EPHA3 as a Novel Therapeutic Target in the Hematological Malignancies

Niamh Keane; Ciara Freeman; Ronan Swords; Francis J Giles


Expert Rev Hematol. 2012;5(3):325-340. 

In This Article

EPH RTKs in Malignancy

Effects of Eph Receptor Expression or Silencing in Malignancy

Ephs are associated with many different malignancies and, as in other settings, adopt varied and often conflicting roles. Ephs may behave in a similar manner to oncogenes or tumor suppressors[52] in different settings.

Eph RTKs have perhaps been best established as markers of aggressive and advanced disease, Eph receptor expression serving as a marker for malignancy quiescent in normal tissues of the cell type in question. For example, EPHA2 is overexpressed in melanoma cell lines, while not expressed in melanocytes,[53] and its expression in melanoma, as with EPHA3, correlates with aggressive and metastatic disease.[53] Both EPHA2 and EPHB4 are highly expressed in breast cancers compared with normal breast tissue.[54]

Initially, a classic RTK role of Ephs as oncogenes was therefore implied. However, loss of expression of Ephs has also frequently been implicated in tumorigenesis.[55] Loss of EPHB2 and EPHB4 receptors in colorectal cancer (CRC) and loss of EPHB6 in breast cancer have been described.[56,57] EPHA2 can also function as a tumor suppressor with inhibition of Akt on activation,[58] in contrast to its putative role in breast cancer above.

Therefore, while Ephs represent promising new targets in cancer treatment, their role in malignancies is complex and varied. Ligand-independent signaling of Eph[59] RTKs may account for apparent contradictory roles of Ephs in malignancy.[60] Loss of Eph forward signaling has been implicated in tumor progression in myriad malignancies.[55] For example, aberrant EPHB4 expression in the absence of its ligand ephrin-B2, and hence the absence of kinase activation, results in increased metastatic potential,[61] while following ligand stimulation EPHB4 exhibits tumor suppressor activity.[56] EPHA2 exhibits tumor-promoting activity independent of ligand activity but functions as a tumor suppressor on stimulation of forward signaling following binding of ephrin-A1.[58]

Eph receptor mutations are likely to play an important role in this setting, with mutations that compromise ligand binding, receptor signaling or kinase activation hindering forward signaling.[12] EPHA3 is widely mutated in cancer with up to 40 mutations identified in solid tumors.[12] Many mutations affect the ephrin-binding, sushi-like and EGF-like domains (all involved in ligand binding) or kinase domains with possible disruption to forward signaling.[12] Conceivably, EPHA3 mutations may result in loss of tumor suppressor activity in these cancers and facilitate ligand/kinase-independent signaling, favoring tumorigenesis.[12] In the absence of forward signaling, ligand-independent Eph signaling appears to favor an EMT-type morphology with the hallmarks of this process noted in these cells – spindle-like morphology with loss of apical–basal polarity and cell–cell junctions[62] are exhibited in these Eph-bearing cells.[63] The opposite process of mesenchymal–epithelial transition occurs as a result of ligand-induced forward signaling with restoration of cell–cell connection and polarity.[63] Forward signaling mediates downregulation of integrin expression and modification of intercellular junctions, disrupting cell–matrix and cell–cell contacts and perhaps facilitating cell motility in advanced stage malignancy.[55] Eph mutations may result in the ability to disinhibit the suppression of cell survival pathways by forward signaling activation of RhoA and inactivation of Rac and cdc42 (Figure 3).[32,63] Eph receptors are also known to regulate and interact with Ras family proteins[32] and EphA receptor forward signaling inhibits this pathway, downregulating MAPK effectors, for example Erk1/2.[64] Additionally, as mentioned above, an activation loop between EphA2 and Akt has been described.[58] With ligand activation of EphA RTK, an antimalignancy effect is seen with inhibition of Akt; however, in the absence of kinase activation, invasive behavior is facilitated and, in this setting, Akt upregulates expression of EphA2 further.[59] Activity of Rac1 and Cdc42 GTPases in the absence of Eph receptor signaling activation of RhoA may also result in the promotion of these survival pathways by their interaction with Akt in the absence of forward signaling.[63]

Silencing of Ephs in malignancy or the expression of Eph RTK without ephrin binding and hence kinase activation may mean loss of forward signaling with resultant promotion of cell–matrix adhesion and invasiveness, net support of cell survival pathways and alteration of cell phenotype with EMT (Figure 3).[55]

Eph RTKs have additionally been implicated in many other processes involved in malignancy, including alteration of the tumor microenvironment such that malignancy is promoted and in angiogenesis,[52] both of which likely play an active role with regard to EPHA3 and hematologic malignancy.

Eph RTK in Stem Cells & Malignancy

Expression of Eph RTK has been documented in numerous stem cell niches; an unsurprising finding given the known involvement of Ephs in embryologic cell position and guidance functions.[65]

Eph receptors function in neurological stem cell niches with typical complementary roles described in different settings.[66] As an example, EphA7 is expressed in ependymal cells, with reverse signaling via Ephrin-A2 resulting in negative regulation of neuro-genesis.[54] EphB receptors expressed in the subventricular zone reduce proliferation of progenitor cells and disrupt migration[54,65,66] and have roles in the maintenance of the phenotype of these cells.[68] By contrast, EphB receptors in the hippocampus positively stimulate proliferation and neurogenesis on binding to ephrin ligand.[66,69]

The role of EphB2 in the intestinal stem cell (ISC) niche has been studied, with expression in early cells, downregulation in more differentiated cells and a role in directing migration of progenitor cells from crypts and control of proliferation.[70] ISCs express Lgr5[71] and are located in the intestinal crypts.

Recently, ISCs in CRC were characterized as highly expressing EPHB2,[72] with low-level or no EPHB2 in better differentiated tumor cell subsets as well as normal epithelium, supporting previous studies of EphB2 in intestinal epithelium.[73] These EPHB2-expressing cells coexpressed Lgr5. This important subset has been shown to have the ability to self-renew and differentiation potential, and following inoculation into NOD/SCID mice, cells with high EPHB2 expression led to tumor formation, with similar proportions of ISC-like and differentiated cells as seen in the original CRC, and no tumor formation in mice inoculated with EPHB2 cells.[73] EPHB2-expressing ISCs are also predictive of relapse of disease in CRC following treatment.[73]

In the development of mammary epithelium, EPHB4 and ephrin-B2 also exhibit complementary expression patterns[74] with upregulation in development and physiological proliferation,[74] and absence in end-differentiated tissue.[74] Increased EPHB4 expression at developmental stages resulted in abnormal tissue architecture,[75] while aberrant EPHB4 expression in the absence of ephrin-B2 capable of initiating reverse signaling resulted in increased metastatic potential;[61] although other studies have indicated that with ligand stimulation EPHB4 exhibits tumor suppressor activity.[56] The combination of aberrant EPHB4 expression in this setting with loss of reverse signaling alters stem cell homeostasis and facilitates breast cancer invasion and metastasis.[76]

Eph receptors are therefore well represented in adult stem cell niches and in some instances their unscheduled presence in the tumor microenvironment is in association with the tumor stem cell population specifically.

As with normal stem cells, cancer stem cells by definition have self-renewal capacity and differentiation potential, and a hierarchical system exists within cancer niches as with, for example, the hematopoietic system, with nonhomogenous populations of cells present[77,78] of which cancer stem cells constitute only a small proportion.[78] Initial studies using acute myeloid leukemia (AML) blasts found low potential for colony formation.[79] In further studies in NOD/SCID mice, unfractionated leukemia cells had the ability to engraft, self-renew and differentiate, with similar cells to the initial leukemia identified.[80] A small proportion of cells termed SCID leukemia-initiating cells or leukemic stem cells (LSCs) possess self-renewal capacity and differentiation potential, and it is thought that this population of cells, although slow growing, account for propagation of the entire tumor.[78] Although first identified and described in hematologic malignancies, there is growing evidence for cancer stem cells in solid malignancies.[81–83] The evidence of Eph receptor expression on putative cancer stem cells suggests that the unscheduled appearance of Eph receptors in malignancy may correspond with their expression in this stem cell population in particular, highlighting a potential role as a cancer stem cell target with potential for use in and prevention of relapsed disease.

LSCs have now been well characterized and reside in the CD34+CD38 population, and in AML, are also positive for CD123 (IL-3Ra).[84]

In fact, in leukemic cells studied, EPHA3 is seen on all CD34CD38+CD123+ cells,[85] making this potentially an exciting target for LSCs.

EPHA3 as a Novel Target in Hematological Malignancies

Evidence in favor of the development of EPHA3 as a therapeutic target in hematological malignancies has been growing in recent years. Aberrant expression of EPHA3 is seen in almost all types of hematolologic malignancies, although it is not expressed ubiquitously ( Table 1 ).[15,40] The role of EPHA3 in these malignancies is less clear, however. In contrast to EPHA3-expressing solid organ malignancy in which the kinase domain or other part of the EPHA3 receptor are frequently mutated,[12] EPHA3 is structurally normal in hematologic malignancies.[36,40]

Lymphoproliferative Disorders EPHA3 in lymphoproliferative disorders was the initial focus of research, with in vitro studies showing expression in ALL and T-cell leukemia (Jurkatt, HSB-2 and Molt4 cells).[40] EPHA3 is highly expressed in T-cell lymphomas, although absent in normal T cells or in Hodgkins lymphoma[31] (see Table 1 for details of expression in lymphoproliferative and myeloproliferative disorders). In these EPHA3-expressing cells of lymphoid lineage, upregulation of the RTK is effected by CD28.[31] CD28 stimulates transcription of EPHA3, via activation of the IGF1 receptor with resultant alteration of cell motility,[31] is associated with progression of disease in myeloma[31,86] and is similar to EPHA3 in that CD28 is expressed on the majority of advanced myeloma samples (relapsed or refractory).[54] Normal plasma or B cells do not express CD28, and it is not expressed at initial diagnosis of myeloma, becoming upregulated later.[54] CD28's role in myeloma may be mediated by alteration of the tumor microenvironment, which supports myeloma cell survival,[87] and is similar to the effect of aberrant EPHA3 expression with relative paucity of ephrins in the tumor microenvironment.[22] Expression of EPHA3 in lymphoma may be regulated by CD28 or a consequence of interplay with growth factors such as IGF-1, signaling advanced disease.

Myeloproliferative Disorders Initially, EPHA3 was highlighted due to its advent from an ALL cell line, as a receptor associated with lymphoid lineage malignancies. Its expression in myeloid malignancy has been recognized subsequently. The integral role of tyrosine kinases in the pathogenesis of the myeloproliferative neoplasms is well established, although the role of the Eph RTK family appears to be quite distinct from that of established kinase targets. Elevated EPHA3 expression occurs in myeloproliferative neoplasms in contrast to bone marrow and peripheral blood cells from healthy controls in which EPHA3 is absent.[41,88] Patients with chronic myeloid leukemia (CML) in the chronic phase also demonstrate low expression of EPHA3 with expression markedly elevated in patients in the accelerated or blastic phases.[88] Phosphorylation of EPHA3 and hence activation of the tyrosine kinase is prominent in transformed CML.[89]

An emerging clinical problem in the treatment of CML is the development of resistance to tyrosine kinase inhibitors (TKIs). While newer TKIs overcome resistance in some cases, non-TKI approaches are required. HSP32/HO1 increases CML cell survival independent of Bcr/Abl.[37] Inhibiting Hsp32/HO1 induces apoptosis in CML cells and does so synergistically with imatinib or nilotinib in imatinib-resistant CML.[37] Expression of Hsp32 in CML is associated with the upregulation of kinases, including EPHA3.[90] Thus, EPHA3 may have a role in rendering CML cells resistant to TKIs via upregulation by survival molecules. Additionally, inhibition of EPHA3 in CML is effective in targeting TKI-resistant and nonresistant CML forms ( Table 1 ).

LSC background and phenotype have already been discussed. It is of considerable interest that EPHA3 is expressed on CD34+CD38CD123+ cells, as these represent a population of leukemia stem cells/initiating cells.[85] In many cases in which EPHA3 was expressed on these leukemia-initiating cells, it was expressed on 100% of stem cells, but less commonly expressed on other leukemic cell fractions.[85] Association of Ephs in general with advanced malignancy, and of EPHA3 in particular with advanced hematologic malignancy has been established previously. A putative role for EPHA3 as a novel target on LSCs is promising and may be useful for treating and preventing relapsed disease. Of note, several studies have indicated that normal hematopoietic stem cells do not express EPHA3 and preclinical data in which EPHA3 was targeted with a monoclonal antibody indicated that, while killing of LSCs was effected, normal bone marrow stem cells were not targeted.[85]

Putative Role(s) of EphA3 in Hematologic Malignancies Upregulation of EPHA3 in hematologic malignancies may be stimulated by growth factors or signaling molecules active in malignancy, for example, IGF in lymphoproliferative and hsp in myeloproliferative malignancies as outlined above. In this regard, appearance of EPHA3 may signal established, advanced disease. Also implicated in the reappearance of EPHA3 in hemato-logic malignancies are hypomethylation of the EPHA3 gene promoter region and copy number variations of the gene.[40] Loss of copy numbers of EPHA3 is the trend in these malignancies,[91] although EPHA3 expression is not silenced; as evidenced by the association of EPHA3 with leukemia and lymphoma, particularly in advanced stages. Therefore, epigenetic regulation, such as changes in gene methylation, may play a dominant role in upregulating the expression of EPHA3 in neoplastic cells[40] and may be the result of already abnormal cell biology, supported by the fact that EPHA3 expression is not regulated by methylation of the promoter in normal tissue.[40]

Studies examining EPHA3-expressing hematologic malignancies have indicated an adherent, mesenchymal-like phenotype of cells.[31,36] In Jurkat cells, binding of ephrin-A5 ligand leads to reduced adherence to matrix and was associated with CrkII, indicating functional EPHA3 kinase,[31] for example. In LK63 pre-B ALL cell lines, PTP1b is highly active and modulates the function of EPHA3 receptor, reducing kinase phosphorylation and downstream responses. It may contribute to the alteration of cell phenotype, which is supportive of invasiveness, and PTP1b itself is associated with advanced disease.[37]

Ephrin-A5 stimulation mediates cellular repulsion by default and results in loss of cell–matrix adhesion via Fak dephosphorylation and integrin inhibition. Contraction of the cytoskeleton with alteration of cell phenotype by signaling through Rho family GTPases, with upregulation of RhoA and downregulation of Rac and cdc42, also results in inhibition of support for the cell survival pathway. In unligated EPHA3 present on malignant hematologic cells, these prosurvival pathways may be active, as is the case in other EphA expressing malignancies.[56,58] Akt is active in this setting and interacts with Rac, and the Ras–Erk pathway is disinhibited in this setting and may promote malignancy (Figures 3 & 4).[22]

The mesenchymal morphology exhibited by these EPHA3-expressing cells in the absence of ligand stimulation may indicate a role of EMT in hematologic malignancies in which EPHA3 is overexpressed. EPHA3 has a known involvement in EMT during formation of the atrioventricular valves in embryogenesis, and it is plausible that EMT may be a feature of its unscheduled reappearance in hematologic malignancies.[50] Furthermore, a functional EPHA3 kinase is a requirement for EMT to take place in embryogenesis, and hematologic malignancies express wild-type EPHA3 receptor with no known mutations affecting the kinase domain in these malignancies. As noted above, there is growing evidence for the expression of EPHA3 on LSCs. Recently, it has been suggested that the process of EMT may result in acquisition of stem cell properties by malignant cells.[92] Eph-expressing cells with properties of stem cells may command plasticity of phenotype, which allows switching between the proinvasive mesenchymal phenotype and a growth-promoting epithelial phenotype.[93] Changes in EPHA3 expression and phosphorylation with later stages of malignancy may correspond with phenotype, switching; for example, the highly phosphorylated EPHA3 detected in blast crisis of CML. The potential involvement of EPHA3-expressing hematologic malignancy with EMT and the possibility of expression of the RTK on all LSCs are important for the emerging role of EPHA3 in therapeutics and merit further investigation.

Tumor microenvironment is likely to play an important role in the setting of EPHA3-expressing hematologic malignancy. Relative expression of ephrin ligand has a bearing on tumor progression, and relative paucity allows Eph RTK an advantage in this setting,[66] without the default repulsion effected by the presence of its ligand. Low expression of ephrin-A5 in bone marrow milieu allows the promalignant effects described downstream of the unligated EPHA3 RTK to proceed unchecked. Furthermore, EPHA3 may be upregulated in response to hypoxia in bone marrow.[94] This suggests that EPHA3 may play a role in survival in the cell expressing the receptor and also may interact with the microenvironment to promote malignancy too. Interestingly, a high proportion of bone marrow stem cells localize to hypoxic areas[95] and hypoxic areas may provide a suitable environment for LSCs.[96] EPHA3 has also been identified as a gene that is upregulated in promoting LSC survival in vivo.[97] EPHA3 may have a role in supporting angiogenesis in the microenvironment too, in keeping with its role in altering tumor microenvironment and its upregulation in response to hypoxia. Preclinical data indicate that EPHA3 may be expressed in the vasculature of the bone marrow, providing further rationale for specifically targeting EPHA3 in hematologic malignancies.[97]

One key aspect of major developmental therapeutics implications in which EPHA3 in hematologic malignancies remains consistent with what is seen in other Eph-expressing cancers is the direct correlation of its upregulation with aggressive and invasive disease.[53]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.