Effects of 6-Aminonicotinic Acid Esters on the Reprogrammed Epigenetic State of Distant Metastatic Pancreatic Carcinoma (2024)

Associated Data

Supplementary Materials

Abstract

Keywords: 6-Aminonicotinamide, 6-aminonicotinic acid esters, epigenetics, metabolomics, pancreatic cancer, 6-phosphogluconate dehydrogenase

6-Aminonicotinamide (1, 6AN) (Figure ​Figure11) is one of the most extensively used 6-phosphogluconatedehydrogenase (6PGD) inhibitors in cell-based studies. 6AN itselfdoes not inhibit 6PGD but rather is converted to 6-amino-NADP⁺ (2) by being processed via the salvage pathwayin place of nicotinamide.¹ 6-Amino-NADP⁺ was found to inhibit 6PGD from rat tissues with a K_i value of 0.1–0.2 μM.¹ Despite its close structural similarity to NADP⁺, 6-amino-NADP⁺ was reported to be a >200-foldweaker inhibitor against two other NADP⁺/NADPH-dependentenzymes, glucose-6-phosphate dehydrogenase (G6PD) and glutathionereductase,¹ alleviating some concerns overindiscriminate inhibition of various NADP⁺/NADPH-dependentenzymes.

Given the essential role played by 6PGD in theoxidative phaseof the pentose phosphate pathway (oxPPP), 6AN has been extensivelyutilized as a tool compound to study the effects of 6PGD inhibitionin many forms of cancer.²⁻⁶ We recently reported epigenetic reprogramming (e.g., reduction inH3K9 di- and trimethylation) concurrent with elevated endogenous 6PGDenzymatic activity and complete depletion of 6-phosphogluconate (6PG),the 6PGD substrate, across pancreatic ductal adenocarcinoma (PDAC)distant metastases and their precursors in the primary tumors, butnot other subclones in the primary tumors and locoregional metastases.^7,8 Although 6AN had no effect on global chromatin modifications inthe peritoneal subclone, it reversed the reprogrammed chromatin stateand gene expression states of the distant metastasis from the samepatient, impeding cell growth and tumorigenesis in 2D and 3D tumorigenicassays. These findings suggest that 6AN has the potential to serveas an anticancer agent targeting 6PGD-dependent PDAC distant metastases.

The therapeutic utility of 6AN is, however, severely limited becauseof its well-documented neurotoxicity observed in both clinical⁹ and preclinical studies.¹⁰⁻¹² It was recentlyreported that 5-(aminomethyl)pyridin-2-amine potently and selectivelyreversed 6PGD-dependent metastatic properties by forming a conjugatewith adenine dinucleotide phosphate in the intracellular compartment.¹³ It remains to be seen, however, whether thismolecule is devoid of the neurotoxic effects observed with 6AN. Weexplored another strategy to design molecules that can intracellularlyinhibit 6PGD. To this end, we investigated various esters derivedfrom 6-aminonicotinic acid (2, 6ANA) with the premisethat they can be first hydrolyzed into 6ANA intracellularly and thenconverted into 6-amino-NADP⁺ via the Preiss–Handlerpathway, another endogenous biosynthetic route to NADP⁺ starting from nicotinic acid. Herein we report the pharmacologicaleffects of 6-aminonicotinic acid esters in comparison to 6AN in cell-basedassays assessing the ability to reverse the reprogrammed epigeneticand gene expression state of the distant metastasis.

Alkyl estersof 6-aminonicotinic acid 5a–c wereeither purchased or synthesized as described previously.¹⁴ Acyloxymethyl and alkoxycarbonyloxymethyl esters 5d–i were synthesized in one step from6ANA via base-mediated reaction with the corresponding chlorides 4 in yields of 23–89% as shown in Scheme 1.

6-Aminonicotinic acid esters 5a–i were first tested for their abilityto reverse the loss of histone3 lysine 9 trimethylation (H3K9me3) in patient-derived PDAC distantmetastasis (A38-5) using hom*ogeneous time-resolved fluorescence (HTRF)assays.¹⁵ Lysates from A38-5 cells treatedwith test compounds for 72 h were analyzed for H3K9me3 and total histoneH3 levels, respectively, using the HTRF assays. In each experiment,the HTRF ratio from the H3K9me3 assay was divided by that of the histoneH3 assay from the same lysate in order to eliminate the potentialinfluence of cell death caused by the treatment. The resulting valueswere normalized to those obtained from the DMSO control group. 6ANwas used as a positive control every time this assay was conductedso that the data obtained from the test compounds could be directlycompared to those from 6AN in the same run. The results are summarizedin Table 1. 6AN increasedthe intracellular H3K9me3 levels in a dose-dependent manner up to∼2.5-fold at the highest concentration tested (250 μM).6ANA, however, showed almost no effect on the intracellular H3K9me3levels at up to 250 μM. This is presumably due to its poor cellmembrane permeability because of the negatively charged carboxylategroup at physiological pH. Alkyl esters 5a–c also showed negligible effects on the intracellular H3K9me3levels. We hypothesized that the lack of activity shown by these alkylesters 5a–c could be attributed tothe inability of intracellular hydrolases to recognize nicotinic acidesters as substrates, failing to produce 6ANA intracellularly. Thisprompted us to explore acyloxymethyl and alkoxycarbonyloxymethyl estersthat can be hydrolyzed at their terminal site followed by spontaneousrelease of a formaldehyde or acetaldehyde molecule, resulting in theformation of 6ANA. In marked contrast to alkyl esters 5a–c, acyloxymethyl esters 5d–f increased the intracellular H3K9me3 levels with a potencysimilar to or greater than that of 6AN. Alkoxycarbonyloxymethyl esters 5g–i also increased the intracellularH3K9me3 levels in a dose-dependent manner. These findings suggestthat intracellular hydrolases can act on the terminal ester moietyaway from the nicotinate ester, eventually releasing 6ANA. Among them,compound 5i appears to be the most potent in increasingthe intracellular H3K9me3 levels with a robust activity at 1.0 μM.It should be noted that an approximately 15-fold higher concentration(15.6 μM) was required to achieve the same degree of increasewith 6AN.

The ability of compound 5i to increasethe intracellularH3K9me3 levels was subsequently validated by Western blotting (Figure ​Figure22A). Compound 5i also exhibited stronger cytotoxicity in A38-5 comparedto 6AN (Figure ​Figure22B).This is consistent with our previous findings that the restorationof the global heterochromatin is closely associated with suppressedcell growth and tumorigenicity.⁷ Theseresults support that compound 5i is cell-permeable, hydrolyzedto 6ANA, and taken up by the Preiss–Handler pathway to generate6-amino-NADP⁺. Indeed, it was recently reported that theDNA copy numbers of nicotinate phosphoribosyltransferase (NAPRT) andNAD synthetase 1 (NADSYN1), the two key enzymes involved in the Preiss–Handlerpathway, were increased in pancreatic cancers.¹⁶

Since quantification of the intracellular concentrationof 6-amino-NADP⁺ derived from compound 5i isunfeasible in theabsence of a reference material, we conducted targeted metabolomicsprofiling to verify intracellular 6PGD inhibition. Lysates from A38-5cells exposed to DMSO, 6AN (250 μM), and compound 5i (15.6 μM) for a period of 72 h were analyzed for key metabolitesinvolved in oxPPP. Notably, the lysates derived from A38-5 cells treatedwith 6AN and compound 5i displayed a modest (2- to 2.5-fold)increase in glucose 6-phosphate (G6P), a substrate for G6PD, and a150- to 180-fold increase in 6PG, a substrate for 6PGD (Figure ​Figure33). These findings are consistentwith the selective inhibition of 6PGD over G6PD achieved by 6-amino-NADP⁺ derived from 6AN or compound 5i. In contrastto the metabolites upstream of 6PGD in oxPPP (G6P and 6PG), both 6ANand compound 5i had little impact on the levels of ribulose5-phosphate (Ru-5-P), a product of the 6PGD-catalyzed reaction. Thismay be due to the reversible nature of the nonoxidative phase of PPP(noxPPP), where Ru-5-P is in equilibrium with ribose-5-phosphate.Taken together, the targeted metabolomic analysis results indicatethat compound 5i is efficiently transformed into 6-amino-NADP⁺ intracellularly and selectively inhibits 6PGD.

Given the well-documented neurotoxicity of 6AN,^12,17−19 we investigated whether compound 5i retainsthe neurotoxic properties of 6AN using primary rat hippocampal neurons.The cultures were subjected to either 6AN or compound 5i at 5 or 50 μM for 24 h. Treatment with 6AN at both 5 and 50μM led to a significant increase in the percentage of cellsthat were positive for TUNEL staining compared to the DMSO control(Figure ​Figure44). This resultis consistent with the previously reported neurotoxicity of 6AN.^12,17−19 In contrast, neither 5 nor 50 μM treatmentof compound 5i led to a significant increase in TUNEL-positivecells compared with DMSO-treated cells. Taken together, these datademonstrate that compound 5i does not cause significantneurotoxicity at concentrations that reverse the epigenetic and metabolicchanges in the distant metastatic pancreatic carcinoma. Although speculative,the stark contrast in neurotoxicity of 6AN and compound 5i could be due to the insignificant contribution to NAD biosynthesisin the brain by the Preiss–Handler pathway, which is essentialfor the conversion of compound 5i to 6-amino-NADP⁺. Indeed, NAPRT responsible for the rate-limiting step inthe Preiss–Handler pathway was previously reported to havethe lowest gene expression levels in normal brain tissue comparedto various tissues analyzed.¹⁶

Our study demonstrated the ability of some 6-aminonicotinicacidesters to reverse the loss of H3K9me3 in distant metastatic pancreaticcarcinoma cells more potently than 6AN. Among them, compound 5i was found to reduce viable A38-5 cells in a dose-dependentmanner. Like 6AN, metabolite analysis indicated that the pharmacologicaleffects of compound 5i appear to be caused by its abilityto inhibit 6PGD after being converted into 6-amino-NADP⁺ intracellularly via the Preiss–Handler pathway. Despite thesimilar pharmacological properties shown by 6AN and compound 5i in the distant metastatic pancreatic carcinoma cells, compound 5i did not display the neurotoxic properties exhibited by6AN, opening up opportunities to explore 6ANA derivatives, includingcompound 5i, as new therapeutic agents targeting distantmetastatic cells. An important next step will be characterizationof the absorption, distribution, metabolism, and excretion (ADME)profile and pharmaco*kinetic properties in preparation for in vivoproof-of-concept studies.

Acknowledgments

This work was supported by the Lustgarten Foundationand the National Institute of Mental Health (P30MH075673).

Glossary

Supporting Information Available

The Supporting Information isavailable free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.2c00404.

• Synthetic procedures, ¹H and ¹³C NMR spectra, and biological experimental methods (PDF)

Author Contributions

^# R.-D.G. and M.M. contributed equally. All of theauthors contributedto the work and approved the final version of the manuscript.

Notes

Theauthorsdeclare no competing financial interest.

Supplementary Material

References