AML drug resistance: c-Myc comes into play

In this issue of Blood, Nanbakhsh et al show that leukemic cell lines rendered resistant to cytarabine display a higher expression of the natural killer (NK) group 2, member D (NKG2D) ligands (NKG2DL) UL-16 binding proteins 1-3 (ULBP1-3). This expression is c-Myc dependent as it is abrogated by c-Myc inhibitors or small interfering RNA. Consequently, upregulation of NKG2DL renders the cell lines more sensitive to NK-cell–mediated lysis.¹ 

Although not described as a major receptor involved in the recognition of acute myeloid leukemic cells (in contrast to others such as natural cytotoxicity receptors [NCRs] NKp30, NKp44, and NKp46 or DNAM-1), NKG2D is one of the main activating receptors expressed by NK cells and displays some role in the recognition of AML after histone deacetylase inhibitor (HDACi) therapy.^2,3  Noteworthy, NKG2DLs include ULBP1-4 and the nonclassical HLA class I (HLA-I) molecules MICA/B.

This work, in line with previous studies, confirms that it is possible to modify leukemic cells to render them more susceptible to NK-cell recognition. Hence, cytarabine, a drug that interferes with DNA synthesis and kills cycling cells, is used as a classical treatment of leukemia. However, especially in acute myeloid leukemia (AML), it is usually used in combination with anthracyclines. It would be interesting to analyze the effect of cotreatment of leukemic cells with cytarabine and anthracyclines (such as daunorubicine) on not only the induction of c-Myc but also the expression of NKG2DL.

As mentioned by Nanbakhsh et al, c-Myc regulates the expression of ∼15% of all cellular genes. In the cellular model used by the authors, induction of c-Myc is involved directly or indirectly in the upregulation of NKG2DL. Cytarabine and c-Myc both trigger the DNA damage response (DDR) pathway, here independently of p53. Although no clear hints were obtained from gene expression from cytarabine-resistant cell lines, it would be interesting to check whether other ligands for activating NK receptors, such as NCR and DNAM-1 ligands, are modified. For instance PVR, a ligand for DNAM-1, can be expressed after DNA damage.⁴  Unfortunately, little is known about the ligands for NCR. On a global point of view, it would be interesting to verify whether cytarabine (alone or in combination with anthracyclines) would also upregulate other ligands of NK receptors (see figure). Another question that rises is whether this is the toxicity of the drugs that induces the expression of NKG2DL or whether this is due to a bystander effect generated consequently to c-Myc induction and other genes linked to DNA damage or cell arrest.⁵ 

Nanbakhsh et al also confirm their finding in primary leukemic cells. Hence, in 8 patients suffering from refractory AML (ie, resistant to chemotherapy treatment), leukemic cells are ULBP1-3 positive and this expression is reduced after ex vivo treatment with c-Myc inhibitor. These data suggest that, although AML blasts are resistant to cytarabine, they are sensitive to in vitro NK-cell–mediated killing thanks to the expression of NKG2DL. As mentioned above, DNAM-1 and NKp46 are also involved in AML killing. Therefore, it would be interesting to test whether ligands for these receptors are also induced by chemotherapy. It is noteworthy that other factors may explain the inability of NK cells to definitely clear the tumor cells. As noticed by the authors, one should remember that many factors might alter the successful recognition of leukemic cells by autologous NK cells. As mentioned above, the other NK receptors may be involved, but the expression of HLA-I molecules may also be a critical parameter. Hence, a major limitation for antitumor activity of autologous NK cells remains the inhibition by HLA-I expressed by leukemic cells. Although HLA downmodulation has been observed in many cancers including AML, unless there is specific HLA loss, HLA expression usually remains high enough to ensure NK-cell inhibition. In that context, it would be then interesting to test whether HLA-matched normal NK cells could kill cytarabine-treated AML blasts (see figure).

Last but not least, another parameter to take into account for the failure of NK cells to clear AML cells positive for NKG2DL, albeit resistant to chemotherapy, is the overall responsiveness of autologous NK cells. Several studies have highlighted defects of NK cells in AML both at diagnosis and in refractory and relapse stages. Among the defects identified, NCR, NKG2D, and DNAM-1 (Chretien et al,⁶  Sanchez-Correa et al,⁷  and C.F. and D.O., unpublished data) are often downregulated in patients. Based on this knowledge, it is easy to understand why the effects generated by cytarabine and c-Myc on leukemic cells in patients may not be sufficient to clear refractory or relapsing AML. As the authors mentioned, the current study emphasizes the potential of combinatorial therapy using, for instance, donor–NK-cell infusion, that is, cells that have not been altered by AML development in drug-resistant patients.

As c-Myc is an oncogene that promotes tumor growth and resistance to drugs, this study raises the question of whether, in refractory AML, inhibiting c-Myc with inhibitors would be a good strategy. Studies are required to decipher the impact of chemotherapies on the immunogenicity of remaining AML cells. This strategy will have to be carefully examined because ongoing clinical studies have embarked novel chemotherapy combinations together with the methyltransferase inhibitor decitabine and HDACi that display complex functional consequences on AML functions in vitro.^8-10