Advances and Challenges in RAS Signaling Targeted Therapy in Leukemia

Leukemia is a type of cancer marked by the uncontrolled growth and impaired differentiation of leukemic cells (1). These cells infiltrate the bone marrow and peripheral blood, disrupting normal hematopoiesis and potentially leading to bone marrow failure. Major forms of leukemia include acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), and chronic myeloid leukemia (CML; ref. 1). Current therapies for leukemia include chemotherapy, targeted therapy, and bone marrow transplantation, with treatment plans often tailored to the specific type of leukemia and its genetic mutations. Notably, some germline drivers of risk have been identified, such as those in the RAS, BCR-ABL, FLT3, and NPM1 genes, which play a significant role in the development and progression of leukemia, influencing both the behavior of the cancer and the choice of treatment strategies. In particular, the prevalence of oncogenic RAS alleles is not equally distributed in leukemia. Thus, advances in understanding the genetic basis of leukemia, such as oncogenic mutations in the RAS gene, will lead to the development of more precise therapeutic strategies, improving patient outcomes, although there are substantial challenges.

The classical RAS gene family comprises KRAS, NRAS, and HRAS and encodes KRAS4A, KRAS4B (referred to hereafter as KRAS), NRAS, and HRAS proteins (2), and the RRAS subgroup (RRAS, TC21/RRAS2, and MRAS/RRAS3) of the RAS family is the closest relatives to the classical RAS oncogenes (2, 3). These small GTPases act as a binary molecular switch to regulate critical cellular events, including cell proliferation, differentiation, and survival (2, 4). They cycle between an active state (GTP bound) and an inactive state (GDP bound; ref. 5), which is mediated by guanine nucleotide exchange factor (GEF)–facilitated nucleotide exchange and GTPase-activating protein (GAP)–stimulated GTP hydrolysis (6–8). Oncogenic mutations, which often occur at hotspots, such as codons 12, 13, or 61, lock RAS in its GTP-bound form and potentially lead to constitutive signaling activation and tumorigenesis (9, 10).

The RAS oncogene family has garnered attention as a therapeutic target in leukemia, due to its frequent mutations and pivotal role in disease progression, as well as its crucial role in cell proliferation and differentiation (11), including efforts on the direct inhibition of RAS function, disruption of downstream signaling nodes (e.g., RAF, MEK, ERK, and PI3K), targeting upstream molecules (e.g., PDEδ, SHP2, and STK19), RNA interference (RNAi) of RAS expression, targeting the distinct metabolic processes associated with RAS mutations (e.g., micropinocytosis and autophagy), and exploitation of synthetic lethality (12). Moreover, small-molecule inhibitors that specifically target mutant RAS isoforms have shown promise in preclinical models and have entered clinical trials. Of note, the recent approval of KRAS^G12C inhibitors, adagrasib (MRTX849; ref. 13) and sotorasib (AMG510; ref. 14), has validated KRAS as a direct target and demonstrated the feasibility of selectively targeting specific RAS mutants. However, so far, KRAS^G12C remains the only RAS mutant successfully targeted in patients with non–small cell lung cancer (NSCLC) with FDA-approved inhibitors, which limits the applicability of these therapies to other oncogenic RAS mutants, such as G12D and G13D, in leukemia. In addition, the intrinsic properties of RAS proteins, such as their high affinity for GTP, the complexity of their effector interactions, activation of alternative signaling pathways, and resistance to RAS inhibitors, render them elusive targets for drug development.

In this review, we have described the recent advances and discussed the challenges in developing therapies targeting RAS proteins. We have covered the structure and pathway of RAS, emphasizing the role of RAS mutation in leukemia. We have detailed recent strategies to directly or indirectly inhibit RAS, including emerging therapeutic strategies, and highlighted the breakthrough therapies that target RAS allele specifically.

The RAS family of GTPases is characterized by a high degree of homology in their primary sequence, particularly in the GTP-binding domain (residues 1–166) and the C-terminal hypervariable region (residues 167–188/189; refs. 2, 15, 16). The GTP-binding domain contains critical motifs, including switch I (residues 30–40), switch II (residues 58–72), and the P-loop (residues 10–14), which are essential for effector interaction and downstream signal transduction (5, 16). The C-terminal region is crucial for membrane association, with the CAAX tetrapeptide acting as a site for posttranslational modifications, such as isoprenylation, proteolysis, and methylation, which collectively facilitate membrane localization and function of RAS proteins (17).

RAS proteins function as a binary molecular switch, alternating between an inactive GDP-bound state and an active GTP-bound conformation. Mutations in the RAS family lead to constitutive activation of RAS signaling pathways (RAS/RAF/MEK/ERK and RAS/PI3K/AKT), resulting in autonomous signal propagation independent of extrinsic ligands (10). This constitutive RAS pathway activation primarily leads to aberrant RAF/MEK/ERK or PI3K/AKT/mTOR signaling, thereby dysregulating cell proliferation, survival, and metabolic processes (18, 19). Moreover, RAS interacts with other downstream effectors, such as RalGDS and PLCε, which participate in various cellular activities, including the modulation of cytoskeletal architecture, vesicular trafficking, and transcriptional regulation. Dysregulation of these pathways is pivotal to the oncogenic potential of RAS mutations in the development and progression of leukemia (11). In summary, RAS GTPase family members serve as crucial molecular determinants in the pathophysiology of leukemia, orchestrating a complex network of signaling pathways that regulate cellular processes (10).

HRAS was initially recognized as an oncogene following the discovery of a point mutation in 1982 (20). Subsequently, mutations in the NRAS and KRAS genes were also identified (20). RAS mutations are frequently detected across a spectrum of leukemia subtypes, including AML, ALL, CML, and myelomonocytic leukemia. Analysis of the Catalogue of Somatic Mutations in Cancer (COSMIC) database indicates that NRAS mutations account for 11.25% (Fig. 1), KRAS mutations account for 6.05% (Fig. 2), and HRAS mutations account for 0.15% of all leukemias (Fig. 3; COSMIC v99; ref. 21). Specifically, mutation frequencies in ALL are 10.57% for NRAS, 8.27% for KRAS, and 0.37% for HRAS; in AML, the frequencies are 13.4% for NRAS, 4.79% for KRAS, and 0.12% for HRAS; and in CML, they are 6.63% for NRAS and 2.62% for KRAS. These data indicate that mutated NRAS is the most prevalent RAS isoform. Notably, the most common NRAS mutations are G12D, G13D, and G12S (Fig. 4A; Supplementary Table S1), whereas the most frequent KRAS mutations include G12D, G13D, and G12V (Fig. 4B; Supplementary Table S2). No dominant mutation is found in HRAS (Fig. 4C; Supplementary Table S3).

Mutations in the three RAS isoforms at G12, G13, and Q61 can substantially attenuate the intrinsic and GAP-stimulated GTPase activity of RAS (22–24). As a result, RAS remained a continuously active GTP-bound state, leading to oncogenic activity. This constitutive activation triggers the two predominant RAS signaling pathways, that is, MAPK and PI3K/AKT pathways, as well as ancillary pathways including RalGDS and the JAK/STAT axis (25, 26). Such activation alters cellular processes critical for maintaining hematologic homeostasis and is implicated in the pathogenesis of various leukemias (18). However, studies suggested a nuanced role of RAS mutations in different types of leukemia. Jerchel and colleagues (27) indicated that NRAS mutation-associated B-cell precursor acute lymphoblastic leukemia (BCP-ALL) may not consistently engage the MAPK pathway. Conversely, Chan and colleagues (28) found that NRAS mutations could potentiate B-cell leukemogenesis through the activation of either the STAT5 or MAPK pathways, highlighting the complexity of the molecular mechanisms associated with NRAS mutations in B-cell acute lymphoblastic leukemia (B-ALL). Moreover, Qian and colleagues (18) demonstrated that not all NRAS mutants uniformly contribute to leukemogenesis. Their research showed that the ectopic expression of NRAS^G12D significantly promoted IL3-independent growth in murine hematopoietic progenitor Ba/F3 cells (18). A comparative analysis of various NRAS G12 mutants revealed a heterogeneity in leukemogenic potential. A subset of NRAS mutants (G12L, G12T, G12I, G12K, G12V, G12Q, and G12R) displayed a more pronounced leukemogenic capacity relative to NRAS^G12D, whereas another cohort (G12W, G12C, G12H, G12E, G12N, G12M, and G12A) exhibited leukemogenic capabilities ranging from comparable to marginally inferior. A group of four NRAS mutants (G12P, G12Y, G12S, and G12F) lacked the capacity to transform Ba/F3 cells, suggesting a nuanced spectrum of pathogenicity among NRAS G12 variants (18). Additionally, the roles of KRAS mutations in the leukemogenesis of CML and their clinical significance remain ambiguous (8). Furthermore, the co-occurrence of RAS mutations with other genetic aberrations is common. Papaemmanuil and colleagues (29) described distinct co-mutation patterns within gene hot spots, noting that NPM1 mutations preferentially co-occurring with NRAS G12/13 mutations but not with other NRAS mutations. Ouerhani and colleagues (8) found that KRAS mutations were absent in BCR/ABL1-positive CML, suggesting that BCR/ABL1 fusion gene and KRAS mutations are mutually exclusive.

Longitudinal mutation analysis in patients with ALL, from diagnosis to relapse, indicated that although initial chemotherapy may eliminate the dominant leukemic clone, relapse is often characterized by the persistence of mutations in critical pathways, such as the RAS signaling pathway, highlighting the importance of RAS mutations in leukemia recurrence (27, 30–34). Previous studies suggested that clonal mutations at relapsed ALL emerge from relapse-favoring subclones that already existed at diagnosis (35). However, Irving and colleagues (31) reported that although the overall incidence of samples with mutations does not significantly change at relapse, about half of the relapse samples with RAS pathway mutations were initially wild type at diagnosis. Relapse-specific mutations in pediatric ALL have been identified in prior studies (36, 37). Qian and colleagues (18) pointed out that RAS mutations correlate with a higher incidence of relapse in patients with pediatric ALL, especially in the KRAS mutation subgroup. They also noted that leukemogenic NRAS mutations may respond to inhibition of multiple signaling pathways, including MEK, autophagy, AKT, EGFR, polo-like kinase, SRC, and TGF-β receptor (18). These mutations, present at diagnosis or acquired during treatment, are associated with high-risk features, such as early relapse and failure to achieve complete remission (38). However, the precise molecular and cellular contributions to therapy resistance and relapse remain incompletely understood.

There is some controversy about the impact of RAS mutations on prognosis and clinical outcomes in several types of hematopoietic malignancies (11). For example, NRAS mutations have been identified as a significant prognostic factor for relapse in B-ALL (39), whereas patients with T-cell acute lymphoblastic leukemia (T-ALL) who relapse generally had a uniformly poor prognosis (40). Although inhibitors targeting KRAS mutations have shown clinical efficacy, targeting NRAS mutations remains challenging due to intricate and compensatory mechanisms (18). NRAS mutations have been linked to a longer median survival compared with wild-type NRAS in AML (41). However, Lee and colleagues (42) found a significantly lower rate of complete remission in patients with AML with NRAS mutations compared with those without RAS mutations. In patients with pediatric ALL, RAS mutations have not been associated with a significant difference in overall survival compared with those who have wild-type RAS (18). Therefore, the significance of RAS mutations may vary across different diseases. For leukemia, the use of RAS mutations as a clinical marker for progression has not been established yet (11). Considering the impact of RAS mutations on downstream pathways, prognosis, relapse, and drug resistance, exploring RAS mutations in leukemia is crucial for the development of targeted therapies. RAS mutations modulate responses to targeted therapies. Agarwal and colleagues (43) demonstrated that KRAS^T58I mutation in patients with CML (in the absence of BCR-ABL kinase domain mutations) correlated with imatinib (400 mg) resistance. Qian and colleagues (18) highlighted the sensitivity of the NRAS^G12D mutation to MEK inhibition.

Recent research on the protein structure, biological function, signaling output, and molecular mechanisms of RAS have provided new insights into the development of strategies targeting RAS (44). Efforts to inhibit oncogenic RAS signaling include direct intervention in RAS itself, targeting of upstream proteins and downstream effectors, as well as the employment of RNAi.

Directly targeting RAS proteins was once considered impossible, but recent advances in structural and functional analyses have revealed new opportunities for therapeutic intervention (44). These advancements have enabled the development of strategies such as disrupting RAS–SOS (son of sevenless) interactions, competitively inhibiting the guanine nucleotide binding site, as well as blocking RAS–effector interactions (12, 45). A significant milestone in RAS pharmacology was the successful targeting of the KRAS^G12C mutant, exemplified by the approval of adagrasib (13) and sotorasib (14) for treating KRAS^G12C-mutant NSCLC in clinic. However, its application in preclinical leukemia studies has yet to be reported.

Targeting the RAS–SOS interaction is a recognized strategy for inhibiting RAS activation. Small molecules, such as 4,6-dichloro-2-methyl-3-aminoethyl-indole, bind to the RAS–SOS interaction surface, effectively impeding nucleotide exchange and thereby inhibiting RAS activation in cancer cells (46). Compound 2C07 alters nucleotide preference, inhibiting SOS-mediated nucleotide exchange, and has been modified into irreversible covalent analogs that target both GDP- and GTP-bound states, thereby inhibiting PI3K activation in vitro (47). This dual-targeting strategy offers potential for inhibiting oncogenic RAS mutants predominantly existing in the GTP-bound state within cells (47). BI-2852 binds to a larger region of KRAS to form unfunctional RAS dimer, inhibiting both SOS1-catalyzed exchange from GDP-KRAS to GTP-KRAS and GAP-catalyzed exchange from GTP to GDP-KRAS (48). In a novel approach to RAS-binding small molecules, merging two distinct chemical series led to the development of compound Ch-3, an inhibitor of RAS protein–protein interactions derived from synthesizing protein–protein interaction networks with antibody fragment series (49). Compound Ch-3’s low molecular weight served as a foundation for further medicinal chemistry efforts to enhance its pharmacologic properties and binding affinity (49). BI-2493 is a structural analog of BI-2865 that was optimized for in vivo administration. These two pan-KRAS inhibitors exhibit a similar binding mode to mutant KRAS, have broad therapeutic implications, and merit clinical investigation in patients with KRAS-driven cancers (50).

Disrupting RAS–effector interactions is also a promising strategy. These inhibitors are designed to disrupt the interface between RAS proteins and their downstream effectors, particularly focusing on conserved residues within the switch I region, which is a critical site for effector binding. By disrupting these interactions, RAS-mediated signal transduction flow that is crucial for cancer cell proliferation and survival can be impeded.

For example, compounds Kobe0062 and Kobe0065 not only inhibit HRAS–RAF interaction that induce apoptosis of H-ras^G12V-transformed NIH3T3 cells but also exhibit antitumor activity on a xenograft of human colon carcinoma SW480 cells carrying the KRAS^G12V mutation by oral administration (51). Rigosertib, a RAS mimetic, competitively binds to RAS effectors (52), resulting in an inhibition of multiple signaling pathways including the RAS/RAF/MEK/ERK pathway. Also, the inhibition of the RAS/RAF/MEK/ERK pathway by rigosertib could be ascribed at least in part to oxidative stress (53). Moreover, Malacrida and colleagues (54) showed a dose- and time-dependent anticancer effect of rigosertib in RPMI 8226 (multiple myeloma) cell line. Compound 3144 binds to KRAS^WT, KRAS^G12D, NRAS, and HRAS and interacts with D38, A59, and Y32, acting as a multivalent inhibitor (55). This indicates that pan-RAS inhibitors could potentially target multiple RAS mutants with antitumor efficacy, although further optimization is needed to enhance specificity and reduce toxicity (56). Nam and colleagues (57) demonstrated that PHI-501 is a novel small-molecule inhibitor intended for the treatment of AML in patients expressing the NRAS-activating mutation. Of note, a novel class of next-generation inhibitors with unique mechanism of action from Revolution Medicines, such as KRAS(ON) inhibitor, RM-007, RM-008, RMC-4998, RMC-5127, RMC-6291, RMC-6236, and RMC-7977 (Fig. 5), covalently binds to KRAS mutants in the GTP-bound state and form a tri-complex with GTP-bound KRAS and immunophilin cyclophilin A, demonstrating anticancer effects in different types of cancer (58–62). Of clinical importance, targeting RAS GTP-bound form with approaches like these could counteract mechanisms of resistance to GDP-bound KRAS^G12C covalent inhibitors.

Alternative strategies targeting KRAS^G12C involve trapping inactive RAS proteins using switch II mutant selective covalent inhibitors. These therapeutic agents are designed to form irreversible bonds with the mutant switch II region of RAS proteins, thereby blocking aberrant signaling crucial for cancer cell proliferation. Shokat and colleagues (63) pioneered the development of covalent inhibitors that target KRAS^G12C, demonstrating that compound six binds switch II pocket, rather than the GDP pocket, which was not observed in native RAS structures (63). The discovery of the switch II pocket paved the way for subsequent inhibitors targeting KRAS^G12C.

For example, ARS853 selectively targets the KRAS^G12C mutant by covalently binding to the GDP-bound KRAS, thereby maintaining KRAS in an inactive state (64, 65). However, the plasma instability of ARS853 limited its application in vivo. 1_AM, a covalent quinazoline-based compound targeting the switch II pocket, effectively suppresses GTP loading of KRAS^G12C in cancer cells harboring this mutation (66). ARS1620, an optimized derivative of ARS853, has shown promising results in preclinical models by overcoming previous limitations (67). JNJ-74699157 (ARS3248), a next-generation KRAS^G12C inhibitor derived from ARS1620, is currently undergoing clinical trials to assess its safety and antitumor efficacy. Despite these advances, the prevalence of RAS proteins in the GTP-bound form poses a challenge for inhibitors such as ARS853 and ARS1620. Of note, ARS-1620 was optimized into the later approved sotorasib (AMG 510), a selective and well-tolerated inhibitor, which has demonstrated clinical benefits in patients with NSCLC (14). Furthermore, combination therapies have shown improved therapeutic outcome (Fig. 5; ref. 68). However, resistance mechanisms have emerged, including the novel KRAS^Y96D mutation that affects the switch II pocket, necessitating the development of new strategies to combat drug resistance (69). Another successful example of KRAS^G12C inhibitor, adagrasib, occupies the switch II pocket through interaction with the mutant cysteine and engages Y64 on KRAS, thereby trapping KRAS in an inactive conformation. It has received FDA breakthrough therapy designation for treating KRAS^G12C-positive NSCLC and is undergoing clinical trials for other solid tumors (70). Notably, the adagrasib scaffold has been adapted to target other oncogenic KRAS variants, such as KRAS G12S (71). Notably, an additional modification of the core scaffold generated MRTX1133 targeting KRAS^G12D, which has shown evident tumor regression in preclinical models (72). We have noticed an advancement in this warhead evolution strategy by incorporating “malolactone” electrophiles into the MRTX1133 core scaffold to selectively react with the Asp12 side chain (73). Meticulous structure-guided design resulted in (R)-7, which labels KRAS^G12D but not other mutants or wild-type KRAS (73). (R)-7 interacts not only with GDP-bound KRAS, typical of other switch II pocket binders, but also with GTP-bound KRAS, potently suppressing downstream signaling and cancer cell growth (73). (R)-7 will require further testing and possible optimization enroute to the clinic. Targeting the GTPase nucleotide binding site, long deemed challenging, has been realized with compounds such as SML-8-73-1, a GDP analog that forms covalent bonds with the site, thereby stabilizing GDP-bound KRAS^G12C (74, 75).

Moreover, Sacher and colleagues (76) demonstrated that treatment with divarasib (GDC-6036) resulted in durable clinical responses across KRAS^G12C-positive tumors, with mostly low-grade adverse events. Inhibition of KRAS^G12C (ON) is necessary for optimal target coverage and prevention of adaptive mechanisms of resistance. FMC-376, a novel inhibitor of the activated, GTP-bound form of KRAS^G12C, also potently inhibits the inactive, GDP-bound form of KRAS^G12C (77, 78). FMC-376 binds KRAS in the switch II pocket, rapidly forming a covalent bond with cysteine 12 in the presence of either GDP or GTP. This results in the potent inhibition of RAF1 and PI3Kα effector interactions in contrast to sotorasib or adagrasib (78). Also, researchers from BridgeBio Pharma Inc. demonstrated that BBO-8520, the novel next-generation KRAS^G12C GTP/GDP dual inhibitor, can drive deep responses following the development of resistance to sotorasib (NCT06343402).

Interestingly, a small-molecule KRAS agonist, KRA-533, binding to nucleotide binding pocket of KRAS and interacting with K117, causes an accumulation of GTP-bound KRAS and activation of the ERK signal (79). Treatment with KRA-533 resulted in increased KRAS activity and suppressed cell growth by inducing apoptosis and autophagy. Furthermore, cell lines with KRAS mutations were more sensitive to KRA-533 compared with those without such mutations (79). Although KRAS mutations are prevalent in leukemia, to date, there have been no reports on the utilization of this molecule in the research of leukemia treatments.

Together, these advancements on small molecules underscore the diverse strategies and ongoing efforts in developing effective therapies targeting KRAS mutants, particularly KRAS^G12C, highlighting their potential in combating various cancers, including those driven by RAS mutations. Continued research and clinical trials are crucial for optimizing these therapies, overcoming challenges posed by resistance mechanisms, and extending the application for other types of cancer, including RAS-mutant leukemia.

An alternative strategy for directly inhibiting RAS is to target its posttranslational modifications. RAS requires three enzymatic posttranslational processing events, including prenylation of the CAAX box catalyzed by farnesyltransferase (FTase) or geranylgeranyltransferase type 1 (GGTase1); cleavage of the terminal AAX residues by RAS-converting enzyme 1 (RCE1); and methylation of the remaining cysteine of the CAAX box catalyzed by isoprenylcysteine carboxyl methyltransferase (ICMT), to associate with the inner leaflet of the plasma membrane (58, 80, 81). Inhibitors targeting the posttranslational processing of RAS can inhibit its membrane association and subsequent signaling. Despite the initial successes of farnesyltransferase inhibitors (FTI), the existence of alternative lipid modification pathways has hindered their clinical efficacy (82). For example, HRAS can only be farnesylated by FTase, whereas KRAS and NRAS can be prenylated by both FTase and by a “bypass” GGTase pathway (11, 82–84). Furthermore, if FTase is inhibited, the KRAS and NRAS isoforms may be targeted by GGTase1 (82). Mutation analysis approaches give FTIs renewed hope in treating cancers. KRAS and NRAS do not respond to FTI; however, HRAS is sensitive (85). Therefore, FTIs are sufficient to inhibit the action of HRAS, but inhibiting both FTase and GGTase1 is required to block the action of KRAS and NRAS. It seems that there has been a resurgence of interest in the strategy of inhibiting HRAS mutants using FTI tipifarnib in cancers (58). Tipifarnib is currently progressing through phase II clinical trials for use in HRAS-mutant head and neck, leukemia, lymphoma, and thyroid cancers (ClinicalTrials.gov). In the context of CML, three FTIs have been investigated: two in phase I clinical trials (tipifarnib and lonafarnib; refs. 86, 87) and one (BMS-214662; ref. 88) at the preclinical stage. Tipifarnib has elicited hematologic responses in most evaluated patients, and lonafarnib and BMS-214662 have yielded promising results. Despite these findings, the clinical progression of these compounds for CML treatment has stalled.

In addition to targeted use of existing FTIs, a new FTI has recently been developed that overcomes KRAS resistance and potentially opens the way to pan-RAS inhibition (89). Consequently, this has prompted a shift in the focus of research and drug discovery efforts, particularly toward targeting the KRAS protein. A geranylgeranyl transferase inhibitor (GGTI) targets an alternative lipid modification pathway, geranylgeranylation, which becomes more active in the absence of farnesylation. However, GGTIs, like FTIs, have also encountered similar challenges in clinical settings. Nevertheless, as geranylgeranylation of normal cellular proteins is more prevalent than farnesylation, a selective FTase inhibitor should have fewer toxic effects (11, 90). Inhibitors that target the RAS palmitoylation cycle, such as palmostatin B, disrupt oncogenic RAS localization. They have shown effectiveness in preclinical models, but clinical validation is still pending (91). PDEδ inhibitors and RCE1 inhibitors also have the potential to hinder RAS activation by hindering RAS trafficking to the membrane (92). FTS, a RAS farnesylcysteine mimetic, selectively disrupts the binding of active RAS proteins to the plasma membrane, functioning as a RAS inhibitor (93). A study revealed that the application of FTS to JD1 cells led to diminished RAS-GTP levels and ERK dephosphorylation (94).

Collectively, inhibiting RAS posttranslational modifications presents a viable strategy to disrupt its oncogenic signaling. Despite challenges posed by alternative lipid modification pathways and compensatory mechanisms, advancements in FTIs, GGTIs, and other inhibitors continue to show promise. Continued research and clinical trials are essential to optimize these strategies and improve therapeutic outcomes for RAS-driven cancers, in particular, leukemia.

RNAi is an effective technology for silencing or inhibiting the expression of target genes. Advances in RNAi have been shown to inhibit growth in both cellular and mouse models of tumorigenesis (95–98). Innovative siRNA delivery systems, such as the Local Drug EluteR (LODER) platform, facilitate sustained siRNA release to tumor sites for more than 70 days. The use of LODER to deliver KRAS^G12D-specific siRNAs (siG12D LODER) has been proven effective in reducing pancreatic cancer cell growth and improving survival in mouse models (99). Clinical trials have shown promise, with siG12D-LODER combined with chemotherapy found to be safe and effective in treating locally advanced pancreatic cancer, indicating a median overall survival rate of more than 15 months (100). Additionally, antisense oligonucleotides offer an alternative approach for targeting RAS. In particular, AZD4785 (101), a chemically modified antisense oligonucleotide with a 2′-4′ constrained ethyl modification, has demonstrated strong KRAS suppression in tumor models without the need for a specialized delivery vehicle, which could potentially reduce immune reactions or rapid clearance. CRISPR/Cas9 genome editing has been employed to target specific oncogenic KRAS mutations, which has resulted in tumor shrinkage (102). EXO NRAS (EXO NRAS exosome nonsolid tumor-targeted therapeutic agent) is a biological drug developed by Anling Biotechnology Co., Ltd. At present, it is in preclinical investigation, which is used to treat gliomas and blood tumors. However, these technologies have not yet been reported in leukemia treatment.

PROteolysis TArgeting Chimeras (PROTAC) represent a novel approach to targeting mutant RAS by promoting the degradation of the protein rather than inhibiting its function directly. These bifunctional molecules consist of a ligand that binds to the target protein and another that recruits an E3 ubiquitin ligase, leading to the ubiquitination and subsequent proteasomal degradation of the target protein. LC-2, a bifunctional molecule, establishes a covalent bond with KRAS^G12C and recruits an E3 ligase, leading to the ubiquitination and subsequent degradation of the protein (103). Furthermore, PROTACs incorporating ARS-1620 have demonstrated the ability to degrade GFP-tagged KRAS^G12C within cells (104). Tag-based PROTACs, employing CRISPR/Cas-mediated approaches, facilitate the degradation of tagged proteins, thus expanding the toolkit for RAS targeting (105–109). KRAS G12C/D/V mutations PROGRAM (AnHorn Medicines) is a PROTAC developed by AnHorn Biopharmaceuticals Co., Ltd. It is a KRAS G12C, G12D, and G12V degrader. Currently, it is in preclinical investigation, which is used to treat lung cancer, melanoma, colorectal cancer, blood tumors, and pancreatic cancer. Despite their promises, PROTACs have not yet been reported in the study of leukemia.

Aberrant proteins derived from mutant RAS can elicit immune responses, underscoring the potential of immunotherapy in treating RAS-mutant cancers. Adoptive cell therapy involves harvesting a patient’s immune cells, such as T cells, and then amplifying these cells after their isolation from blood or tumor tissue. Subsequently, these T cells are modified or activated ex vivo to enhance their tumor-fighting abilities (110). Several ongoing clinical trials are evaluating the effectiveness and safety of immunotherapies that target mutant RAS proteins. For example, T cells engineered from the peripheral blood to target KRAS^G12D and KRAS^G12V mutants are undergoing trials for rectal and pancreatic cancers (NCT03745326 and NCT03190941). A phase I/II clinical trial is currently active, utilizing G12V-specific T-cell therapy for advanced pancreatic cancer (NCT04146298). A case report described the complete regression of lung metastases in a patient with a metastatic colorectal cancer following treatment with CD8⁺ T cells specific to the KRAS^G12D mutant (111).

Cancer vaccines operate by stimulating the immune system to target RAS-mutant cancer cells. Peptide-based vaccines employ peptides derived from RAS-mutant proteins to provoke an immune response against these particular cancer cells. Two phase I clinical trials are underway, utilizing KRAS-targeted peptide vaccines; one is prophylactic for patients at risk of developing pancreatic cancer (NCT05013216), whereas the other targets patients with resected microsatellite stable pancreatic or colorectal cancer (NCT04117087). Additionally, a clinical trial explored an mRNA cancer vaccine (V941) targeting common KRAS mutants (G12D, G12V, and G12C) in participants with advanced or metastatic NSCLC, colorectal cancer, or pancreatic adenocarcinoma (NCT03948763). Combining cancer vaccines with immune checkpoint inhibitors or other immunomodulatory agents may amplify their therapeutic efficacy. Furthermore, Sahin and colleagues (112) introduced the concept of individualized mutanome vaccines and implemented an RNA-based poly-neoepitope approach to mobilize immunity against a spectrum of cancer mutations. Although promising, the clinical application of adoptive cell therapy and cancer vaccines targeting RAS mutations remains in an exploratory phase, necessitating additional research and clinical trials to refine and evaluate their potential.

Another strategy involves identifying synthetic lethal interactors that are specific to cells with mutant RAS (113). Mutationally activated RAS genes represent one genetic alteration, and the ablation of a second gene’s expression in cancer cells, mediated by RNAi, constitutes the second hit. Because normal cells do not possess mutant RAS, genes identified through this approach should, in principle, be selectively lethal to tumor cells rather than to normal cells (82, 113, 114). Studies identified that genes encoding the SYK and RON tyrosine kinases and the STK33 and TBK1 serine/threonine kinase as synthetic lethal partner of mutated KRAS (115, 116). A recent study utilizing CRISPR/Cas9 screening techniques has identified synthetic lethal interactions between genes implicated in RAS processing and MAPK signaling pathways and oncogenic RAS mutations in both human and mouse leukemia cell lines (117). Targeting mutant RAS involves various strategies beyond direct inhibition, including PROTACs for protein degradation, immunotherapies like adaptive cell therapy and cancer vaccines, and identification of synthetic lethal interactors. These approaches hold promise but require further research and clinical trials to optimize their efficacy and safety, particularly in the context of leukemia.

The challenge of directly targeting RAS proteins has stimulated extensive research efforts to modulate RAS-associated effector pathways in tumors. Targeting molecules of the RAS signaling pathway, including downstream effectors and upstream regulators, has been explored as a potential therapeutic approach for leukemias with RAS mutations (118). PI3K inhibitors have shown therapeutic potential in specific neoplasms characterized by RAS mutations (119). However, combined inhibition of MEK and PI3K in a mouse model of NRAS-mutant AML failed to induce leukemic cell death, suggesting that other RAS effectors provide critical support to leukemic cells (120, 121).

MEK inhibitors, such as trametinib and cobimetinib (120, 122), have been designed to inhibit the MEK/ERK signaling cascade. However, the administration of ERK or MEK inhibitors in patients with NRAS-mutant leukemia has not consistently yielded the anticipated therapeutic outcomes. Jain and colleagues (123) reported an absence of response to the MAPK inhibitor selumetinib (AZD6244) in three patients with NRAS-mutant AML. Additionally, numerous in vitro and in vivo studies have indicated that myeloma and leukemia cells with NRAS mutations demonstrate resistance to small molecules targeting KRAS^G12C (13, 14, 56, 67). Collectively, this evidence underscores the complexity of NRAS downstream signaling and the role of compensatory mechanisms as significant obstacles to precise molecular targeting.

RAS-GEF inhibitors constitute an active area of research, although developing selective and potent RAS-GEF inhibitors continues to pose a significant scientific challenge. Additionally, BAY-293 (124), BI 1701963 (125), and BI-1701963 (NCT04111458) were found to be inhibitors binding to the catalytic site of SOS1 and thereby blocking the interaction with KRAS-GDP, thereby hindering the exchange from RAS-GDP (inactive form) to RAS-GTP (active form).

Indirect RAS-targeting strategies exploit inhibitors of key proteins in the RAS pathway to diversify the approaches for combating RAS-driven malignancies. At the same time, the complexity and dynamic nature of signaling networks also require attention. Understanding how cancer cells adapt to the inhibition of specific signaling proteins will guide the focus of future efforts on strategies that target specific signaling networks at multiple levels (126). Combination therapies incorporating RAS pathway inhibitors alongside other targeted agents or chemotherapeutics are currently being investigated to overcome resistance and enhance therapeutic efficacy in leukemias harboring RAS mutations (127). Lastly, ongoing advancements in unbiased genome-wide functional screening are expected to yield novel and practical strategies (128).

Targeting the RAS pathways in leukemia involves both direct and indirect strategies. Although inhibitors of the RAS signaling pathway, such as MEK and PI3K inhibitors, have shown potential, their effectiveness is often limited by compensatory mechanisms and resistance. RAS-GEF inhibitors and combination therapies represent promising approaches to enhance therapeutic efficacy. Continued advancements in functional screening and a deeper understanding of signaling networks will be crucial in developing effective treatments for RAS-driven leukemias.

The RAS gene family encodes distinct isoforms, each possessing unique biochemical characteristics (126). RAS’s picomolar affinity for GTP and the challenge of designing small molecules capable of restoring mutant RAS’s defective GTPase activity have thwarted the successful development of direct inhibitors of oncogenic RAS (118). The effectiveness of isoform-specific therapeutic agents may be influenced by these isoforms.

RAS mutants exhibit distinct biochemical characteristics, especially in their affinities for the active or inactive states. Recent research has concentrated on developing drugs that target the GTP-bound state of RAS. RAS mutations introduce variability into RAS protein function and downstream signaling, exhibiting unique biochemical properties and signaling profiles that can significantly alter the response to therapies designed to target specific RAS mutants. For example, some RAS mutants may preferentially activate the MAPK pathway, whereas others might exert a more pronounced effect on the PI3K/AKT pathway. Qian and colleagues (18) observed that not all NRAS mutants contribute to leukemogenesis, suggesting that a subset of NRAS mutants may serve as targets. They discovered that various NRAS G12 mutants activated distinct downstream signaling pathways, potentially elucidating the observed heterogeneity in responses among these mutants. Specifically, they reported that Ba/F3 cells transformed by NRAS mutants exhibited differential response to daunorubicin, trametinib, and ruxolitinib (18). Qian and colleagues (18) found that NRAS^G12D ALL cells were more resistant to daunorubicin than those with wild-type NRAS, which was in line with previous reports (31, 38). Furthermore, these cells also demonstrated variable resistance to a pan-RAS inhibitor (Pan-Ras-IN-1) and other RAS-targeted compounds. These findings underscore the complexity of targeting strategies across different NRAS G12 mutants (18). The development of pan-RAS inhibitors, aiming to target a broad spectrum of RAS mutants regardless of their distinct biochemical properties, is a focus. These inhibitors are designed to disrupt common molecular interactions, including those involving downstream signaling partners or membrane associations.

Specific to the KRAS^G12C mutant, sotorasib represents a significant breakthrough, although it presents challenges. Its application is limited to single mutational profile and drug resistance (129). Intercellular variability or intratumoral heterogeneity is the main factor leading to KRAS^G12C inhibitor resistance (130). Increased tumor heterogeneity and cancer stem cells are associated with resistance and relapse in many tumor types, including AML and ALL (131, 132). Xue and colleagues (133) demonstrated that a subpopulation of cells synthesized the new KRAS^G12C protein, rather than the wild-type KRAS protein, and the EGFR or aurora kinase signals can maintain the newly expressed KRAS^G12C protein in the active GTP-bound form, thereby evading KRAS^G12C inhibitor ARS1620 treatment in KRAS^G12C-mutant NSCLC cell line. Another study suggested that wild-type RAS activation mediated by multiple RTKs, rather than a single RTK, is responsible for the acquired resistance of KRAS^G12C inhibitors (ARS1620 and sotorasib) in various types of cancer cell lines (134). Many independent studies demonstrated the importance of PTPN11/SHP2 as a common downstream of RTKs to activate the wild-type or mutant KRAS protein in mediating acquired drug resistance (130, 133–136). In addition, the activation of the PI3K/AKT/mTOR pathway contributes to the development of sotorasib resistance in human PDAC cell line in vitro and in xenograft mouse model (137). Similarly, acquired resistance to adagrasib has been reported in a patient with NSCLC, which is related to the reactivation of the RAS/RAF/MEK/ERK pathway (69). Induction of epithelial-to-mesenchymal transition promotes resistance to sotorasib or ARS1620 through the activation of the PI3K or ERK pathway in NSCLC cells in a cell type-dependent manner (135, 138). Leukemia cells may activate alternative signaling pathways in response to RAS inhibition, enabling cellular survival and proliferation despite therapeutic intervention. For example, the upregulation of pathways such as PI3K/AKT/mTOR or MAPK, which are independent of RAS signaling, can lead to therapeutic resistance. In addition, intrinsic or adaptive resistance may be caused by concurrent genetic changes, such as secondary KRAS mutations and other genetic mutations, which are not targeted by KRAS^G12C inhibitors (130). Awad and colleagues (139) studied the adagrasib resistance mechanisms of a cohort of 38 patients including 27 with NSCLC, 10 with colorectal cancer, and one with appendiceal cancer. Resistance mechanisms were found in 17 patients of whom seven exhibited multiple coincident mechanisms. Several had new KRAS mutations including G12D, G12R, G12V, G12W, G13D, Q61H, H95D, H95Q, H95R, and Y96C. Several had acquired bypass mechanisms including MET amplification and NRAS, BRAF, MAP2K1, and RET mutations. Several developed oncogenic fusion proteins involving ALK, BRAF, CRAF, FGFR3, and RET (139). Owing to the variety of resistance mechanisms involving KRAS mutations and several bypass pathways, a single antagonist cannot be developed to counteract the resistance to adagrasib (80). Especially, KRAS^Y96D mutant directly affects the binding of adagrasib to the switch II pocket, thereby conferring resistance to sotorasib, adagrasib, or ARS-1620 in multiple cancer cell lines (H358, MIAPaCa2, and BaF3; ref. 69). In contrast, RM-018 [KRAS^G12C(ON) inhibitor] retains potent inhibitory activity against tumor cells harboring dual KRAS^G12C/Y92D mutation, which provides a novel insight to tackle KRAS (69). Awad and colleagues (139) found that the H95D, H95Q, or H95R mutations, which are insensitive to adagrasib, do not confer resistance to sotorasib in vitro. In addition to KRAS^Y96D, Y96S and Y96C also contribute to the resistance to sotorasib or adagrasib in BaF3 cells, and this process can be reversed by the combined use of SOS1 inhibitors (BI-3406; refs. 125, 140). To surmount resistance and improve treatment outcomes, combination therapies targeting multiple signaling pathways or exploiting specific vulnerabilities may be necessary (133, 141). Elucidating resistance mechanisms specific to leukemia subtypes involving RAS mutations is crucial for developing strategic therapeutic approaches.

siRNA directed against mutant RAS mRNA constitutes a viable strategy for RAS-targeted therapy. Although siRNA-based therapies targeting RAS are under active investigation, they encounter challenges, including efficient tumor cell delivery, avoidance of off-target effects, and the potential to trigger immune responses. Researchers are also exploring the combination of siRNA with other treatments to enhance therapeutic efficacy and overcome resistance. As research advances, refining siRNA delivery and enhancing their specificity and clinical efficacy continue to be focal points.

Drugging RAS remains a formidable challenge due to the unique biochemical properties of RAS isoforms, the complexity of downstream signaling pathways, and the adaptive resistance mechanisms in cancer cells. Developing effective pan-RAS inhibitors, combining therapies, and refining siRNA-based strategies are critical areas of ongoing research. Understanding the specific resistance mechanisms in leukemia and other cancers will be crucial for advancing therapeutic approaches and improving patient outcomes.

Although there are many RAS inhibitors in preclinical and clinical settings, they are primarily targeted at solid tumors, and research on leukemia is lacking. Given some significant advantages in zebrafish over other models, many models of leukemia and other hematologic malignancies have been developed in the zebrafish, which provide a powerful resource for investigation of the molecular basis of human leukemia. In particular, the high conservation between human and zebrafish hematopoiesis has stimulated the development of zebrafish models for human hematopoietic malignancies to elucidate molecular pathogenesis and to expedite the preclinical investigation of novel therapies. Although T-ALL was the first transgenic cancer model in zebrafish, a wide spectrum of zebrafish models of human hematopoietic malignancies has been established since 2003, largely through transgenesis and genome-editing approaches. Liu and colleagues (142) showed that KRAS knockdown resulted in specific hematopoietic and angiogenic defects. Alghisi and colleagues (143) used a zebrafish genetic model to induce the expression of oncogenic RAS in endothelial cells and observed an increased number of immature hematopoietic cells and arrest of myeloid maturation in kidney marrow in few surviving juveniles. Shen and colleagues (144) found that MYCN overexpression in zebrafish promoted cell proliferation and enhanced the repopulating activity of myeloid cells and the accumulation of immature hematopoietic blast cells and identified that MAPK pathway was upregulated. Zebrafish has been used to identify new oncogenic cofactors that drive T-ALL pathogenesis by different mechanisms (145–147). Gutierrez and colleagues (148) used zebrafish to identify new therapeutics for T-ALL. Pruvot and colleagues (149) demonstrated that xenografted leukemic cells in zebrafish embryos are pharmacologically relevant models for screening nonteratogenic drugs. Among antileukemic drugs, they found that imatinib and oxaphorines decrease the leukemic burden in xenografted animals. After screening 26,400 molecules, Ridges and colleagues (150) found that Lenaldekar eliminates immature T cells in genetically engineered, T-cell reporting zebrafish and induces long-term remission in adult zebrafish with T-ALL. Lang and colleagues (151) thought that miR-124 is a potent suppressor of NRAS in a transgenic zebrafish model. They found that Ectopic Viral Integration Site-1 upregulated NRAS expression, consequently activating the MAPK pathway by epigenetically silencing miR-124 (151).

Zebrafish models have significantly advanced our understanding of leukemia and hematopoietic malignancies. These models have not only elucidated the molecular pathogenesis of these diseases but have also provided a platform for preclinical drug screening. The continued development and refinement of zebrafish leukemia models will likely yield further insights into disease mechanisms and therapeutic strategies, ultimately contributing to better clinical outcomes for patients with hematopoietic malignancies.

The targeting of RAS proteins has long posed a significant challenge due to their “undruggable” nature, attributed to the lack of apparent drug-binding pockets and the high GTP-binding affinity. Initial strategies to inhibit RAS involved developing GTP antagonists and small molecules to mimic RAS-GAP, but these efforts were unsuccessful (118). However, extensive research has yielded various strategies to target RAS, including direct engagement with chemical compounds, interference with upstream activators, inhibition of downstream effectors, RNAi to reduce RAS expression, exploitation of unique metabolic processes, immune system engagement, and discovery of synthetic lethal interactions. A notable advancement in RAS targeting is the FDA approval of KRAS^G12C inhibitors for the treatment of patients with NSCLC, demonstrating the potential for directly targeting previously “undruggable” targets. Also, inhibitors of both active and inactive forms of KRAS^G12C, such as FMC-376 and BBO-8520, provide a differentiated mechanism of action with the potential for broader and more durable response in the clinic (78).

Current research continues to explore novel strategies for modulating RAS signaling pathways. Chimeric antigen receptor T-cell therapy and cancer vaccines are in clinical trials, whereas adoptive cell therapy and siRNA show promise in certain cancers. However, the application of these strategies specifically for RAS mutations in leukemia remains under investigation. Pan-RAS inhibitors and PROTACs offer innovative methods for targeting RAS-driven cancers, but their clinical development requires further research to mitigate resistance and ensure safety. The therapeutic efficacy of KRAS inhibitors in leukemia has yet to be confirmed in preclinical and clinical settings, though activating RAS mutations are linked to disease progression and drug resistance in AML, suggesting potential future research directions.

Over the past two decades, genetically modified zebrafish have been instrumental in modeling human hematopoietic malignancies, including T-ALL, B-ALL, myeloproliferative neoplasms, myelodysplastic syndromes, and AML. Zebrafish offers a cost-effective system for experimental tumor therapy and large-scale anticancer drug screening. Their capacity to image cancer cells and niche biology in an endogenous tumor context makes them indispensable for understanding RAS-driven tumorigenesis and screening RAS inhibitors to evaluate potential antitumor efficacy.

In leukemia, RAS-directed therapy aims to disrupt the aberrant signaling driven by RAS mutations. RAS signaling varies across tissues and cell types, thereby differentially influencing the role of RAS mutations on cancer biology, complicating RAS-targeting drug development (152). In RAS-targeted leukemia treatment, the emergence of new mutations has a significant impact on therapeutic efficacy and clinical outcomes. RAS mutations are prevalent across various leukemia subtypes, contributing to disease progression and resistance to treatment. An in-depth understanding of the role of RAS mutation in leukemogenesis is imperative for therapeutic interventions. However, the association between RAS mutations and the onset and progression of leukemia remain a subject of debate. The relationship between RAS mutations and relapse in leukemia is complex, involving clonal evolution, resistance mechanisms, and interactions with other genetic alterations, posing challenges for targeted therapies. The presence of new mutations can either compromise or enhance the response to RAS-targeted interventions. Understanding these dynamics is critical for targeting RAS therapies in leukemia. Genomic profiling and molecular diagnostics are crucial for identifying new mutations and guiding subsequent therapeutic choices. It is important to consider the genetic alterations present and the mechanisms underlying treatment resistance in identifying the most suitable treatment approaches based on the individual tumor’s molecular signature. Identifying genomic characteristics and following specific mutation targeting strategies should be the treatment approach for leukemia in the future.

K.D. Westover reports grants from the NIH and CPRIT during the conduct of the study, personal fees and other support from Vibliome Therapeutics, other support from Velorum Therapeutics, and personal fees from Sanofi, Amgen, and AstraZeneca outside the submitted work. Z. Zhou reports grants from CPRIT RP220145 and NIH R01CA244341 during the conduct of the study. L. Shu reports grants from the National Natural Science Foundation of China, the Guizhou Province’s Science and Technology Major Project, Doctor Start-up Fund of Affiliated Hospital of Guizhou Medical University, the Guizhou Medical University Key Laboratory Project, and Young Talents Plan of Guizhou Medical University during the conduct of the study. The other authors reported no disclosures.

This project is supported by CPRIT RP220145 and NIH R01CA244341 (to K.D. Westover), the National Natural Science Foundation of China (82303191 to L. Shu, 32160168 to Y. Chen), Doctor Start-up Fund of Affiliated Hospital of Guizhou Medical University (gyfybsky-2023-16 to Y. Chen), and Young Talents Plan of Guizhou Medical University (22QNRC20 and XBHJ[2023]016 to Y. Chen), and in part by the Guizhou Province’s Science and Technology Major Project (QianCXTD[2021]002 and GCC[2023]034 to L. Shu) and the Guizhou Medical University Key Laboratory Project ([2024]004, to L. Shu).