Cancer-Stem-Cell Phenotype-Guided Discovery of a Microbiota- Inspired Synthetic Compound Targeting NPM1 for Leukemia

Cancer-Stem-Cell Phenotype-Guided Discovery of a Microbiota- Inspired Synthetic Compound Targeting NPM1 for Leukemia Sergio Algar, # Henar Vázquez-Villa, # Pedro Aguilar-Garrido, Miguel Angel Navarro-Aguadero, María Velasco-Estévez, Anabel Sánchez-Merino, Iván Arribas-Alvarez, Alberto Paradela, Rafael L. Giner-Arroyo, Joaquín Tamargo-Azpilicueta, Irene Díaz-Moreno, Joaquín Martínez-López, Miguel Gallardo,*María L. López-Rodríguez,*and Bellinda Benhamu* Cite This: JACS Au 2024, 4, 1786−1800 Read Online ACCESS Metrics & More Article Recommendations * sı Supporting Information ABSTRACT: The human microbiota plays an important role in human health and disease, through the secretion of metabolites that regulate key biological functions. We propose that microbiota metabolites represent an unexplored chemical space of small druglike molecules in the search of new hits for drug discovery. Here, we describe the generation of a set of complex chemotypes inspired on selected microbiota metabolites, which have been synthesized using asymmetric organocatalytic reactions. Following a primary screening in CSC models, we identified the novel compound UCM-13369 (4b) whose cytotoxicity was mediated by NPM1. This protein is one of the most frequent mutations of AML, and NPM1-mutated AML is recognized by the WHO as a distinct hematopoietic malignancy. UCM-13369 inhibits NPM1 expression, downregulates the pathway associated with mutant NPM1 C+, and specifically recognizes the C-end DNA-binding domain of NPM1 C+, avoiding the nucleus-cytoplasm translocation involved in the AML tumorological process. The new NPM1 inhibitor triggers apoptosis in AML cell lines and primary cells from AML patients and reduces tumor infiltration in a mouse model of AML with NPM1 C+ mutation. The disclosed phenotype-guided discovery of UCM-13369, a novel small molecule inspired on microbiota metabolites, confirms that CSC death induced by NPM1 inhibition represents a promising therapeutic opportunity for NPM1- mutated AML, a high-mortality disease. KEYWORDS: drug discovery, small molecule, microbiota, organocatalysis, stem cells, NPM1, leukemia ■INTRODUCTION The human microbiota is a complex ecosystem of symbiotic microorganisms that play an important role in human health and disease, in part through the secretion of metabolites that can regulate host proteins. 1,2 Microbiota can synthesize, transform, and degrade a large repertoire of organic molecules that are key for maintaining homeostasis providing beneficial functions to the host. Consequently, microbiota dysbiosis, i.e., alterations in its composition that result from exposure to environmental factors, such as diet, xenobiotics, drugs, and pathogens, eventually contributes to the pathogenesis of various diseases, such as obesity, type 2 diabetes, cardiovascular disease, cancer, and neurological disorders. 3−6 In this context, several metabolites produced by the microbiota have been shown to confer protection against disease due to their antioxidant and anti-inflammatory properties. 7−9 For example, equol exhibits strong antiproliferative effects on breast neoplasm and hepatocellular carcinoma; 10,11 urolithins have been suggested to target specific tumorigenesis pathways in prostate cancer; 12,13 niacin (vitamin B3) has been shown to suppress colitis and colon cancer in mice; 14,15 indole appears to modulate the important antihyperglycemic hormone glucagon-like peptide (GLP1), with a substantial impact on pancreatic function, insulin release, and food intake regulation; 16 protocatechiuc acid prevents esophageal cancer; 17 and butyrate has been reported to modulate the activity of the nervous system in mouse models. 18 Hence, microbiota metabolism produces a wide and diverse range of bioactive small molecules that are available to the host to regulate key biological functions. The compatibility of the microbiome with healthy human systems suggests that microbiota metabolites represent drug-like molecules that are highly Received: November 3, 2023 Revised: January 19, 2024 Accepted: January 22, 2024 Published: February 9, 2024 Articlepubs.acs.org/jacsau © 2024 The Authors. Published by American Chemical Society 1786 https://doi.org/10.1021/jacsau.3c00682 JACS Au 2024, 4, 1786−1800 This article is licensed under CC-BY 4.0 Downloaded via UNIV DE SEVILLA on June 5, 2024 at 07:20:46 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. specific modulators with no risk of adverse effects. 8 This supports the hypothesis that the metabolites can be considered an unexplored chemical space of small organic molecules in the search of initial hits toward the identification of novel drug candidates for the treatment of a wide range of pathological conditions. Focused on cancer, in the present work, we have designed and synthesized a set of new small molecules inspired on human microbiota metabolites and explored their potential in cancer cellular models with the ultimate goal to identify novel antitumor drug candidates. Following a primary screening in cancer stem cells (CSCs), we have identified the synthetic compound UCM-13369 that has been subsequently characterized as a nucleophosmin 1 (NPM1) inhibitor that specifically recognizes the C-end DNA-binding domain of the protein. The capacity of the new NPM1 inhibitor to trigger stem cell apoptosis has been translated to ex vivo and in vivo efficacy in models of acute myeloid leukemia (AML). 19 Our results support phenotypic drug discovery as a valuable approach to develop active molecules in a pathologically relevant model, which together with the validation of the target protein involved in the disease, contributes to enhance innovation and deliver truly novel therapeutics for unmet medical needs. 20,21 Figure 1. Design of new chemotypes inspired on human microbiota metabolites. (A) Selected small-molecule metabolites produced by the microbiota. (B) Synthetic approach via asymmetric aminocatalysis to generate microbiota-inspired molecules 1−13. (C) Overview of the synthetic strategy based on a reaction between an α,β-unsaturated aldehyde and a nitro compound or a second aldehyde. PrSc: privileged scaffold. Untapered lines represent the relative stereochemistry; a and b are used to designate the two enantiomers of each compound. JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.3c00682 JACS Au 2024, 4, 1786−1800 1787 Figure 2. (A) Synthesis of compounds 1−3containing 2-piperidinone or piperidine scaffolds. Reagents and conditions: (a) (i) 10 mol % chiral cat., 20 mol % benzoic acid, DCM, 0 °C to rt, 5 h; (ii) PDC, rt, on, 45−58%; (b) TFA, DCM, rt, 45 min-4 h, 73−93%; (c) (i) NiCl2·6H2O, NaBH4, MeOH, 0 °C, 1 h; (ii) Boc2O, rt, on, quantitative; (d) (i) NaH, DMF, 0 °C, 1 h; (ii) (bromomethyl)cyclopropane, NaI, rt, on, 36−42%; (e) LiAlH4, THF, rt, on, 77−85%; (f) TFA, rt, 2 h, 84−87%; (g) NiCl2·6H2O, NaBH4, MeOH, 0 °C to rt, on, 53−60%; (h) (i) cyclopropranecarbaldehyde, MeOH, rt, 4 h; (ii) NaBH4, rt, 2 h, 63−64%. (B) Synthesis of compound 4containing a tetrahydrocarbazole scaffold. Reagents and conditions: (a) 20 mol % chiral cat., 20 mol % benzoic acid, toluene, MW, 110 °C, 5.5 h, 78%; (b) (i) (cyclopropylmethyl)amine, MeOH/DCM, rt, 3 h; (ii) NaBH4, rt, 2 h, 83−86%; (c) Zn, AcOH/MeOH, 2 h, quantitative; (d) HCl, MeOH, rt, 4 h, 40−50%. (C) Synthesis of compounds 5−12 containing chromane and tetrahydrobenzo[c]chromene scaffolds. Reagents and conditions: (a) 20 mol % chiral cat., 20 mol % AcOH, CHCl3, rt, on, 21−26% for 17, 22− 31% for 19, 52−61% for 20; (b) 20 mol % chiral cat., 20 mol % AcOH, CHCl3, rt, 1 h then 24 h, 43−48%; (c) (i) (cyclopropylmethyl)amine, MeOH/ DCM, rt, 3 h; (ii) NaBH4, rt, 2 h, 17%-quantitative; (d) Zn, AcOH/MeOH, rt, 2 h, 41−93%; (e) (cyclopropylmethyl)amine, DCM, rt, 3 h, quantitative; (f) NaBH4, MeOH, rt, 3 h, 30−33%; (g) HSiCl3, DCM, DMF, rt, on, 80−89%; (h) NaNO2, AcOH, MeOH/H2O, rt, on, 30−40%; (i) NaClO2, KH2PO4, 2-methylbut-2-ene, t-BuOH/THF, 30 °C, on, 79%-quantitative; (j) (i) (COCl)2(2 M in DCM), cat. DMF, DCM, rt, 1 h; (ii) JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.3c00682 JACS Au 2024, 4, 1786−1800 1788 ■RESULTS Design and Synthesis of New Microbiota-Inspired Compounds We propose that microbiota metabolites represent an interesting starting point for the design of new compounds for drug discovery. Toward this end, we have analyzed the structures of identified small-molecule metabolites 22 and explored the repertoire of organocatalytic reactions, to generate a set of microbiota-inspired compounds that combine metabolite structural diversity with synthetic accessibility using a key methodology that allows to obtain chemotypes with high structural complexity in a straightforward and stereocontrolled manner. This strategy offers the advantage of combining two or more reagents in a one-pot fashion to effectively yield a target chemotype. Thus, based on privileged scaffolds (PrSc) found in the metabolites selected in Figure 1A and amino catalytic reactions starting with a small group of aldehydes and nitro derivatives (Figure 1B), we designed new chemotypes that contain a PrSc-based central core linked to the common PrSc pmethoxyphenyl as well as to a drug-like moiety that could favor good pharmacokinetic properties. In the designed molecules 1− 13 (Figure 1C), the central cores are piperidin-2-one, piperidine, tetrahydro-1H-carbazole, chromane, and dihydropyrido[2,3-b]pyrazine. These systems are inspired on metabolites including 3-amino-2-piperidone, pipecolic acid, methyl 3-carbazolecarboxylate, equol, γ-carboxyethylhydroxy- chroman (γ-CEHC) related to vitamin E, folic acid, and trimethylpyrazine shown in Figure 1A. The central cores are constructed in a single step and with high diastereoselectivities by reaction between an α,β-unsaturated aldehyde and a nitro compound or a second aldehyde promoted by diphenylprolinol trimethylsilyl ether catalyst (Figure 1C). Further transformation of aldehyde and/or nitro groups following a general sequence allows us to obtain the desired compounds 1−13 that contain the drug-like moiety (cyclopropylmethyl)amine and an amino group. For each chemotype, we have synthesized both enantiomers�designated as compounds aand b�with high enantioselectivity by using the corresponding catalyst. The synthesis of final compounds 1−3(Figure 2A) containing a disubstituted piperidin-2-one or a piperidine scaffold started with the reaction between N-Boc-2-amino-1- nitroethane and p-methoxycinnamaldehyde to obtain intermediates 14 and 15. 23 Michael addition catalyzed by the Sor R enantiomer of 2-{diphenyl[(trimethylsilyl)oxy]methyl}- pyrrolidine in the presence of benzoic acid, and subsequent in situ oxidation of the resulting hemiaminal afforded piperidin-2- one scaffold 14 (diastereoisomeric ratio (dr) = 9:1, determined by 1H-nuclear magnetic resonance (NMR) analysis of the crude). Intermediate 14 was then deprotected before one-pot reduction of the nitro group and protection of the resulting amine to give N-Boc-5-aminopiperidin-2-one (22), which was alkylated at the amide nitrogen with (bromomethyl)- cyclopropane and treated with trifluoroacetic acid (TFA) to obtain compound 1a or 1b (enantiomeric ratio (er) = 97:3 for 1a, 96:4 for 1b). Reduction of the carbonyl group of 1yielded target compound 2a or 2b (er = 99:1). On the other hand, piperidine scaffold 15 (Figure 2A) was prepared via the same asymmetric organocatalytic Michael addition to p-methoxycin- namaldehyde, but this time the resulting cyclic hemiaminal was in situ dehydrated (dr = 8:2, determined by high-performance liquid chromatography−mass spectrometry (HPLC-MS) analysis of the crude). Subsequent reduction of both the nitro group and the double bond yielded 5-aminopiperidine 23, which afforded final compound 3a or 3b via reductive amination and deprotection (er = 98:2). For the synthesis of the tetrahydrocarbazole core present in compound 4(Figure 2B), we applied an asymmetric Diels− Alder reaction between a β-indolyl α,β-unsaturated aldehyde and trans-p-methoxy-β-nitrostyrene catalyzed by the Sor Rform of the diphenylprolinol silyl ether. 24 In this case, we optimized the reported thermal Diels−Alder reaction by using microwave (MW) conditions (irradiation at 110 °C for 5.5 h) to obtain cycloadduct 16 (dr = 98:2, 1H NMR; er = 96:4 for 16a, 98:2 for 16b). These new conditions increased the yield and reduced considerably the time with respect to the thermal reaction without affecting the stereoselectivity outcome. Compound 16 was further derivatized via reductive amination to introduce a (cyclopropylmethyl)amino moiety, subsequent reduction of the nitro group and Boc deprotection, to afford final compound 4a or 4b (er = 94:6 for 4a, 95:5 for 4b). The chromane and tetrahydrobenzo[c]chromene scaffolds 17−20 (Figure 2C), precursors of compounds 5−12, were synthesized via an asymmetric organocatalytic domino reaction between an o-nitrovinylphenol and a properly substituted α,βunsaturated aldehyde. 25,26 Thus, intermediates 17 and 19 were both obtained using p-methoxycinnamaldehyde and o-nitro- vinylphenol through a domino oxa-Michael−Michael reaction and a quadruple cascade oxa-Michael−Michael−Michael-aldol condensation reaction, respectively. Upon optimization of the reported conditions, tetrahydrobenzo[c]chromene 17 (single diastereoisomer) and chromane 19 (dr 8:2, 1H NMR) were obtained in a 1:1 ratio. Regarding tetrahydrobenzo[c]chromene and chromane intermediates 18 and 20 (Figure 2C), bearing a geminal dimethyl group, the organocatalytic cascade between 3,3-dimethylacrolein and trans-2-hydroxy-5-methoxy-β-nitro- styrene afforded only the chromane derivative 20 (dr = 8:2, 1H NMR), through a domino oxa-Michael−Michael reaction. In this case, the corresponding tetrahydrobenzo[c]chromene scaffold was not formed, probably due to the steric hindrance of the geminal dimethyl group, which may prevent the second Michael addition to a third molecule of 3,3-dimethylacrolein. Tetrahydrobenzo[c]chromene 18 was then obtained following a two-step procedure. Initial reaction between trans-2-hydroxy-β- nitrostyrene and 3,3-dimethylacrolein resulted in a chromane intermediate, which was treated in situ with less hindered Michael acceptor p-methoxycinnamaldehyde to induce a subsequent Michael-aldol condensation that provided 18 as a single diastereoisomer. Next, reductive amination with (cyclopropylmethyl)amine followed by nitro reduction of tetrahydrobenzo[c]chromenes 17 and 18 provided final compounds 5a or 5b and 6a or 6b, respectively (er = 99:1 for 5a, 98:2 for 5b, 99:1 for 6a,b) (Figure 2C). However, when the one-pot reductive amination was applied to chromane Figure 2. continued (cyclopropylmethyl)amine, 0 °C to rt, 3−18 h, 55%-quantitative; (k) LiAlH4, THF, reflux, 24 h, 20−22%. (D) Synthesis of compound 13 containing a dihydropyrido[2,3-b]pyrazine scaffold. Reagents and conditions: (a) 20 mol % chiral cat., 20% mol AcOH, CHCl3, rt, 7 days, 40−45%; (b) CH3I, Cs2CO3, DMF, rt, 1 h, 95%-quantitative; (c) (i) (cyclopropylmethyl)amine, MeOH/DCM, rt, 3 h; (ii) NaBH4, 0 °C to rt, 1−2 h, 49−50%. JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.3c00682 JACS Au 2024, 4, 1786−1800 1789 Figure 3. Compound UCM-13369 targets NPM1 protein, binding mainly through the C-end domain of the C+ mutant form. (A) Dose−response curves of UCM-13369 in AML cell lines MOLM13 (NPM1 WT) and OCI-AML3 (NPM1 C+). (B) Confocal microscopy images with staining of NPM1 (Alexa Fluor-488, green), UCM-13369-Cy5 (red), and nuclei (DAPI, blue) in OCI-AML3 cells. (C,D) 1D 1H NMR spectra of the C-end domain of NPM1 WT (C) and C+ (D). (E) ITC thermograms and binding isotherms of the interaction of UCM-13369 with the WT (blue) and C+ JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.3c00682 JACS Au 2024, 4, 1786−1800 1790 intermediate 20, the expected amine was not formed, but oxime 7was obtained instead. The procedure in two steps did not yield the expected product either but the cyclic amine 8resulting from the reduction of the corresponding imine. Alternative reduction of the imine with trichlorosilane combined with N,N- dimethylformamide (DMF) as a soft Lewis base afforded again compound 7, which was hydrolyzed to 9. Unexpected chemotypes 7a,b-9a,b (Figure 2C, er = 99:1 for 7a,b, 96:4 for 8a,b, 98:2 for 9a,b) were considered as additional final compounds to be tested in the phenotypic assays. The synthesis of desired compounds 10−12 was achieved following an alternative synthetic route starting from chromanes 19 and 20 (Figure 2C). The aldehyde group was oxidized to the corresponding carboxylic acid through the Pinnick reaction, and the (cyclopropylmethyl)amine moiety was then incorporated by coupling via the acid chloride. Next, reduction of the nitro group under standard conditions afforded compounds 10a or 10b and 11a or 11b, respectively (er = 99:1 for 10a,b and 11a,b). Finally, the amide group in the latter was reduced to obtain analogue 12a or 12b (er = 99:1). The synthesis of the dihydropyrido[2,3-b]pyrazine scaffold 21 (Figure 2D) was performed via an asymmetric domino aza- Michael-aldol reaction between 3-aminopyrazine-2-carbalde- hyde and p-methoxycinnamaldehyde. N-Methylation followed by reductive amination with (cyclopropylmethyl)amine led to the final compound 13a or 13b (er = 98:2). Cancer-Stem-Cell (CSC) Phenotype Screening: Identification of UCM-13369 Once microbiota-inspired compounds 1a−13a and 1b−13b were synthesized, we carried out a cellular phenotypic screening in cancer and CSC models. The cytotoxic effect was determined in breast (MCF-7) and colon (HCT-116) tumor cell lines using MTT assays. Then, the formation of MCF-7-derived mamospheres and HCT-116-derived colonospheres was evaluated in the presence of a noncytotoxic concentration of the compounds, using bright-field microscopy. As shown in Table S1, compounds 4b, 5a, 5b, 6a, and 6b were able to completely inhibit the formation of tumorspheres, a characteristic growth in CSCs, without affecting the viability of differentiated tumor cells. To select those compounds endowed with a good balance between efficacy and safety, we subsequently determined if the compounds that are able to act on CSCs do not affect the viability of IMR90 fibroblasts, which are commonly used as control normal cells. From the data in Table S1, tetrahydrocarbazole 4b was not toxic in nontumor cells at 5 μM, while tetrahydrobenzo[c]chromenes 5a, 5b, 6a, and 6b exhibited toxicity (20−57% cell viability @5 μM). Among them, compound 4b (UCM-13369) induced low levels of cytotoxicity in tumor cells (96 and 70% cell viability), complete inhibition of tumorsphere formation, and no toxicity in fibroblasts (100% cell viability) (Table S1 and Figure S1A), and was selected for further studies. Compound UCM-13369 Targets Nucleophosmin 1 (NPM1) Protein One of the most paradigmatic diseases with critical involvement of CSCs is AML. Among the AML stem cell population, CD34+/CD38- are the only ones capable of inducing leukemia in vivo. 27 Hence, with the aim to decipher the molecular targets of UCM-13369, 1D-nano LC ESI-MSMS proteomic analysis was performed in CD34+ cells before and after 24 h of treatment with 5 μM UCM-13369. The results showed multiple differentially expressed proteins. We focused on proteins that disappeared completely after treatment with the compound and whose concentration could be quantified in untreated samples. According to this profile, 340 proteins, including targets related with stemness and cancer, were identified (see Supplementary Excel file). Among them, nucleophosmin 1 (NPM1) especially attracted our attention, as it is one of the most common mutated proteins in AML. 28 To support that the inhibitory capacity of UCM-13369 observed in CSC cells (see Table S1) is mediated by NPM1, the compound was tested on OCI-AML3 and MOLM13 cell lines, as in vitro AML models bearing mutant (NPM1 C+) and wild-type (WT) protein, respectively. As shown in Figure 3A, the compound exhibits cytotoxicity against both cell lines in the μM range. In addition, different efficacies of UCM-13369 were observed between both AML models, being higher in NPM1 C+ expressing cells (OCI-AML3: IC50 = 7.26 μM, MOLM13: IC50 = 10.89 μM, p-value ≤0.01, Figure 3A). To further support NPM1 as a molecular target of UCM- 13369, we synthesized UCM-13369-Cy5 (Figure S2), a derivative of UCM-13369 containing the fluorescent tag Cy5, as a chemical probe for use in confocal microscopy. As shown in Figure 3B, we visualized the overlapping of UCM-13369-Cy5 (in red) with an Alexa Fluor-488-conjugated NPM1 primary antibody (in green), indicating colocalization of the compound with the protein in OCI-AML3 cells (see video in the Supporting Information). To confirm a direct interaction of UCM-13369 with NPM1 and to address which domain of the protein would be responsible for binding the compound, pure protein samples of the pentameric N-end domain of NPM1, along with the C- terminal DNA-binding domains of WT and mutant NPM1 C+, were used in calorimetry assays. As a confirmation of its native state, analysis of the 1D 1H NMR spectrum of NPM1 C-end WT showed well-dispersed sharp peaks corresponding to proper protein folding (Figure 3C). In contrast, the spectrum obtained for the C+ mutant variant exhibited drastically less dispersed and broader signals, particularly in the amide region (6−10 ppm) (Figure 3D). This is indicative of a lower diversity in the chemical environment of amide protons consistent with the previously reported protein unfolding. 29 The lower signal intensities in this region also suggest more pronounced relaxation effects, which might be related to the higher aggregation tendency of the mutant. 30 Isothermal titration calorimetry (ITC) experiments revealed that the interaction between the NPM1 protein and UCM-13369 occurs mainly through the DNA-binding domain C-end of NPM1, with the C- Figure 3. continued mutant (red) forms of NPM1. ITC-derived thermodynamic parameters of UCM-13369−NPM1 interaction. Gibbs free energy (ΔG), enthalpy (ΔH), entropic term (-TΔS), equilibrium dissociation constant (KD), and binding stoichiometry (n). (F) Visualization of UCM-13369 binding interfaces to NPM1 C-end domain. The centroids of UCM-13369 in the four lowest-energy structures of each 10 clusters are plotted as balls colored according to the HADDOCK score, as indicated in the color legend. The structure of UCM-13369 of the top-ranked structure of each cluster is represented in the insets (i-v). JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.3c00682 JACS Au 2024, 4, 1786−1800 1791 end of the NPM1 C+ variant showing a 2-fold higher affinity toward UCM-13369 than the WT domain (Figure 3D). On the other hand, the pentameric construct of the NPM1 N-end, which lacks the AML-related region, yielded a much lower binding affinity (KDca. 100 μM range, Figure 3D). Interestingly, the interaction is entropically driven in all cases. Altogether, these data suggest that the new NPM1 inhibitor UCM-13369 binds specifically to the C-enddomain of the C+ mutant protein. To identify putative binding sites of UCM-13369 in this region of NPM1, ab initio docking predictions were performed using HADDOCK software. 31,32 Docking solutions were clustered in 10 groups distributed across the surface of this domain, suggesting that UCM-13369 explores the surface of NPM1 C-end domain stochastically in the WT construct (Figure 3E, Figure S4, and Table S2). Notably among these, cluster 3 matched with a previously reported NPM1 binding site of a small molecule that would allow the conformation stabilization of the C+ mutant. 33 In line with this finding, this cluster exhibited the lowest desolvation energy and harbored the second largest buried surface area of all analyzed docking clusters (Table S2). Further work should aim to confirm whether enhanced disorder in the NPM1 C+ species results in the unraveling of cryptic binding sites that lead to (i) a higher binding affinity to UCM-13369 and (ii) the stabilization of NPM1 conformational states. Figure 4. UCM-13369 inhibits gene and mutant NPM1 C+ protein expression and restores nucleolar localization. (A) Schematic representation of the molecular signaling pathways of NPM1 mutation in AML. (B) Dot plot showing decreased NPM1 expression with UCM-13369 treatment (10 μM) in OCI-AML3 cells. (C) Western blot showing decreased NPM1 and c-Myc expression and increased FBXW7 expression with UCM-13369 treatment (10 μM) in OCI-AML3 cells at 24 and 48 h. Densitometry values for NPM1: reduction of 50% and 70%, respectively. (D) Top: Confocal microscopy images with NPM1 (Alexa Fluor-647, red) and nuclei (DAPI, blue) staining in MOLM13, OCI-AML3, and UCM-13369-treated OCI-AML3 cells (10 μM). Bottom: Violin plot of NPM1 nucleus/cytoplasm ratio in MOLM13 and OCI-AML3 cell lines, and in untreated and UCM-13369-treated OCIAML3 cells. JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.3c00682 JACS Au 2024, 4, 1786−1800 1792 Unraveling the Molecular Mechanism of the New NPM1 Inhibitor UCM-13369 We next studied the impact of UCM-13369 in the genes and/or expression levels of NPM1 and NPM1 C+ pathway components (Figure 4A), including the positive regulation of c-Myc 34 and the negative regulation of FBXW7 and the pro-apoptotic CASP8. 35 Upon treatment with the compound in OCI-AML3 cell line, qPCR and Western blot revealed reduced levels for NPM1 and c-Myc, while FBXW7 was found to be upregulated (NPM1 p-value ≤0.0001, Figure 4B,C). The inhibition of NPM1 was also observed in MOLM13 cells expressing WT NPM1 (Figure S1B). To further elucidate the mechanism of action of UCM-13369, we studied the influence of the compound on the subcellular localization of the protein by confocal microscopy, using an Alexa Fluor-647-conjugated NPM1 primary antibody (Figure 4D). In OCI-AML3 cells, we visualized that UCM-13369 induced the reduction of the cytoplasmic form of the protein, typically associated with the NPM1 mutation (NPM1 C+), 36 restoring the nucleolar localization phenotype of the protein, more similar to the WT state of NPM1 in MOLM13 cells. Validation of the Therapeutic Potential of UCM-13369 in AML Patient Cells Considering the inhibitory capacity of UCM-13369 on AML cell lines (shown in Figure 3A), the therapeutic potential was assessed in primary CD34+ cells from AML patients. Two ex vivo models were used from AML patients and healthy donors: CD34+ liquid HSC (hematopoietic stem cell) culture as a shortterm model and CFU-E (colony-forming unit−erythroid) as a midterm model. Primary CD34+ cells were sensitive to UCM- 13369, exhibiting decreased viability, as shown in Figure 5A for CFU-E (IC50 = 0.30 μM for AML patients, IC50 = 1.73 μM for healthy donors, p-value ≤0.0001) and in Figure 5B for liquid HSC culture. Interestingly, the compound displayed higher efficacy against primary cells in AML patients than in healthy donors (IC50 = 1.877 and 6.399 μM, respectively, p-value ≤0.0001). Moreover, the compound showed higher efficacy in patients with NPM1 C+ AML (IC50 = 1.725 μM, n= 4) than in WT NPM1 AML patients (IC50 = 3.774 μM, n= 3, p-value ≤0.007) (Figure 5B). The inhibition of NPM1, especially NPM1 C+ pathway (as shown in Figure 4), by UCM-13369 correlates with the increase Figure 5. UCM-13369 has efficacy against AML cell lines and primary cells. (A) Dose−response curve of UCM-13369 in colony-forming unit (CFU) assays performed with primary HSCs bone marrow mononuclear fraction from healthy donors and AML patients. (B) Left: Dose−response curve of UCM-13369 in short-term cultures of HSCs performed with primary HSCs from bone marrow mononuclear fraction from healthy donors and AML patients. Right: Dose−response curve of the same assay with samples from AML patients with NPM1 WT vs NPM1 C+. (C) Dot plot of a representative sample from flow cytometry analysis using CD34-PE/Annexin V-FITC staining of short-term cultures of primary HSCs from AML patients performed with primary HSCs from bone marrow mononuclear fraction with and without treatment. JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.3c00682 JACS Au 2024, 4, 1786−1800 1793 of apoptosis observed in treated primary CD34+ cells upon staining with Annexin V (Figure 5C). Efficacy in an AML Animal Model Based on the encouraging results obtained in ex vivo models of AML, we further assessed the therapeutic potential of the new NPM1 inhibitor in vivo. Prior to the evaluation of the compound in an animal model of AML, in vivo pharmacokinetic studies were addressed. Following a single intraperitoneal administration of UCM-13369 at a dose of 25 mg/kg, levels of 200−600 μg/L were detected in mouse peripheral blood up to 6 h (Figure 6A). Plasma protein binding of the compound in mouse serum was measured using an equilibrium dialysis assay, indicating a bound fraction of 83 ±1%. Next, the efficacy of the compound was evaluated in immunodeficient mice after injection of OCIAML3-ffLuc-GFP cells as an AML animal model. Notably, as shown in Figure 6B, the UCM-13369-treated group (50 mg/kg, tail vein injection) showed a reduction of tumor infiltration compared to untreated group at end point (the reduction in tumor mass was negligible in 25-mg/kg-treated animals, data not shown). However, this effect did not result in an increase in survival as no difference was observed between treated and untreated animals (Figure S1C). Hematoxylin-eosin (H&E) staining of the bone marrow revealed less tumor cell infiltration in treated mice, supporting the efficacy of the compound against malignant cells (Figure 6C). Caspase-3 pathology analysis of treated mice revealed the absence of toxicity in bone marrow and liver (data not shown), while some caspase-3 cleavage was observed in the kidney, pointing to some toxicity in this organ, although most part of this staining was unspecific (Figure S1D). Indeed, the dose−response viability assay in human renal proximal tubule epithelial cells (hRPT Epi Cells) revealed toxicity (Figure S1E), suggesting that renal toxicity is one of the probable causes of lower-than-expected survival of treated AML animals. ■DISCUSSION In the past decade, the human microbiota has emerged as one of the most important topics in cancer research, as multiple associations with cancer development, progression, and response to treatment have been reported. 4,37,38 Considering the cancer protective effect of several microbiota metabolites, 39−41 in the present work, we describe the design, screening, and validation of novel small molecules inspired on microbiota metabolites in order to identify new drug candidates for cancer treatment. We designed a set of new chemotypes containing privileged scaffolds present in selected metabolites 22 (Figure 1) and obtained them following synthetic routes based on asymmetric organocatalytic reactions (Figure 2). This valuable methodology provides complex chemotypes characterized by the presence of chiral carbon centers, which also address the demand for compounds with an increased number of sp3carbon atoms, a parameter that has achieved great significance in Figure 6. UCM-13369 shows efficacy against AML cells in vivo. (A) Values of UCM-13369 concentration in vivo at different time points after injection (25 mg/kg) analyzed by HPLC-MS. (B) Bioluminescence images acquired using an in vivo imaging system of NSG mice after injection of OCI-AML3 cell line with and without UCM-13369 treatment (50 mg/kg). (C) Microscopy analysis of the OCI-AML3 infiltration by H&E staining slides of bone marrow samples from NSG mice with paired time points after engraftment with and without UCM-13369 treatment (50 mg/kg). JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.3c00682 JACS Au 2024, 4, 1786−1800 1794