Acute myeloid leukemia (AML) is a heterogeneous hematopoietic neoplasm, characterized by immature myeloid cells (myeloblasts) in bone marrow of blood. Standard AML therapies, including chemotherapy and allogeneic hematopoietic stem cell transplantation (HSCT) have been unsatisfactory with a cure rate less than 30% and even worse in elderly patients. Although targeted therapies have improved the treatment outcome in specific AML subtypes, the benefit from the treatment has remained dismal. Therefore, a more effective AML therapy is urgently needed.
Alkaline intracellular pH (pHi) plays an important role in oncogenesis. Monocarboxylate transporters (MCT), Na+/H+ exchangers (NHE), vacuolar ATPase (V-ATPase), Cl−/HCO3− anion exchangers (AE), Na+/HCO3− co-transporters (NBC) and carbonic anhydrases (CA) have been reported to be cancer-associated pHi regulators. However, clinically approved and potent pHi-regulator inhibitor has been lacking except amiloride, which is a safe and effective NHE1 inhibitor.
NHE1 maintains acid-base balance by exporting H+ in exchange for Na+. NHE1 can be activated by phosphorylation in order to alkalize the intracellular compartment of tumor cells. As kinase mutation is common in normal karyotype AML, we hypothesized that NHE1 is activated by gain-of-function kinase mutant in AML and resulted in intracellular alkalinization. Combined inhibition of NHE1 and NHE1-activating kinase might improve the treatment outcome and a new, safe and effective regime can be identified.
A kinase screen (195 human kinases) was performed in normal human cord blood (CB) progenitor cells (CD34+). Thirteen kinases were found to significantly alkalize pHi in NHE1-dependent manner. We further discovered that FLT3-ITD, KRASG12D, and BTK could enhance intracellular alkalinization and promoted cell growth in the presence of NHE1 expression in CB progenitors. Increased phosphorylation of NHE1 protein could be demonstrated by overexpression of FLT3-ITD, KRASG12D, and BTK.
Kasumi-1, MOLM-13, and MV4-11, carrying KIT mutation and FLT3-ITD showed higher NHE1 phosphorylation and were significantly more sensitive to NHE1 inhibitor (amiloride) demonstrated by lower cellular viability, stronger intracellular acidification and higher apoptotic induction. Similar results were observed in primary human AML samples carrying FLT3-ITD, KIT, or RAS mutations compared to AML samples without the 3 mutations (Figure 1). The dependency of NHE1 was further confirmed by genetic approach that NHE1-KD in MV4-11, MOLM-13, Kasumi-1 and primary AML with FLT3-ITD and RAS mutations significantly lowered pHi, induced apoptosis, and suppressed proliferation in amiloride-sensitive. Further, NHE1-KD reduced primary AML leukemic burden in vivo and prolonged animal survival suggesting that NHE1 activity is essential in supporting the function and survival of the leukemic initiating cells.
These results led us to investigate the molecular mechanisms of NHE1 phosphorylation in AML. We demonstrated that the intracellular acidification effects of FLT3-ITD and BTK are mediated by phosphorylation of NHE1 at S648, S703 and S766/S770/S771 (SSS). FLT3-ITD can also phosphorylate NHE1 at T718/S723/S726/S729 (TSSS) to exert its effect. Co-immunoprecipitation was also observed between NHE1 and kinases such as FLT3, BTK and ERK, indicating that NHE1 might be phosphorylated and activated by direct binding to them.
We then demonstrated the therapeutic potential of the kinase activated NHE1 pathway in AML. Combination of amiloride and quizartinib (FLT3-ITD/KIT inhibitor) or ibrutinib (BTK inhibitor) synergistically decreased pHi, suppressed growth, increased apoptosis in AML cell lines. Overexpression of NHE1 or MCT4 resulted in intracellular alkalization and therefore diminished the growth inhibitory effect of these kinase inhibitors indicating the alkaline pHi plays a role in translating the anti-leukemic effect of the inhibitors in AML. Additionally, amiloride was found to cooperate with weakly basic therapeutic agents, such as doxorubicin and ibrutinib by increasing the intracellular drug distribution.
To strengthen the potential of using amiloride for AML, amiloride-containing serum (ACS) from patients was obtained after 4 hours of amiloride administration for diuresis. The ACS significantly decreased pHi and increased apoptosis with kinase inhibition or chemotherapy in FLT3-ITD AML cell lines. Similar effects were observed in primary AML samples with FLT3-ITD and RAS mutations (Figure 2). The results supported that the amiloride concentration is sufficient for inhibiting NHE1 in AML. Importantly, the estimated effective amiloride concentrations in ACS were demonstrated to be comparable with that used in our in vitro study.
Taken together, our findings revealed a novel and fundamental gain-of-function kinase mutants conferring proliferation signal, by phosphorylating and activating NHE1 directly or indirectly through ERK and p90-RSK, and thus increasing the pHi, which is crucial in promoting leukemic growth and reducing the cellular uptake of therapeutic agents such as quizartinib and ibrutinib (Figure 3). However, this can be suppressed by targeting NHE1 with amiloride or its upstream activators with quizartinib/ibrutinib. Using them in combination shifts the pHi to acidic, which consequently leads to leukemia arrest and increased uptake of therapeutic agents. Our finding has opened a new angle to examine how modulating pH of tumor cell can improve current therapeutic regime. However, the manner of pH regulation might be disease-specific or individual-specific. For instance, in AML, NHE1 activation might be associated with gain-of-function kinase mutant and MCT4 dependence might be correlated with aberrant epigenetic regulation. Therefore, a comprehensive analysis of pH regulation in AML, particularly the potential of subtype-specific pH regulation will be essential to better examine the therapeutic value of pH modulation in different disease models.