulate ICAM-1 [24], inhibit SOCS1 [25], or activate JAK1, such as IFNα or STING agonists [26, 27, 28], in AML may enhance the efficacy of DNTs, leading to better clinical outcomes.

Overall, ex vivo expanded DNTs meet the criteria for an off-the-shelf therapy: (1) they can be expanded to clinically relevant numbers (∼1.1x108 DNTs/mL of blood and ∼1550x-fold expansion by day 17 of culture) under good manufacturing practices (GMP), (2) they do not cause GvHD, (3) they target a broad spectrum of cancers in a donor-independent manner, (4) they are not rejected by the host’s immune response, and (5) they can be cryopreserved under GMP conditions without compromising function [13].

As a result, a phase I trial was conducted in patients with AML who relapsed after allo-HSCT to assess the feasibility, safety, and efficacy of allogeneic DNT therapy. None of the treated patients experienced adverse events greater than grade 2 related to allogeneic DNT treatment and no signs of GvHD or neurotoxicity were observed in any treated patients. All patients received three complete DNT infusions as planned. Among the ten patients treated, five achieved complete remission (two with incomplete hematologic recovery and one with minimal residual disease). DNTs remained detectable in some patients for at least 21 days after the last DNT infusion. Patients treated with DNT therapy had a prolonged one-year overall survival compared to those receiving conventional or hypomethylating agent-based salvage therapy (60 % vs 35.2 % and 35.2 %, respectively) [29]. Furthermore, four out of the five patients, who achieved complete remission, continue to remain in complete remission after 3 years at the last follow-up. Cases of treatment inefficacy may be due to a combination of AML-intrinsic evasion mechanisms against DNTs relating to SOCS1 [22] and CD64 [20] expression. Nevertheless, these promising results among a high-risk AML population support the need for a next-phase clinical trial with a larger patient cohort.

Conventional therapies have been successfully demonstrated to enhance DNT therapy and increase its therapeutic efficacy. Common first-line chemotherapeutic agents for AML, such as cytarabine and daunorubicin, can render AML cells more susceptible to DNTs, including those previously resistant to DNTs. These chemotherapeutic agents upregulate NKG2D/DNAM-1 ligands, allowing DNTs to better engage AML cells both in vitro and in vivo [30]. Recent molecularly targeted chemotherapeutics have also shown synergistic effects when combined with DNTs. Favorable clinical outcomes using a combination of the BCL-2 inhibitor venetoclax and the hypomethylating agent azacitidine in elderly patients with AML [31] may be partly attributed to T cell-mediated activities [32]. Lee et al. showed that venetoclax could enhance the potency of DNTs and conventional T cells against AML by inducing reactive oxygen species production, while azacitidine sensitized AML cells to DNT killing by evoking viral mimicry response through the STING pathway [32].

Beyond direct cytotoxic mechanisms, cellular persistence is crucial in maintaining robust anti-leukemic effects and avoiding disease relapse [33]. The addition of the PI3Kδ inhibitor idelalisib during ex vivo expansion drastically skewed DNTs toward a central memory phenotype by increasing levels of SELL (CD62L), IL7R, TCF7, and LEF1, while reducing levels of exhaustion markers such as LAG3, TIM3, and TOX. As a result, idelalisib-treated DNTs remained detectable in the peripheral blood of immune-deficient mice for at least 74 days post-infusion, compared to 34 days for untreated DNTs. AML-engrafted mice infused with idelalisib-treated DNTs experienced the eradication of AML cells from both lymphoid and non-lymphoid organs for at least 100 days by the conclusion of the experiment, while low levels of AML remained detectable in various organs of AML-engrafted mice infused with untreated DNTs [14]. Since PI3K inhibitors are currently being investigated as monotherapies for AML and solid tumors [34,35], their use with DNTs may directly target cancer cells and simultaneously increase DNT persistence, leading to synergistic effects to improve patient outcomes.

The unique properties of DNTs may also increase the efficacy of other cellular therapies. In the context of allo-HSCT, DNTs can mitigate lethal GvHD through CD18-dependent mechanisms while maintaining the GvL effects of allo-HSCT against AML in xenograft mouse models [36]. DNTs may possess professional antigen-presenting cell features similar to conventional γδ T cells [37], albeit with greater innate anti-leukemic effects [15], thereby potentiating the effector function of donor-derived conventional T cells. Combining allo-HSCT with DNT therapy may induce synergistic anti-leukemic effects and favorable outcomes, similar to those observed in a phase I trial that sequentially administered allo-HSCT followed by adoptive T-cell therapy [38], while reducing incidences of lethal GvHD and other unwanted immune-related reactions. Collectively, these results demonstrate the adaptable nature of DNTs to synergize with other anti-cancer therapies in treating patients with AML.

To bolster the effectiveness of adoptive T-cell therapy, patient-derived T cells can be genetically modified with a chimeric antigen receptor (CAR) to specifically target and activate autologous T cells against tumor-associated antigens. Autologous CAR-T cells have demonstrated impressive clinical efficacy in inducing prolonged remission in patients with relapsed/refractory B-cell malignancies using CARs directed at B-cell restricted antigens [39]. However, due to the lack of restricted antigens for AML, CAR-T cells targeting leukemia-associated antigens overexpressed on AML but also found on healthy tissue can lead to lethal on-target off-tumor toxicities [40, 41, 42, 43, 44]. Multiple strategies to reduce off-tumor toxicities are being studied, such as bispecific CAR-T cell designs that activate only upon the simultaneous recognition of multiple antigens on AML [45]. Despite this progress, the highly individualized nature of autologous CAR-T cell therapy inherently faces many challenges, such as high treatment costs, product variability, and long delivery times due to lengthy manufacturing processes, all of which limit patient accessibility and overall therapeutic effectiveness. The average estimated cost for autologous CAR-T cell administration exceeds $350,000, with a significant portion attributed to CAR-T cell manufacturing [46,47]. Thus, allogeneic off-the-shelf therapies such as DNTs have become an attractive alternative for CAR-based therapies.

Interestingly, three independent clinical trials using CD19-and BCMA-directed CAR-T cells have reported that DNTs become the dominant population in CAR-expressing cells in patients with durable remission, despite the low frequency of DNTs present in the infusion product (Table 1). In those trials, four pediatric patients with B-cell acute lymphoblastic leukemia and one adult patient with multiple myeloma were those who achieved the longest remission status in their respective ∼5-year follow-up studies, and one adult patient with chronic lymphocytic leukemia experienced at least 10 years in remission. Persistent CAR-DNTs rose up to 91.6 % within CAR-expressing T cells and remained detectable for 3–6 years [48∗∗, 49∗∗, 50∗∗]. These CAR-DNTs were identified as γδ T cells in one of the studies [48], while the pediatric study ruled out the CAR-DNTs as γδ T cells, natural killer cells, or immature thymocytes based on transcriptomic data [49]. Nevertheless, this suggests that an intrinsic therapeutic property unique to DNTs may be driving the robust clinical outcome, in which the initial use of CAR-DNTs without the phenotypic transition period could yield durable clinical remission.

Recently, we successfully transduced allogeneic DNTs with CD19-and CD4-directed CARs (CAR19 and CAR4). While maintaining their off-the-shelf therapy properties, CAR-DNTs exhibited improved and expanded cytotoxic capacity against B-cell leukemia and T-cell cancers [18,19]. CAR19-DNTs demonstrated better safety, stem-cell memory, and exhaustion profile compared to conventional CAR19-T cells in preclinical models [18]. In addition, the lack of CD4 and CD8 expression on DNTs allow CAR4-DNTs to effectively target T-cell leukemic blasts without experiencing fratricide, unlike conventional CAR4-T cells that express CD4. CAR-DNTs can also be combined with idelalisib to promote a central memory phenotype, improving cellular persistence and anti-leukemic activity in vivo [19]. The safety and efficacy of CAR19-DNTs were examined in a phase I clinical trial involving 12 patients with relapsed/refractory large B-cell lymphoma. No severe adverse events, such as GvHD, grade 2 or higher cytokine release syndrome, or immune effector cell-associated neurotoxicity syndrome were observed, indicating that the maximum tolerated dose was not reached and higher or additional CAR-DNT doses may be warranted. At the highest dose, all three patients achieved an objective response with one complete remission and two partial remissions [51]. Although the development of CAR-DNTs for AML treatment remains in progress, these studies provide a proof-of-concept for equipping allogeneic DNTs with an AML-targeting CAR to elicit enhanced anti-leukemic effects while capitalizing on the strong safety profile of DNTs relative to current conventional CAR-T cell therapies.