Small Molecule Inhibitors of Arylamine N-Acetyltransferase 1 Attenuate Cellular Respiration

Abstract

Arylamine N-acetyltransferase 1 (NAT1) expression has been shown to attenuate mitochondrial function, suggesting it is a promising drug target in diseases of mitochondrial dysfunction. Here, several second-generation naphthoquinones have been investigated as small molecule inhibitors of NAT1. The results show that the compounds inhibit both in vitro and in whole cells. A lead compound (Cmp350) was further investigated for its ability to alter mitochondrial metabolism in MDA-MB-231 cells. At concentrations that inhibited NAT1 by over 85%, no overt toxicity was observed. Moreover, the inhibitor decreased basal respiration and reserve respiratory capacity without affecting ATP production. Cells treated with Cmp350 were almost exclusively dependent on glucose as a fuel source. We postulate that Cmp350 is an excellent lead compound for the development of NAT1-targeted inhibitors as both experimental tools and therapeutics in the treatment of hypermetabolic diseases such as amyotrophic lateral sclerosis, cancer cachexia, and sepsis.

Arylamine N-acetyltransferase 1 (NAT1) was first identified as a drug-metabolizing enzyme that catalyzes the acetylation of arylamines and heterocyclic amines. More recently, the enzyme has been associated with changes in cell proliferation and colony formation in soft agar,^1,2 mitochondrial oxidative phosphorylation³ and pyruvate dehydrogenase regulation,⁴ and production of reactive oxygen species.⁵ Deletion of the NAT1 gene in cancer cells decreases mitochondrial cytochrome C release following treatment with cytotoxins.⁶ These effects have led several investigators to suggest NAT1 is a potential drug target.^7,8 Inhibitors of NAT1 may be useful in subpopulations of breast cancer patients with elevated tumor NAT1 activity who consistently have poorer outcomes⁹ as well as diseases with hypermetabolic states such as sepsis.

Small molecule inhibitors of enzymes are important experiment tools and pivotal to our understanding of many cellular processes. They are also essential therapeutics for the treatment of human diseases. The first systematic study of small molecule NAT1 inhibitors investigated rhodamine and thiazolidine-2,4-dione analogues, which showed IC₅₀ concentrations in the low micromolar range.¹⁰ These compounds suffered from poor solubility and poor uptake in cells.¹¹ One analogue (rhodo-hp) was subsequently tested in intact cells and showed weak inhibitory activity with evidence of toxicity.² A series of naphthoquinones was developed as colorimetric probes for NAT1^12,13 but failed to detect NAT1 in intact biological systems.¹⁴ A thiozolidine derivative (5E-[5-(4-hydroxy-3,5-diiodobenzylidene)-2-thioxo-1,3-thiazolidin-4-one), with an IC₅₀ of 0.7 μM for recombinant NAT1 and 50–60 μM for intact cells, significantly reduced cell proliferation and anchorage-dependent colony formation in MDA-MB-231 cells.¹⁵ Lastly, a virtual screen of 2 million chemicals identified several anthraquinones that selectively inhibited NAT1. These compounds also showed efficacy in whole cells.¹⁶ A priority in this field is to identify a lead compound that could be developed for further in vitro and in vivo studies. Such a compound would be a valuable tool for investigating the role of NAT1 in cell biology. It might also provide a lead for preclinical studies into the safety and efficacy of NAT1 inhibition.

Here, we investigated the effects of second-generation naphthoquinone inhibitors that were synthesized as part of a large study into the development of novel NAT1 inhibitors.¹⁴ The compounds were chosen based on their solubilities and potencies compared to an initial hit (Cmp32, Table 1) which has been reported previously as “naphthoquinone 1”,¹⁷ “compound 1” ¹³ and “compound 4”.¹⁸ However, Cmp32 lacks favorable solubility in aqueous solutions for future development. Cmp350, bearing a carboxylic acid substituent on the aniline moiety, and Cmp353, a furanyl amine, were chosen because of their superior solubility and potential cell-penetrating activities (flux) compared to Cmp32 (Table 1). Cmp220 is a trimethoxy substituted biarylamine derivative that showed excellent inhibitory activity in preliminary studies using ZR-75-1 cell extracts.¹⁴

Table 1. List of Compounds Investigated along with Their Solubility, Flux Across Caco-2 cells, IC₅₀ and Effects on PABA Metabolism by NAT1 (Data are mean ± SEM).

^aData from Egleton.¹³

^bKinetic solubility was determined by turbidimetry in 0.01 M PBS (pH 7.4) with 1% DMSO.

^cFlux measured across confluent Caco-2 cells (apical to basolateral) using 10 μM compound for 2 h.

^dp-Aminobenzoic acid (PABA) kinetics were determined in the presence of each inhibitor at their respective 50% inhibitory concentrations (IC₅₀). Km = Michaelis–Menten constant, Vmax = maximum velocity, Cmp = compound. *Significantly different compared to PABA alone (Km = 115 ± 8 μM, Vmax = 9.7 ± 0.2 nmol/min/mg protein), p < 0.01 (Student’s t test).

Results

Initially, concentration-response curves were constructed for each compound using cell cytosols from MDA-MB-231 cells (Figure 1A). Cmp32 showed an IC₅₀ of 2–3 μM (Table 1), which agrees with previously published data.¹⁸ Cmp220 was the most potent inhibitor (IC₅₀ = 7.4 nM). However, its flux across confluent Caco-2 cells was poor (3.0 ± 2.0 × 10^–6 cm^–1) and its solubility was similar to that for Cmp32. By contrast, Cmp350 and 353 have higher kinetic solubilities and better permeability activities. Cmp350 was significantly more potent than Cmp32 (IC₅₀ 44 ± 3 vs 2650 ± 360, p < 0.001, Mann–Whitney test), but Cmp353 showed weak inhibitory activity toward NAT1 (IC₅₀ > 9 μM).

Figure 1.

Inhibition of NAT1 in vitro and in intact cells and effect of inhibitors on mitochondrial respiration. (A) Inhibition of PABA acetylation by each inhibitor using MDA-MB-231 cell lysates. (B) Eadie Scatchard plot of data in A. (C) Effect of 100 μM inhibitor on NAT1 activity in MDA-MB-231 cells treated for 24 h (p values calculated by one-way ANOVA). (D) Relationship between in vitro inhibition (IC₅₀) and in situ inhibition for the inhibitors. (E) Effect of NAT1 inhibitors on mitochondrial respiration after 24 h treatment. Each compound was removed by washing before Seahorse traces were measured. Cells were treated sequentially with oligomycin (O), carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) (F) and rotenone/antimycinA (RA) which allowed quantification of basal respiration, ATP production, reserve respiratory capacity and maximum respiration (see Supporting Information for details). (F) Effect of NAT1 inhibitors on mitochondrial respiration after 24 h treatment without removal of each compound. (G) Effect of NAT1 inhibitors on mitochondrial respiration after 1 h treatment without subsequent removal. Data are mean ± SD, n = 4 except for A-C where n = 3.

To identify the possible mechanism of inhibition for each compound, p-aminobenzoic acid (PABA) kinetic curves were constructed in the absence and presence of each inhibitor (Figure 1B). In the absence of inhibitor, the Km and Vmax for PABA acetylation were 115 ± 8 μM and 9.7 ± 0.2 nmol/min/μg protein, respectively. Apart from Cmp350, each inhibitor decreased Vmax, suggesting these compounds may not be simple competitive inhibitors of NAT1 (Table 1). Although not investigated here, it is possible the inhibitors affect the interaction of acetyl-coenzyme A (AcCoA) with the enzyme, which can produce an “uncompetitive” inhibitor profile, as has been reported for ATP.¹⁹

Next, the in situ inhibition of NAT1 was determined by incubating MDA-MB-231 cells with PABA. Preliminary studies using 0–20 μM of each compound showed dose-dependent inhibition (Supporting Information Figure S1). However, the decrease in NAT1 activity was less than 30%. Therefore, 100 μM was tested. All four inhibitors reduced NAT1 activity by 56–91% suggesting they were taken up into the cells (Figure 1C, one-way ANOVA, p < 0.001). The decrease in NAT1 activity was not due to overt toxicity (Supporting Information Figure S2). Moreover, the decrease in enzyme activity in situ strongly correlated with the IC₅₀ values seen in vitro (r = −0.98, p = 0.02, Figure 1D). Taken together, these results indicate all four compounds can cross the plasma membrane and inhibit NAT1 in intact cells.

To evaluate the effect of each inhibitor on mitochondrial respiration, cells were treated using three different protocols. First, each compound was added to the cells for 24 h and then respiration was measured without removing the drug. This experiment established the effects of prolonged exposure (Figure 1E). Second, cells were treated for 24 h and then washed before respiration was measured (Figure 1F). This experiment determined whether any chronic effects on mitochondrial respiration could be reversed by removing the inhibitor. Finally, cells were treated for only 1 h to determine whether acute exposure to each inhibitor affected respiration (Figure 1G). For prolonged treatment without subsequent washing, all compounds altered oxidative phosphorylation, primarily by decreasing basal respiration and reserve respiratory capacity (Figure 1E, two-way ANOVA p < 0.01)). The oxygen consumption curves for Cmp32 and 220 showed marked inhibition with a profound decrease in basal respiration and ATP production that suggested mitochondrial dysfunction. As these changes were not observed in MDA-MB-231 cells devoid of NAT1,^3,4 the results suggest off-target effects by both compounds.

When the cells were washed following treatment to remove the inhibitor, the effects on mitochondrial respiration were attenuated for Cmp32 and 353, and completely reversed for Cmp220 and 350 (Figure 1F). Additional washing steps may have also reversed the effects of Cmp32 and 353. Finally, when cells were treated for only 1 h, both basal respiration and reserve respiratory capacity decreased in a similar manner following chronic treatment for all compounds except Cmp32, which was less effective (Figure 1G, two-way ANOVA p < 0.01). These results demonstrate that both acute and chronic treatments reduce mitochondrial function, which can be reversed by removing the inhibitors.

Based on its potency both in vitro and in whole cells, its ability to inhibit mitochondrial respiration, and the rapid and complete reversal following washing, Cmp350 was chosen for more in-depth study in whole cells. First, a concentration-response curve following 24 h treatment in MDA-MB-231 cells showed an IC₅₀ of 39 μM (Figure 2A). Again, the decrease in activity was not due to acute cell toxicity since Cmp350 (100 μM) had no effect on viable cell numbers after 24 h (Figure 2B, Student’s t test, p > 0.05). Because NAT1 substrates such as p-aminosalicylic acid, ethyl-p-aminobenzoate and p-aminophenol can induce NAT1 protein degradation,^20,21 changes in protein stability may account for the loss of activity in situ. To test this, the level of NAT1 protein in the absence and presence of Cmp350 was determined (Figure 2C). There was a significant increase in protein, although this varied in experiments with the inhibitor (Student’s t test, p = 0.03). To identify if this apparent induction of NAT1 protein affected activity, cells were washed to remove Cmp350 after 24 h treatment and cytosols were prepared to measure NAT1 activity (Figure 2D). There was a significant increase in activity at 0 and 4 h post-wash, which agrees with the increased protein expression. By 24 h, activity returned to control levels. These results show that inhibition of NAT1 using a small molecule inhibitor may lead to increased expression or stability of the protein.

Figure 2.

Effect of Cmp350 on NAT1 in MDA-MB-231 cells. (A) Concentration-response curve for Cmp350 inhibition of NAT1 for 24 h. Results are mean ± SD, n = 3. (B) Effect of 100 μM Cmp350 on viable cell number after 24 h treatment. For no compound, n = 4; for Cmp350 n = 8. (C) Effect of Cmp350 on NAT1 protein after 24 h treatment. Western blot is representative of 6 replicates. Results are mean ± SD, n = 6. (D) NAT1 activity following removal of Cmp350. Cells were treated with 100 μM Cmp350 for 24 h, washed and then NAT1 activity measured at 0, 4, and 24 h. Results are mean ± SD (n = 3, Student’s t test).

The effects of Cmp350 on mitochondrial oxidative phosphorylation following 24 h treatment (100 μM) are shown in Figure 3A. There was a slight decrease in basal oxygen consumption (p < 0.01), but no change in ATP production. Most notable was a significant decrease in reserve respiratory capacity (p < 0.01) and maximum respiration (p < 0.01). The mitochondria in cells treated with Cmp350 had little or no reserve indicating that, under basal conditions, they respired at maximum rate. This is similar to what was seen following only 1 h treatment (Figure 1G).

Figure 3.

Effect of Cmp350 on mitochondrial respiration and fuel usage. (A) Oxidative phosphorylation in parental MDA-MB-231 cells in the absence (−) or presence (+) of 100 μM Cmp350, n = 8. (B) Oxidative phosphorylation data for NAT1 knockout MDA-MB-231 cells under the same conditions as A. None of the data were significantly different, n = 8. (C) Fuel dependency in MDA-MB-231 cells in the absence (−) or presence (+) Cmp350, n = 4. (D) Fuel capacity in MDA-MB-231 cells in the absence (−) or presence (+) of Cmp350, n = 4 except for residual n = 20. All data are mean ± SD (p values calculated by Student’s t test test).

Mitochondrial respiration was also measured in MDA-MB-231 cells where the NAT1 gene was deleted by CRISPR-Cas9 gene editing²² (Figure 3B). The NAT1-deleted line had a decreased reserve and maximum capacity compared to the parental line (compare Figure 3A and B), consistent with a requirement for NAT1 expression. However, Cmp350 had no effect on either basal respiration, ATP production, reserve respiratory capacity or maximum respiration. Interestingly, maximum respiratory capacity in the parental line treated with Cmp350 was similar to that in the NAT1-deleted line. Taken together, these results suggest NAT1 activity is required for both optimum reserve respiratory capacity and maximum respiratory capacity, and that the effects of Cmp350 on mitochondrial respiration are the result of NAT1 inhibition.

Finally, the dependence of mitochondrial respiration on glucose, fatty acids and glutamine was determined by blocking the use of each with UK5099, etomoxir and BPTES, respectively. MDA-MB-231 cells primarily use glucose for oxidative phosphorylation, followed by glutamine with little or no dependence on fatty acids (Figure 3C). Cmp350 increased glucose dependence by more than 50% and decreased glutamine dependency completely (Figure 3C). Thus, the cells were almost entirely dependent on glucose for mitochondrial basal respiration. However, the cells had the capacity to use all three fuels if required (Figure 3D). Cmp350 enhanced the capacity for cells to use glucose, but not fatty acids or glutamine. At the end of each experiment, a residual oxygen consumption rate is often observed. This represents non-mitochondrial oxygen consumption (see Supporting Information Figure S4), and the use of other carbon substrates such as amino acids and ketone bodies.²³Figure 3D, last panel, shows that Cmp350 caused a small but significant decrease (p < 0.001) in the use of the alternative carbon substrate(s). Thus, inhibition of NAT1 in MDA-MB-231 cells not only switched mitochondrial fuel dependency to glucose exclusively but it also increased the capacity of these cells to use glucose. There are numerous factors that dictate mitochondrial carbon substrates in both normal and transformed cells.²⁴ It will be important to investigate the effects of Cmp350 in other cell lines for a more complete understanding of the role of NAT1 in fuel usage.

Discussion

In this study, the ability of three novel NAT1 inhibitors to affect mitochondrial respiration have been compared with the initial lead compound Cmp32. All compounds inhibited NAT1 both in vitro and in whole cells, albeit at very different concentrations. Cmp350 showed excellent inhibition in cells up to 100 μM without toxicity. Moreover, it decreased maximum respiration rates in a reversible manner. The inhibitor also induced a switch to an exclusive glucose-dependent phenotype. The molecular mechanism for this is currently unknown. However, it is similar to that seen following p31 inhibition in glioma cells,²⁵ Complex I inhibition in fibroblasts,²⁶ and hypoxia treatment in stem cells.²⁷

An unexpected observation in the current study was the induction of NAT1 protein with Cmp350. This may be a form of positive feedback loop where the cell senses the loss in NAT1 activity. One possible mechanism for this might involve changes to AcCoA levels, which reportedly increased in NAT1-deleted cells.²⁸ Increasing AcCoA increases protein acetylation, including histones, which affects gene expression.²⁹ Increasing protein acetylation using sirtuin inhibitors such as trichostatin A or vorinostat also increases NAT1 expression and activity.^30,31

Aside from their role as experimental tools to study NAT1 function, NAT1 inhibitors may be useful in the future for the treatment of various hypermetabolic diseases. For example, high NAT1 expression in patients with amyotrophic lateral sclerosis correlates with a hypermetabolic state,³² which predicts poorer outcomes.³³ Patients with high NAT1 show a significantly shorter survival,³⁴ suggesting that inhibition of NAT1 may have therapeutic potential. We postulate that Cmp350 may also be useful in diseases where fatty acid oxidation is prevalent since it can switch cell metabolism to a more glucose-dependent phenotype. Fatty acid oxidation increases reactive oxygen production in mitochondria, which can disrupt Ca²⁺ homeostasis.³⁵ An over-reliance on fatty acid oxidation by mitochondria is observed in central nervous system disorders, such as multiple sclerosis, Parkinson’s disease and Alzheimer’s disease,³⁶ as well as in heart disease.³⁷ An important extension of the current study is to investigate the effects of Cmp350 in appropriate animals models of these diseases.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.4c00282.

• Experimental procedures including reagents, synthesis of inhibitors, NAT1 assays, mitochondrial respiration and fuel use assays, Western blotting, cell number quantification and data analysis (PDF)

Author Contributions

^† CC and JEE contributed equally. CC, JEE, ARJ and RFM planned the experiments; CC, JEE and NJB performed experiments; NJB, AJR and RFM provided supervision; CC and RFM wrote the manuscript; all authors reviewed and edited the manuscript.

The authors declare no competing financial interest.