Inhibition of Mitochondrial Translation As a Therapeutic Strategy for Acute Myeloid Leukemia (AML)

Abstract 233

To identify novel therapeutic strategies that can eliminate AML and AML stem cells, we screened a library of on and off-patent drugs for candidates that could reduce the viability of engineered human AML cell lines that display the stem cell properties of differentiation, self-renewal, and marrow repopulation. This screen identified the anti-microbial agent tigecycline (TIG) as a top candidate with an LD₅₀ of 3 to 8 uM, on 5 human AML cell lines. A lethal action was also demonstrated on 13 of 20 1°AML samples with similar potency (LD₅₀ <5 uM). In contrast, normal hematopoietic cells, including the CD34+ subset, were more resistant (LD₅₀ >10 uM). We also found that 5 mM TIG reduced the clonogenic growth of 1°AML samples by 93±4% and was effective in reducing the ability of AML cells to regenerate disease in transplanted immunodeficient mice. In contrast, 5 uM TIG had no effect on the clonogenic growth or repopulating potential of normal human hematopoietic cells.

To determine the mechanism of action of TIG, we used Haplo-Insufficiency Profiling, a functional chemical genomic screen, in S. cerevisiae. The Gene Ontology component that was the most enriched for TIG was the mitochondrial ribosome. We subsequently demonstrated that TIG inhibited mitochondrial but not cytoplasmic translation in AML cell lines and in 1°AML samples. Consistent with the inhibition of mitochondrial translation, TIG decreased the enzyme activity of Complex I and IV, which contain mitochondrially-translated subunits, but not complex II (nuclear-encoded subunits only). TIG also decreased oxygen consumption and decreased mitochondrial-membrane potential in AML cell lines and 1°AML samples, but not normal hematopoietic cells. Interestingly, unlike many mitochondrial inhibitors, TIG did not increase ROS production in AML cells. Additional experiments demonstrated that inhibition of mitochondrial translation was functionally important for the anti-leukemia activity of TIG.

Next, we asked whether genetic strategies in leukemia cells would produce similar anti-leukemic effects as obtained with TIG. Knockdown of the mitochondrial-elongation factor EF-Tu mimicked the ability of TIG to inhibit mitochondrial translation, decrease mitochondrial membrane potential, decrease complex I and IV activity and induce cell death in AML cells. Also, EF-Tu knockdown did not increase ROS production.

To investigate the basis of the hypersensitivity of AML cells to mitochondrial translation inhibition, we assessed baseline mitochondrial characteristics of 1°AML cells and their normal counterparts. 1°AML cells (including CD34+CD38- AML cells) had higher intrinsic mitochondrial-biogenesis (mtDNA copy number, mitochondrial mass) than normal CD34+ hematopoietic cells. Furthermore, rates of oxygen consumption were higher in 1°AML cells as compared to normal hematopoietic cells. Baseline mitochondrial-mass in AML cells also predicted in vitro toxicity to TIG, as 1° AML cells with higher mitochondrial mass were more sensitive to TIG (r = −0.71, p <0.05).

To assess the anti-leukemia efficacy of mitochondrial translation inhibition in vivo, we investigated human AML cells in mouse xenograft models. TIG significantly delayed tumor growth of OCI-AML2 xenografts in SCID compared to untreated control mice. We then assessed the effect of TIG on AML stem cells defined by their ability to sustain leukemic cell growth in vivo. NOD/SCID mice engrafted with human AML cells and then treated with TIG showed a decrease in human AML cells by up to 77% without toxicity including alterations in liver and muscle enzymes. In contrast, NOD/SCID mice engrafted with normal cord blood did not show reduced engraftment after TIG treatment. Importantly, the human AML cells harvested from the bone marrow of the TIG-treated 1° mice generated fewer leukemic cells in secondary mice, compared to the AML cells harvested from control (untreated) primary mice, thus demonstrating an in vivo effect on the AML stem cells.

In conclusion, mitochondrial translation inhibition selectively kills AML vs. normal cells, including those defined functionally as AML progenitors and stem cells. This selectivity appears attributable to the higher rate of mitochondrial biogenesis found in AML cells. Given these results and the known pharmacology and toxicology of TIG in humans, targeting mitochondrial translation inhibition as a therapeutic strategy in AML is attractive.