Introduction

T cell acute lymphoblastic leukemia (T-ALL) has traditionally been thought of as a higher risk leukemia that is more resistant to therapy than its B cell counterpart. Recent advances in therapy, including intensification of combination chemotherapy in the adolescent and young adult (AYA) population, have improved overall survival (OS), and T-ALL is no longer an independent risk factor for worse outcomes [1••, 2]. Outside of combination chemotherapy, which is associated with significant morbidity and late effects, there have been few advances specific to the treatment of T-ALL. Recently, the Children’s Oncology Group (COG) demonstrated improved event-free survival (EFS) with the addition of nelarabine to an intensive pediatric chemotherapy backbone, but an improvement in OS was not demonstrated at the time of publication [3••]. With the addition of nelarabine to frontline therapy because of these data, there are now even fewer options for the treatment of relapsed or refractory (R/R) T-ALL. New agents and targets are desperately needed for T-ALL, similar to the influx of options now available for B-ALL. In this review, we describe potential novel agents and pathways for T-ALL and drugs currently in development that could impact the future of T-ALL therapy.

Signal Transduction Inhibitors

Signal transduction pathways regulate normal cell growth and homeostasis. They are frequently dysregulated in malignant cells, leading to molecular and cellular survival advantages. As important parts of malignant transformation, perturbation of these pathways offers potential opportunities for targeted therapy. We review in this section select novel pathways that may be important in future targeted therapy for T-ALL.

Interleukin-7 Receptor Pathway

Steroids are a foundation in the treatment of ALL, but steroid resistance is common in R/R T-ALL. Synthetic glucocorticoids appear to act through modulating the apoptotic pathway involving protein B cell lymphoma-2 (BCL-2) [4]. Resistance to glucocorticoid therapy has previously been tied to the Janus kinase/signal transducer and activator of transcription (JAK-STAT) and the phosphatidylinositol 3-kinase/protein kinase B (PI3K-AKT) pathways [5, 6]. The JAK-STAT pathway is linked to interleukin-7 receptor (IL7R) signaling, and its effect on cell growth and survival is potentially mediated through upregulation of BCL-2 activity [5, 6]. Recent work has demonstrated the importance of the canonical mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK-ERK) signal transduction pathway in IL7R-mediated steroid resistance [7]. Different upstream pathways have been shown to activate MAPK/ERK kinase (MEK) within this cascade and are oncogenic drivers for many malignancies [8]. Targeted inhibitors have now been developed to disrupt activation of this pathway and are clinically efficacious in other cancers, making MAPK-ERK an attractive target in T-ALL [8]. In T-ALL cell lines and patient-derived xenograft (PDX) models, mutant IL7R signaling was associated with steroid resistance and increased downstream activation of MAPK-ERK signaling, in both models whose growth was dependent on increased exogenous IL7 signaling and in models who were independent of increase exogenous IL7 signaling [7]. One downstream target of the MAPK-ERK pathway is MEK which phosphorylates BCL-2-like protein 11 (BIM). BIM is a pro-apoptotic protein that binds anti-apoptotic proteins, such as BCL-2 and others, and is inactivated by phosphorylation by MEK [5, 7, 8]. The combination of glucocorticoids and MEK inhibition resulted in reversal of steroid resistance in cell lines, PDX models, and patient samples, suggesting the promising potential of combination therapy with MEK inhibitors [5, 7]. There is a phase I/II clinical trial currently enrolling in the UK examining the safety and early efficacy of the MEK inhibitor selumetinib in combination with dexamethasone (NCT03705507).

Either from mutations leading to increased activation or through overexpression of wild-type IL7R, stimulation of IL7R also leads to JAK-STAT signaling through a pathway separate from MAPK-ERK, which seems to coalesce in phosphorylation of signal transducer and activator of transcription 5 (STAT5) and induces the upregulation of proviral integration site for Moloney murine leukemia virus-1 (PIM1), a protein kinase [9, 10]. PIM1 is involved in cell cycle regulation and apoptosis and is susceptible to small-molecule inhibitors [11, 12]. Preclinical studies have demonstrated the importance of PIM1 in IL7R signaling for the survival of T-ALL cells, both in cell lines and from patient-derived samples [9, 13, 14]. PIM1 inhibition reversed cell cycle progression and cell growth in T-ALL cells in vitro and led to lower leukemic burden and longer survival in PDX models [9, 14]. PIM1 inhibition also synergizes with standard T-ALL therapy, such as corticosteroids, making it a potential addition to standard chemotherapy in the future [9]. Multiple PIM1 inhibitors are in early-phase clinical trials for both solid tumors and hematologic malignancies [15].

NOTCH Pathway Inhibitors

NOTCH receptors, specifically NOTCH1, play an important role in thymocyte development, and activating mutations are found in well over half of all T-ALL cases [16, 17]. NOTCH mutations resulting in increased activity exert at least part of their oncogenesis through upregulation of MYC transcription [18]. Significant work has been done to target this pathway with some early successes, but nothing has yet resulted in a Food and Drug Administration (FDA)-approved therapy. NOTCH1 is activated through cleavage, and gamma secretase inhibitors (GSIs) can inhibit NOTCH1 activation by preventing this cleavage [19]. GSIs were originally developed for the treatment of Alzheimer’s disease. Rather than directly cause apoptosis, GSIs arrest the cell cycle and may reintroduce steroid sensitivity in steroid-resistant T-ALL [19]. In combination with steroids, GSIs work synergistically to induce T-ALL cell death partly through transcriptional upregulation of the nuclear receptor subfamily 3 group C member 1 (NR3C1) gene and increased expression of pro-apoptotic regulators such BIM and BCL2 [19]. Unfortunately, GSIs cause considerable gastrointestinal (GI) toxicity, including diarrhea. Severe diarrhea may arise from disrupted NOTCH regulation of transcription of the gene KLF4 (Krüppel-like factor 4), which is involved in goblet cell differentiation in the GI tract [20]. While steroids seem to abrogate this side effect in vivo, diarrhea was the dose-limiting toxicity in early-phase studies [19, 21, 22]. A multicenter, nonrandomized, open-label, dose escalation phase I study of adult patients with R/R T-ALL was performed combining crenigacestat, a GSI, with dexamethasone, but the dose escalation of the GSI was limited by grade 3 GI toxicity despite steroid treatment [23]. This combination did result in stable disease in 17% of patients and one complete response (CR), which lasted 10.5 months, and established a dose to be investigated in further trials [23]. In a separate multicenter phase I trial, another GSI, BMS-906024, demonstrated similar toxicity patterns and some early evidence of response with a decrease in percentage of blasts in the bone marrow, one CR, and one partial response (PR) [21]. The patient who obtained a CR had an early T cell precursor ALL (ETP-ALL) and subsequently achieved a deep molecular response, underwent an allogeneic hematopoietic cell transplant (HCT), and remains without disease [22]. This patient’s ETP-ALL also carried mutations in protein tyrosine phosphatase non-receptor type 11 (PTPN11), DNA methyltransferase 3 alpha (DNMT3A), and colony-stimulating factor 3 receptor (CSF3R) in addition to a mutation in NOTCH1, possibly explaining this exceptional response, but these findings require further investigation [22]. Preclinical studies have demonstrated that resistance mutations will commonly emerge in the setting of GSI treatment in part through alternative activation of MYC [24, 25]. Two of these pathways identified through preclinical testing involve upregulation of BCL-2 and BRD4, which are targetable and therefore lend themselves to future clinical testing [25].

Cyclin-Dependent Kinase (CDK) Inhibitors

CDKs are a heterogenous group of proteins that play a critical role in cell cycle regulation and act as transcriptional cofactors. Within T-ALL, cyclin D3 is integral to NOTCH-driven tumorigenesis [26•]. Mouse cyclin D3 knockout models and mice treated with a CDK4/6 inhibitor both demonstrated NOTCH-driven T-ALL regression [27]. In other mouse models, CDK4/6 inhibitors with mammalian target of rapamycin (mTOR) inhibitors or steroids demonstrated synergism and disease response [28, 29]. Preclinical work has also targeted other CDKs, such as through a novel agent THZ1, which irreversibly covalently binds CDK7 outside of its kinase domain but still prevents its kinase activity [30]. Since THZ1 irreversibly binds to CDK7, on-target toxicity similar to GSIs would be expected, but THZ1 preferentially impacts oncogenes driven by super-enhancers, which suggests the possibility of a clinically relevant therapeutic window that could be further explored [30, 31]. Finally, targeting CDK9 has been attractive as it is involved in promoting transcription of myeloid cell leukemia-1 (MCL1). AZD4573 targets CDK9 through an alternative PROteolysis TArgeting Chimeras (PROTAC) approach [32, 33]. Tested on MOLT-4 T-lymphoblasts, the PROTAC approach induced apoptosis even after drug washout, suggesting a prolonged effect in vitro [33]. This preclinical evidence has led to multiple clinical trials, including one sponsored by the COG that combines palbociclib with reinduction chemotherapy (NCT03792256) and an institutional trial examining the combination of ribociclib, dexamethasone, and everolimus (NCT03740334).

BCL-2 Family Inhibitors

BCL-2 is an anti-apoptotic protein that was first discovered in B cell lymphoma. Along with B cell lymphoma-extra-large (BCL-XL) and MCL1, BCL-2 is a member of an anti-apoptotic family of proteins that share BCL-2 homology (BH) domains including BH3. These proteins bind to pro-apoptotic proteins such as BCL-2 associated X (BAX) and BCL-2 homologous antagonist/killer (BAK), sequestering them and thereby inhibiting apoptosis [34]. This family of proteins is important in the pathogenesis of T-ALL as evidenced by their involvement in other targetable pathways discussed above (section “Interleukin-7 Receptor Pathway”), but also by BH3 profiling of T-ALL cell lines, which have demonstrated different levels of dependence on BCL-2 or BCL-XL based on their level of cellular differentiation [35].

There is already significant experience targeting BH3 proteins in clinical practice. A retrospective case series descried the potential utility of venetoclax, a BCL-2 inhibitor, in combination with various chemotherapy regimens, including hyperfractionated cyclophosphamide, vincristine, doxorubicin, dexamethasone (hyper-CVAD), asparaginase, nelarabine, decitabine, and others, for the treatment of 13 patients with R/R T-ALL (5 of whom had ETP-ALL) [36]. Of the 10 evaluable patients, 60% achieved a composite CR (3 with a CR, 1 with a CR with incomplete hematologic recovery (CRi), and 2 with a morphologic leukemia-free state (MLFS)) [36]. The median venetoclax dose was 200 mg daily for 21 days, and prolonged cytopenias were seen with daily doses of 400 mg or durations of 14 days or more per cycle [36]. A phase I study of venetoclax in combination with a modified hyper-CVAD protocol, mini-hyper-CVD, in older adults with untreated ALL and adult patients with R/R ALL demonstrated promising results with manageable toxicities and recommended a phase II venetoclax dose of 600 mg daily [37••]. All newly diagnosed patients in this study achieved a response with 9 of the 10 (90%) achieving a CR with undetectable minimal residual disease (MRD) based on multicolor flow cytometry, while one achieved a PR [37••]. Of these patients, 3 had T-ALL (2 had ETP-ALL) [37••]. In the R/R ALL cohort, 3 out of 8 (38%) achieved a CR/CRi with 2 of the 3 becoming MRD negative [37••]. A different phase I study evaluated the use of venetoclax in combination with low-dose navitoclax, a BCL-XL inhibitor, and chemotherapy to treat patients with both B-ALL and T-ALL with the hypothesis that the combination would limit the thrombocytopenia from navitoclax while still achieving clinically significant BCL-2 family inhibition [38•]. A total of 44 patients with R/R ALL were treated (57% B-ALL and 43% T-ALL), and the CR rate was 64% for B-ALL and 53% for T-ALL [38•]. The study recommended phase II doses of navitoclax 50 mg daily (reduced to 25mg if less than 45 kg) and venetoclax 400 mg daily continuously, while response is maintained [38•]. Protocols are now being designed to incorporate these BH3 mimetics (venetoclax and navitoclax) into frontline treatment regimens.

Epigenetic Modifiers

There are two main classes of epigenetic modifying agents: histone deacetylase inhibitors (HDACi) and hypomethylating agents (HMAs). Histone deacetylases regulate chromatin structure, and HDACi have been approved for the treatment of lymphomas and myeloma. One HDACi, panobinostat, has shown some efficacy in the MOLT-4 T-ALL cell line in vitro and in xenografts in vivo, in contrast to little or no efficacy in B-ALL cell lines [39]. In PDX models, panobinostat in combination with chemotherapy prolonged mouse survival over either chemotherapy or HDACi alone [39]. This has led to several small clinical trials including a phase I trial in both older, newly diagnosed ALL patients, and those with R/R ALL that combined entinostat (a HDACi) with clofarabine [40]. In this study, 5 of 28 patients had T-ALL, but responses were not reported separately for this subset; CR or CRi was achieved in 4 patients in total [40].

HMAs play an important role in the modern treatment of acute myeloid leukemia (AML) but are not routinely used in ALL. Preclinical work on adult T-ALL samples has identified subgroups of T-ALL based on DNA methylation profiles [41]. A hypermethylated T-ALL subgroup had a worse outcome than other subgroups [41]. Using PDX models of this hypermethylated subgroup, treatment with azacitidine, a HMA, prolonged survival and delayed leukemia progression [41]. Case reports show clinical response to decitabine, another HMA, in combination with venetoclax [40,41,42]. Two cases of T-ALL achieved MRD negative CR, and one case of T cell lymphoblastic lymphoma also achieved a CR [42,43,44].

Proteasome Inhibitors

The proteasome degrades and processes proteins within cells and when inhibited can induce apoptosis due to the collection of unfolded or misfolded proteins [45]. Available proteasome inhibitors include bortezomib, carfilzomib, and ixazomib, and this class of drugs is FDA-approved and commonly used for the treatment of multiple myeloma. They selectively inhibit the 26S proteasome, leading to several anti-cancer mechanisms: increased pro-apoptotic proteins such as phorbol-12-myristate-13-acetate-induced protein 1 (NOXA), suppressed nuclear factor kappa B (NF-ĸB) signaling, and decreased anti-apoptotic proteins BCL-XL and BCL-2 [46]. Pediatrics has more readily embraced the potential role of proteasome inhibitors in the treatment of ALL to date. In a large study, 135 pediatric patients with relapsed ALL received reinduction chemotherapy plus bortezomib, yielding a promising CR rate of 68% for R/R T-ALL (22 of the 135 patients) [47]. This study led to a large COG study of T-ALL using bortezomib in combination with chemotherapy in the upfront setting (AALL1231) [48••]. The study randomized 824 patients to a standard pediatric chemotherapy backbone with or without bortezomib [48••]. The intervention arm had a 4-year EFS of 83.8% and 4-year OS of 88.3%, compared to 80.1% and 85.7%, respectively, in the control arm. These differences were not statistically significant, but there was also no increase in toxicity on the bortezomib arm [48••]. There was a statistically significant difference favoring bortezomib in the T lymphoblastic lymphoma subset, with a 4-year EFS of 86.4% and 4-year OS of 89.5%, compared with 76.5% and 78.3%, respectively, in the control arm (p=0.009) [48••]. The next upfront study from the COG for T-ALL may randomize a proteasome inhibitor with a different chemotherapy backbone.

Other studies, primarily involving patients with B-ALL, have evaluated the use of bortezomib and ixazomib. One study of 10 children with R/R ALL (9/10 with B-ALL) gave mitoxantrone, dexamethasone, pegaspargase, vincristine, intrathecal methotrexate, and bortezomib (1.3 mg/m2 for 4 weekly doses), and 80% achieved CR or CRi, including a CR in the one patient with T-ALL [49]. A phase I dose-finding study of 19 adults (51–75 years old) with B-ALL evaluated the use of ixazomib together with induction therapy with prednisone, vincristine, and doxorubicin (plus dasatinib if BCR-ABL1 positive); 79% had CR or CRi [50]. The recommended phase II dose for ixazomib was 2.3 mg weekly [50].

JAK/STAT Inhibitors

The JAK/STAT signal transduction pathway is an important driver in the development and propagation of T-ALL and has been discussed already in relationship to other pathways in a previous section (section “Interleukin-7 Receptor Pathway”). In addition to these associations, recurring mutations in T-ALL have been described in genes encoding different JAK and STAT proteins, including STAT5b, JAK1, and JAK3 [51,52,53]. This pathway is directly inhibited by ruxolitinib, a JAK 1/2 inhibitor. In preclinical models, the JAK/STAT pathway was hyperactive with increased IL-7 signaling that can be effectively inhibited in vitro with ruxolitinib [54]. This inhibition was more effective in ETP-ALL than non-ETP-ALL subtypes and was even seen without mutations in the IL7-JAK-STAT pathway [54]. Ruxolitinib can also decrease glucocorticoid resistance in ETP and T-ALL preclinical samples, possibly through IL-7 pathway hyperstimulation [55]. Because of this synergism, ruxolitinib is currently being investigated in R/R T-ALL in combination with glucocorticoids and chemotherapy (NCT03613428).

Tyrosine Kinase Inhibitors (TKIs)

Mutations and amplification of the ABL1 (Abelson tyrosine–protein kinase 1) gene in T-ALL lead to aberrant tyrosine kinase activation [26•, 56]. Fusions involving ABL1 also occur in T-ALL, but these fusions are more commonly with NUP214 (nucleoporin 214) than BCR (breakpoint cluster region, which when fused with ABL1 is referred to as the Philadelphia (Ph) chromosome) [57]. In NUP214::ABL1 translocated T-ALL, TKIs such as imatinib, nilotinib, and dasatinib have demonstrated efficacy both in vitro and in PDX models, leading to increases in mouse OS [57,58,59]. In some NUP214::ABl1 T-ALL, co-occurring TLX1 (T cell leukemia homeobox 1) mutations have been shown to be drivers of proliferation and survival, making these T-ALL cells sensitive to targeting with a combination of TKIs and BCL-2 inhibitors [59]. The NUP214::ABL1 translocation is not common in T-ALL; thus, the larger family of available TKIs has only limited utility in this disease [57]. Dasatinib does have significant activity outside of targeting ABL kinases and can induce cell death in 30% of T-ALL in vitro, regardless of the mutation profile, which has been confirmed in mouse models [60, 61]. Case reports have described dramatic responses to dasatinib [61, 62]. Besides the NUP214::ABL1 translocation, there are currently no predictable mutational patterns for sensitivity to TKIs. It is possible that phosphoproteomic profiling of T-ALL can help identify targetable kinases such as insulin receptor (InsR) and insulin-like growth factor-1 receptor (IGF-1R), leading to increased sensitivity to dasatinib [63]. In R/R T-ALL, novel combinations with dasatinib are appealing, and more data should be gathered about predictive patterns in T-ALL.

PI3K/AKT/mTOR Inhibitors

Phosphoinositide-3 kinase (PI3K) plays an important role in normal T cell development and in T-ALL leukemogenesis. PI3K is activated by multiple different growth factors and G protein-coupled receptors and then activates the protein kinase B (AKT), which increases mTOR activity and drives cell proliferation and survival [64]. Nearly half of T-ALL have alterations in this pathway, including mutations in PTEN (phosphatase and tensin homolog; PI3K inhibition), USP7 (ubiquitin specific protease 7; PTEN inhibition), PIK3R1 (phosphoinostidide-3-kinase regulatory subunit 1), and PIK3CD (phosphatidylinositol-4, 5-bisphosphonated 3-kinase catalytic subunit delta) [26•]. These mutations seem to cluster in T-ALL that also carry TAL1 (T cell acute lymphocytic leukemia protein 1) mutations, possibly because of interaction with AKT within the PI3K pathway [26•, 65]. Mutations in this pathway contribute to resistance to chemotherapy and corticosteroids [66]. Numerous studies have demonstrated the efficacy of PI3K and AKT inhibition in various T-ALL models, which resensitizes leukemic blasts to corticosteroids [5, 13, 67]. Perhaps the most attractive target in this pathway is mTOR, for which multiple FDA-approved inhibitors exist for other diseases, including sirolimus/rapamycin, everolimus, and temsirolimus. T-ALL cell death is induced in vitro and in vivo through either mTOR inhibition alone or both mTOR and PI3K inhibition [13, 68]. In a phase I/II study of everolimus combined with hyper-CVAD for R/R ALL, a response rate of 50% was noted in a very heavily pretreated population of T-ALL (median of 4 prior lines of therapy) [69]. A number of clinical trials that are either actively recruiting or recently closed include an mTOR inhibitor in combination with other chemotherapy and targeted agents for R/R ALL.

Anti-CD38 Therapy

CD38 (cluster of differentiation 38) is typically expressed on the surfaces of immune cells as a signal transducer. It is also expressed on the surface of lymphoblasts from patients with T-ALL, making it a potential target for therapeutic intervention [70•, 71]. CD38 expression persisted on T-ALL cells despite 1 month of exposure to induction chemotherapy of dexamethasone, vincristine, daunorubicin, and PEGylated asparaginase with or without bortezomib [70•, 71]. This was not true for B-ALL, which downregulated CD38 after exposure to induction chemotherapy [70•]. Similar to proteasome inhibitors, the anti-CD38 monoclonal antibodies daratumumab and isatuximab are more commonly used in multiple myeloma than in leukemia, likely because plasma cells have the highest expression of CD38 [71]. However, PDX models of T-ALL showed reduction of leukemia burden after daratumumab [70•, 71]. A phase II study evaluating daratumumab combined with chemotherapy for the treatment of pediatric and young adult patients with R/R ALL, including those with T-ALL, has enrolled 47 participants to date (NCT03384654). A phase II study of isatuximab monotherapy in 11 patients with T-ALL was not encouraging as no patients had CR or CRi, most had progressive disease, and over half had an infusion reaction [72]. A novel approach using a bispecific T cell-engaging antibody targeting CD38 and CD3 (XmAb 18968 CD3-CD38) has preclinical evidence of efficacy in multiple myeloma and is currently undergoing clinical trials in T-ALL (NCT05038644) [73].

CAR T Cell Therapy

Chimeric antigen receptor (CAR) T cell therapy has been used successfully for patients with B-ALL, but there is now early evidence for its use in T-ALL [74, 75]. The CARs typically consist of an extracellular antigen binding domain, a transmembrane domain, costimulatory molecules, and immunoreceptor tyrosine-based activation motifs [76]. Currently, designs with modified costimulatory domains are in their fourth generation, but a second generation CAR T cell with CD3ζ and 4-1BB domains (tisagenlecleucel) was the first-in-class FDA-approved CAR T cell therapy for R/R CD19-positive B-ALL in children and young adults up to 25 years of age [77]. Additional CAR T cell products have been approved in adult R/R B-ALL, large B cell lymphoma, follicular lymphoma, mantle cell lymphoma, and, most recently, multiple myeloma. CAR T cell therapy is much more at its infancy in the treatment of T-ALL because of the difficulty of a target that also does not lead to untenable toxicity and lifelong late effects [78]. Treatment of T cell malignancies with CAR T cell therapy is also complicated by fratricide, which is the destruction of CAR T cells by other CAR T cells [78]. In a phase I trial, 20 patients with R/R T-ALL received anti-CD7 CAR T cell therapy with 90% (18/20) achieving CR and 35% (7/20) going on for an allogenic HCT [75]. CD1a is another CAR T cell target being explored in T-ALL [79]. Methods to address fratricide and future studies to evaluate effective CAR T cell targets in T-ALL will likely lead to further advancements in this therapy line.

Conclusion

A variety of targets are under investigation to improve the treatment and clinical outcomes for patients with T-ALL. We have summarized important medications in Table 1 that may reshape the way we approach T-ALL treatment and reduce disparities in health outcomes between B-ALL and T-ALL. Current and future studies will help better delineate which promising treatments are most advantageous for patients depending on their disease biology and genetics.

Table 1 Drug classes and agents under investigation for T-ALL
Full size table