A Novel Approach to Induce ATRA Mediated Differentiation in NPM1 Mutant Acute Myeloid Leukemia

Introduction: Retinoic acid (RA), an active metabolite of vitamin A, activates the retinoic acid receptors (RARs), ligand-controlled transcription factors that regulate expression of target genes affecting homeostasis and cellular differentiation. The success of RA in the form of all-trans retinoic acid (ATRA) to treat acute promyelocytic leukemia (APL) has stimulated considerable interest in the development of small molecules to modulate RARs in AML cells. Besides RARs, RA can also activate the pro-survival peroxisome proliferator-activated receptor β/δ (PPARβ/δ), a member of the nuclear hormone receptor superfamily of transcription factors that regulate several metabolic pathways in a tissue-selective manner. In the aqueous intracellular milieu, RA is transported by two alternative carriers, cellular retinoid-binding protein CRABP2, or fatty-acid-binding protein 5 (FABP5). These carriers preferentially deliver RA to RARs or PPARβ/δ respectively. In cells, expressing high CRABP2 and low FABP5, RA is delivered to and activates RARs, whereas in the presence of the reverse ratio, RA is delivered to and activates PPARβ/δ. Here we demonstrate that these two different modes of RA delivery leads to opposite cellular outcomes: CRABP2-mediated RA delivery to RAR prompts differentiation and growth arrest, while in cells containing high levels of FABP5, RA delivery to PPARβ/δ promotes survival and exponential proliferation. Inhibiting FABP5 carrier-function could be a useful method, therefore, to extend RA therapeutic activity to non-APL acute myeloid leukemia (AML).

Methods: Genetic as well as pharmacologic in vitro and in vivo models of AML were used to validate FABP5 as a target for therapy to, increase RA stimulation of RARs instead of PPARβ/δ. Knockdown and overexpression of key regulators of these pathways were rationally used to understand pro-survival effects and proliferative mechanism of malignant myeloid cells. The efficacy of small molecule pharmacophore targeting FABP5 was evaluated in vitro in cell free and cell culture model as well as in vivo small animal model systems as a proof of therapeutic concept.

Results: In public databases of AML gene expression, the majority of non-APL AMLs demonstrated upregulation of FABP5 and downregulation of CRABP2, potentially explaining limited benefit of RA to treat non-APL AMLs. On the other hand, ATRA sensitive APL samples demonstrated low FABP5, as expected from the established efficacy of RA to treat APLs. The greatest elevations of pro-survival FABP5 was in AMLs with mutated-NPM1. Consistent with the elevated FABP5 in these cells, we confirmed that RA activates pro-survival PPARβ/δ rather than pro-differentiation RAR in NPM1-mutated AML cells. We thus evaluated if inhibiting FABP5 could produce ATRA sensitization in NPM1 mutant AML cells. Genetic ablation of FABP5 using shRNA sensitized AML cells to ATRA treatment, while this effect was not seen in AML cells transfected with non-targeting scrambled shRNA. Cells stably expressing shRNA targeting FABP5 has significant growth perturbation in vivo in murine AML model. To target FABP5 using a small molecule approach, we used in silico drug-design to develop a novel small molecule FABP5 inhibitor (iFABP5): the inhibitor bound to recombinant FABP5 (K_d= 2x10 ^-7) with a 56-fold higher affinity than ATRA (K_d=11x10 ^-6 Mole), demonstrated in a cell free fluorescence reporter binding assay using 8-Anilino-1-naphthalenesulfonic acid (ANS). We then treated RA-resistant AML cells with iFABP5 alone and in combination with RA: neither iFABP5 nor RA alone had anti-proliferative activity as expected, but combination iFABP5 + RA downregulated PPARβ/δ target genes and upregulated RAR target genes, terminating AML cells growth.

Conclusion: The lack of effective ATRA response in AML might be, in part, due to high expression of FABP5 and delivery of RA to pro-survival PPARβ/δ. A small molecule inhibitor of FABP5 could redirect RA to pro-terminal-differentiation RARs, and could be an approach to extending the remarkable activity of RA in APL to other AMLs. Reprogramming of the RA delivery to RAR pathway with the addition of a small molecular inhibitor of FABP5 could potentially restore therapeutic effects of ATRA in FABP5 over expressing cases of AML subtypes.