MERTK in Cancer Therapy: Targeting the Receptor Tyrosine Kinase in Tumor Cells and the Immune System

Abstract

The receptor tyrosine kinase MERTK is aberrantly expressed in numerous human malignancies, and is a novel target in cancer therapeutics. Physiologic roles of MERTK include regulation of tissue homeostasis and repair, innate immune control, and platelet aggregation. However, aberrant expression in a wide range of liquid and solid malignancies promotes neoplasia via growth factor independence, cell cycle progression, proliferation and tumor growth, resistance to apoptosis, and promotion of tumor metastases. Additionally, MERTK signaling contributes to an immunosuppressive tumor microenvironment via induction of an anti-inflammatory cytokine profile and regulation of the PD-1 axis, as well as regulation of macrophage, myeloid-derived suppressor cell, natural killer cell and T cell functions. Various MERTK-directed therapies are in preclinical development, and clinical trials are underway. In this review we discuss MERTK inhibition as an emerging strategy for cancer therapy, focusing on MERTK expression and function in neoplasia and its role in mediating resistance to cytotoxic and targeted therapies as well as in suppressing anti-tumor immunity. Additionally, we review preclinical and clinical pharmacological strategies to target MERTK.

Introduction to MERTK

MERTK (myeloid-epithelial-reproductive tyrosine kinase) is a receptor tyrosine kinase (RTK) that is frequently abnormally expressed in a broad range of human cancers (Graham, DeRyckere, Davies, & Earp, 2014; Linger, Keating, Earp, & Graham, 2008). Although this RTK, like others, can promote tumor cell proliferation to some extent, MERTK primarily lends tumor cells crucial survival advantages while promoting invasion, migration and metastasis, drug resistance and, in the innate immune system, suppressing anti-tumor immunity. Additionally, MERTK overexpression can be transforming in certain contexts: transgenic expression in the hematopoietic lineage induces lymphoblastic leukemia or lymphoma in mice (Keating, et al., 2006) and retroviral transduction of the chicken orthologue v-eyk is associated with cancer development in chickens (Jia, Mayer, Hanafusa, & Hanafusa, 1992). Hence, MERTK is an attractive potential target for cancer therapy and clinical trials targeting MERTK have now been initiated.

MERTK is a member of the TAM family of RTKs, which additionally comprises TYRO3, and AXL. These RTKs are characterized by an extracellular, N-terminal structure composed of two immunoglobulin-like and two fibronectin III domains (Graham, Dawson, Mullaney, Snodgrass, & Earp, 1994; Graham, et al., 2014; Lai, Gore, & Lemke, 1994; Linger, et al., 2008; O’Bryan, et al., 1991). A hydrophobic, single pass transmembrane domain is followed by a cytoplasmic tail containing a tyrosine kinase domain with a KWIAIES sequence that is characteristic of the TAM family (Lai & Lemke, 1991). MERTK is activated by ligand binding-induced auto-phosphorylation and influences a range of downstream signaling pathways and cellular functions, described below. The best described MERTK ligands are GAS6 and PROS1 (J. Chen, Carey, & Godowski, 1997; Nagata, et al., 1996; Prasad, et al., 2006). Although both ligands can activate MERTK on their own, the most efficient activation requires the formation of a unique ternary complex in which one of the protein ligands functions as a bridge between MERTK and the lipid phosphatidylserine (PtdSer) exposed on the surface of a second cell (Kasikara, et al., 2017; Tsou, et al., 2014; Zagorska, Traves, Lew, Dransfield, & Lemke, 2014). Both GAS6 and PROS1 contain a γ-carboxylglutamic acid-rich (GLA) domain and vitamin K-dependent carboxylation of this domain is needed for binding and bridging to PtdSer (M. Huang, et al., 2003). While PtdSer is usually not present on the surface of normal steady state cells, apoptotic cells expose the lipid to their extracellular surface and ligand-dependent bridging of MERTK to PtdSer mediates the clearance of apoptotic cells by phagocytic cells in a process termed efferocytosis (Ishimoto, Ohashi, Mizuno, & Nakano, 2000; Scott, et al., 2001). Additionally, MERTK interacts with PtdSer exposed on aggregating platelets (Angelillo-Scherrer, et al., 2008; C. Chen, et al., 2004), ectosomes (Eken, et al., 2010), invading virus envelopes (Bhattacharyya, et al., 2013; Mercer & Helenius, 2008) and activated T cells, which expose a patch of externalized PtdSer without undergoing apoptosis in a recently described novel mechanism of cell-to-cell signaling (Carrera Silva, et al., 2013).

GAS6 and PROS1 also function as ligands for TYRO3, while AXL is only activated by GAS6 (Kasikara, et al., 2017; Lew, et al., 2014; Stitt, et al., 1995; Tsou, et al., 2014). This is consistent with the overlapping physiologic functions mediated by TAM kinases (Graham, et al., 2014; Linger, et al., 2008), although the roles for individual family members can be highly context and cell-type dependent (Seitz, Camenisch, Lemke, Earp, & Matsushima, 2007; Zagorska, et al., 2014). Additional but less studied ligands include Tubby (interacts with all TAM members), Galectin-3 (LGALS3) (interacts with MERTK and TYRO3), and tubby-like protein 1 (TULP-1) (interacts selectively with MERTK) (Caberoy, Alvarado, Bigcas, & Li, 2012; Caberoy, Alvarado, & Li, 2012; Caberoy, Zhou, & Li, 2010; Sassan & Nour, 2019). All three of these proteins interact with MERTK, stimulate MERTK phosphorylation, and promote efferocytosis.

MERTK is physiologically expressed in a wide range of cell types, including hematopoietic lineage cells, such as monocytes, macrophages, microglial and dendritic cells, natural killer (NK) cells, natural killer T (NKT) cells and platelets, as well as epithelial cells in the retina, lung, testes, ovary, prostate and kidney (Linger, et al., 2008). MERTK is also expressed at low levels in heart, brain and skeletal muscle. Additionally, the extracellular domain of MERTK can be cleaved from the cell surface by metalloproteases, such as ADAM17 (Y. J. Lee, et al., 2012; Thorp, et al., 2011; Y. Zhang, et al., 2019), to release a soluble MERTK-cleavage product that can act as a ligand sink, thereby inhibiting MERTK-functions such as efferocytosis (Sather, et al., 2007; Y. Zhang, et al., 2019) and possibly innate immune control (Thorp, et al., 2011). In the context of nonalcoholic steatohepatitis progression, ADAM17-mediated MERTK cleavage is suppressed, leading to increased TGF-ß production (Cai, et al., 2019).

Although it is not known how MERTK expression is regulated in normal or cancerous cells, it has been demonstrated that expression is significantly induced as monocytes differentiate into macrophages (Graham, et al., 1994) and in macrophages and epithelial cells in response to phagocytosis of apoptotic cells (Graham, et al., 1994; N, et al., 2009; Sandahl, Hunter, Strunk, Earp, & Cook, 2010; Shao, et al., 2010; Zizzo, Hilliard, Monestier, & Cohen, 2012). In macrophages, phagocytosis-induced MERTK expression is mediated by direct binding of the transcription factor liver X receptor (LXR) to the Mertk promoter (A-Gonzalez, et al., 2009). LXR is likely activated through binding of modified cholesterols derived from ingested apoptotic membranes.

Additionally, Mertk is induced in macrophages and dendritic cells in response to corticosteroids (Zagorska, et al., 2014; Zahuczky, Kristóf, Majai, & Fésüs, 2011) and by lytic transcription factors of the Epstein-Barr virus (Y. Li, et al., 2004).

Physiologic MERTK Functions

Physiologically, MERTK is centrally involved in regulating tissue homeostasis and repair as well as innate immune control. In many cases these functions are linked to MERTK’s role in mediating efferocytosis by monocyte-derived immune cells, such as macrophages, and by epithelial cells.

Tissue Homeostasis and Repair

In contrast to their wild-type counterparts, Mertk knock-out (Mertk^−/−) macrophages do not clear apoptotic cells and debris efficiently (Scott, et al., 2001). Interestingly, Mertk^−/− macrophages recognize and bind apoptotic cells but fail to engulf them, indicating that MERTK is needed to trigger phagocytosis (Scott, et al., 2001). Indeed, MERTK activation induces changes in cytoskeletal architecture and cell shape that are needed for engulfment of apoptotic cells (Guttridge, et al., 2002). Additionally, upon ligand-dependent activation MERTK interacts with and phosphorylates the guanine nucleotide exchange factor VAV1 (Mahajan & Earp, 2003). VAV1 then activates RHOA family members, which are well-described regulators of the cytoskeleton.

Important biological processes that require MERTK-dependent efferocytosis by epithelial cells during tissue modeling and repair include clearance of dead germ cells by Sertoli cells in the testes (Sun, et al., 2010), clearance of apoptotic material by mammary epithelial cells during weaning-induced involution (Sandahl, et al., 2010) and efferocytosis by podocytes in the renal glomerulus following nephrotoxic injury (Shao, et al., 2010). In the retina, MERTK is required for pigmented epithelial cell-mediated efferocytosis of apoptotic material shed by rods and cones (Duncan, et al., 2003; Prasad, et al., 2006) and Mertk^−/− mice exhibit scarring and retinal degeneration (Duncan, et al., 2003). Mertk mutations are causative for retinitis pigmentosa in rats (D’Cruz, et al., 2000) and have been associated with human retinitis pigmentosa in multiple family cohorts (Bhatia, Kaur, Singh, & Vanita, 2019; Charbel Issa, et al., 2009; Gal, et al., 2000; Jinda, et al., 2016; S. Liu, et al., 2019; M. Yang, et al., 2018). These findings have led to the testing of MERTK correcting gene therapies for retinitis pigmentosa (Artero Castro, et al., 2019; Conlon, et al., 2013; DiCarlo, Mahajan, & Tsang, 2018), including an ongoing phase I trial (NCT01482195).

MERTK-dependent efferocytosis by murine microglia, the brain-resident macrophage, is important during adult neurogenesis and for the pruning of neurite endings at non-productive synapses (Caberoy, Alvarado, & Li, 2012; Chung, et al., 2013; Fourgeaud, et al., 2016; Ji, et al., 2013). Additionally, MERTK is required for phagocytosis of myelin debris in the brain and reduced MERTK expression in macrophages correlating with reduced phagocytotic capacity has been linked to multiple sclerosis (Healy, et al., 2017; Healy, et al., 2016).

Innate Immune Control

In addition to regulating tissue homeostasis, MERTK signaling is involved in innate immune control where it contributes to resolving inflammation and wound healing. MERTK thereby crucially assists in maintaining the delicate and tightly regulated balance between pro-inflammatory responses to pathogens or tumor cells and prevention of autoimmune reactions that could lead to tissue damage or promote tumorigenesis (Grivennikov, Greten, & Karin, 2010). While Mertk^−/− mice are viable without apparent developmental defects, they develop splenomegaly and succumb to endotoxic shock in response to much lower doses of lipopolysaccharide (LPS) than wild-type mice (Camenisch, Koller, Earp, & Matsushima, 1999; Lu & Lemke, 2001). Mertk^−/− macrophages respond to LPS treatment with significantly higher production of the pro-inflammatory cytokine tumor necrosis factor-α (TNF-α) (Camenisch, et al., 1999) and loss of TAM receptors in murine microglia leads to elevated inflammatory responses to LPS in the hippocampus (Ji, et al., 2013). Additionally, Mertk^−/− mice display increased levels of autoantibodies (Cohen, et al., 2002; Scott, et al., 2001). In a murine model of cardiac infarct, Mertk^−/− led to delayed inflammatory resolution (Wan, et al., 2013) and MERTK cleavage by ADAM17 during bleomycin-induced lung damage led to increased inflammatory response and damage (Y. J. Lee, et al., 2012).

One way that MERTK dampens inflammation is by preventing the release of intracellular antigens from apoptotic cells through mediating efferocytosis (Cohen, et al., 2002; Scott, et al., 2001). Additionally, MERTK signaling can polarize macrophages toward a wound-healing M2 phenotype associated with anti-inflammatory gene expression, cytokine profiles, and functions. Macrophage phenotypes can be broadly categorized as pro-inflammatory M1 or anti-inflammatory M2, although the transition between the two states is fluid and many different stages of polarization can occur (Murray, 2017). Circulating monocytes largely lack MERTK (Behrens, et al., 2003; Lee-Sherick, et al., 2013; Zizzo, et al., 2012), but they strongly upregulate its expression as they migrate into the tissue and differentiate into macrophages (Graham, et al., 1994; Zizzo, et al., 2012).

MERTK-mediated clearance of apoptotic material induces M2 polarization (Filardy, et al., 2010; Tibrewal, et al., 2008). As such, MERTK signaling increases the expression of anti-inflammatory M2 cytokines including IL-10, TGF-β, and hepatocyte growth factor (HGF) (Alciato, Sainaghi, Sola, Castello, & Avanzi, 2010; Eken, et al., 2010; H. J. Park, Baen, Lee, Choi, & Kang, 2012; Wallet, et al., 2008; Zizzo, et al., 2012). Interestingly, M2-polarized macrophages produce increased levels of the MERTK ligand GAS6, indicating the potential for an autocrine feedback loop that enhances anti-inflammatory polarization (Zizzo, et al., 2012). Furthermore, MERTK signaling in macrophages and dendritic cells decreases the production of pro-inflammatory cytokines, such as TNF-α, IL-1, IL-6 and IL-12p70 by inhibiting nuclear factor κ-light-chain-enhancer of activated B cells (NF-κB) and phosphatidylinositol 3-kinase (PI3K)/AKT signaling and activating suppressor of cytokine signaling 1 (SOCS1) (Alciato, et al., 2010; Camenisch, et al., 1999; Eken, et al., 2010; Sen, et al., 2007; Tibrewal, et al., 2008; B. Zhang, et al., 2019). Additionally, MERTK signaling promotes production and release of specialized pro-resolving mediators (SPMs) by macrophages. These fatty acid derived factors contribute to the resolution of inflammatory processes. MERTK cleavage suppressed SPM release and mice expressing cleavage-resistant MERTK had increased circulating SPMs in several inflammation models (Cai, et al., 2018; Cai, et al., 2017; Cai, et al., 2016).

Clearance of apoptotic material is not the only MERTK-mediated mechanism that can induce anti-inflammatory phenotypes in monocyte-derived immune cells. Activated T cells can express both PROS1 and PtdSer, which can directly bind MERTK on dendritic cells to suppress the production of pro-inflammatory cytokines (Carrera Silva, et al., 2013). In dendritic cells, TAM kinase signaling dampens the inflammatory response to toll-like receptor (TLR) engagement by promoting expression of SOCS1 and SOCS3 (Rothlin, Ghosh, Zuniga, Oldstone, & Lemke, 2007). Even though induction of this regulatory mechanism is primarily dependent on AXL, lack of either TAM kinase renders dendritic cells hyper-reactive to TLR stimulation. In NK cells, TAM kinase signaling decreases anti-cancer immunity through action of the E3 ubiquitin ligase casitas B-lineage lymphoma-b (CBL-B), which also mediates a negative feedback loop to down regulate TAM RTKs, although it is not yet understood what role MERTK plays compared to the other family members (Chirino, et al., 2019; Paolino, et al., 2014).

Platelet Aggregation

MERTK plays a physiologic role in regulating platelet aggregation. During the initial aggregation phase, platelets externalize PtdSer and secrete GAS6 and all three TAM kinases are needed to facilitate integrin signaling, adherence to fibrinogen, platelet spreading, and the release of aggregation-promoting factors (Angelillo-Scherrer, et al., 2008; Angelillo-Scherrer, et al., 2001; Branchford, et al., 2018; C. Chen, et al., 2004). Importantly, while Mertk^−/− mice and mice treated with a MERTK kinase inhibitor displayed platelet defects, tail bleeding times did not significantly differ from wild-type mice (Branchford, et al., 2018; C. Chen, et al., 2004) and impaired GAS6-TAM kinase signaling did not affect formation of an initial platelet monolayer (Angelillo-Scherrer, et al., 2001), suggesting that bleeding-related side effects of MERTK targeting therapies could be safely managed. Interestingly, treatment with the MERTK-selective inhibitor UNC2025 blocked pulmonary embolism and arterial thrombosis in murine models, indicating a potential clinical application for MERTK targeting agents (Branchford, et al., 2018).

MERTK in Cancer Cells

MERTK Expression in Human Malignancy

MERTK is aberrantly expressed in a variety of malignancies including acute myeloid leukemia (AML) (Lee-Sherick, et al., 2013), acute lymphoblastic leukemia (ALL) (Graham, et al., 1994; Graham, et al., 2006; Linger, Lee-Sherick, et al., 2013), lymphoma (Shi, et al., 2018), lung cancer (Linger, Cohen, et al., 2013), astrocytoma (Keating, et al., 2010), breast cancer (Nguyen, et al., 2014), gastric cancer (Yi, et al., 2017), melanoma (Schlegel, et al., 2013), rhabdomyosarcoma (Khan, et al., 1999), prostate cancer (Y. M. Wu, Robinson, & Kung, 2004), colon cancer (Frejno, et al., 2017), glioblastoma multiforme (GBM) (Y. Wang, et al., 2013), head and neck squamous cell cancer (HNSCC) (von Massenhausen, et al., 2016), neuroblastoma (Y. Li, Wang, Bi, Zhao, & Yu, 2015), and schwannoma (Ammoun, et al., 2014), implicating MERTK as a potential therapeutic target.

MERTK was initially identified from a B lymphoblast library and designated c-mer (Graham, et al., 1994). Ectopic expression of MERTK was demonstrated in ALL and AML cells and subsequent studies demonstrated MERTK protein expression in greater than 80% of pediatric and adult AML patient samples, compared with little or no expression in normal myeloid bone marrow precursor cells (Lee-Sherick, et al., 2013). MERTK protein was also ectopically expressed in 30% of pediatric B-cell ALL (B-ALL) and 40–50% of T-cell ALL (T-ALL) samples (Graham, et al., 2006; Linger, Lee-Sherick, et al., 2013). Although MERTK is expressed across biologic subsets of leukemia, in B-ALL abnormal expression is uniformly found in subsets expressing the E2A-PBX1 fusion protein associated with the t(1;19) translocation (Sawczyn, Porter, DeGregori, & Graham, 2007; Shiozawa, Pedersen, & Taichman, 2010) and in T-ALL, MERTK is preferentially expressed in subsets with an immature phenotype. In mantle cell lymphoma (MCL), a clinically aggressive B cell lymphoma, MERTK protein was evident in approximately half of patient samples evaluated (Shi, et al., 2018).

Aberrant MERTK expression has also been demonstrated in numerous solid malignancies and in some cases has been correlated with poor prognosis and/or disease progression. MERTK protein overexpression was detected in 69% of non-small cell lung cancers (NSCLCs) relative to normal lung tissue adjacent to tumors (Linger, Cohen, et al., 2013). MERTK is also highly expressed in astrocytoma patient samples and cell lines (Keating, et al., 2010). Of 34 breast carcinoma samples analyzed, 70% had highly elevated levels of MERTK protein compared with all normal samples evaluated (Nguyen, et al., 2014). MERTK RNA was also induced in three PAX3-FKHR-transduced cell lines, a fusion which characterizes a majority of alveolar rhabdomyosarcoma cases (Khan, et al., 1999) and MERTK is significantly overexpressed in schwannoma tissue (Ammoun, et al., 2014). In gastric cancer, high levels of MERTK mRNA or protein were associated with poor prognosis (Yi, et al., 2017) and in colorectal cancer, high MERTK expression was associated with worse overall and disease-free survival (Frejno, et al., 2017). Finally, MERTK expression in melanoma correlated with disease progression, with the highest MERTK protein and mRNA levels in metastatic melanoma and little expression in nevi (Schlegel, et al., 2013).

MERTK and Survival Signaling in Cancer Cells

MERTK expression is associated with growth factor independence and decreased apoptosis, particularly in stressful conditions such as limited nutrients or hypoxia (Figure 1) (Graham, et al., 2014). Apoptosis induced by IL-3 withdrawal was prevented by MERTK expression in IL-3-dependent cell lines (Georgescu, Kirsch, Shishido, Zong, & Hanafusa, 1999; Guttridge, et al., 2002). MERTK expression in the Jurkat T-ALL cell line led to decreased apoptosis, whereas MERTK knock-down using short-hairpin RNA (shRNA) resulted in increased apoptosis in neuroblastoma cell line cultures (Y. Li, et al., 2015), in serum starved cultures of AML, T-ALL, NSCLC and astrocytoma cell lines (Fan, et al., 2007; Keating, et al., 2010; Linger, Cohen, et al., 2013) and in B-ALL, T-ALL, MCL, NSCLC and glioblastoma cell line cultures treated with cytotoxic chemotherapy (Keating, et al., 2010; Linger, Cohen, et al., 2013; Linger, Lee-Sherick, et al., 2013; Shi, et al., 2018; Y. Wang, et al., 2013). MERTK shRNA also reduced colony-forming potential in B-ALL, AML, melanoma and astrocytoma cell line cultures (Keating, et al., 2010; Lee-Sherick, et al., 2013; Linger, Lee-Sherick, et al., 2013; Schlegel, et al., 2013; Tworkoski, et al., 2013). Moreover, MERTK is critical for oncogenesis in murine models. Aberrant expression of MERTK in the hematopoietic system led to development of T-ALL and in some cases B-ALL in transgenic mice (Keating, et al., 2006) and MERTK inhibition in tumor cells using shRNA decreased tumorigenesis in xenograft models of NSCLC, melanoma, glioblastoma, B-ALL, T-ALL, and AML (Knubel, et al., 2014; Lee-Sherick, et al., 2013; Linger, Cohen, et al., 2013; Linger, Lee-Sherick, et al., 2013; Schlegel, et al., 2013; Tworkoski, et al., 2013; Xue, et al., 2017). All of these data point to critical oncogenic roles for MERTK in cancer cells, particularly as a pro-survival mechanism in the context of cell stress.

Figure 1: MERTK Regulating Hallmarks of Cancer.

MERTK mediates many hallmarks of cancer including resistance to cell death, evasion of growth suppressors, induction of drug resistance, avoidance of immune destruction, sustaining of proliferative signaling, as well as activation of invasion and metastases. In some contexts, MERTK signaling may be transforming (not shown). Hence, targeting MERTK to inhibit these oncogenic functions represents an attractive strategy for cancer therapies.

BAX (Bcl-2-associated X),

PUMA (p53 upregulated modulator of apoptosis),

BCL-XL (B-cell lymphoma-extra-large),

BCL-2 (B-cell lymphoma 2),

MCL1, (myeloid cell leukemia 1),

ACK1 (activated CDC42 kinase 1),

WWOX (WW domain-containing oxidoreductase),

YAP (yes-associated protein 1),

TAZ (transcriptional coactivator with PDZ-binding motif),

PD-1 (programmed cell death protein 1)

PD-L1 (programmed cell death 1 ligand 1)

M1/2 (macrophage polarization 1/2)

Ligand binding leads to homodimerization and auto-phosphorylation of MERTK and MERTK activation leads to signaling through canonical oncogenic pathways such as MEK/ERK, PI3K/AKT and JAK/STAT, resulting in cell growth and proliferation, evasion of apoptosis, cell cycle progression, and tumor growth (Figure 2). Early studies used chimeric receptors containing the MERTK transmembrane and intracellular domains to investigate MERTK signaling in the absence of a known ligand and identified the RAF/MEK/ERK, PI3K/AKT/p70 S6K, p38, PLCγ, and NF-κB pathways downstream of MERTK, with the SHC and GRB2 adapter proteins playing a role in activation of several of these pathways (Georgescu, et al., 1999; Guttridge, et al., 2002; Ling & Kung, 1995; Y. M. Wu, et al., 2004).

Figure 2: MERTK Signaling in Cancer.

MERTK activation and subsequent signaling regulates numerous oncogenic pathways. Depending on cancer type, context, and the specific ligand (GAS6, PROS1, Tubby, TULP-1 or Galectin-3), MERTK signaling results in increased proliferation, anti-apoptosis and survival, migration and metastasis, anchorage-independent growth, cancer stem cell (CSC) maintenance, as well as expression of the immune checkpoint protein PD-L1. In certain contexts, MERTK signaling may be transforming. MERTK has also been found in the nucleus of leukemic cells but the function remains unknown.

GAS6 (growth arrest-specific 6),

PROS1 (vitamin K-dependent protein S),

TULP-1 (tubby-like protein 1),

MEK (mitogen-activated protein kinase),

ERK (extracellular signal-regulated kinases),

PI3K (phosphatidylinositol 3-kinases),

SAV1 (protein salvador homolog 1),

BAD (BCL-2 associated agonist of cell death),

mTOR (mammalian target of rapamycin),

BCL-2 (B-cell lymphoma 2),

MCL1, (myeloid cell leukemia 1),

BCL-XL (B-cell lymphoma-extra large),

PUMA (p53 upregulated modulator of apoptosis),

BAX (Bcl-2-associated X),

FAK1 (focal adhesion kinase 1),

RHOA (Ras homolog family member A),

RAC (Ras-related C3 botulinum toxin substrate),

NF-ĸB (nuclear factor-ĸB),

CREB (cAMP-responsive element-binding protein),

ATF1 (activating transcription factor 1),

JAK (Janus kinase),

STAT 3/5/6 (signal transducer and activator of transcription 3/5/6),

PD-L1 (programmed cell death 1 ligand 1)

ACK1 (activated CDC42 kinase 1),

WWOX (WW domain-containing oxidoreductase),

YAP (yes-associated protein 1),

TAZ (transcriptional coactivator with PDZ-binding motif),

SRC (sarcoma),

p-SRC (phospho-SRC),

KRAS (Kirsten rat sarcoma viral oncogene homolog),

SOX2 (Sry-related HMG box),

Zeb1 (Zinc finger E-box-binding homeobox 1),

Later studies defined more specific roles for several of these pathways in cancer cells. RAS activation downstream of MERTK promotes gene transcription via the RAS/RAF/MEK/ERK pathway, leading to cell proliferation, survival, and metastases (Knight & Irving, 2014; Linger, et al., 2008; Linger, Lee-Sherick, et al., 2013; Roberts & Der, 2007). Stimulation with GAS6 ligand led to activation of ERK1/2 and shRNA mediated Mertk knockdown led to decreased ERK activation in T-ALL, B-ALL, AML, NSCLC, melanoma and astrocytoma cell lines (Keating, et al., 2010; Lee-Sherick, et al., 2013; Linger, Cohen, et al., 2013; Linger, Lee-Sherick, et al., 2013; Schlegel, et al., 2013). MERTK activation also led to pro-survival signaling via regulation of PI3K/AKT in B-ALL, NSCLC, melanoma, and astrocytoma cell lines (Keating, et al., 2010; Linger, Cohen, et al., 2013; Linger, Lee-Sherick, et al., 2013; Schlegel, et al., 2013).

Furthermore, PI3K activation has been linked to MERTK activation in NIH 3T3 fibroblasts (Ling & Kung, 1995) and AKT has been implicated downstream of PI3K in macrophage activation and polarization, cell growth and proliferation, and cell survival (Graham, et al., 2014; Myers, Amend, & Pienta, 2019). MERTK knockdown also leads to decreased PI3K expression in B-ALL cells relative to control cells (Linger, Lee-Sherick, et al., 2013). More recently, a novel role for MERTK as an activator of AKT was identified. Salvador family WW domain containing protein 1 (SAV1), which plays roles in transcription, RNA splicing and protein degradation and has been implicated as a tumor suppressor, binds AKT and prevents its localization to the cell membrane, thereby blocking AKT activation (Jiang, et al., 2019). MERTK phosphorylates the AKT pleckstrin-homology domain to prevent SAV1 binding, thereby enhancing AKT activity.

Numerous other signaling pathways implicated in tumor cell proliferation and survival and oncogenic transformation are also regulated downstream of MERTK. MERTK regulates ACK1, which subsequently leads to the degradation of tumor suppressor WWOX, promoting tumor development (Linger, Keating, Earp, & Graham, 2010). JAK-STAT signaling was activated downstream of MERTK in T-ALL and melanoma cell lines and was responsive to GAS6 in AML cells (Lee-Sherick, et al., 2013; Linger, Lee-Sherick, et al., 2013; Schlegel, et al., 2013). Phosphorylation of p38 was activated downstream of MERTK in B-ALL and AML cells and CREB protein levels were decreased in NSCLC and AML cells in response to MERTK inhibition (Lee-Sherick, et al., 2013; Linger, Cohen, et al., 2013; Linger, Lee-Sherick, et al., 2013). NF-kB is activated by MERTK to promote anti-apoptotic and survival signaling in tumor cells and also has a distinct role as a key regulator of the normal immune response (Camenisch, et al., 1999; Crittenden, et al., 2016; Georgescu, et al., 1999; Karin, 2006; Sen, et al., 2007; Tibrewal, et al., 2008). In this context, MERTK inhibits NF-kB signaling and inflammatory cytokine production in innate immune cells, thereby suppressing the immune response (Sen, et al., 2007; Tibrewal, et al., 2008). Consequently, MERTK inhibition may both decrease pro-survival signaling through NF-kB in cancer cells and promote an immune response to tumor cells. Silencing of MERTK in glioblastoma cells led to down-regulation of SRC proteins and the STAT3/KRAS pathway, important in the maintenance of glioblastoma stem-like cells, as well as downstream targets Musashi-1, SOX2, ZEB1, and slug (Eom, et al., 2018). MERTK also directly phosphorylates YAP/TAZ, which are central mediators of increased cell proliferation, transformation, motility, and angiogenesis in multiple tumor types (T. Azad, et al., 2019). More extensive signaling networks have also been defined downstream of GAS6 stimulation in AML (ERK1/2, AKT, p38, MSK1, CREB, ATF1, STAT6, SRC-family kinases), NSCLC (ERK1/2, AKT, p38, MSK1/2, CREB, MEK1/2, GSK3α/β, mTOR, FAK) and melanoma (ERK1/2, AKT, p38, STAT6, GSK3α/β, FAK, AMPKα) (Lee-Sherick, et al., 2013; Linger, Cohen, et al., 2013), although the contribution of MERTK relative to the other TAM-family kinases has only been confirmed for a subset of these pathways.

Ultimately numerous of these pathways converge on pro-survival processes, including BCL-2 family members and other regulators of apoptosis. In NSCLC cells, shRNA-mediated knock-down of MERTK led to reduced AKT/CREB signaling and reduced levels of the anti-apoptotic proteins BCL-XL and survivin (Linger, Cohen, et al., 2013). AKT is a known activator of BCL-XL expression (Pugazhenthi, et al., 2000) and also inhibits phosphorylation of the pro-apoptotic BCL-2 family member BAD (Datta, et al., 1997). In B-ALL cell lines, shRNA-mediated MERTK inhibition led to upregulation of the pro-apoptotic proteins BAX, PUMA/BBC3, and NOXA/PMAIP1 and down-regulation of the anti-apoptotic proteins BCL-XL, PI3K, and protein kinase C (PKC) (Linger, Lee-Sherick, et al., 2013). Similarly, shRNA-mediated MERTK inhibition increased expression of pro-apoptotic BAX and decreased expression of anti-apoptotic BCL-2, MCL-1, and BCL-XL in MCL cells (Shi, et al., 2018) and decreased expression of BCL-2 in T-ALL cells (Fan, et al., 2007).

MERTK can also localize to the nucleus. While the role of MERTK in the nucleus has not been defined, long-term GAS6 exposure increased the fraction of chromatin-bound nuclear MERTK in leukemia cells, consistent with a potential role for MERTK in regulation of gene expression (Migdall-Wilson, et al., 2012), which would provide an entirely new mechanism of downstream signaling for MERTK. Other oncogenic mechanisms that have been associated with MERTK but are not well-explored include induction of autophagy and cellular senescence (Rogers, et al., 2012; Sufit, et al., 2016).

MERTK and Metastasis

Metastatic disease is a leading cause of morbidity and mortality, and MERTK has been implicated in tumor metastasis (Figure 1). Metastasis involves numerous processes that have been associated with MERTK, including immune evasion, cytoskeletal alterations, increased cell motility and invasive potential, and facilitation by macrophages. Remodeling of the extracellular matrix, for example, is promoted by M2 tumor-associated macrophages, and GAS6 mediated MERTK signaling promotes pro-tumor M2 macrophage polarization (Finkernagel, et al., 2016; Myers, et al., 2019). MERTK is also expressed at higher levels in M2 macrophages relative to M1 macrophages (Myers, et al., 2019). Mouse models of postpartum breast cancers lacking MERTK or treated with a MERTK inhibitor demonstrated decreased M2 macrophages and decreased tumor metastases (Stanford, et al., 2014). Similarly, inhibition of Mertk in the tumor microenvironment in Mertk^−/− mice decreased lung metastases in syngeneic murine breast cancer, colon cancer, and melanoma models (Cook, et al., 2013). MERTK activation also leads to cytoskeletal changes by disrupting an inhibitory interaction between MERTK and VAV1, thereby allowing VAV1 to activate RHO family members (RAC1, RHO, CDC42) (Mahajan & Earp, 2003), mediate production of the motility factor HGF (H. J. Park, et al., 2012), and phosphorylate FAK, critical to cell migration (Tang, et al., 2015). In a murine leukemia cell line, activation of an EGFR/MERTK chimeric receptor led to cell flattening, increased cell adherence, extension of dendritic-like processes, and cytoskeletal alterations (Guttridge, et al., 2002). Morphologic changes also occurred in GBM cell lines with shRNA knockdown of MERTK, leading to decreased cell migration and invasion in vitro (Rogers, et al., 2012; Y. Wang, et al., 2013) and in murine GBM xenografts (Eom, et al., 2018). MERTK overexpression also increased migration and invasion in an oral squamous cancer cell line (von Massenhausen, et al., 2016) and shRNA-mediated MERTK inhibition decreased migration and/or invasion in HNSCC (von Massenhausen, et al., 2016), MCL (Shi, et al., 2018), neuroblastoma (Y. Li, et al., 2015), and NSCLC (Xie, et al., 2015) cell lines. Potential mechanisms include altered FAK signaling, crucial to cellular movement, and changes in RHOA signaling, known to regulate cytoskeletal actin, ultimately leading to cellular fixation and decreased motility (Y. Li, et al., 2015; Rogers, et al., 2012; von Massenhausen, et al., 2016; Xie, et al., 2015). Expression and phosphorylation of myosin light chain 2, which is required for actomyosin ATPase activity and contractility, was also reduced in GBM cells with MERTK knock-down (Y. Wang, et al., 2013). MERTK activation upregulated factors associated with tumor progression and metastasis in prostate cancer cell lines, including IL-8, angiogenic factors, CXC chemokines, and bone morphogenic factors (Y. M. Wu, et al., 2004). As mentioned previously, MERTK expression was upregulated in metastatic melanoma relative to localized melanoma or benign nevi (Schlegel, et al., 2013). These roles for MERTK in remodeling of the extracellular matrix, mitigation of cellular adhesion, and promotion of immune suppression as well as the decreased incidence of metastasis in multiple tumor types upon MERTK inhibition, demonstrate important roles for MERTK in metastatic disease.

MERTK in Chemoresistance

MERTK plays several roles in chemoresistance, including enhancing survival signaling in tumor cells treated with traditional cytotoxic chemotherapies and providing escape mechanisms for tumors treated with various targeted agents (Figure 1). Numerous studies have demonstrated increased sensitivity to cytotoxic chemotherapies in response to MERTK inhibition in both leukemias and solid tumors. MERTK knockdown with shRNA resulted in decreased pro-survival signaling through AKT and ERK1/2 in B-ALL cells treated with methotrexate and increased apoptosis in B-ALL cells treated with vincristine, methotrexate, dexamethasone or L-asparaginase (Linger, Lee-Sherick, et al., 2013). MERTK inhibition also increased the sensitivity of T-ALL cells to cytarabine, etoposide, mercaptopurine and/or methotrexate (Linger, Lee-Sherick, et al., 2013) and increased the sensitivity of MCL cells to vincristine and doxorubicin (Shi, et al., 2018). Conversely, ectopic expression of MERTK in lymphocytes from Mertk transgenic mice, recapitulating the expression of MERTK in T-ALL, conferred resistance to dexamethasone-induced apoptosis compared to lymphocytes from wild-type mice (Keating, et al., 2006). In solid tumor cell lines, MERTK inhibition sensitized GBM cells to treatment with carboplatin, temozolomide, vincristine or etoposide (Keating, et al., 2010; Y. Wang, et al., 2013), sensitized NSCLC cells to treatment with cisplatin or carboplatin (Linger, Cohen, et al., 2013), and decreased IC50 values by 2–40 fold in neuroblastoma cells treated with vincristine or cisplatin relative to control cells (Y. Li, et al., 2015). Numerous tyrosine kinase inhibitors have been employed in the management of various malignancies with generally less toxic side effects compared with standard chemotherapy, and MERTK inhibition may offer a unique therapeutic target to sensitize human cancers to cytotoxic chemotherapy.

MERTK has also been implicated in resistance to molecularly targeted agents. In particular, MERTK shares downstream signaling pathways with other TAM family kinases and common oncogenic drivers such as EGFR in NSCLC and B-RAF in melanoma and can provide bypass signaling to compensate for these proteins in the context of therapeutic inhibitors. For instance, although AXL is expressed in numerous aggressive epithelial cancers, AXL inhibition had limited efficacy in tumor cell lines with high levels of MERTK expression (McDaniel, et al., 2018). In this context, MERTK expression and activity was induced following AXL inhibition and overexpression of MERTK was sufficient to confer resistance to AXL inhibitors. Moreover, MERTK inhibition sensitized MERTK-expressing tumor cells to AXL inhibition, both in vitro and in HNSCC and triple-negative breast cancer xenograft models. These data provide rationale for development of therapeutic strategies targeting MERTK and AXL together. Similarly, overexpression of MERTK in a NSCLC cell line was sufficient to confer resistance to the EGFR inhibitor erlotinib (Xie, et al., 2015) and more recent data demonstrated upregulation of MERTK in NSCLC cell line derivatives with acquired resistance to osimertinib and in NSCLC xenografts treated with osimertinib, the preferred front-line EGFR inhibitor in NSCLC patients (D. Yan, et al., 2019). Moreover, MERTK inhibitors synergized with osimertinib and other third-generation EGFR inhibitors to inhibit expansion of EGFR-expressing NSCLC cells, irrespective of EGFR mutational status, implicating MERTK as a mediator of bypass signaling in the context of EGFR overexpression or mutation (D. Yan, et al., 2019; D. Yan, et al., 2018). Indeed, downstream signaling through ERK1/2 and AKT was not effectively blocked in response to treatment with either single agent, but was completely abrogated in response to the combination. Consistent with the biochemistry, treatment with the combination blocked tumor growth in a NSCLC xenograft model, even after treatment was stopped (D. Yan, et al., 2018). In contrast, while treatment with EGFR inhibitor alone controlled tumor growth during therapy, it did not provide a durable response. A similar interaction was observed in NSCLC cell cultures treated with osimertinib in combination with ONO-7475, a dual MERTK and AXL inhibitor, and tumor volume was significantly reduced in response to treatment with osimertinib and ONO-7475 relative to treatment with osimertinib alone (Okura, et al., 2020). Similarly, MAPK/ERK and PI3K/AKT have been implicated in resistance to BRAF inhibitors in BRAF mutant melanomas(Atefi, et al., 2011; Johannessen, et al., 2010; Nazarian, et al., 2010; Poulikakos, et al., 2011; Schlegel, et al., 2013; Xue, et al., 2017) and MERTK upregulation has been demonstrated in BRAF mutant melanomas with acquired resistance to BRAF inhibitors or BRAF and MEK inhibitor combination therapy (Xue, et al., 2017). While treatment with the BRAF inhibitor vemurafenib was sufficient to block ERK1/2 signaling in BRAF mutant melanoma cell lines, even in the presence of GAS6 ligand, combined treatment with MERTK and BRAF inhibitors was necessary to effectively block downstream signaling through AKT and STAT6 (Sinik, et al., 2019). MERTK inhibition using shRNA sensitized BRAF mutant melanomas to BRAF inhibition, leading to decreased tumor volume in vivo (Xue, et al., 2017). Treatment with pharmacologic MERTK and BRAF inhibitors together also decreased colony-formation in BRAF-mutant melanoma cell line cultures and reduced tumor growth in a BRAF-mutant patient derived melanoma xenograft model more effectively than BRAF inhibition alone (Sinik, et al., 2019). Colony formation was similarly decreased in response to combined MERTK and MEK inhibition, suggesting that MEK is the critical downstream mediator for BRAF in this context. Thus, in several different tumor types, MERTK inhibition combined with targeted therapy may provide enhanced therapeutic efficacy, while limiting the toxicity associated with classical chemotherapeutic agents.

MERTK in Anti-Tumor Immunity

MERTK signaling controls innate immunity as part of a regulatory feedback mechanism that limits the extent of inflammatory responses. Thus, under physiologic conditions, MERTK prevents chronic inflammation and auto-immunity. However, in the context of cancer, MERTK can be subverted and contributes to an immune-suppressive microenvironment that promotes cancer growth and progression.Therefore, inhibiting MERTK signaling in the tumor microenvironment represents a potential immunotherapeutic strategy for cancer patients.

MERTK Signaling Induces Immunosuppressive Cytokines

The first evidence that TAM RTKs play a role in suppressing anti-tumor immunity came from a study that demonstrated significantly reduced tumor growth in immunocompetent mouse models of pancreatic, colon and breast cancer lacking the TAM kinase ligand GAS6 (Gas6^−/− mice) (S. Loges, et al., 2010). In mice with colon cancer grafts, the reduction in tumor growth could be overcome by transplant with wild-type bone marrow cells, implicating MERTK inhibition in hematopoietic cells in the tumor microenvironment as a mechanism of anti-tumor activity. In this model, colon cancer cells secreted IL-10 and macrophage colony-stimulating factor (M-CSF) to induce GAS6 production by macrophages, thereby presumably potentiating tumor-promoting TAM-signaling. In later studies, Mertk^−/− in the tumor microenvironment in immune-competent breast cancer and melanoma models significantly reduced tumor incidence, size and lung metastases (Cook, et al., 2013). In this case, transplant of Mertk^−/− bone marrow into lethally irradiated wild-type mice was sufficient to provide protection from breast cancer, indicating that MERTK signaling on hematopoietic cells promotes tumor growth and metastasis. Mertk^−/− mice had a more pro-inflammatory cytokine profile than wild-type mice, characterized by increased levels of the pro-inflammatory cytokines IL-6 and IL-12p40 and decreased levels of anti-inflammatory IL-10, accompanied by increased lymphocyte proliferation and an increased incidence of cytotoxic CD8⁺ T cells in tumors. Following tumor antigen recognition, CD8⁺ T cells are capable of releasing cytokines and lytic molecules that mediate killing of tumor cells (Shankaran, et al., 2001). Of note, the observed tumor latency in Mertk^−/− mice was dependent on CD8⁺ T cells in antibody depletion experiments (Cook, et al., 2013). However, MERTK is not expressed on murine T cells, even after T cell receptor (TCR) stimulation (Behrens, et al., 2003; Graham, et al., 1995), suggesting thatMERTK signaling in other immune cells, such as tumor-associated macrophages, suppresses CD8⁺ T cell anti-tumor activity by inducing anti-inflammatory cytokine production to promote an immune-suppressive microenvironment.

A subsequent study linked MERTK’s role in inhibiting anti-tumor immunity to its function as a mediator of efferocytosis (Stanford, et al., 2014). During postpartum mammary gland involution, cell death is widespread and cell debris is cleared by MERTK-dependent efferocytosis, which activates MERTK and induces anti-inflammatory cytokines. Thus, in an immune-competent postpartum breast cancer model, mammary gland involution led to an anti-inflammatory phenotype and increased tumor metastasis in wild-type mice. In contrast, metastasis was significantly decreased in Mertk^−/− mice and in wild-type mice treated with the pan-TAM inhibitor BMS-777607. In this context, MERTK-dependent efferocytosis by macrophages was decreased, leading to decreased M2 polarization and decreased levels of wound healing cytokines, such as IL-4, IL-10, IL-13 and TGF-β. Thus, MERTK-dependent induction of an immune-suppressive microenvironment is at least in part a consequence of macrophage M2-polarization following efferocytosis of apoptotic cells in the tumor tissue. Consistent with these findings, co-culture of irradiated and presumably apoptotic breast cancer cells with a macrophage cell line increased MERTK-dependent release of anti-inflammatory IL-10 and radiation therapy increased mRNA expression of MERTK and its ligands PROS1 and GAS6 in tumor-associated macrophages in a colorectal cancer model (Crittenden, et al., 2016). Additionally, blockade of the apoptotic “eat-me” signal with the anti-PtdSer antibody 2aG4 shifted macrophages to a pro-inflammatory phenotype, further supporting the notion that MERTK-dependent efferocytosis and immunosuppression are connected functions (Yin, Huang, Lynn, & Thorpe, 2013).

Immunotherapeutic targeting of MERTK could benefit patients with a broad range of cancers. Melanoma, breast and prostate cancer cells induce MERTK and TYRO3-dependent inhibition of macrophage M1 polarization, and at least melanoma cells achieve this by secreting PROS1 (Ubil, et al., 2018). Tumor-derived PROS1 mediates MERTK dimerization on primary murine peritoneal macrophages and activated MERTK forms a complex with PTP1b and p38α, leading to p38α inhibition and decreased p38-dependent expression of pro-inflammatory M1 cytokines such as IL1, IL6, and TNF-α. Thus, PROS1 activates MERTK to suppress the immune response in the tumor microenvironment. In this context, tumor cell production of PROS1 is stimulated by IFN-γ, reminiscent of the induction of the immune checkpoint protein PD-1 by IFN-γ, which provides feedback suppression of the adaptive immune response. Interestingly, treatment with the toll-like receptor 7/8 (TLR7/8) agonist resiquimod prolonged survival in mice bearing PROS1-deficient tumors but not PROS1-secreting tumors, indicating that therapies targeting the PROS1-MERTK axis can synergize with other agents that activate innate immunity.

In human GBM, MERTK expression is upregulated in tumor cells and associated macrophages and microglia and expression is further increased at disease recurrence (J. Wu, et al., 2018). Additionally, treatment with the MERTK-selective inhibitor UNC2025 decreased M2-polarized macrophages in a murine GBM model and improved survival in a subgroup of mice receiving radiation therapy. Furthermore, MERTK expression increased with tumor progression on both tumor cells and macrophages in a murine colon cancer model and treatment with the pan-TAM inhibitor RXDX-106 increased numbers of tumor-infiltrating lymphocytes and M1-polarized macrophages, enhanced NK cell and CD8⁺ T cell activity, and significantly reduced tumor burden (Yokoyama, et al., 2019). Decreased tumor growth was also observed in immunocompetent models of renal cortical and breast cancers treated with RXDX-106. Importantly, this effect was decreased in immune-deficient mice, indicating that the anti-tumor activity is at least partially mediated through immunomodulatory mechanisms. Additionally, in a recent study using a syngeneic colon cancer model treatment with a MERTK blocking antibody induced a type I IFN response and enhanced anti-tumor T cell activity resulting in decreased tumor growth (Zhou, et al., 2020). Interestingly, the authors provided evidence for a mechanism involving the cGAS-STING pathway, a known driver of anti-tumor type I IFN responses. Antibody-induced inhibition of MERTK-mediated efferocytosis leads to accumulation of cytosolic DNA in tumor cells, which activates the DNA sensor cGAS (cGAMP synthase) to produce cGAMP (cyclic GMP-AMP). Subsequently, this second messenger enters surrounding immune cells and transactivates STING (stimulator of interferon genes) resulting in induction of the type I IFN response.

MERTK Signaling Regulates the PD-1 axis

An increasing body of research delineates an additional mechanism by which MERTK signaling suppresses anti-tumor immunity. MERTK regulates components of the immune checkpoint PD-1 axis, including the receptor protein programmed cell death protein 1 (PD-1) and its ligands programmed cell death 1 ligand 1 (PD-L1) and programmed cell death 1 ligand 2 (PD-L2). Under physiologic conditions, these immune-inhibitory proteins constitute another important regulatory system to prevent chronic inflammation and auto-immune reactions. PD-1 is predominantly expressed on activated T cells (Ishida, Agata, Shibahara, & Honjo, 1992), while PD-L1 and PD-L2 can be expressed on different cell types, including myeloid immune cells, but their expression levels are low in normal tissue (Dong, et al., 2002; Latchman, et al., 2001; Tseng, et al., 2001). Binding of PD-L1 or PD-L2 to the PD-1 receptor inhibits T cell activity and induces anergy and tolerance (Blank, et al., 2004; Dong, et al., 2002). Many cancer types induce expression of these checkpoint proteins in their microenvironment, leading to suppression of the T cell anti-cancer immune response. Therapeutic agents targeting the PD-1 axis are in clinical use and have revolutionized the treatment of multiple typesof cancer for a subset of patients, including melanoma, HNSCC, NSCLC and renal cell carcinoma (Alsaab, et al., 2017).

A regulatory relationship between MERTK signaling and the PD-1 axis was first revealed by experiments demonstrating increased PD-L1 and PD-L2 mRNA and protein levels in 293T cells with ectopic expression of constitutively activated MERTK (Nguyen, et al., 2014). Subsequent studies demonstrated regulation of PD-L1 downstream of MERTK in tumor cells. While constitutively activated constructs of all three TAM kinases induced PD-L1 in breast cancer cells, this effect was most pronounced for MERTK (Kasikara, et al., 2017; Nguyen, et al., 2014). Importantly, ectopic expression of native TAM constructs alone was not sufficient to increase PD-L1 surface levels, indicating that TAM activation is required (Kasikara, et al., 2017). Indeed, treatment with GAS6 in combination with apoptotic cells or PtdSer liposomes induced MERTK phosphorylation and PD-L1 protein levels in breast and cervical cancer cell lines, but GAS6 treatment alone was not sufficient to induce PD-L1 surface levels (Kasikara, et al., 2017; Nguyen, et al., 2014). Conversely, PD-L1 expression was reduced in response to Mertk shRNA knock-down, the pan-TAM kinase inhibitor BMS-777607, Ptd-Ser-blocking agents or a PI3K-inhibitor. These findings suggest a mechanism in which PtdSer and GAS6 binding induces MERTK kinase activity to signal via the PI3K-AKT pathway, finally resulting in increased PD-L1 expression.

While initial studies demonstrated a role for MERTK in regulation of PD-L1 and PD-L2 directly in cancer cells, more recent data revealed a similar role for MERTK in regulation of PD-L1 and PD-L2 expression on myeloid monocytes and macrophages in the tumor microenvironment. In a murine B-ALL model, Mertk^−/− mice had significantly reduced expression of PD-L1 and PD-L2 on myeloid cells and decreased PD-1 on T cells in the leukemia microenvironment, accompanied by significantly decreased disease burden and prolonged survival (Lee-Sherick, et al.,2018). Similar effects were achieved in wild-type mice using the MERTK-selective inhibitor MRX-2843, with only minimal effects on healthy immune cell development, but not in immune-compromised mice, implicating an immune mechanism. Furthermore, genetic or pharmacologic MERTK inhibition led to increased T cell activation in an ex vivo co-culture system. Thus, MERTK regulates the PD-1 signaling axis to enhance T cell activation and promote anti-leukemia immunity.

Combination therapies targeting multiple nodes of a signaling pathway can provide synergistic therapeutic activity (Atefi, et al., 2011; Cooper, et al., 2014; Hu-Lieskovan, et al., 2015; Zanardi, Bregni, de Braud, & Di Cosimo, 2015). Thus, the biochemical interaction between PD-L1 and PD-L2 and MERTK suggests that therapies targeting MERTK and PD-1 signaling could be particularly effective. Interestingly, initial data demonstrated decreased tumor growth when a PtdSer-targeting antibody was administered in combination with an anti-PD-1 antibody in an immune-competent murine breast cancer model (Gray, et al., 2016). While PtdSer-targeting agents also show combinatory effects with radiation and chemotherapy (X. Huang, Bennett, & Thorpe, 2005; X. Huang, et al., 2009), it is probably more desirable to target MERTK selectively. Blocking PtdSer would broadly affect all of its sensors, including TIM-4, BAI1, Stabilin-2, integrin α_Vβ₅, RAGE and others, thereby increasing the risk of side effects (Akalu, Rothlin, & Ghosh, 2017; Hochreiter-Hufford & Ravichandran, 2013). In a more recent study using a syngeneic model of triple-negative breast cancer, combined treatment with the pan-TAM inhibitor BMS-777607 and anti-PD1 antibody led to significantly greater reductions in tumor volume and metastasis than either monotherapy (Kasikara, et al., 2019). Additionally, the combination therapy induced a more pro-immunogenic cytokine and immune cell profile compared to either single agent. Of importance for potential translation to the clinic, the authors reported no evidence of weight loss or apparent toxicity when combining the pan-TAM inhibitor and PD-1 antibody. Counterintuitively, the combination induced an increase in Pd-l1mRNA in tumors in this context, possibly in response to increased levels of the pro-inflammatory cytokine IFN-γ. Combined treatment with the pan-TAM inhibitors RXDX-106 or UNC4241 and a PD-1 targeted antibody also significantly decreased tumor burden and prolonged survival relative to single agents in murine colon cancer and melanoma models, respectively (Holtzhausen, et al., 2019; Yokoyama, et al., 2019). Similarly, in a NSCLC model, triple therapy with radiation, anti-PD-1 and anti-MERTK antibodies results in decreased MDSCs at abscopal tumor sites (Caetano, et al., 2019). Moreover, M1-polarization of macrophages was increased at both the primary and abscopal tumor sites and similar effects were observed in response to treatment with the triple therapy in a mouse model of pancreatic cancer. In another study, sitravatinib, a spectrum-selective inhibitor targeting several kinases including TAM kinases, potentiated the anti-tumor effects of anti-PD1 in murine models of lung, colon and pancreatic cancer, potentially by inhibiting macrophage M2 polarization and promoting a pro-inflammatory microenvironment that favors PD-1 therapy (Du, Huang, Sorrelle, & Brekken, 2018). In two of the herein reviewed studies, the combination therapy provided a durable immunologic response that prevented tumor engraftment upon re-challenge in animals that were tumor-free at the end of therapy (Du, et al., 2018; Gray, et al., 2016). Notably, combined treatment with sitravatinib and the anti-PD1 antibody nivolumab is currently being tested in a phase III trial in patients with advanced NSCLC (NCT03906071). Impressive combinatory effects were also achieved when combining a MERTK-blocking antibody with checkpoint inhibitors targeting the PD-1 axis in in vivo models of colon cancer and breast cancer (Zhou, et al., 2020).

MERTK in T Cell Activation

While murine T cells do not express MERTK, recent publications demonstrated induction of Mertk expression in primary human CD8⁺ and CD4⁺ T cells upon activation of the TCR (Cabezon, et al., 2015; Peeters, et al., 2019). In T cells, MERTK appears to act in an immuno-stimulatory fashion, contraryto its inhibitory function in myeloid monocytes and macrophages. Treatment with a soluble MERTK extracellular domain (Mer-Fc) that sequesters MERTK-ligands inhibited activation of naïve CD4⁺ T cells and antigen-specific response of memory CD4⁺ T cells (Cabezon, et al., 2015). In human CD8⁺ T cells, PROS1-activated MERTK acts as a co-stimulator to enhance proliferation and production of effector and memory cytokines in response to TCR activation (Peeters, et al., 2019). Treatment with the MERTK-inhibitor UNC2025 or Mertk shRNA knock-down prevented PROS1-induced CD8⁺ T cell proliferation and activation, implicating MERTK as a critical downstream effector of PROS1 in this context.

The differential roles for MERTK in macrophages and human T cells suggest a complex and tightly regulated system in which MERTK’s immuno-regulatory function is highly context dependent. Consistent with this idea, at lower concentrations PROS1 inhibited CD8⁺ T cell activation (Peeters, et al., 2019). More research into the roles of MERTK in the context of a complete tumor microenvironment is needed to elucidate the intricate roles for and regulation of MERTK in this context and advance the development of MERTK inhibitors as anti-cancer immunotherapies.

MERTK in Other Immune Cells

In addition to macrophages and human T cells, MERTK regulates immunomodulatory functions in MDSCs, regulatory T cells (Tregs), NK cells, and NKT cells. In most of these cell types, MERTK plays an immune-suppressive role.

MERTK signaling has been implicated in mediating the immunosuppressive function of MDSCs, a heterogeneous population of immature myeloid cells that promote tumor growth and metastasis (Veglia, Perego, & Gabrilovich, 2018). Expression of MERTK and the other TAM kinases was dramatically increased in MDSCs from melanoma-bearing mice compared to non-tumor bearing mice (Holtzhausen, et al., 2019). The TAM RTK ligands GAS6 and PROS1 were also substantially induced in MDSCs from tumor-bearing mice providing an autocrine or paracrine source of activating ligand for MDSCs or tumor cells and other immune cells, respectively. Increased TAM signaling keeps MDSCs in an immature, highly immunosuppressive state with high levels of immunosuppressive T cell enzymes such as arginase, indoleamine 2,3-dioxygenase (IDO), and nitric oxide synthase (iNOS). Single knockout of any TAM kinase or treatment with the pan-TAM inhibitor UNC4241 reduced the T cell immunosuppressive action and migratory capacities of MDSCs, increased CD8⁺ T cell infiltration into the tumor microenvironment, and decreased tumor growth. Similarly, in a NSCLC model, triple therapy with radiation, anti-PD-1 and anti-MERTK antibodies resulted in decreased MDSCs at abscopal tumor sites (Caetano, et al., 2019). In human melanoma patients, expression of all three TAM members was increased in circulating MDSCs, with MERTK levels being considerably higher than the others, implicating MERTK-positive MDSCs as a potential biomarker for MERTK and/or PD-1 targeted therapies (Holtzhausen, et al., 2019). Additionally, MERTK is required for expansion of MDSCs following allogenic transfer of donor apoptotic cells to promote transplant tolerance (L. Zhang, et al., 2019).

MERTK signaling in the tumor microenvironment also increases the number of Tregs, an immunosuppressive T cell subpopulation, in B-ALL and breast cancer models (Lee-Sherick, et al., 2018; Werfel & Cook, 2018). Whether this effect is cell-autonomous or mediated through other MERTK-expressing cells is unknown.

TAM kinase signaling is required for NK cell development (Caraux, et al., 2006; I. K. Park, et al., 2009; Walzer & Vivier, 2006) and additionally suppresses NK cell activation (Chirino, et al., 2019; Paolino, et al., 2014) at least partially by phosphorylating the E3 ubiquitin ligase CBL-B (Chirino, et al., 2019). Interestingly, CBL-B also functions as an E3 ligase for TAM receptors in NK cells, suggesting a feedback mechanism (Paolino, et al., 2014). Importantly, treatment with the pan-TAM inhibitor LDC1267 leads to increased NK cell anti-tumor activity and decreased metastasis in a murine melanoma model. However, it is not currently understood what roles MERTK plays in NK cells relative to the other family members.

In contrast, MERTK signaling enhances NKT cell activity. NKT cells from Mertk^−/− mice secrete decreased pro-inflammatory cytokines in response to TCR stimulation with α-galactosylceramide (Behrens, et al., 2003), reminiscent of the co-stimulatory role for MERTK in T cells.

Further Considerations

The potential utility of MERTK-targeted immunotherapy is not limited to tumors that express MERTK (Lee-Sherick, et al., 2013). However, the success of MERTK-targeted immunotherapy in any individual patient will likely be dependent on other factors, such as the immunogenicity of their cancer. In a murine colorectal cancer model, survival following radiation therapy was significantly increased in Mertk^−/− mice compared to wild-type mice (Crittenden, et al., 2016). However, in the case of a poorly immunogenic pancreatic cancer model, inhibition of TGF-β was required in addition to Mertk^−/− to prolong survival following radiation therapy. Other mechanisms may also contribute to resistance to MERTK-targeted immunotherapies. Apoptotic HER2-positive breast cancer cells resulting from lapatinib treatment are cleared by MERTK-dependent efferocytosis, resulting in an immunosuppressive tumor microenvironment (Werfel & Cook, 2018). Surprisingly, treatment with the pan-TAM inhibitor BMS-777607 was not sufficient to prevent immunosuppression in this model and expression and activity of IDO1 (a known suppressor of anti-cancer immunity (Lob, Konigsrainer, Rammensee, Opelz, & Terness, 2009; Prendergast, 2008; Smith, et al., 2012)), immunosuppressive cytokine levels, and tumor-associated MDSCs and Tregs remained elevated despite inhibition of efferocytosis. However, combined treatment with BMS-777607 and the IDO1 inhibitor epacadostat reduced immunosuppressive cytokines and leukocytes and dramatically decreased tumor volume and lung metastases. The development of precision medicine approaches could help to identify optimal strategies for application of MERTK-targeted immunotherapies in individual patients.

Additionally, inflammation can be a double-edged sword in the context of cancer. While acute inflammation is a threat to cancer cells and is therefore often suppressed through mechanisms such as aberrant PD-1 signaling, a chronic inflammatory environment can, on the other hand, induce neoplasia and apply an evolutionary pressure that selects tumor cells that evolve to escape the acute anti-tumor immune response (Shalapour & Karin, 2015). MERTK’s physiologic function as a built-in negative regulator of innate immunity to prevent sustained inflammation suggests that in some contexts loss of MERTK signaling in the tumor microenvironment might promote tumor progression. Indeed, in murine models, loss of GAS6-dependent TAM signaling can promote colitis-associated colon cancer (Rothlin, Leighton, & Ghosh, 2014). While increasing evidence indicates favorable immunotherapeutic potential, these factors should be considered during development and application of MERTK-targeted cancer therapies.

MERTK Inhibition: Preclinical and Clinical Results

Agents in Preclinical Development

Inhibitors that target MERTK, AXL and/or TYRO3 have been developed, including both biologic agents and small molecules. Biologic agents include Mer590, a monoclonal antibody directed against the extracellular domain of human MERTK that inhibited MERTK phosphorylation and downstream signaling, reduced colony formation, and increased sensitivity to treatment with carboplatin chemotherapy in NSCLC cultures (Cummings, et al., 2014). ELB031, a monoclonal antibody that targets TYRO3 and MERTK, is being developed by ElsaLys Biotech (ElsaLys Biotech, 2018). Determination of key amino acid residues impacting the AXL/GAS6 interaction have been identified, leading to generation of AXL-Fc constructs that function as ligand traps and inhibit Gas6-induced cancer cell migration (Duan, et al., 2019). Additionally, with the recent focus on MERTK as an immune-oncologic target, numerous companies have reported novel MERTK-targeted antibodies in early stages of development (Alvarado, et al., 2019; Takeda, et al., 2019; White, et al., 2019; Zhou, et al., 2020). In in vivo models of colon and breast cancers, tumor volume was significantly decreased in response to treatment with anti-MERTK antibody, both as a single agent and in combination with immune checkpoint inhibitors targeting the PD-1 axis (Zhou, et al., 2020). Interestingly, in some cases these antibodies inhibit MERTK-dependent immune functions, but do not inhibit MERTK kinase. This uncoupling of MERTK kinase and functional activities highlights the possibility that MERTK kinase inhibition may not be sufficient to block all MERTK-dependent oncogenic functions and, at least in some cases, down-regulation of total MERTK using antibody-mediated approaches or novel developing technologies, such as biochemical degraders, may be a more effective therapeutic strategy. Finally, while in the case of malignancy it is advantageous to inhibit MERTK, there may be conditions in other disease states in which MERTK agonism is beneficial. For instance, bispecific antibodies that recognize both MERTK on human monocyte-derived macrophages and the aggregated amyloid beta found in Alzheimer’s patients have been used to mediate targeted phagocytosis of aggregates in preclinical models (Kedage, et al., 2020). Despite the conserved sequences in the kinase domain and overall structural similarity of the TAM family RTKs, there are numerous small molecule inhibitors with selectivity for one or more of the family members and demonstrated biochemical and/or functional activity in cell-based cancer models in preclinical development (Table 1). Several series of compounds have been developed using structure-based design starting from the known structure of MERTK with compound 52, which was identified in a kinase inhibitor library screen (X. Huang, et al., 2009; Suarez, et al., 2013). The first was UNC569 (J. Liu, et al., 2012), a pyrazolopyrimidine with potent activity against MERTK (Christoph, et al., 2013; Koda, Itoh, & Tohda, 2018; J. Liu, et al., 2012), and then UNC1062, a derivative with improved potency and selectivity (J. Liu, et al., 2013), followed by UNC2025 and MRX-2843, which share very similar structures and activities and are the best-characterized compounds in this series (W. Zhang, et al., 2014). Pyrrolopyrimidine and pyridinepyrimidine-based MERTK-selective inhibitors, UNC1666 (Lee-Sherick, et al., 2015) and UNC2250 (Shi, et al., 2018; W. Zhang, et al., 2013), were also discovered, followed by several other structural derivatives (W. Zhang, et al., 2013; Zhao, et al., 2018). These compounds inhibited MERTK phosphorylation and had myriad functional effects in cell-based assays, including altered cell cycle progression in ALL, AML, MCL, and GBM cell lines (UNC2025), induction of apoptosis in ALL (UNC569, UNC2025), AML (UNC569, UNC1666, UNC2025, UNC4042), melanoma (UNC1062, UNC2025), NSCLC (UNC2025), and GBM (UNC2025) cells, inhibition of clonogenic growth in NSCLC, melanoma and GBM (UNC2025), induction of senescence and reduced neurosphere diameter in GBM (UNC2025), decreased invasive capacity in MCL (UNC2250), and decreased anchorage-independent colony formation in ALL (UNC569, UNC2025), AML (UNC1666, UNC2025, UNC4042), rhabdoid tumor (UNC569, UNC2250), NSCLC (UNC2250, UNC2025, UNC4042) and melanoma (UNC1062) cell line cultures (see references in Table 1). Several compounds also reduced anchorage-independent colony formation (UNC1666, UNC2025) and induced apoptosis (UNC1666) in AML patient sample cultures, while mononuclear cells from normal human cord blood samples were not affected until 20-fold higher concentrations were used (DeRyckere, et al., 2017; Minson, et al., 2016). Additionally, treatment with UNC2025 reduced cell density in 30% of primary leukemia patient sample cultures (n=261), including 40% of AMLs (DeRyckere, et al., 2017). Numerous compounds inhibited MERTK phosphorylation in mice (UNC2025, UNC2250, UNC4042, UNC4203) (Da, et al., 2019; DeRyckere, et al., 2017; Lee-Sherick, et al., 2018; Shi, et al., 2018; Zhao, et al., 2018) and reduced tumor burden and/or prolonged survival in cell line and/or patient-derived murine ALL (UNC2025) (DeRyckere, et al., 2017; Lee-Sherick, et al., 2018), AML (UNC2025) (DeRyckere, et al., 2017; Minson, et al., 2016), MCL (UNC2250) (Shi, et al., 2018), NSCLC (UNC2025) (Cummings, et al., 2015), and melanoma (UNC2025) (Sinik, et al., 2019) xenograft models. Of note, most of the compounds in these series are potent and orally-available type II dual MERTK and FLT3 inhibitors that are selective for MERTK over the other TAM kinases. The exceptions are UNC4241, which targets MERTK, FLT3 and TYRO3, and UNC4042 and UNC4203, which are dual MERTK and TYRO3 inhibitors and are selective for MERTK over FLT3 (Holtzhausen, et al., 2019).

Table 1.

MERTK inhibitors in preclinical development

Name	Primary Target(s)*	Other Targets	Therapeutic Modality	Malignancies Targeted
Mer590	MERTK		Monoclonal antibody	NSCLC (Cummings, et al., 2014)
ELB031	MERTK, TYRO3		Monoclonal antibody	Not reported
UNC569	MERTK	AXL, TYRO3, FLT3, MAPKAPK2, RET	Small-molecule kinase inhibitor	ALL, AML, AT/RT (Christoph, et al., 2013; Koda, et al., 2018; J. Liu, et al., 2012)
UNC1062	MERTK	AXL, TYRO3, FLT3	Small-molecule kinase inhibitor	Melanoma, NSCLC, rhabdoid tumors, AML (Koda, et al., 2018; J. Liu, et al., 2013; Schlegel, et al., 2013)
UNC1666	MERTK, FLT3	AXL, TYRO3	Small-molecule kinase inhibitor	AML (Lee-Sherick, et al., 2015)
UNC2025	MERTK, FLT3	AXL, TYRO3, TRKA, TRKC	Small-molecule kinase inhibitor	ALL, AML, NSCLC, GBM, melanoma (Cummings, et al., 2015; DeRyckere, et al., 2017; McDaniel, et al., 2018; Sinik, et al., 2019; Sufit, et al., 2016; W. Zhang, et al., 2014)
UNC2250	MERTK	AXL, TYRO3	Small-molecule kinase inhibitor	Rhabdoid tumors, NSCLC, MCL (Shi, et al., 2018; W. Zhang, et al., 2013)
UNC2541	MERTK	AXL, TYRO3, FLT3	Small-molecule kinase inhibitor	Not reported
UNC3133	MERTK, FLT3, AXL	TYRO3, FGFR1, ITK, KDR, PDGRα, TRKA, AURKA, p70S6K	Small-molecule kinase inhibitor	Not reported
LDC1267	MERTK, AXL, TYRO3, MET, Aurora B	LCK, SRC	Small-molecule kinase inhibitor	Metastatic melanoma, metastatic breast cancer (Paolino, et al., 2014)
CT413	MERTK, AXL, MET, RON		Small-molecule kinase inhibitor	Not reported
Compound-52	MERTK		Small-molecule kinase inhibitor	Not reported
6g	MERTK, AXL, MET, TYRO3		Small-molecule kinase inhibitor	Not reported
GSK2606414	MERTK, AXL, PERK, C-KIT, Aurora B, BRK, MLK2, DDR2, MLCK2, IKKe, TRKC, MLK3, RET, LCK, NEK4, KHS, MLK1, TRKA, TRKB, YES, WNK2		Small-molecule kinase inhibitor	Pancreatic cancer (Axten, et al., 2012)
JNJ-28312141	MERTK, FLT3, AXL, MCSFR, KIT, TRKA, LCK	TYRO3	Small-molecule kinase inhibitor	Adenocarcinoma, AML (Manthey, et al., 2009)
BMS-777607	MERTK, FLT3, AXL, TYRO3, RON, MET, AURB, LCK		Small-molecule kinase inhibitor	Prostate cancer, breast cancer, ovarian cancer, glioblastoma, pancreatic cancer (Dai & Siemann, 2010; Kasikara, et al., 2019; Onken, et al., 2016; Sharma, et al., 2014; C. C. Wu, Weng, Hsu, & Chang, 2019; Zeng, et al., 2014)
RXDX-106	MERTK, TYRO3, AXL, c-MET		Small-molecule kinase inhibitor	Gastric cancer (Kim, et al., 2017)
SA4488	MERTK		Small-molecule kinase inhibitor	Lung cancer (Pyo, et al., 2019)
SGI-7079	MERTK, FLT3, TRKB, RET, TRKA, YES, AXL, MET, JAK2, KDR, JNK3, ABL	TAO1, AURORA, TAK1, SYK, RSK1, CHK2, EPHA1, LCK, GSK3β, EPHB1, FYN, FGFR1	Small-molecule kinase inhibitor	Breast Cancer (X. Wang, et al., 2013)
NPS-1034	AXL, MET	MERTK, FLT3, KIT, DDR1, TYRO3	Small-molecule kinase inhibitor	Lung Cancer (Rho, et al., 2014)

^*Primary targets included are no less selective than 10-fold MERTK activity based on biochemical/enzymatic IC50

Several additional structurally-distinct MERTK inhibitors have also been developed and have demonstrated activity in preclinical cancer models. CT413, an oral AXL/MERTK inhibitor, reduced tumor growth in subcutaneous NSCLC, ovarian and chronic myeloid leukemia (CML) xenograft models (Xi, et al., 2015). Similarly, GSK2606414, which primarily targets KIT and PERK, a critical transducer of the unfolded protein response, and also targets the TAM kinases, decreased tumor volume in human pancreatic (Axten, et al., 2012) and CML (X. H. Zhang, et al., 2017) xenograft models and had preferential activity against chemotherapy-resistant breast cancer cells relative to parental cell line controls (Alasiri, et al., 2019). RXDX-106, a type II pan-TAM inhibitor, reduced cell density in a gastric cancer cell line culture (Kim, et al., 2017). Compound 6g was also developed in an effort to make type II TAM kinase inhibitors (X. Huang, et al., 2009; Suarez, et al., 2013) and other novel structures include the macrocyclic MERTK-selective inhibitors UNC2541 and UNC3133 (McIver, et al., 2017; X. Wang, et al., 2016). Additional compounds that are being developed as AXL inhibitors, such as SGI-7079, also target MERTK and TYRO3 and have activity in preclinical cancer models (Byers, et al., 2013; X. Wang, et al., 2013). In many cases, the degree to which inhibition of individual TAM-family members contributes to therapeutic activities mediated by agents that target more than one of the TAM kinases is not clear.

MERTK kinase inhibitors also exhibited activity in pre-clinical immune-based models. As indicated above, UNC4241 altered MDSC suppressive function and slowed tumor growth in an immune-competent murine melanoma model (Holtzhausen et al CIR 2019) and more recent data demonstrated alterations in immune checkpoint pathway components (PD-L1, PD-L2, and PD-1), increased T cell activation, and prolonged survival in response to treatment with MRX-2843 in a mouse B-ALL model (Lee-Sherick, et al., 2018). In a murine GBM model, treatment with UNC2025 decreased the fraction of tumor-associated macrophages expressing CD206, an anti-inflammatory cell surface marker, consistent with induction of a more inflammatory phenotype in the tumor microenvironment (J. Wu, et al., 2018). JNJ-28312141 is a CSF1R inhibitor with anti-inflammatory properties that also targets AXL and, to a lesser degree, MERTK. Treatment with JNJ-28312141 decreased tumor growth in a murine lung cancer xenograft model, even though the tumor cells did not express CSF1R and were not affected by treatment with JNJ-28312141 in vitro, suggesting an immune-mediated effect (Manthey, et al., 2009). Treatment with the pan-TAM kinase inhibitor RXDX-106 increased tumor-infiltrating lymphocytes, CD8⁺ T cell infiltration, M1 macrophage polarization, and activation of both T cells and NK cells, leading to decreased tumor volume (Yokoyama, et al., 2019) in a murine colon cancer model. Treatment with JNJ-28312141 also reduced tumor volume in mice with renal cell or breast cancer, but in all cases, tumor volume was only reduced when the immune system was intact. In murine melanoma and breast cancer models, treatment of NK cells with LDC1267, a pan-TAM kinase inhibitor, followed by adoptive transfer into mice reduced metastasis (Paolino, et al., 2014). While the degree to which the individual TAM kinases and other targets contribute to the immune effects of these agents is also not clear, the known roles for MERTK in the affected immune cell types and processes implicate MERTK inhibition as a mechanism of anti-tumor immunity mediated by these agents.

Enhanced sensitivity to cytotoxic therapies and targeted agents has also been reported in cancer cells treated with MERTK kinase inhibitors. UNC2025 sensitized ALL cells to treatment with methotrexate in a murine xenograft model (Cummings, et al., 2015; DeRyckere, et al., 2017; Sufit, et al., 2016; W. Zhang, et al., 2014). Similarly, UNC2250 sensitized MCL cells to vincristine and doxorubicin in vitro and reduced tumor volume in combination with doxorubicin relative to either agent alone in a murine MCL xenograft model (Shi, et al., 2018; W. Zhang, et al., 2013). In a murine GBM model, a subset of mice treated with UNC2025 in combination with radiation exhibited dramatic tumor regression and prolonged survival (J. Wu, et al., 2018), whereas none of the mice treated with single agents survived. The improved chemosensitivity mediated by several MERTK-targeted small-molecule inhibitors reinforces the potential of MERTK inhibition as a therapeutic goal, particularly in combination with cytotoxic therapies.

Agents in Clinical Development

Numerous agents with MERTK inhibitory activity are currently in clinical development (Table 2). MRX-2843 (Minson, et al., 2016; W. Zhang, et al., 2014), ONO-7475 (Okura, et al., 2020), and INCB081776 (Favata, et al., 2018) target MERTK and FLT3 or MERTK and AXL with high specificity, characterized both by selectivity for the indicated targets over unintended off-targets and by off-target activity against only a small fraction of the kinome. In preclinical studies, treatment with the orally-available type II dual MERTK and FLT3 inhibitor MRX-2843 reduced MERTK phosphorylation/activation in leukemia cells in vitro and in vivo, induced apoptosis and decreased colony formation in AML cell line and/or patient sample cultures, and reduced tumor burden and/or prolonged survival in cell line and/or patient-derived murine ALL, AML, and NSCLC xenograft models (Lee-Sherick, et al., 2018; Minson, et al., 2016; D. Yan, et al., 2019). As noted above, MRX-2843 also inhibited PD-1 signaling in the leukemia microenvironment, leading to enhanced T cell activation, reduced leukemia burden, and prolonged survival in a murine ALL model (Lee-Sherick, et al., 2018), and MRX-2843 synergized with the third generation EGFR TKI osimertinib to provide durable inhibition of tumor growth in a NSCLC xenograft model (D. Yan, et al., 2018). Similarly, ONO-7475 reduced proliferation of FLT3-driven AML cells in vitro, induced leukemia cell apoptosis, and prolonged survival in a FLT3-driven murine AML model (Ruvolo, et al., 2017). Additionally, like MRX-2843, ONO-7475 synergized with osimertinib to provide enhanced therapeutic benefit in a NSCLC xenograft model (Okura, et al., 2020). INCB081776 has primarily been characterized in immune models. Treatment with INCB081776 decreased the percentage of intratumoral M2 macrophages and monocytic MDSCs and increased M1 macrophages, increased T cell infiltration and activation, and inhibited tumor growth in a mouse NSCLC model, but only when an intact immune system was present (Favata, et al., 2018).

Table 2.

MERTK inhibitors in clinical development

Name	Target(s)	Malignancies Targeted	Development Phase
MRX-2843	MERTK, FLT3	Relapsed/Refractory Advanced and/or Metastatic Solid Tumors (Minson, et al., 2016; D. Yan, et al., 2018)	Phase 1
ONO-7475	MERTK, AXL	Acute Leukemias (Ruvolo, et al., 2017)	Phase 1
INCB081776	MERTK, AXL	Advanced Solid tumors (Favata, et al., 2018)	Phase 1
SU-14813	FLT3, VEGFR, MERTK, AXL, TYRO3, PDGFR, KIT	Advanced Solid Malignancies (Fiedler, et al., 2011)	Phase 1
ASLAN002	MET, MERTK, AXL, TYRO3, RON, AURKB, FLT3, AXL	Advanced or Metastatic Solid Tumors (Roohullah, et al., 2018)	Phase 1
MGCD265	MERTK, AXL, TYRO3, MET, VEGFR2	Non-small cell lung cancer (NSCLC) (Padda, Neal, & Wakelee, 2012; J. Wang, et al., 2018), malignant pleural mesothelioma (MPM) (J. Wang, et al., 2018), triplenegative breast cancer (Linklater, et al., 2016)	Phase 1
Bemcentinib (BGB324)	RET, MERTK, AXL, TYRO3, VEGFR, FLT3, ABL, TIE2	NSCLC (Felip, et al., 2019), Metastatic or Recurrent Pancreatic Cancer (Beg, et al., 2019), AML/MDS (Sonja Loges, et al., 2019), recurrent GBM (clinicaltrials.gov), Breast Cancer (Yule, et al., 2018), Mesothelioma (Krebs, et al., 2018)	Phase 1, Phase 2
S49076	MET, FGFR, AXL, MERTK	Advanced solid tumors (Rodon, et al., 2017), Recurrent GBM (Hoang-Xuan, et al., 2016), NSCLC (Viteri, et al., 2018)	Phase 1, Phase 2
Merestinib (LY2801653)	MET, MERTK, AXL, TYRO3, RON, FLT3, MST1R, ROS1, TEK, DDR1/2, MKNK1/2	Advanced solid tumors including adenocarcinoma of colon/rectum, head and neck SCC, uveal melanoma (Cheng, et al., 2017), cholangiocarcinoma (Barat, et al., 2016), gastric adenocarcinoma. NSCLC (Kawada, et al., 2014; W. Wu, et al., 2013; S. B. Yan, et al., 2018) (Ghanaatgar-Kasbi, Khorrami, Avan, Aledavoud, & Ferns, 2018; He, et al., 2019; S. B. Yan, et al., 2013)	Phase 1, Phase 2
MK2461	MET, MERTK, RON, VEGFR	Advanced solid tumors (Inoue, et al., 2017; Padda, et al., 2012)	Phase 1, Phase 2
Foretinib (GSK1363089, XL880)	MET, VEGFR2, AXL, MERTK, TYRO3, RON, PDGFRβ, KIT, FLT3, TIE2	Hepatocellular Carcinoma (Yau, et al., 2017), Recurrent/metastatic Breast Cancer (Simiczyjew, et al., 2018), NSCLC (Leighl, et al., 2017), Advanced/metastatic solid tumor, Squamous Cell Cancer of the Head/Neck (G. Z. Chen, et al., 2017), Papillary Renal Cell Carcinoma (Logan, 2013), Metastatic Gastric Cancer (Shah, et al., 2013)	Phase 1, Phase 2
AT9283	AURKA, AURKB, JAK, MERTK	Relapsed/refractory Multiple Myeloma (Hay, et al., 2016), Advanced/metastatic solid tumor (Arkenau, et al., 2012; Dent, et al., 2013; Moreno, et al., 2015), Advanced/metastatic non-Hodgkin’s Lymphoma (Qi, et al., 2012), Relapsed/refractory AML/ALL/CML or high risk MDS (Foran, et al., 2014)	Phase 1, Phase 2
Dubermatinib TP-0903	TYRO3, AXL, MERTK, AURKA, AURKB, JAK2, ALK, ABL1	Lung cancer (Taverna, et al., 2020), Neuroblastoma (Aveic, et al., 2018), Chronic Lymphocytic Leukemia (Patel, Keating, Wierda, & Gandhi, 2016; Sinha, et al., 2018)	Phase 1, Phase2
Sitravatinib (MGCD516)	VEGFR, PDGFR, KIT, MET, RET, AXL, MERTK, TYRO3	Squamous Cell Carcinoma, Liposarcoma, Urothelial Carcinoma, Clear Cell Renal Cell Carcinoma, Non-squamous NSCLC, Advanced/Metastatic Kidney Cancer, Advanced/Metastatic Hepatocellular Carcinoma/Gastric Cancer, Advanced Solid Tumors (Dolan, et al., 2019; Du, et al., 2018)	Phase 1, Phase 2, Phase 3
Gilteritinib ASP2215	TYRO3, AXL, MERTK,	AML (Ha, et al., 2020; Sidaway, 2020; Yun, et al., 2019), Colorectal Cancer (L. Li, Lin, Li, & Li, 2019)	Phase 1, Phase 2, Phase 3, Approved
Neratinib	HER2, AXL, MERTK, TYRO3	Advanced Solid Tumors (Deeks, 2017), Metastatic Breast Cancer (Abraham, et al., 2019; Echavarria, Lopez-Tarruella, Marquez-Rodas, Jerez, & Martin, 2017; Martin, et al., 2017), Recurrent/refractory Pediatric Solid Tumor/Leukemia/Lymphoma, Metastatic Colorectal Cancer (Kavuri, et al., 2015), NSCLC (Bose & Ozer, 2009), Glioblastoma	Phase 1, Phase 2, Phase 3, Approved
Lestaurtinib (CEP-701)	FLT3, ACL, MERTK, TYRO3, JAK2, TRKA TRKB, TRKC	Primary Myelofibrosis/Essential Thrombocythemia related Myelofibrosis/Polycythemia Vera related Myelofibrosis (Mascarenhas, et al., 2019), Relapsed/Refractory AML (Knapper, et al., 2017; Levis, et al., 2011), Recurrent/Refractory High-Risk Neuroblastoma (Minturn, et al., 2011; Norris, Minturn, Brodeur, Maris, & Adamson, 2011), ALL (Brown, Levis, McIntyre, Griesemer, & Small, 2006), Advanced Multiple Myeloma, Prostate Cancer (Festuccia, et al., 2007; Kohler, et al., 2012),	Phase 1, Phase 2, Phase 3, Approved
Vandetanib (ZD6474)	VEGFR2, VEGFR3, EGFR, AXL, MERTK, TYRO3, RET	Metastatic Medullary Thyroid Cancer (Wells, et al., 2012), Hepatocellular Carcinoma (Hsu, et al., 2012), NSCLC (Morabito, et al., 2010; Yoh, et al., 2017), Gliosarcoma, Glioblastoma (E. Q. Lee, et al., 2015), Breast Cancer (Hatem, et al., 2016), GIST (Glod, et al., 2019), Advanced Pheochromocytoma and Paraganglioma, Advanced Solid Tumors, DIPG (Broniscer, et al., 2013), Papillary Renal Cell Carcinoma (Drevs, et al., 2004), Pancreatic Carcinoma (Kessler, et al., 2016), Ovarian/Fallopian Tube/Peritoneal Cancer, Metastatic Urinary Tract Cancer, Prostate Cancer (A. A. Azad, et al., 2014), Metastatic Colorectal Cancer, High-risk Head/neck Cancer, Biliary Tract Cancer (Kessler, et al., 2016), Esophageal/GE Junction Cancer, Mesothelioma	Phase 1, Phase 2, Phase 3, Approved
Sunitinib	PDGF, VEGF, AXL, MERTK, TYRO3, RET, FLT3	Pancreatic Neuroendocrine Tumors (Blumenthal, et al., 2012), Advanced/Metastatic Renal Cell Carcinoma (Knox, et al., 2017)	Phase 3, Approved
Crizotinib (PF2341066)	MET, AXL, MERTK, TYRO3, ALK, RON	Metastatic NSCLC (Landi, et al., 2019; Masuda, et al., 2019), ALCL (Sekimizu, et al., 2018; R. Wang, et al., 2019), Myofibroblastic tumors (Mosse, et al., 2017), Breast Cancer (Ayoub, Al-Shami, Alqudah, & Mhaidat, 2017), Gastric Cancer (Okamoto, et al., 2012), High-Risk Uveal Melanoma (Surriga, et al., 2013), Metastatic/Recurrent Endometrial Cancer, Metastatic Prostate Cancer (Kato, et al., 2018), Lymphoma (Gambacorti Passerini, et al., 2014), Glioblastoma (Junca, et al., 2017; Nehoff, Parayath, McConnell, Taurin, & Greish, 2015), ALK positive tumors, Metastatic Urothelial Cancer, DIPG/HGG (Broniscer, et al., 2013), Recurrent Neuroblastoma (Schulte, et al., 2013; Sekimizu, et al., 2018), Advanced Solid Tumors, Locally advanced/metastatic Kidney Cancer (Schoffski, et al., 2017)	Phase 1, Phase 2, Phase 3, Phase 4, Approved
Cabozantinib (XL184, BMS-907351)	AXL, MERTK, TYRO3, VEGFR2, MET, MEK, KIT, RET	Renal cell carcinoma (Bergerot, Lamb, Wang, & Pal, 2019; Lazaro, et al., 2020), Recurrent/metastatic endometrial cancer (Dhani, et al., 2020), Metastatic Prostate Cancer (Corn, et al., 2020), Adrenocortical carcinoma (Kroiss, et al., 2020), Thryoid Cancer (Ancker, Kruger, Wehland, Infanger, & Grimm, 2019), Advanced Hepatocellular Carcinoma (Abou-Alfa, et al., 2018; Personeni, Rimassa, Pressiani, Smiroldo, & Santoro, 2019), Urothelial Carcinoma (Bergerot, et al., 2019), Esophageal Squamous Cell Carcinoma (P. W. Yang, et al., 2019)	Phase 1, Phase 2, Phase 3, Phase 4, Approved
Bosutinib (SKI-606, PF-5208763)	SRC, ABL, AXL, MERTK, TYRO3	Ph+ CML (Assi, et al., 2017; Shen, Wilson, Gleason, & Khoury, 2014), Metastatic Solid Tumors – non-squamous non-small cell lung cancer/pleural malignant mesothelioma/bladder cancer/urethral cancer/ovarian cancer/peritoneal cancer/thymic cancer/uterine cervical cancer, Glioblastoma (Taylor, et al., 2015), CML (Cortes, et al., 2018; Doan, Wang, & Prescott, 2015), Ph+ ALL with CD22 expression, Pancreatic Cancer (Daud, et al., 2012), Breast Cancer (Campone, et al., 2012; Vultur, et al., 2008), Colorectal Cancer (Daud, et al., 2012), Cholangiocarcinoma, Advanced/Recurrent Solid Tumors (Isakoff, et al., 2014)	Phase 1, Phase 2, Phase 3, Phase 4, Approved

Additional less selective agents with TAM family kinase inhibitory effect are also in development or have been approved, with various primary targets (Table 2).

Potential Side Effects of MERTK Inhibition

MERTK’s physiological functions in tissue repair, innate immune control, and platelet aggregation may portend potential side effects that require consideration during the development of MERTK-targeting cancer therapies. Importantly, mice with single Mertk^−/− or triple knock-out of all TAM family members are viable without apparent developmental defects (Lemke & Lu, 2003; Lu, et al., 1999; Lu & Lemke, 2001). Susceptibility to LPS-induced inflammation as well as circulating autoantibody levels are much less pronounced in single Mertk^−/− mice compared to triple knock-out mice lacking all three family members (Camenisch, et al., 1999; Lu & Lemke, 2001), providing rationale that targeting MERTK specifically rather than all TAM kinases, could minimize potential side effects. In addition, defects in tissue homeostasis, such as blindness caused by MERTK inhibition in the retina, have been corrected using gene therapy to provide MERTK, suggesting that some side effects of MERTK inhibition would be reversible (Conlon, et al., 2013; DiCarlo, et al., 2018). Furthermore, in murine leukemia models, MERTK inhibition by selective small molecule inhibitors is well tolerated at therapeutic doses (DeRyckere, et al., 2017; Lee-Sherick, et al., 2018; Minson, et al., 2016). Treatment strategies which minimize length of MERTK inhibition may further mitigate toxicity compared with tyrosine kinase inhibitors used on a long-term basis.

Summary and Perspectives

MERTK is an emerging target for cancer therapy. MERTK is aberrantly expressed in a wide variety of malignancies and has numerous roles in oncogenesis. MERTK promotes growth factor independence, survival signaling, and tumor cell motility, leading to oncogenic transformation, enhanced tumor growth, therapeutic resistance, and metastasis. In addition, MERTK is expressed in the innate immune system, where it can be subverted to suppress anti-tumor immunity. Thus, therapeutic strategies targeting MERTK are expected to have both direct and immune-mediated mechanisms. Pre-clinical studies demonstrated the utility of targeting MERTK to induce tumor cell apoptosis, decrease tumor burden, and prolong survival in cell culture and murine cancer models and MERTK inhibitors have been successfully combined with both traditional cytotoxic and molecularly targeted agents to enhance therapeutic efficacy. Induction of immune-mediated anti-tumor activity in response to MERTK inhibition has also been demonstrated in both solid tumor and leukemia models. These preclinical data support development of MERTK inhibitors and clinical trials are currently ongoing, including the first studies testing agents that selectively target MERTK or MERTK and AXL. While MERTK is not a classical oncogene, neoplastic cells depend on non-oncogenes for survival (i.e. non-oncogene addiction), particularly in the context of cell stress due to the altered tumor microenvironment, rapid tumor cell proliferation driven by tumor suppressor loss or oncogene expression, or treatment with other therapies. Future clinical application may therefore employ MERTK inhibitors in conjunction with classical cytotoxic chemotherapy or newer targeted therapies and these could be further augmented with innate and/or T cell immune checkpoint inhibitors, depending on the tumor type and/or specific genetic and phenotypic characteristics. Ultimately, the broad spectrum expression of MERTK across tumor types and ability to inhibit essential signaling in tumor cells and enhance the immune response in the tumor microenvironment by targeting a single protein make MERTK a particularly attractive therapeutic target.

Figure 3: MERTK Immune Functions in Cancer.

MERTK signaling in tumor and immune cells contributes to an immunosuppressive tumor microenvironment that favors tumor growth.

LEFT: MERTK signaling and immune functions in cells of the tumor microenvironment

MERTK signaling in macrophages induces an immunosuppressive M2 phenotype (Stanford, et al., 2014). Tumor cells can release the MERTK ligands GAS6 and PROS1 (Ubil, et al., 2018) and tumor cell-derived PROS1 inhibits the expression of M1 polarization-associated genes in macrophages (Ubil, et al., 2018).

Additionally, tumor cells can release the cytokines IL-10 and M-CSF, which induce autocrine GAS6 production in macrophages (S. Loges, et al., 2010). Furthermore, MERTK-dependent efferocytosis of apoptotic tumor cells induces macrophages to release the anti-inflammatory cytokines IL-10, IL-4, TGF-β (Crittenden, et al., 2016; Stanford, et al., 2014) and to produce increased Mertk, Gas6 and Pros1 mRNA, forming a positive feedback loop (Crittenden, et al., 2016).

Activation of MERTK through binding of its ligands and PtdSer (which may be exposed on apoptotic tumor cells) induces expression of the immune checkpoint protein PD-L1 on cancer cells (Kasikara, et al., 2017). Furthermore, MERTK signaling can induce PD-L1 and PD-L2 expression on myeloid cells (shown here: macrophages) (Lee-Sherick, et al., 2018). Binding of PD-L1 or PD-L2 to PD-1 on T-cells inhibits T-cell anti-tumor activity. TAM (TYRO3, AXL, MERTK) signaling suppresses NK cell function but the individual role of MERTK is not understood (Chirino, et al., 2019; Paolino, et al., 2014). Furthermore, MERTK signaling induces immunosuppressive and migratory capacities in MDSCs (Caetano, et al., 2019; Holtzhausen, et al., 2019). In contrast, MERTK signaling enhances NKT cell activity (Behrens, et al., 2003) and may have an immune-stimulatory role in activated human T cells (Cabezon, et al., 2015; Peeters, et al., 2019).

RIGHT: Immune effects of MERTK signaling in the tumor microenvironment.

Effects of MERTK signaling include macrophage M2 polarization (Crittenden, et al., 2016; Stanford, et al., 2014), induction of an immunosuppressive cytokine profile (Cook, et al., 2013; Stanford, et al., 2014) and altered immune cell infiltration (Caetano, et al., 2019; Cook, et al., 2013; Holtzhausen, et al., 2019; Lee-Sherick, et al., 2018; Werfel & Cook, 2018).

GAS6 (growth arrest-specific 6),

PROS1 (vitamin K-dependent protein S),

IL-4/6/10/12p40/13 (interleukin 4/6/10/12p40/13),

M-CSF (macrophage colony-stimulating factor),

M1/2 (macrophage polarization 1/2)

PtdSer (phosphatidylserine),

PI3K (phosphatidylinositol 3-kinases),

PD-L1/2 (programmed cell death 1 ligand 1/2),

PD-1 (programmed cell death protein 1),

TGF-β (transforming growth factor-β),

NKT cell (natural killer T cell),

NK cell (natural killer cell),

MDSCs (myeloid-derived suppressor cells),

MΦ (macrophage),

DC (dendritic cells),

CD8⁺ (CD8⁺ T cell),

Tregs (regulatory T cells)

Abbreviations:

ALL

acute lymphoblastic leukemia

AML

acute myeloid leukemia

CBL-B

casitas B-lineage lymphoma-b

CML

chronic myeloid leukemia

EGFR

epidermal growth factor receptor

ERK

extracellular signal-regulated kinase

FLT3

FMS-like tyrosine kinase 3

GAS6

growth arrest-specific 6

GBM

glioblastoma multiforme

GEF

guanine nucleotide exchange factor

GLA

γ-carboxylglutamic acid-rich

HGF

hepatocyte growth factor

HNSCC

head and neck squamous cell carcinoma

IDO1

indoleamine 2,3-dioxygenase 1

IC50

inhibitory concentration 50

IFNγ

interferon gamma

IL

interleukin

JAK

Janus kinase

LPS

lipopolysaccharides

MCL

mantle cell lymphoma

MEK

MAPK/ERK kinase

M-CSF

macrophage colony-stimulating factor

MDSC

myeloid-derived suppressor cell

MERTK

myeloid-epithelial-reproductive tyrosine kinase

NF-κB

nuclear factor κ-light-chain-enhancer of activated B cells

NK

natural killer

NKT cell

natural killer T cell

NSCLC

non-small cell lung cancer

PD-1

programmed cell death protein 1

PD-L1

programmed cell death ligand 1

PD-L2

programmed cell death ligand 2

PI3K

phosphatidylinositol 3-kinases

PROS1

vitamin K-dependent protein S

PtdSer

phosphatidylserine

RTK

receptor tyrosine kinase

shRNA

short-hairpin RNA

SOCS1

suppressor of cytokine signaling 1

SPMs

specialized proresolving mediators

STAT

signal-transducer and activator of transcription

TAM

TYRO3, AXL, MERTK

TCR

T cell receptor

TGF-β

transforming growth factor-β

TLR

toll-like receptor

TNBC

triple-negative breast cancer

TNFα

tumor necrosis factor-α

Tregs

regulatory T cells

TULP-1

tubby-like protein 1

WT

wild-type