AXL at the nexus of therapeutic resistance, tumor vascularization, and immune evasion in cancer

Summary

AXL, a member of the TAM (TYRO3, AXL, and MERTK) receptor tyrosine kinase family, has emerged as a high-priority therapeutic target in oncology. Unlike canonical oncogenic kinase drivers, AXL activation is rarely attributable to direct genetic alterations but instead arises through adaptive, non-genetic mechanisms including hypoxia, inflammation, epithelial-to-mesenchymal transition, and therapeutic selection pressure. In this review, we synthesize the current understanding of AXL biology and the translational opportunities for AXL inhibition across multiple malignancies, with a focus on glioblastoma. We explore the expanding landscape of AXL-directed therapeutics, discuss the rationale for combinatorial strategies, and highlight emerging areas of AXL regulation including the AXL-ferroptosis axis. The current understanding of AXL suggests it functions not as a bystander but as an active mediator of tumor adaptation. Successful targeting of AXL has the potential to simultaneously suppress cancer cell survival, angiogenesis, and epithelial-to-mesenchymal transition while reinvigorating antitumor immunity.

Graphical abstract

Therapeutics; Microenvironment; Cancer

Introduction

Distinct from most oncogenic receptor tyrosine kinases, AXL, a member of the TAM (TYRO3, AXL, and MERTK) family, is rarely activated by a primary driver mutation. Instead, AXL becomes upregulated under therapeutic pressure or other stressors, positioning it as a uniquely adaptive mediator of treatment failure and thus a promising therapeutic target across diverse tumor types.¹ AXL activation by its high-affinity ligand, growth arrest-specific protein 6 (GAS6), in conjunction with phosphatidylserine (PtS), plays essential roles in the clearance of apoptotic debris and negative regulation of innate immune responses.² In cancer, abundant GAS6 and PtS combine with aberrant AXL expression to drive malignant transformation and immune evasion in tumor types that include glioblastoma (GBM),^1,3,4 non-small cell lung cancer (NSCLC),^1,5 acute myeloid leukemia (AML),^1,6,7 triple-negative breast cancer (TNBC),^1,8 gastric and colorectal adenocarcinomas,^1,9,10 and prostate cancer.^1,11 AXL-activating PtS is prevalent in cancer due to tumor-intrinsic cell death induced by nutrient limitation as well as therapy-induced cytotoxic responses. In parallel to tumor release of PtS, mechanisms such as hypoxia, epithelial-mesenchymal transition (EMT), and inflammation can upregulate AXL expression.^12,13

AXL activates canonical downstream signaling cascades, including phosphatidylinositol 3′-kinase (PI3K)/AKT, MAPK/ERK, JAK/STAT, and NFκB, to promote tumor cell proliferation, survival, migration, and invasion.¹⁴ Functionally, these signaling pathways coordinate cancer progression through cytoskeleton remodeling, manipulation of transcriptional and post-translational programs, and tumor microenvironment (TME) alterations.^5,13,15,16 Further, AXL drives immunosuppressive cellular phenotypes in the TME through tumor cell-intrinsic mechanisms and modulation of various immune cell populations.² Ultimately, AXL coordinates multiple tumor adaptation programs, enabling cancer cells to cope with stress caused by immune surveillance, tumor progression, and therapeutic targeting. Here, we highlight key mechanisms of AXL-driven therapeutic resistance and immune evasion, describe exciting recent advances that link AXL to ferroptosis and sex-dependent biology, and conclude with an analysis of barriers and opportunities for AXL-targeting in oncology. This review organizes AXL signaling across cancers, tumor, immune, stromal cellular compartments, resistance contexts, and sexes into one cohesive conceptual model.

AXL structure and activation

AXL is a single-pass transmembrane protein composed of an extracellular domain, a transmembrane domain, and an intracellular tyrosine kinase domain. The extracellular domain contains pairs of immunoglobulin (Ig)-like domains and fibronectin type III (FNIII) domains.¹⁷ The Ig-like domains of AXL are the principal mediators of interactions with the GAS6 protein ligand. FNIII domains are thought to assist with dimerization and link the Ig-like domains to the transmembrane region, which anchors AXL within the cell membrane (Figure 1).^18,19 The GAS6 protein ligand consists of a γ-carboxyglutamic acid (Gla) domain, four epidermal growth factor (EGF) repeats, and two laminin-G (LG) domains, which are presented in more depth below.

Figure 1.

Structure and functions of AXL

Phosphatidylserine is exposed on the outer leaflet of the cell membrane of apoptotic cells and binds to the gamma-carboxylated Gla domain of GAS6. Four epidermal growth factor (EGF) repeats and two laminin-G (LG) domains compose the rest of the GAS6 protein. The LG domains bind to AXL, which is composed of two immunoglobulin (Ig-like) domains, two fibronectin type III (FNIII) domains, a transmembrane region, and a tyrosine kinase domain. Dimerization and autophosphorylation induce downstream signaling cascades through MMP9, RAC, PI3K/AKT/mTOR, RAS/RAF/MEK/ERK, SRC/FAK, DKK3, Ang-2, JAK/STAT, and NFκB, which induce migration/invasion, proliferation and survival, angiogenesis, immunomodulation, and EMT.

Full AXL activation requires assembly of GAS6 with PtS, illuminating the key function of AXL as a sensor of cell death in the TME. In the course of apoptosis, PtS is exposed on the outer leaflet of the plasma membrane of apoptotic cells. AXL does not directly bind to PtS, as the activating protein ligand GAS6 forms a “bridge” between PtS and AXL. It is worth noting that while GAS6 is the canonical activator of AXL, there is some evidence that other protein ligands such as protein S1 (PROS1) can bind to AXL with lower affinity.⁴ Importantly, the glutamic acid residues in the Gla domain of GAS6 must be γ-carboxylated through a vitamin K-dependent reaction to permit the Gla domain of GAS6 to bind to PtS.²⁰ Then, the tandem globular LG domains of GAS6 bind to the Ig-like domains of AXL, resulting in dimerization between two GAS6-AXL complexes.^17,20,21 Dimerization permits trans-autophosphorylation of tyrosine residues within the intracellular domain, leading to recruitment of SH2-domain containing proteins such as GRB2 and the p85 subunit of PI3K.²² Binding of these proteins facilitates downstream signaling cascades through pathways including PI3K/AKT, STAT, NFκB, MMP9, RAC1, DKK3, Ang-2, and RAS/RAF/MEK/ERK.^17,23,24

Beyond direct ligand activation of AXL, some findings suggest that the AXL receptor tyrosine kinase can partner with alternative receptor tyrosine kinases to initiate or amplify signal transduction. AXL heterodimerization with MET, EGFR, HER2, or HER3 has been shown to increase signaling in response to ligand-induced and oncogenic activation of signal transduction in GBM,²⁵ squamous cell carcinomas of the head and neck,²⁶ and breast cancer.²⁷ However, it is worth noting that these studies did not rule out possible contributions of GAS6 binding to activation of the heterodimerized receptor tyrosine kinases. Further experiments using agents that can reduce or neutralize endogenous GAS6 and PtS, such as warfarin,²⁸ are necessary to rigorously test whether ligand-independent activation of AXL is mediated via receptor tyrosine kinase heterodimerization. Nevertheless, the potential ability to heterodimerize with other activated receptor tyrosine kinases suggests the importance of direct kinase inhibition in strategies to counteract AXL function in tumors.

After activation, AXL can shed its soluble extracellular domain upon cleavage by ADAM family metalloproteases.²⁹ Shedding results in signal dampening via two routes: it removes AXL from the plasma membrane, and the resulting cleaved extracellular domain can serve as a decoy receptor that binds free GAS6, inhibiting the activation of other AXL receptors.³⁰ Furthermore, the soluble extracellular domain of AXL (sAXL) can function as a biomarker for disease progression.³¹ If uncleaved, GAS6:AXL complexes can be internalized via endocytosis.³² Internalized GAS6:AXL complexes can traffic to the lysosomes for degradation, or they may be recycled to the plasma membrane to sustain AXL signaling capability.³³

Post-translational modifications such as phosphorylation and ubiquitination modulate AXL activity.³⁴ The intracellular kinase domain of AXL contains six tyrosine phosphorylation sites, which are phosphorylated upon receptor dimerization and activation to enact downstream signaling. Within the autoregulatory loop, autophosphorylation of Tyr698, Tyr702, and Tyr703 occurs. In the distal cytoplasmic domain, phosphorylation of Tyr779, Tyr821, and Tyr866 results in these residues becoming docking sites for downstream effectors such as the p85 subunit of PI3K, GRB2, and SRC.² Activation of AXL by GAS6 supports ubiquitination by E3 ubiquitin ligases such as c-Cbl and the LZTR1-CRL3 complex, which mark AXL for lysosomal degradation.^35,36,37 A better understanding of the complex web of regulatory mechanisms that govern AXL structure and activation will greatly aid the development of therapeutic strategies to target this receptor tyrosine kinase.

AXL in immunomodulation

Immune homeostasis and disease

Under normal physiologic conditions, AXL and its paralogs TYRO3 and MERTK induce the production of anti-inflammatory mediators while simultaneously suppressing active inflammatory signals.³⁸ In complex with the ligand GAS6, TAM RTKs form a “bridge” between antigen-presenting cells (APCs) and dying cells by binding to PtS on the outer leaflet of the plasma membrane.^20,39 This association promotes engulfment of apoptotic cells by APCs, a process known as efferocytosis (from the Latin effere, “to carry to the grave”). Clearance of apoptotic cells via efferocytosis promotes resolution of inflammation through the release of immunosuppressive cytokines such as interleukin 10 (IL-10) and transforming growth factor β (TGF-β).²¹ Furthermore, AXL signaling acts as a negative regulator of Toll-like receptor (TLR) signaling. TLR activation of type I interferon receptor (IFNAR) and its transcription factor STAT1 upregulates AXL and other TAM RTKs, which inhibits this inflammatory signaling axis by inducing the cytokine and TLR suppressors SOCS1 and SOCS3.^23,40 Thus, AXL acts as a responder to cell death that suppresses inflammation and supports wound healing.

Since AXL signaling enhances efferocytosis and dampens inflammatory immune responses, it follows that AXL deficiency or loss results in the inverse. Chronic inflammation stemming from the buildup of apoptotic cells leads to autoimmune inflammation as exhibited by increased proliferation of innate and adaptive immune cells, and production of autoreactive antibodies.⁴¹ Human autoimmune diseases involving AXL deficiency or dysregulation include multiple sclerosis, Sjögren’s syndrome, systemic lupus erythematosus, and rheumatoid arthritis.³⁸

Relatedly, AXL activation is exploited by infectious agents and promotes the evolution of cardiovascular and neurodegenerative disease. AXL functions as a cellular receptor that mediates entry of Ebola⁴² and Zika viruses.⁴³ Interestingly, preclinical experiments suggest that AXL overexpression promotes SARS-CoV-2 infection through a spike protein interaction, and there is a correlation between SARS-CoV-2 spike protein levels and AXL expression in bronchoalveolar lavage fluid cells from COVID-19 patients.⁴⁴ Furthermore, AXL signaling can influence inflammation, apoptosis, and oxidative stress in the cardiovascular system, exacerbating cardiovascular disease by promoting atherosclerosis, aneurysm rupture, and other pathological processes.⁴⁵ AXL has also been implicated in neurodegenerative disease. sAXL was significantly elevated in Alzheimer’s disease subjects with pathological cerebrospinal fluid amyloid/tau profiles and was negatively associated with brain structural integrity and cognitive function, positioning sAXL as a peripheral inflammatory biomarker that may correspond to specific stages of neurodegeneration.⁴⁶ These findings place AXL at the nexus of autoimmune, cardiovascular, and neurodegenerative disease.

Tumor immune microenvironment

AXL plays critical roles in cell survival and immune regulation. However, in the context of cancer, AXL signaling on both tumor cells and immune cells can reinforce tumor progression and immune evasion. AXL is detected on regulatory T cells, natural killer (NK) cells, myeloid-derived suppressor cells, dendritic cells, monocytes, macrophages, and microglia within tumors, supporting a diverse array of immunosuppressive and tumor-supporting functions.^17,47

AXL signaling on tumor cells results in the downregulation of MHCI and thus interferes with endogenous antigen presentation, resulting in dampened antitumor immune responses.^23,48 Loss of MHCI enables tumor cells to more easily avoid elimination by cytotoxic T cells.⁴⁹ However, this leaves cancer cells vulnerable to NK cells, which survey the body for cells exhibiting “missing self”, or lack of MHCI. In efforts to exploit this cellular mechanism, a clinical trial assessing safety and preliminary efficacy of AXL targeting CAR-NK cells in patients with AXL-positive advanced solid tumors is currently recruiting (NCT05410717).

Tumor cell-intrinsic AXL signaling suppresses NK cell accumulation, activation, and cytotoxic activity.⁴¹ AXL loss in head and neck cancer (HNC) tumor cells encouraged infiltration of NK cells in vivo and heightened NK cell cytotoxicity. Furthermore, NK cell depletion in the AXL knockout (KO) HNC mouse model exacerbated tumor growth, indicating that antitumor NK cell function is restored upon AXL KO. Mechanistically, AXL KO reduced expression of CD73 on HNC cells, which converts AMP into the immunosuppressive molecule adenosine. AXL KO increased expression of the chemokines CXCL10 and CCL5, which are chemoattractants for NK cells that can also increase NK cell-mediated cytotoxicity.⁵⁰ These results suggest that AXL promotes tumor growth by suppressing NK cell recruitment and activity, opening doors for AXL targeting combined with NK cell-based immunotherapies.

AXL signaling in myeloid cells promotes macrophage recruitment and immunosuppressive phenotypes through increased release of anti-inflammatory cytokines and modulation of immune checkpoint receptors.⁴¹ Macrophage AXL signaling upregulates IL-10 and TGF-β expression and secretion, sometimes via efferocytosis of dead cells within the TME.^51,52 This process prevents release of apoptotic cellular contents that could stimulate inflammation and antitumor immune response. Interestingly, PD-L1 expression increases in macrophages following TAM RTK-mediated efferocytosis.^53,54 Inhibiting AXL reduced immunosuppressive CD206⁺ macrophage populations and decreased release of multiple immunosuppressive cytokines, including IL-13, via the transcription factor STAT6 in inflammatory breast cancer.⁵⁵ Similarly, AXL targeting in macrophages blocked CSF-1 production and reduced the frequency of immunosuppressive macrophages in the NSCLC and TNBC TME.⁵⁶ An AXL-IL-11-pSTAT3 axis in macrophages supports lung cancer progression by recruiting CD163⁺ immunosuppressive macrophages.⁵⁷ AXL signaling also upregulated GM-CSF in nasopharyngeal carcinoma, increasing myeloid-derived suppressor cell recruitment.⁵⁸ Thus, AXL signaling promotes pro-tumorigenic macrophage polarization and recruitment of myeloid cells to the TME (Figure 2).

Figure 2.

AXL signaling in both tumor cells and immune cells drives immunosuppression in cancer

GAS6 and phosphatidylserine from dead cells in the tumor microenvironment drive AXL signaling. AXL signaling inhibits release of inflammatory cytokines and promotes release of immunosuppressive cytokines such as IL-10 and TGF-β, resulting in immunosuppressive macrophage polarization and impaired T cell function. AXL upregulates expression of immune checkpoint molecules including PD-1, upregulates forkhead box P3 (FOXP3) to support T regulatory cell function, and downregulates expression of MHC1 to inhibit cytotoxic T cell activation. Chemokines that drive NK cell recruitment and activation are also downregulated by AXL signaling. TME, tumor microenvironment; KIR, killer cell immunoglobulin-like receptor, TCR, Toll-like receptor.

AXL signaling on T cells supports T regulatory cell (Treg) function, inhibits cytotoxic T cell function, and reinforces T cell exhaustion in the tumor environment. Treg functions are supported by immunosuppressive cytokines and chemokines, including IL-10 and TGF-β, released downstream of TAM signaling.^41,59 Furthermore, AXL activation on Tregs upregulates their expression of FOXP3 and enhances Treg suppression of CD4⁺ lymphocytes by altering IL-2 homeostasis.⁶⁰ AXL activation upregulates the expression of immune checkpoint ligands such as PD-L1 on macrophages and tumor cells, which hampers cytotoxic T cell-mediated killing of tumor cells.⁴¹ In terms of suppression of inflammatory mediators, AXL inhibition with bemcentinib resulted in increased tumor cell-intrinsic expression of chemokines associated with CD4⁺ T cell activity, including CXCL9, CXCL10, and IFNγ, in a mouse model of ovarian cancer.⁶¹ Thus, AXL signaling suppresses antitumor functions of cytotoxic T cells and supports immunosuppressive activities of Tregs via upregulation of immune checkpoint ligands and release of anti-inflammatory cytokines (Figure 2).

GBM TME

The role of AXL in the TME is particularly important for the tumor growth of GBM, where the microenvironment is uniquely primed for AXL activation. GBM tumors are highly necrotic at the core due in part to the limited availability of nutrients including oxygen, resulting in prevalent dead cells together with nutrient-limited live cells that exhibit increased levels of PtS on their cell surfaces.⁶² AXL is a key therapeutic target in GBM due to both tumor cell-intrinsic and -extrinsic mechanisms. Blood-derived macrophages and brain-resident microglia are recruited to the site of the tumor and populate the necrotic center, in sufficient numbers that myeloid cells can compose up to 30% of the tumor mass.⁶³ As GBM tumors develop, activated/inflammatory macrophages are repolarized into immunosuppressive/anti-inflammatory macrophages that promote tumor progression.^64,65 Multiple studies have found that macrophage phenotypes within GBM tumors correlate with poor prognosis.^66,67 Altogether, prevalent tumor cell death, macrophage infiltration, and high expression of AXL indicate a therapeutic opportunity keyed on the suppression of AXL function in GBM.

Preclinical experiments suggest that AXL targeting in GBM induces both tumor cell-intrinsic and -extrinsic antitumor outcomes. Knocking down or inhibiting AXL in mesenchymal glioma stem cell (GSC), xenograft mouse models extended survival and reduced infiltration of CD11b⁺ macrophages into tumors.⁴ Incubating AXL-targeting CAR-KHYG-1 NK cells with GBM cells resulted in IFN-γ secretion and specific killing of antigen-positive GBM cells.⁶⁸ Interestingly, a Zika virus live-attenuated vaccine (ZIKV-LAV) can utilize AXL for GBM cell entry, resulting in selective oncolytic activity in GBM cells with minimal effects on non-cancer cells.⁶⁹ These approaches indicate the versatility and benefits of AXL targeting in GBM and its immune compartment. The use of multi-kinase inhibitors such as Q702, which targets TAM RTKs as well as the macrophage-specific target CSF1R, shows great promise for blocking tumor cell-intrinsic and macrophage-intrinsic tumorigenic mechanisms but has not yet been examined in brain tumors.⁷⁰ Further experiments optimizing efficacy and minimizing adverse effects of kinase targeting and biologic agents are necessary to enable the use of these therapeutic strategies in the clinic.

Immune checkpoint blockade

AXL signaling promotes immune escape by increasing PD-L1 expression.^71,72 High AXL expression is correlated with high PD-L1 expression and shorter progression-free survival after anti-PD-1 treatment in clear cell renal cell carcinoma (ccRCC) patients.⁷³ Interestingly, treatment with anti-PD-L1 antibodies stabilized an interaction of PD-L1 with AXL, promoting cell proliferation in some subtypes of liver cancer and indicating that AXL is activated to mediate resistance to PD-L1 neutralizing therapy.⁷⁴ Considered together with the function of AXL in efferocytosis-associated PD-L1 expression noted above, targeting both AXL and immune checkpoint ligands or receptors is a rational approach for combination therapy in cancer.

Accordingly, AXL inhibition has been combined with immune checkpoint blockade in various preclinical and clinical studies. CTLA-4 blockade lengthened survival in mesenchymal-like GBM in vivo.⁷⁵ Tumor suppression was mediated by CD4⁺ T cells and microglial phagocytosis of GBM cells resulting from AXL/MERTK signaling.⁷⁵ In contrast, combining the AXL inhibitor bemcentinib with an anti-PD-1 antibody resulted in extended overall survival in an immunocompetent GBM mouse model.⁴ Consistent with its role in promoting therapy resistance, AXL inhibition re-sensitized hepatocellular carcinoma (HCC) tumors to anti-PD-1 antibodies in vivo.⁷⁶

The AXL inhibitor bemcentinib was tested in combination with pembrolizumab in phase I and II trials for NSCLC and TNBC. Though the TNBC study was terminated due to lack of clinical efficacy (NCT03184558), the NSCLC trial reported a 26% overall response rate for patients that received bemcentinib and pembrolizumab compared to a platinum-containing chemotherapy control arm, with durable responses (>12 months) in AXL-positive tumors⁷⁷ (NCT03184571). These findings indicate that AXL targeting combined with immune checkpoint blockade is a promising therapeutic strategy in specific contexts.

AXL in angiogenesis

Aside from its role in immunomodulation, AXL promotes angiogenesis through both tumor cell and stromal cell compartments in multiple cancer types, including breast cancer and GBM (Table 1). In breast cancer, the HIF-1α/VEGF, Ang2/Tie2, and DKK3 pathways connect AXL to angiogenesis. A prospective study examined 261 invasive breast cancers immunohistochemically and found that strong HIF-1α expression was significantly associated with increased AXL and VEGF expression, a higher proliferating microvessel density, greater tumor proliferation as measured by Ki-67 staining and correlated with a worse prognosis.¹⁰⁹ Interestingly, AXL has been reported as a mediator of tumor angiogenesis via multiple mechanisms beyond VEGF signaling. AXL inhibition in TNBC decreased secretion of pro-angiogenic factors thrombospondin-1, endothelin-1, and VEGF and impaired endothelial cell functional properties both in vitro and in vivo.¹¹⁰ Additional work provided early in vivo evidence that AXL knockdown in umbilical vein endothelial cells and commercially available lung and breast cancer lines impaired tube formation, and this effect was additive with anti-VEGF therapy.¹¹¹ Further analysis demonstrated that AXL regulates endothelial cell functions through modulation of the angiopoietin/Tie2 and DKK3 pathways, suggesting that blocking AXL could enhance antiangiogenic therapy.¹¹¹ This additive effect of AXL inhibition with anti-VEGF therapy indicates that AXL functions through mechanisms both overlapping and distinct from VEGF-mediated angiogenesis.

Table 1.

AXL functions stratified by cancer type

Cancer type	AXL interaction	Resistance context	Treatment implication	Reference(s)
Non-small cell lung cancer	AXL dimerizes with EGFR and HER3; osimertinib paradoxically activates AXL by inhibiting a negative feedback loop; AXL/CDCP1/SRC axis; cancer stem cell self-renewal via SCD1; AXL promotes immunosuppressive TME by suppressing cGAS/IFN signaling via AKT; associated with STK11/KRAS co-mutations	acquired resistance to osimertinib and other EGFR-tyrosine kinase inhibitors (TKIs); reduced benefit from anti-PD-1 therapy (HR 4.15 for PFS); T cell exclusion	AXL inhibitor + osimertinib to restore EGFR-TKI sensitivity; AXL inhibitor + anti-PD-1 ± chemotherapy to convert immune-cold tumors	Taniguchi et al.,78; Murakami et al.,79; Pan et al.,80; Dhakal et al.,81; Ancel et al.82
Melanoma (BRAF-mutant)	low MITF/high AXL ratio predicts early resistance to multiple targeted drugs; AXL/AKT axis mediates BRAF inhibitor (BRAFi) resistance in PTEN-WT tumors; AXL-high clones resistant to both BRAF and PD-1 therapy	intrinsic and acquired resistance to BRAF/MEK inhibitors; phenotype switching between MITF-high (proliferative) and AXL-high (invasive) states	AXL inhibitor (e.g., bemcentinib) + BRAFi/MEKi to target heterogeneous tumor populations; AXL-ADC (EnaV) to eliminate AXL-high resistant clones	Müller et al.,83; Zuo et al.,84; Boshuizen et al.85
Acute myeloid leukemia (FLT3-ITD)	AXL sustains constitutive FLT3 phosphorylation; AXL upregulated within hours of FLT3 TKI treatment via STAT5 and hypoxia in the bone marrow niche	adaptive and acquired resistance to FLT3 inhibitors (quizartinib, midostaurin, and AC220)	dual FLT3/AXL inhibition (e.g., gilteritinib) or AXL inhibitor + FLT3 TKI to overcome niche-mediated resistance	Seale et al.,86
Triple-negative breast cancer	AXL dimerizes with EGFR in TNBC; promotes EMT and invasiveness; AXL dimerizes with EGFR to activate PLC-γ/PKC cascade, sustaining mTORC1 activity	resistance to anti-EGFR therapies and PI3K inhibitors	AXL inhibitor + anti-EGFR to restore drug sensitivity, AXL inhibitor + PI3K inhibitor to block/ bypass mTORC1 signaling	Asiedu et al. 87; Scaltriti et al.12
Pancreatic ductal adenocarcinoma	AXL activates TBK1-NFκB pathway; promotes immune suppression and fibrotic TME; sustains tumor cell plasticity	resistance to gemcitabine and targeted therapies; immunosuppressive microenvironment	AXL inhibitor (e.g., bemcentinib) + gemcitabine to drive epithelial differentiation and enhance chemosensitivity	Cruz et al. 88; Liu et al.89; Kirane et al.28
Ovarian cancer (high-grade serous)	AXL blockade restores DNA damage (γ-H2AX)	resistance to PARP inhibitors (olaparib)	AXL inhibitor + PARPi to overcome PARPi resistance	Kariolis et al.90
Clear cell renal cell carcinoma	High AXL expression associated with lower response rates to nivolumab; AXL co-expressed with PD-L1, especially in VHL-inactivated tumors	resistance to anti-PD-1 therapy; worst OS in AXL+/PD-L1+ subgroup	AXL inhibitor or AXL decoy receptor (batiraxcept) + cabozantinib ± nivolumab	Terry et al.73
Head and neck squamous cell carcinoma	AXL overexpressed and tightly associated with EGFR in resistant cells; AXL-EGFR positive feedback loop via MAPK/c-Jun; AXL regulates NRG1 transcription to activate HER3; AXL dimerizes with EGFR to activate PLCγ/PKC-mTOR axis; AXL-PI3K-PD-L1 signaling axis elevated in radiation-resistant patient-derived xenografts (PDXs); reduced miR-34a drives AXL upregulation with IL-6/IL-8 inflammatory signaling	resistance to cetuximab (via tyrosine 821 and c-ABL kinase pathway), erlotinib (EMT phenotype; HR 1.66 for high AXL mRNA, p = 0.007), PI3Kα inhibitors (in PIK3CA-mutant/amplified SCC), and radiation (PD-L1-high tumors: 50%–70% local failure rates vs. 20%–25% in PD-L1-low)	AXL inhibitor (bemcentinib) + cetuximab or erlotinib; dual AXL/c-ABL inhibition; combined PI3Kα + AXL or PKC inhibitors; AXL inhibitor + radiation ± anti-PD-1; miR-34a restoration to abrogate AXL expression	McDaniel et al.91;Brand et al.,92; Brand et al. 93; Iida et al.,94; Gileset al.,95; Skinner et al.,96; Elkabets et al.,26
Esophageal squamous cell carcinoma	AXL dimerizes with EGFR to activate PLCγ-PKC-mTOR signaling	resistance to PI3Kα inhibition	AXL or EGFR or PKC inhibitor + PI3Kα inhibitor	Elkabets et al.26
Hepatocellular carcinoma	AXL phosphorylation increased 4-fold as adaptive feature of chronic sorafenib treatment; AXL expression correlates with EMT; AXL upregulation in sorafenib- and lenvatinib-resistant HCC suppresses TNF-α and STING-type I IFN pathways; AXL synergizes with ErbB2-4 receptors in regorafenib-resistant cells; Galectin-1 promotes AXL/MET phosphorylation to inhibit ferroptosis; circulating AXL levels correlate with shorter survival in sorafenib-treated patients	acquired resistance to sorafenib, lenvatinib, and regorafenib; reduced response to anti-PD-1 immunotherapy after TKI progression; resistance to ferroptosis	AXL inhibitor (bemcentinib) + sorafenib to restore TKI sensitivity; AXL inhibitor + anti-PD-1 to re-sensitize TKI-resistant HCC to immunotherapy; pan-ErbB inhibitor (afatinib) + regorafenib as synthetic lethal combination in AXL-expressing HCC; AXL inhibitor + PARPi based on AXL-RPA2-CHK1 HR repair axis	Pinato et al.97; Xie et al.,76; Breitenecker et al. 98; Hsu et al.99; Li et al.,100
Gastric cancer	cancer-associated fibroblasts (CAFs) are a major source of GAS6, which activates AXL on gastric cancer cells promoting mesenchymal differentiation and cell survival; CAF-derived GAS6 induces chemoresistance via AXL/STAT3/ABCG1 drug efflux pathway; GAS6/AXL pathway contributes to immunosuppressive TME by reducing T-cell activation; in HER2-aberrant gastric cancer, AXL is activated by HER2-TKIs and maintains cell survival via SHC1 axis; MET and AXL co-activation mediates afatinib resistance	resistance to chemotherapy via ABCG1-mediated drug efflux; resistance to immunotherapy (GAS6 negatively correlated with immune activation genes); adaptive resistance to HER2-TKIs in HER2-aberrant tumors	AXL inhibitor (bemcentinib) + chemotherapy to overcome CAF-mediated chemoresistance; AXL inhibitor + anti-PD-1 to restore immunogenic TME; AXL inhibitor + HER2-TKI in HER2-aberrant gastric cancer with high AXL expression; cabozantinib (MET/AXL inhibitor) + afatinib for dual MET/AXL-driven resistance	Bae et al.101; Ishida et al.,102; Kim et al.,10; Zhai et al. 11; Yoshioka et al.103
Brain metastases	lung adenocarcinoma: TAZ-AXL-ABL2 autocrine feedforward signaling loop; nuclear TAZ drives AXL and ABL2 expression; ABL2 phosphorylates TAZ to sustain nuclear localization. Colorectal cancer: AXL expressed in 72.7% of brain metastases with stable or increased expression compared to paired primary tumors; AXL localized primarily to endothelial cells; GAS6 highly expressed in brain microenvironment (microglia). HER2+ breast cancer: AXL required at multiple steps of the metastatic cascade including intravasation, extravasation, and growth at metastatic site; AXL sustains EMT and TGF-β-induced invasion.	AXL required for brain metastasis colonization in lung adenocarcinoma. AXL expression maintained or increased from primary CRC to brain metastasis. AXL expression correlates with poor survival in HER2+ breast cancer.	BBB-penetrant ABL allosteric inhibitor or AXL inhibitor to disrupt TAZ-AXL-ABL2 loop. TAM receptor inhibition + immune checkpoint blockade for CRC brain metastases. pharmacological AXL inhibition to decrease metastatic burden in HER2+ breast cancer.	Noblanc et al.,104; Hoj et al. 105; Goyette et al.106
Glioblastoma	AXL/GAS6 co-expressed in tumor cells and vasculature; AXL drives mesenchymal subtype; upregulated during proneural-to-mesenchymal transition	resistance to bevacizumab (anti-VEGF); mesenchymal phenotype associated with therapeutic resistance and poor survival	AXL inhibitor + anti-VEGF therapy to overcome mesenchymal-driven anti-angiogenic resistance	Hutterer et al.,3; Cheng et al.,107; Piao et al.108

In the context of GBM, AXL drives angiogenesis through secretion of pro-angiogenic VEGF-A,¹¹⁰ through maintenance of endothelial tube integrity,^111,112 and through VEGF-A-dependent activation of PI3K/Akt signaling.¹¹³ AXL is expressed not only in GBM cells but also in the tumor vasculature. Elevated AXL/GAS6 co-expression was observed in both glioma cells and tumor vessels, with AXL staining most pronounced within pseudo-palisading cells adjacent to the hypoxic, necrotic core of GBM. Tumors with greater AXL expression displayed worse overall survival and shorter time to progression.³ These observations have been extended by mapping phosphorylated AXL (P-AXL) in 90 newly diagnosed GBM patients, finding P-AXL detectable in 74% of cases across three distinct patterns—exclusively in tumor vasculature (13%), in hypercellular areas (35%), or both (52%)—with the combined vascular/hypercellular pattern associated with significantly decreased overall survival.¹¹⁴ Functionally, it was demonstrated that dominant-negative inhibition of AXL suppressed experimental oncogenesis by greater than 85% and resulted in long-term survival in mice (AXL-WT: 10 days vs. AXL-DN: >72 days), with tumor angiogenesis identified as a key hallmark affected.¹¹⁵ More direct anti-angiogenic evidence was provided using the AXL inhibitor BMS-777607, achieving 56% tumor volume reduction in SF126 xenografts and greater than 91% remission in U118MG xenografts, with phosphorylated AXL localized to sites of active neovascularization.¹¹⁶ AXL is also significantly upregulated in the mesenchymal GBM subtype, where AXL was identified as a key regulator of mesenchymal GBM stem-like cells via kinome-wide shRNA screen,¹⁰⁷ consistent with results from another group, who demonstrated that acquired resistance to bevacizumab is associated with a proneural-to-mesenchymal transition.¹⁰⁸ This suggests that AXL inhibition may be combined with anti-VEGF therapy to overcome resistance mediated by the mesenchymal transcriptional program. One caveat to consider is that because these studies were conducted before the WHO 2021 reclassification of gliomas, IDH mutational status was not assessed when classifying tumors as GBM or IDH-mutant grade IV astrocytoma. Given AXL’s dual role in promoting tumor angiogenesis and mediating resistance to anti-VEGF therapies in GBM, future studies should investigate whether selective AXL inhibitors can restore sensitivity to bevacizumab and other anti-angiogenic agents in recurrent GBM, where anti-VEGF resistance remains a major clinical challenge. Additionally, the development of dual AXL/VEGFR inhibitors, or combinatorial strategies targeting both the AXL and VEGF signaling axes within the GBM TME, may represent a promising therapeutic approach to simultaneously disrupt tumor angiogenesis while overcoming the adaptive mesenchymal transition that drives treatment failure.

AXL in ferroptosis

Adding to the expanding functional repertoire of AXL in GBM, a novel role for AXL in ferroptosis resistance was recently uncovered via lipocalin-2 (LCN2), a canonical iron-binding protein. LCN2 was elevated in GBM cells relative to lower-grade tumors and non-transformed cells, and LCN2 suppressed ferroptosis through phosphorylated AXL signaling. Knockdown of LCN2 impaired GBM cell fitness in vitro and in vivo, and the resulting cell death was specifically rescued by ferroptosis inhibitors, confirming ferroptosis as the operative death modality. Mechanistically, LCN2 knockdown reduced AXL phosphorylation, and the combination of LCN2 knockdown with pharmacological AXL inhibition extended survival beyond LCN2 knockdown alone¹¹⁷.

These findings position AXL not only as an oncogenic signaling hub and immune modulator, but also as a gatekeeper of ferroptosis resistance. Notably, the AXL-ferroptosis axis may have broader implications beyond GBM. AXL activation via GAS6 has been shown to suppress ferroptosis through PI3K/AKT signaling in hepatic ischemia/reperfusion injury.¹¹⁸ In BRAF-mutant melanoma, therapy-resistant high-mesenchymal cell states marked by AXL upregulation exhibit a GPX4-dependent vulnerability to ferroptosis.¹¹⁹ Furthermore, AXL inhibition with bemcentinib has been shown to directly stimulate ferroptosis in vemurafenib-resistant melanoma cells.¹²⁰ Given that AXL is overexpressed across numerous solid malignancies—including lung, breast, hepatocellular, and pancreatic cancers—and ferroptosis sensitivity is increasingly recognized as a therapeutically exploitable vulnerability in these tumors, the LCN2-AXL-ferroptosis axis may represent a conserved resistance mechanism with broad translational relevance. Future studies interrogating whether AXL-mediated ferroptosis suppression operates in these other cancer types could open new avenues for combinatorial strategies pairing AXL inhibitors with ferroptosis inducers.

AXL in EMT

EMT enables tumor cells to acquire mesenchymal features that enhance motility, invasiveness, and survival under therapeutic stress.^34,121 Activation of AXL can result in a reversible shift from an epithelial to a mesenchymal phenotype, exemplified by AXL increasing Slug and N-cadherin through NF-κB and PI3K signaling pathways in breast cancer.^87,122 AXL activation also upregulates other mesenchymal markers including Snail, vimentin, and α-catenin and downregulates the epithelial marker E-cadherin to reduce cell-cell adhesion.^34,122,123 Importantly, AXL can also be a downstream effector of EMT; overexpression of EMT transcription factors resulted in enhanced AXL expression in breast epithelial cells.¹²⁴ Furthermore, in HER2-positive (HER2⁺) breast cancer, AXL is required at multiple steps of the metastatic cascade, including intravasation, extravasation, and growth at the metastatic site. AXL sustained EMT and TGF-β-induced invasion, and pharmacological AXL inhibition specifically decreased metastatic burden in mice developing HER2⁺ breast cancer.¹⁰⁶ AXL-mediated induction of EMT has also been noted in additional cancer types, as noted in Table 1. Ultimately, AXL activation drives EMT and helps maintain the mesenchymal state.

AXL as a driver of therapeutic resistance

A central feature that distinguishes AXL from many other oncogenic receptor tyrosine kinases is that its increased expression is rarely attributable to an initiating oncogenic driver event but more frequently the result of pharmacologic selective pressure—arising as an adaptive response to radiation and to a broad range of chemotherapies and targeted therapies.^12,125,126 Critically, AXL-mediated resistance is not limited to a single drug class or tumor type. It has been documented across conventional cytotoxic chemotherapy, molecularly targeted agents (including EGFR, HER2, BRAF/MEK, and PI3K inhibitors), PARP inhibitors, and immune checkpoint inhibitors, establishing AXL as a high-priority target for cross-class therapeutic combinations in oncology.^13,127 The causative nature of this relationship is underscored by the consistent observation that genetic or pharmacologic suppression of AXL restores sensitivity to the agents to which resistance had developed.¹² For example, AXL knockdown or pharmacologic inhibition with bemcentinib restored erlotinib sensitivity in EGFR-mutant NSCLC,¹²⁷ while in BRAF-mutant melanoma, AXL shRNA and the small-molecule inhibitor bemcentinib rescued vemurafenib sensitivity through a PTEN-dependent mechanism.⁸⁴ Similarly, AXL inhibition with restored paclitaxel and carboplatin sensitivity in chemo-resistant ovarian cancer both in vitro and in patient-derived xenografts,¹²⁸ and more recently, AXL inhibition via the small-molecule inhibitor dubermatinib or knockout resensitized cabozantinib-resistant renal cell carcinoma cells to c-Met inhibitor-induced cell death.¹²⁹ We focus our below-mentioned in-depth discussion on several cancer types, including GBM, NSCLC, breast cancer, and brain metastases, which exemplify the role of AXL in therapeutic resistance. However, AXL promotes therapy resistance and other pro-tumor functions in an extensive variety of cancer types (Table 1).

GBM

AXL and its ligand GAS6 have been documented to be overexpressed in human high-grade glioma with moderate-to-high AXL mRNA expression in 61% and AXL protein in 55% of GBM samples.³ Notably, patients with high AXL expression and AXL/GAS6 co-expression exhibited significantly shorter time to tumor progression and poorer overall survival, with AXL staining most pronounced in pseudopalisading glioma cells and tumor vasculature.³ Supporting this, it was shown that AXL is preferentially upregulated in the mesenchymal GBM subtype, where tumor-associated macrophage/microglia-derived PROS1 activates AXL to induce NFκB signaling. Importantly, anti-PD-1 therapy with nivolumab paradoxically increased intratumoral macrophages and AXL activation, and combinatorial treatment with the AXL inhibitor bemcentinib plus nivolumab effectively prolonged survival in preclinical GBM models, suggesting that AXL inhibition may be critical for overcoming immune checkpoint resistance.⁴ AXL also functions as a critical mediator of compensatory signaling in GBM. EGFR inhibition in glioma cells triggers a rapid adaptive response through a TNF-JNK-AXL-ERK signaling axis. Increased TNF secretion activates JNK, which in turn upregulates AXL and sustains ERK-mediated cell survival effectively bypassing EGFR blockade and providing a mechanistic explanation for the failure of anti-EGFR monotherapy in GBM.¹³⁰ This compensatory upregulation of AXL under pharmacologic selective pressure appears to be a conserved resistance mechanism, as AXL overexpression and dimerization with EGFR can circumvent the antitumor effects of both anti-EGFR and anti-PI3K therapies in GBM.^37,130 More recently, a mechanistic link between AXL and the tumor suppressor LZTR1 was identified, demonstrating that LZTR1 functions as a substrate-specific adaptor of a CUL3-dependent ubiquitin ligase that targets both EGFR and AXL for lysosomal degradation. Pathogenic cancer-associated LZTR1 mutations abrogate this degradation, leading to aberrant accumulation of EGFR and AXL and dysregulated growth factor signaling, with LZTR1-mutant tumors exhibiting specific vulnerability to concurrent EGFR and AXL co-inhibition.³⁷

Further feedback control of AXL is exerted by the ribosomal protein S6 kinases. In PTEN-deficient GBMs, elevated signaling downstream of PI3K can be counteracted by the inhibition of Akt, mTORC1, or S6K1.^131,132 Inhibitors of all three of these kinases have been tested, most notably in a series of trials targeting mTORC1. Significant issues related to therapy penetrance and target-specific feedback mechanisms have likely contributed to the lack of therapeutic efficacy. In our own work, we identified a brain-penetrant inhibitor of S6K1, which nonetheless is ineffective as a single agent for suppressing GBM growth. Indeed, inhibitor suppression of S6K1 triggered partial inhibition of its paralog, S6K2, which in turn enabled elevated “rebound” activation of AXL. This novel connection between S6K2 inactivation and AXL activation defines a feedback mechanism, similar to the long-established S6K1 feedback control of insulin receptor signaling to PI3K. Coordinated inhibition of AXL with S6K1 was sufficient to circumvent signaling redundancy and reduce pyrimidine biosynthesis in GBM sphere cultures in vitro and in an orthotopic xenograft mouse model.¹³³ These observations support AXL as both a prognostic biomarker associated with poor clinical outcomes and a compelling therapeutic target in combination with PI3K pathway inhibitors, RTK inhibitors, and other targeted therapeutic strategies.¹³⁴

NSCLC

In NSCLC, AXL has emerged as a major mechanism of resistance to third-generation EGFR tyrosine kinase inhibitors (TKIs), particularly osimertinib. Osimertinib paradoxically stimulates AXL activation by inhibiting a negative feedback loop, and activated AXL associates with EGFR and HER3 to maintain cell survival and promote the emergence of drug-tolerant cells.⁷⁸ Subsequent studies identified the AXL/CDCP1/SRC signaling axis as a distinct mechanism of acquired osimertinib resistance, with increased AXL and transmembrane glycoprotein, CDCP1, expression confirmed in refractory tumor samples from patients treated with osimertinib.⁷⁹ More recently, palmitoylation of AXL at Cys869 via the palmitoyl transferase, ZDHHC11, was shown to induce plasma membrane retention and constitutive AXL activation, triggering downstream PI3K/AKT signaling that drives osimertinib resistance in EGFR-mutant NSCLC.¹³⁵ AXL also enhances cancer stem-like cell self-renewal and contributes to osimertinib chemoresistance through increased upregulation of SCD1.⁸⁰ With respect to immunotherapy, AXL inhibition sensitizes STK11/LKB1 mutant NSCLC tumors to anti-PD-1 therapy by increasing type I interferon secretion from dendritic cells and expanding tumor-associated TCF1⁺/PD-1⁺ CD8⁺ T cells, with objective clinical responses observed in NSCLC patients with STK11/LKB1 mutations receiving bemcentinib and pembrolizumab.¹³⁶ Accordingly, a retrospective analysis of NSCLC patients identified that AXL-positive NSCLC tumors were associated with significantly reduced benefit from anti-PD-1 therapy (HR 4.15 for PFS, p = 0.013), with KRAS and STK11/LKB1 mutations more frequent in AXL-positive tumors.⁸²

Given these converging lines of evidence, future studies should investigate whether combining selective AXL inhibitors with both osimertinib and immune checkpoint blockade can simultaneously overcome TKI resistance and restore antitumor immunity in EGFR-mutant NSCLC, particularly in the STK11/LKB1-mutant subset where AXL-driven immunosuppression and therapeutic resistance intersect.

Breast cancer

In basal-like breast cancer (BLBC), a single cycle of cytotoxic combination therapy was sufficient to induce AXL-driven cancer stem cell properties and resistant tumor phenotypes in vivo. AXL inhibition sensitized BLBC cells to cytotoxic cyclophosphamide, adriamycin, and 5-fluorouracil in combination.¹³⁷ In HER2⁺ breast cancer, AXL overexpression drives resistance to both trastuzumab and lapatinib through ligand-independent heterodimerization with HER2, leading to activation of PI3K/AKT and MAPK pathways. Genetic depletion or pharmacological inhibition of AXL restored trastuzumab response in vitro and achieved complete regression in trastuzumab-resistant patient-derived xenograft models.^27,138

Brain metastases

AXL plays a context-dependent role in brain metastasis that varies by cancer type and cellular compartment. In lung adenocarcinoma, an autocrine TAZ-AXL-ABL2 feedforward signaling loop was found to be required for brain metastasis colonization, whereby nuclear localization of the transcription factor TAZ drives AXL and ABL2 expression, and ABL2 phosphorylates TAZ to sustain its nuclear localization. Knockdown of AXL, ABL2, or TAZ markedly decreased lung adenocarcinoma brain metastases, and treatment with a blood-brain barrier (BBB)-penetrant ABL allosteric inhibitor was effective in preclinical models.¹⁰⁵ In HER2⁺ breast cancer, AXL is required at multiple steps of the metastatic cascade, including intravasation, extravasation, and growth at the metastatic site. Furthermore, AXL sustained EMT and TGF-β-induced invasion, and pharmacological AXL inhibition specifically decreased metastatic burden in mice developing HER2⁺ breast cancer.²⁷ In colorectal cancer brain metastases, AXL was expressed in 72.7% of brain metastases compared to paired primary tumors, with AXL localized primarily to endothelial cells rather than tumor cells. GAS6 was identified as an independent prognostic factor in metastatic CRC, with patients harboring low AXL/high GAS6 tumors having significantly shortened overall survival.¹⁰⁴

More recently, it was demonstrated that breast cancer subtype fundamentally determines the architecture of brain colonization and the functional role of AXL in the brain TME.¹³⁹ TNBC (cell line MDA-MB-231) forms perivascular sheaths with diffusive contact with astrocytes and microglia, whereas HER2⁺ breast cancer (HCC1954) forms compact spheroids driven by autonomous tenascin C production that segregate stromal cells to the periphery. Single-cell transcriptomics revealed that these distinct architectures evoke differential Alzheimer’s disease-associated microglia (DAM) responses and differential engagement of AXL. In TNBC, where AXL and GAS6 expression are high, AXL knockdown inhibited brain colonization activity, suggesting that an autocrine AXL/GAS6 loop promotes TNBC brain metastasis. In contrast, in HER2⁺ brain metastases where overall tumor cell AXL expression is low, AXL⁺ CD68⁺ microglia accumulate at the rim of spheroids, and heightened GAS6 in the TME was associated with reduced brain metastatic activity, suggesting that AXL⁺ microglia may exert a phagocytic, antitumor function. These findings collectively indicate that AXL’s role in brain metastasis is not uniformly pro-tumorigenic: it can be tumor cell-intrinsic and pro-metastatic (as in TNBC and lung adenocarcinoma) or microglial and potentially antitumorigenic (as in HER2⁺ breast cancer). These findings indicate that therapeutic strategies targeting AXL in brain metastases must account for cancer subtype, cellular compartment, and the unique GAS6-rich brain microenvironment to avoid paradoxical tumor growth.

Clinical trials

Small molecule TKIs

The receptor tyrosine kinase AXL has emerged as a compelling therapeutic target across a broad range of cancers due to its pivotal role in EMT, immune evasion, and resistance to conventional therapies.^34,81 Multiple targeted therapeutics currently used in oncology have activity against AXL (Table 2). For example, cabozantinib is FDA-approved for activity against VEGFR and RET, and gilteritinib is approved for inhibition of FLT3; both have substantial AXL-inhibitory activity (Figure 3).^86,146 In addition, novel chemotherapeutic agents have been developed specifically for therapeutic targeting of AXL in oncology. Considering the function of AXL at the interfaces of therapy-resistance, immune function, microenvironment, and sex hormone interactions in cancer, integrated analysis of AXL-dependent effects is required to identify agents that will deliver optimal immune-compatible therapy for women and men affected by cancer. A comprehensive review published in 2025 highlighted the full landscape of AXL-targeting agents in early-phase cancer trials and noted that while safety has been established for several agents, patient selection and companion biomarker development remain critical for clinical success.¹⁵⁵

Table 2.

Relevant clinical trials

Agent	Cancer type	Combination partner(s)	Phase	NCT number	References
Small molecule inhibitors
Bemcentinib (BGB-324) +A5:F12	NSCLC	docetaxel	phase I	NCT02922777	Bhalla et al.140
Bemcentinib (BGB-324)	EGFR-mutant NSCLC	erlotinib	phase I/II	NCT02424617	–
Bemcentinib (BGB-324)	metastatic melanoma	pembrolizumab or dabrafenib/trametinib	phase Ib/II	NCT02872259	–
Bemcentinib (BGB-324)	relapsed mesothelioma	pembrolizumab	phase II	NCT03654833	–
Bemcentinib (BGB-324)	AML	monotherapy or low-dose cytarabine	phase Ib/II	NCT02488408	Loges et al.141
Bemcentinib (BGB-324)	NSCLC (STK11-mutant focus)	pembrolizumab + carboplatin + pemetrexed	phase Ib/IIa	NCT05469178	–
Bemcentinib (BGB-324)	glioblastoma	–	phase I	NCT03965494	–
Bemcentinib (BGB-324)	metastatic pancreatic cancer	paclitaxel	phase Ib/II	NCT03649321	–
FC084CSA	advanced malignant solid tumors	tislelizumab	phase Ib/IIa	NCT06499350	–
NTQ2494	advanced hematological malignancies	–	phase I	NCT06049667	–
Q702	AML	azacytidine + venetoclax	phase I	NCT06445907	–
	hematologic malignancies	–	phase I	NCT06712810	–
	advanced solid tumors	pembrolizumab	phase Ib/II	NCT05438420	–
	advanced solid tumors	–	phase I	NCT04648254	–
SLC-391	solid tumors	–	phase I	NCT03990454	Zhang et al.142
	advanced lung cancer	pembrolizumab	phase Ib/IIa	NCT05860296	–
ST-1898	advanced solid tumors	–	phase I	NCT07252661	–
	thyroid cancer	–	phase II	NCT06359847	–
	melanoma	–	phase Ib/II	NCT06359860	–
	RCC	–	phase Ib/II	NCT06127238	–
XZB-0004	solid tumors	–	phase I	NCT05772455	–
	NSCLC	–	phase I	NCT05740917	–
Monoclonal antibodies
YW327.6S2	NSCLC, breast cancer	anti-VEGF or erlotinib	preclinical	–	Ye et al.143
20G7-D9	triple-negative breast cancer	–	preclinical	–	Leconet et al.144
Decoy receptors
Batiraxcept (AVB-S6-500)	platinum-resistant ovarian cancer	paclitaxel	phase Ib	NCT03639246	Fuh et al.145
Batiraxcept (AVB-S6-500)	platinum-resistant ovarian cancer	durvalumab	phase Ib/II	NCT04019288	Knisely et al.146
Batiraxcept (AVB-S6-500)	advanced ccRCC	monotherapy or cabozantinib ± nivolumab	phase Ib/II	NCT04300140	Beckermann et al.147
Batiraxcept (AVB-S6-500)	advanced pancreatic adenocarcinoma	gemcitabine + paclitaxel	phase Ib	NCT04983407	–
Batiraxcept (AVB-S6-500)	endometrial cancer	paclitaxel	preclinical	–	Bruce et al.148
Batiraxcept (AVB-S6-500)	metastatic prostate cancer	docetaxel or carboplatin	preclinical	–	Chiu et al.149
Batiraxcept (AVB-S6-500)	HER2+ endometrial cancer	trastuzumab	preclinical	–	Tankou et al.150
Antibody-drug conjugates
Enapotamab vedotin (AXL-107-MMAE)	melanoma	BRAF/MEK inhibitors	preclinical	–	Boshuizen et al.151
Enapotamab vedotin (AXL-107-MMAE)	NSCLC	–	preclinical	–	Koopman et al.152
Enapotamab vedotin (AXL-107-MMAE)	melanoma, lung cancer	pembrolizumab	preclinical	–	Boshuizen et al.85
Enapotamab vedotin (AXL-107-MMAE)	multiple solid tumors	–	phase I/II	NCT02988817	–
Mipasetamab uzoptirine (ADCT-601)	lung, ovarian, pancreatic cancers	monotherapy or olaparib (ovarian)	preclinical	–	Zammarchi et al.153
Mipasetamab uzoptirine (ADCT-601)	adenoid cystic carcinoma	–	preclinical	–	Humtsoe et al.154
Mecbotamab vedotin (BA3011)	advanced solid tumors	–	phase I	NCT03425279	–

Figure 3.

Therapeutic approaches for PtS/GAS6/AXL targeting

Kinase inhibitors listed in red have received FDA approval, whereas inhibitors and other therapeutics listed in black have not received approval at this time. Phosphatidylserine can be targeted using PtS-binding peptides, PtS-targeted nanoparticles/nanovesicles, or bavituxumab. GAS6 can be targeted using batiraxcept, warfarin, or AXL decoy receptors. AXL itself can be targeted with monoclonal antibodies, AXL antibody-drug conjugates, or small molecule inhibitors. Furthermore, AXL can be used as a target for cell killing via AXL NK CARs or ZIKV-LAV oncolytic virus.

Among small-molecule inhibitors, bemcentinib is the most clinically advanced selective AXL inhibitor. In previously treated advanced NSCLC, bemcentinib combined with docetaxel achieved an objective response rate (ORR) of 35% among evaluable patients, though grade ≥3 neutropenia was observed in 76% of participants (NCT02922777).¹⁴⁰ In relapsed mesothelioma, the MiST3 trial demonstrated that bemcentinib plus pembrolizumab met its primary endpoint with a 12-week disease control rate of 46.2% and an ORR of 15.4% (NCT03654833). In AML, bemcentinib monotherapy and its combination with low-dose cytarabine were well tolerated in patients unfit for intensive chemotherapy, establishing proof of concept for AXL inhibition in hematologic malignancies.¹⁴¹ A phase Ib/II trial tested bemcentinib in combination with either pembrolizumab or dabrafenib/trametinib in advanced non-resectable melanoma, with results yet to be published. A phase Ib/IIa trial of bemcentinib combined with pembrolizumab, carboplatin, and pemetrexed in first-line non-squamous NSCLC with a focus on STK11/LKB1-mutant tumors remains ongoing (NCT05469178). Bemcentinib has also been trialed in EGFR-mutant NSCLC in combination with erlotinib (NCT02424617), in GBM as monotherapy (NCT03965494), and in metastatic pancreatic cancer combined with paclitaxel (NCT03649321).

Other AXL-specific and multi-kinase inhibitors with high AXL affinity are also under clinical investigation. Q702 (adrixetinib) is an AXL/MER/CSF1R inhibitor being tested in phase I trials in relapsed/refractory AML in combination with azacitidine and venetoclax (NCT06445907), and in hematologic malignancies including histiocytic neoplasms, MDS, myelofibrosis, and lymphomas (NCT06712810). In the context of solid tumors, Q702 is being tested in a phase Ib/II trial in combination with pembrolizumab in advanced esophageal, gastric/GEJ, hepatocellular, and cervical cancers that progressed on prior anti-PD-1/PD-L1 therapy (NCT05438420). FC084CSA is a selective AXL inhibitor that demonstrated tolerability in a phase I monotherapy trial in advanced solid tumors (NCT06231550). NTQ2494 is a potent AXL kinase inhibitor currently under phase I clinical evaluation for advanced hematological malignancies and solid tumors.¹⁵⁶ SLC-391, a selective AXL inhibitor, completed phase I trials demonstrating tolerability and a clinical benefit rate of 34.3% in advanced solid tumors (NCT03990454),¹⁴² providing a rationale for a phase Ib/IIa trial in combination with pembrolizumab in advanced NSCLC (NCT05860296). Future clinical trials using these agents will require positive results from preclinical investigations to define rational combinations with specific inclusion criteria and tumor types.

AXL decoy receptors

The sAXL decoy receptors represent a distinct therapeutic strategy that functions by competitively sequestering the AXL ligand GAS6, thereby preventing its interaction with membrane-bound AXL and abrogating downstream signaling. The foundational work in this area involved the engineering of a first-generation AXL “decoy receptor”—an sAXLectodomain fused to an Fc domain named MYD1—which bound GAS6 with femtomolar affinity (80-fold improvement over wild-type AXL) and potently inhibited metastasis and disease progression in vivo.¹⁵⁷ A second-generation high-affinity variant, MYD1-72, achieved an apparent affinity of 93 femtomolar to GAS6 and demonstrated superior antitumor efficacy compared with the most advanced anti-AXL small molecules in preclinical models of pancreatic and ovarian cancer, while exploiting a relationship between AXL signaling and the DNA damage response to augment the therapeutic index of standard-of-care chemotherapies.⁹⁰ These preclinical efforts led to the clinical development of batiraxcept (AVB-S6-500/AVB-500), a recombinant AXL-Fc fusion protein. In a phase Ib study in platinum-resistant ovarian cancer (NCT03639246), batiraxcept combined with paclitaxel yielded an ORR of 34.8% (including 2 complete responses), with median PFS and OS of 3.1 and 10.3 months, respectively; subgroup analyses identified patients without prior bevacizumab or with AVB-500 trough levels >13.8 mg/L as having the best clinical response (ORR ≥50%, OS ≥19 months).¹⁴⁵ Batiraxcept was also evaluated in combination with PD-L1 neutralizing durvalumab in platinum-resistant ovarian cancer (NCT04019288), where the combination was safe and tolerable but did not demonstrate objective responses (median PFS 1.81 months).¹⁴⁶ In advanced ccRCC, a phase Ib/II study (NCT04300140) tested batiraxcept as monotherapy and in combination with cabozantinib ± nivolumab; the doublet arm (batiraxcept + cabozantinib) achieved an ORR of 43% with a median PFS of 9.2 months, while monotherapy showed no objective responses, and the triplet arm demonstrated a 54% ORR, though the trial was discontinued early due to a sponsor decision.¹⁴⁷ In advanced pancreatic adenocarcinoma, a phase Ib dose-escalation study (NCT04983407) evaluated batiraxcept combined with gemcitabine and nab-paclitaxel, reporting an ORR of 29%, median PFS of 5.4 months, and median OS of 12.3 months, with exposure-response modeling suggesting that patients achieving trough levels above the minimally efficacious concentration had longer PFS and OS. Preclinical studies have also demonstrated the potential of batiraxcept in endometrial cancer, where it synergized with paclitaxel to decrease tumor burden in uterine serous cancer models by impairing AKT signaling and basal glycolysis,¹⁴⁸ and in prostate cancer bone metastasis models, where batiraxcept alone or combined with docetaxel or carboplatin significantly suppressed intratibial tumor growth and lung metastasis.¹⁴⁹ Importantly, because GAS6 requires binding to externalized PtS via its γ-carboxylated Gla domain to achieve maximal TAM receptor activation,¹⁵⁸ sAXL decoy receptors that sequester GAS6 have the potential to not only inhibit AXL signaling but also to saturate and neutralize PtS-bound GAS6, thereby reducing the broader immunosuppressive effects of PtS in the TME. Similarly, drug repurposing efforts have utilized warfarin, a vitamin K antagonist that blocks γ-carboxylation of the Gla domain on GAS6, to inhibit GAS6-dependent AXL activation. Low-dose warfarin treatment resulted in AXL inhibition, decreased EMT, metastasis, and tumor burden in several models of pancreatic ductal adenocarcinoma (PDAC).²⁸ Importantly, we note that warfarin is a brain-penetrant therapeutic, circumventing the limited brain permeability of antibody fusions and decoy receptors for brain tumor therapy.

PtS is a dominant immunosuppressive signal that is prevalent in tumors, where, in addition to GAS6-AXL, it engages multiple PtS-responsive receptors, including TIM-3, MERTK, and Tyro3, on infiltrating immune cells to promote efferocytosis, pro-tumorigenic (M2-like) macrophage polarization, and suppression of antitumor immunity.^159,160 Recent work has further established that viable, antigen-specific CD8⁺ T cells externalize PtS during chronic antigen stimulation and that PtS-targeting antibodies can expand CD8⁺ T cell responses and synergize with anti-PD-L1 to improve viral and potentially tumor control, thereby identifying exposed PtS as a non-classical inhibitory molecule in T cell exhaustion.¹⁶¹ Thus, the therapeutic class of sAXL decoy receptors may extensively reprogram the TME by circumventing PtS-mediated immunosuppression across multiple PtS-responsive receptors and cell types. For a comprehensive discussion of PtS-targeting therapeutic strategies in cancer, including PtS-binding peptides, nanoparticles, and nanovesicles, readers are referred to recent reviews on this topic.^162,163

Monoclonal antibodies

Several AXL-targeting monoclonal antibodies and antibody-based therapeutics have been developed, though no selective AXL-targeting therapy has yet received FDA approval.¹⁶⁴ The preclinical anti-AXL monoclonal antibody, YW327.6S2, demonstrated tumor growth inhibition and enhanced efficacy of anti-VEGF therapy, erlotinib, and chemotherapy in NSCLC and breast cancer xenograft models.¹⁴³ 20G7-D9, another anti-AXL monoclonal antibody, inhibited tumor growth and bone metastasis in TNBC patient-derived xenograft (PDX) models by inducing AXL degradation and blocking GAS6-dependent EMT signaling, though neither of these antibodies have advanced to human clinical trials.¹⁴⁴ However, monoclonal antibodies targeting AXL that are conjugated to cytotoxic compounds are significantly further along in development.

AXL-targeting antibody-drug conjugates

Several AXL-targeting antibody-drug conjugates (ADCs) have been developed to exploit AXL overexpression on tumor cells for selective delivery of cytotoxic payloads. The most extensively studied is enapotamab vedotin (EnaV/AXL-107-MMAE), a humanized anti-AXL antibody conjugated to the microtubule-disrupting agent monomethyl auristatin E (MMAE). Enapotamab vedotin demonstrated potent single-agent antitumor activity in PDX models of melanoma, where it cooperatively targeted AXL-high, MAPK inhibitor-resistant tumor populations in combination with BRAF/MEK inhibitors.¹⁵¹ In NSCLC, enapotamab vedotin induced tumor regression or stasis in 28% (17/61) of PDX models, with activity correlating with AXL mRNA expression levels, and showed efficacy in EGFR inhibitor-resistant models including an osimertinib-resistant PDX.¹⁵² Enapotamab vedotin also enhanced sensitivity to immune checkpoint blockade in immunotherapy-resistant melanoma and lung cancer models by inducing immunogenic cell death and promoting a memory-like phenotype in cytotoxic T cells.⁸⁵ The safety and preliminary efficacy of enapotamab vedotin were evaluated in a phase I/II clinical trial across multiple solid tumor types (NCT02988817). Mipasetamab uzoptirine (ADCT-601) is a second AXL-targeting ADC comprising a humanized anti-AXL antibody site-specifically conjugated to a pyrrolobenzodiazepine (PBD) dimer cytotoxin (SG3199). In preclinical studies, ADCT-601 demonstrated potent and durable antitumor activity across a wide variety of xenograft models, including lung and ovarian cancers, and was markedly superior to an auristatin-based comparator ADC in PDX models with heterogeneous AXL expression. ADCT-601 also synergized with the PARP inhibitor olaparib in a BRCA1-mutated ovarian cancer model and retained activity in an MMAE-resistant lung cancer model.¹⁵³ More recently, ADCT-601 showed potent and selective cytotoxicity in adenoid cystic carcinoma (ACC) models, with a single administration at 1.0 mg/kg inducing complete tumor eradication in xenograft models and therapeutic response correlating with AXL expression levels across a panel of ACC PDX models.¹⁵⁴ A third AXL-targeting ADC, mecbotamab vedotin (BA3011), utilizes a conditionally active biologic (CAB) anti-AXL humanized monoclonal antibody conjugated to MMAE, designed to preferentially bind AXL in the TME. BA3011 has been evaluated in a phase I clinical trial in patients with advanced solid tumors including sarcoma (NCT03425279), with population pharmacokinetic and exposure-response safety analyses demonstrating statistically significant correlations between ADC exposure and key safety endpoints. The activity of AXL ADCs supports that AXL expression is increased on transformed cells compared to non-transformed cells, supporting AXL as a priority target in oncology.

Barriers to clinical efficacy

Despite robust preclinical rationale, no selective AXL-targeting therapy has achieved FDA approval, and AXL monotherapy has demonstrated limited clinical activity across trials to date. Several converging factors account for this translational gap. First, AXL is not a potent oncogenic driver itself but rather functions as a context-dependent mediator of resistance, plasticity, and immune evasion, meaning its inhibition alone is unlikely to produce meaningful tumor regression in unselected populations.¹ Second, the absence of validated biomarkers has hampered patient selection. AXL expression is heterogeneous, dynamically regulated by non-genetic mechanisms (e.g., EMT, hypoxia, and pharmacologic selective pressure), and does not correlate with genomic alterations in the way that EGFR mutations or ALK fusions define targetable populations.^12,14 Emerging biomarker candidates, such as sAXL and serum AXL/GAS6 ratios, remain investigational and have not yet been prospectively validated for patient enrichment.³¹ Third, many AXL inhibitors exhibit kinase promiscuity, targeting multiple kinases beyond AXL (e.g., MET, VEGFR, and FLT3), which confounds the attribution of clinical effects to AXL inhibition specifically.⁷⁸ Fourth, compensatory signaling and adaptive rewiring represent fundamental biological barriers: AXL inhibition induces upregulation and activation of alternative RTKs that restore downstream survival signaling and rescue tumor cell viability.¹⁶⁴ This phenomenon mirrors the adaptive resistance observed with FLT3 inhibitors in AML, where AXL itself is rapidly upregulated within hours of TKI treatment as a compensatory escape mechanism.⁸⁶ Finally, AXL’s role in sustaining a mesenchymal and thus therapeutic resistant phenotype means that AXL inhibition may need to be paired with agents targeting the downstream consequences of EMT to achieve durable responses.¹⁵⁵

Safety is another important guideline when designing combination therapy strategies. Based on the physiological roles of AXL and common side effects of other TKIs, potential adverse effects of AXL targeting include diarrhea, neutropenia, fatigue, anemia, and nausea.¹⁶⁴ Dose-limiting toxicities (DLTs) in response to AXL targeting monotherapies are minimal, according to a recent review.¹⁶⁴ Tolerable safety profiles have been observed for multiple AXL targeting agents, such as batiraxcept and mecbotamab vedotin, but adverse effects such as neutropenia have been noted in some combination therapies. However, it is worth noting that these adverse events may be more attributable to the combination agent compared to the AXL targeting agent.¹⁶⁵ Regardless, the development of more specific TKIs, as well as the incorporation of monoclonal antibodies and well-tolerated ADCs into combination therapies, may further reduce adverse effects.

These challenges and advances collectively argue that the future of AXL-directed therapy lies not in monotherapy but in rationally designed combinations pairing AXL targeting agents with TKIs, immune checkpoint inhibitors, or cytotoxic chemotherapy where AXL’s role as a resistance mediator, angiogenesis supporter, and promoter of immune evasion can be safely and specifically exploited.

Outlook

AXL occupies a unique position among therapeutic targets in oncology. Unlike the canonical kinase drivers that have dominated precision medicine—BCR-ABL, EGFR, KRAS, BRAF, and BTK—activation of AXL is infrequently the result of direct cancer driver alterations. In this sense, AXL is less of a fixed vulnerability and more an induced survival strategy—a signaling program that cancer cells deploy when they are under stress, enabling phenotypic plasticity and the emergence of resistant clones.⁸⁷ This distinction carries a critical therapeutic implication: because AXL activation is most often a consequence of other therapeutic pressures rather than an initiating oncogenic event, AXL is unlikely to be effective as a monotherapy target. The consistent pattern of AXL upregulation under therapeutic stress followed by resensitization upon AXL inhibition argues that clinical trial designs should prioritize rational combinations in which AXL inhibitors are paired with the specific agents driving AXL-mediated escape, rather than deploying AXL inhibitors as standalone therapies.

However, it is important to emphasize that AXL targeting is not restricted to situations in which therapeutic resistance is prerequisite or imminent. AXL targeting in combination therapies has shown promising results in multiple preclinical studies in the absence of noted resistance induction by reducing angiogenesis^109,110,111 and re-invigorating antitumor immunity.^41,48 Furthermore, strategies such as AXL NK-CAR,⁶⁸ oncolytic tumor viruses targeting AXL,⁶⁹ and ADCs^151,152,154 specifically recognize and target AXL-expressing cells to enact antitumor response regardless of therapy resistance status.

AXL exhibits promise as both an indicator of disease state and as a therapeutic target, but several important questions regarding implementation remain unresolved. The potential of sAXL as a circulating biomarker for disease progression, treatment response, or the emergence of therapeutic resistance is biologically plausible, but requires prospective validation in well-designed clinical studies. More importantly, the field needs robustly validated biomarkers that can identify the patients most likely to benefit from AXL-directed therapies. However, given the context-dependent nature of AXL signaling, conventional single-analyte biomarkers may prove insufficient; future integration of transcriptomic, proteomic, and spatial profiling approaches may be necessary to capture the complexity of AXL biology in individual patients’ tumors.¹⁶⁶

Several practical barriers also impede clinical translation. AXL shares kinase domain homology with other receptor tyrosine kinases, making selectivity a persistent challenge for small-molecule inhibitors. Many agents in clinical development inhibit AXL alongside other kinases such as MET, VEGFR, or MER, complicating the attribution of clinical effects specifically to AXL inhibition.³⁴ Drug resistance to AXL inhibitors themselves is an emerging concern, as cancer cells may develop compensatory mechanisms to evade AXL blockade.¹⁶⁵ For ADCs, heterogeneous AXL expression within tumors and limited penetration into stromal-dense microenvironments represent additional obstacles.¹⁶⁶ Furthermore, the lack of a standardized assay for AXL expression, whether by IHC, circulating sAXL, or transcriptomic profiling, hampers cross-trial comparisons and patient stratification. Addressing these challenges will require not only improved pharmacologic tools but also a deeper mechanistic understanding of how AXL isoforms, proteolytic cleavage, and ligand-dependent versus ligand-independent activation differentially contribute to tumor biology.^14,30

Emerging opportunities

Beyond its established roles in EMT, angiogenesis, and immune evasion, AXL is now being implicated in two additional biological domains with significant therapeutic potential. First, AXL has recently been identified as a gatekeeper of ferroptosis resistance. Given that ferroptosis sensitivity is increasingly recognized as a therapeutically exploitable vulnerability across solid malignancies, the convergence of AXL signaling and ferroptosis resistance may represent a conserved mechanism with broad translational relevance, opening new avenues for combinatorial strategies pairing AXL inhibitors with ferroptosis inducers.

Second, sex-dependent differences in AXL expression and function remain poorly characterized in oncology but may meaningfully influence both therapeutic responsiveness and toxicity. Plasma GAS6 concentrations are significantly higher in males than females and are independently associated with sex hormones such as free androgen index in males and estradiol in females.¹⁶⁷ In the context of cancer, male aged fibroblasts drive therapy resistance and promote metastasis by increasing expression of AXL in melanoma cells.¹⁶⁸ AXL expression in normal and malignant breast tissue is estrogen-dependent, with AXL protein expression correlating with estrogen receptor status in both normal breast epithelium and breast carcinomas.¹⁶⁹ AXL/GAS6 signaling also plays a key role in the hypothalamic-pituitary-gonadal axis, regulating GnRH neuron migration and spermatogenesis,¹⁷⁰ and AXL has been implicated in androgen deprivation therapy resistance in prostate cancer, where AXL knockdown in RON-overexpressing tumors completely reversed castration resistance.¹⁷¹ Sexually dimorphic expression and activity of AXL has also been noted in the immune compartment. Integrated immune profiling revealed that males exhibit higher frequencies of AXL⁺ dendritic cells and classical monocytes, whereas females display enhanced AXL^- plasmacytoid dendritic cell frequencies with elevated interferon-stimulated gene expression and superior antiviral priming.¹⁷² Whether sex hormone-driven differences in AXL expression translate into differential responses to AXL-targeted therapies, as well as how sexually dimorphic immune cell expression patterns of AXL interact with sex hormone effects in the TME, remain open questions that warrant prospective investigation.

Taken together, AXL rarely functions as a classical initiating oncogenic driver through recurrent activating mutations but instead acts predominantly as an adaptive amplifier of malignant phenotypes characteristically induced by therapeutic and microenvironmental stress. This reinforces the principle that AXL-directed therapies will achieve their greatest impact not as monotherapies, but as rational components of combination strategies designed to intercept the adaptive programs through which tumors escape treatment. Clinical application remains in its infancy, and the path forward will require more selective inhibitors, resolution of challenges including heterogeneous target expression and lack of standardized biomarkers for patient selection, and prospective investigation of underexplored dimensions such as sex-dependent differences in AXL biology and the emerging AXL-ferroptosis axis.³⁴ Successfully targeting this axis has the potential to simultaneously undermine cancer cell survival, disrupt the supportive TME, overcome ferroptosis resistance, and reinvigorate antitumor immunity: a convergence of effects that few other single targets can offer.

Acknowledgments

The manuscript is supported by NIH (National Institutes of Health) T32 Training Grant ES-09250. This work is supported by the University of Cincinnati Graduate College and Office of Research, the University of Cincinnati Cancer Center, the Seal Family Foundation, the University of Cincinnati Brain Tumor Center, and the Anna and Harold W. Huffman Endowed Chair for Glioblastoma Experimental Therapeutics. This work is also supported by Cleveland Clinic Research and the Case Comprehensive Cancer Center. We thank the anonymous reviewers of this manuscript for their insightful feedback.

Declaration of interests

The authors declare no competing interests.