Multifactorial resistance to aminopeptidase inhibitor prodrug CHR2863 in myeloid leukemia cells: down-regulation of carboxylesterase 1, drug sequestration in lipid droplets and pro-survival activation ERK/Akt/mTOR

Abstract

Aminopeptidase inhibitors are receiving attention as combination chemotherapeutic agents for the treatment of refractory acute myeloid leukemia. However, the factors determining therapeutic efficacy remain elusive. Here we identified the molecular basis of acquired resistance to CHR2863, an orally available hydrophobic aminopeptidase inhibitor prodrug with an esterase-sensitive motif, in myeloid leukemia cells. CHR2863 enters cells by diffusion and is retained therein upon esterase activity-mediated conversion to its hydrophilic active metabolite drug CHR6768, thereby exerting amino acid depletion. Carboxylesterases (CES) serve as candidate prodrug activating enzymes given CES1 expression in acute myeloid leukemia specimens. We established two novel myeloid leukemia sublines U937/CHR2863(200) and U937/CHR2863(5uM), with low (14-fold) and high level (270-fold) CHR2863 resistance. The latter drug resistant cells displayed: (i) complete loss of CES1-mediated drug activation associated with down-regulation of CES1 mRNA and protein, (ii) marked retention/sequestration of the prodrug, (iii) a substantial increase in intracellular lipid droplets, and (iv) a dominant activation of the pro-survival Akt/mTOR pathway. Remarkably, the latter feature coincided with a gain of sensitivity to the mTOR inhibitor rapamycin. These finding delineate the molecular basis of CHR2863 resistance and offer a novel modality to overcome this drug resistance in myeloid leukemia cells.

Keywords: aminopeptidase; carboxylesterase; lipid droplets; mTOR; rapamycin.

Figure 1. A. Chemical structure of the aminopeptidase inhibitor prodrug CHR2863 with an esterase motif and its acid metabolite CHR6768. B. Time line for acquisition of resistance to CHR2863 in U937 cells. Two isolates (indicated by arrows) were selected for further characterization; U937 cells grown in the presence of 200 nM CHR2863 (U937/CHR2863(200) and U937 cells grown in the presence of 5 μM CHR2863 (U937/CHR2863(5μM). C. Dose response curve of for growth inhibition by CHR2863 for U937/WT, U937/CHR2863(200), and U937/CHR2863(5μM) cells. Results depicted are the mean ± SD of 7-10 separate experiments

Figure 2. A.Conversion of CHR2863 to CHR6768 and B. retention of CHR2863 in U937/WT, U937/CHR2863(200), and U937/CHR2863(5μM) cells after 6 hr exposure to 6 μM CHR2863. Results are expressed as ng/10⁶ cells and represent the mean ± SE of 7-9 separate experiments. (*): p < 0.001

Figure 3. Carboxylesterase 1 and 2 expression in acute myeloid leukemia (AML) cells and CHR2863-resistant U937 cells

A. CES1 expression (Western blot) in FAB sub-classified (M0, M1, M2, M4, M5) acute myeloid leukemia cells at diagnosis. Superscript symbols refer to matched peripheral blood (PB) and bone marrow (BM) samples. CES1 expression in peripheral blood of healthy controls is depicted for reference. Hep-G2 hepatoma cells serve as control for CES1 and CES2 protein expression. CES2 expression was not observed in the indicated AML samples (not shown). B. mRNA expression of CES1 and CES2 in multiple isolates of CHR2863-resistant U937 cells, including U937/WT, U937/CHR2863(200), and U937/CHR2863(5μM) cells. U937/CHR2863(w/o 1 μM) without CHR2863 for 2 weeks and U937/CHR2863 (w/o 1μM/+1μM) rechallenged with 1 μM CHR2863 for 2 weeks. Mean (± SD) of 3-4 experiments performed in triplicate). (*): p < 0.05. C. Western blots of CES1 (by two different antibodies), CES2 and CES3 protein expression in U937/WT, U937/CHR2863(200), and U937/CHR2863(5μM) cells. KG1 cells served as negative control for CES1 expression [21].

Figure 4. CES1 expression and identification of lipid droplets in U937/WT and CHR2863-resistant cells and CES1-lipid droplet co-localization

A. Nile Red staining of lipid droplets and CES1 immunofluorescence detection by 3D digital imaging fluorescence microscopy. The inset depicts colocalization of CES1 with lipid droplets in U937/WT cells. B. Live cell 3D digital imaging microscopy of CES1 expression in U937/WT, U937/CHR2863(200), U937/CHR2863(5μM) cells and KG1 cells (as CES1-negative control). Left row represents control conditions, right row represent images after pulse exposure to CHR2863 (6 hr, 6 μM CHR2863).

Figure 5. A. Lipid droplets staining in U937/WT cells with Nile Red and representative examples of image stream analysis sorting cells with 1, 4 or 7 lipid droplets per cell. B. Quantification of the distribution of lipid droplets in U937/WT, U937/CHR2863(200), and U937/CHR2863(5μM) cells by Nile Red staining and image stream analysis allowing assessment of numbers of lipid droplets per individual cell (as in (A)). Mean of two separate experiments performed in duplicate. All experiments included 0.06% DMSO solvent controls

Figure 6. A. Expression levels of total and phosphorylated Erk, Akt, mTOR and S6K in U937/WT, U937/CHR2863(200) and U937/CHR2863(5μM) cells. B. Inhibition of cell growth by rapamcyin of U937/WT, U937/CHR2863(200) and U937/CHR2863(5μM) cells. Cell growth inhibition was assessed after 72 hrs drug exposure. Results represent the mean ± SD of 5-7 separate experiments. C. Inhibition of cell growth by rapamcyin of U937/WT, U937/CHR2863(200) and U937/CHR2863(5μM) cells upon co-incubation with non-toxic concentrations of the dual PI3K/mTOR inhibitor BEZ235 (10 nM) or Akt inhibitor MK2206 (100 nM). Cell growth inhibition was assessed after 72 hrs drug exposure and depicted as the mean of two separate experiments

Figure 7. Composite model depicting mechanism of action of CHR2863 sensitivity in U937 cells (left) and mechanism of acquired resistance to CHR2863 in U937/CHR2863(5μM) cells (right)

Parts of the (poly)peptides produced by proteasome-mediated protein degradation are processed for MHC class I presentation involving ERAP (ER-associated aminopeptidase). It is unknown whether CHR2863 or CHR6768 exert inhibitory effects on ERAP. Most polypeptides will be subject to full degradation to amino acids, involving aminopeptidase (AP) action, for renewed protein synthesis. Due to its hydrophobic nature, the cyclopentyl-ester conjugated compound CHR2863 can freely diffuse into cells and has potential to inhibit several APs [14]. The AP-inhibitory potency, however, is significantly improved upon conversion of CHR2863 to its acid metabolite CHR6768 which is accumulated and retained in cells. A likely candidate for CHR2863 conversion includes carboxylesterase 1 (CES1), which has a physiologic function in regulating cholesterol homeostasis, in particular in lipid droplet (LD) cell organelles. Conceivably, CES1 associated with LDs may provide a microenvironment that promotes CHR2863 conversion. CHR6768-induced inhibition of multiple APs will provoke an amino acid depletion which is sensed by mTOR leading to suppression of protein synthesis and inhibition of cell growth. In CHR2863-resistant cells, at least two adaptations took place. One involves a marked down-regulation of CES1 which may convey two effects; (i) rather than conversion, CHR2863 is sequestered in LD. Concomitantly, increased cholesteryl esters in LD may act as cell proliferation regulator, and (ii) loss of CHR6768-induced AP inhibition relieves part of the amino acid deprivation pressure, which could contribute to reactivation of mTOR activity. Second, mTOR reactivation may also be initiated separate from CES1 down-regulation (as in U937/CHR2863(200) cells) to promote protein synthesis and cell growth consistent with a resistant phenotype. This is achieved by Erk activation, as an early response in low resistant cells, and Akt/mTOR activation. Activation of mTOR in CHR2863-resistant cells is a targetable entity for its inhibitor rapamycin.