Sage Journals: Discover world-class research

Abstract

Glucose-6-phosphate dehydrogenase (G6PD) is the key enzyme of the pentose phosphate pathway, converting glucose-6-phosphate to 6-phosphoglucono-δ-lactone with parallel reduction of NADP⁺. Several human diseases, including cancer, are associated with increased G6PD activity. To date, only a few G6PD inhibitors have been available. However, adverse side effects and high IC₅₀ values hamper their use as therapeutics and basic research probes. In this study, we developed a high-throughput screening assay to identify novel human G6PD (hG6PD) inhibitors. Screening the LOPAC (Sigma-Aldrich; 1280 compounds), Spectrum (Microsource Discovery System; 1969 compounds), and DIVERSet (ChemBridge; 49 971 compounds) small-molecule compound collections revealed 139 compounds that presented ≥50% hG6PD inhibition. Hit compounds were further included in a secondary and orthogonal assay in order to identify false-positives and to determine IC₅₀ values. The most potent hG6PD inhibitors presented IC₅₀ values of <4 µM. Compared with the known hG6PD inhibitors dehydroepiandrosterone and 6-aminonicotinamide, the inhibitors identified in this study were 100- to 1000-fold more potent and showed different mechanisms of enzyme inhibition. One of the newly identified hG6PD inhibitors reduced viability of the mammary carcinoma cell line MCF10-AT1 (IC₅₀ ~25 µM) more strongly than that of normal MCF10-A cells (IC₅₀ >50 µM).

Keywords

glucose-6-phosphate dehydrogenase pentose phosphate pathway redox metabolism high-throughput screening MCF10

Introduction

Glucose-6-phosphate dehydrogenase (G6PD) is the first and rate-limiting enzyme of the oxidative branch of the pentose phosphate pathway (PPP). G6PD catalyzes the conversion of glucose-6-phosphate (G6P) to 6-phosphoglucono-δ-lactone (6PGδL), which is accompanied by NADPH production. The product 6PGδL is hydrolyzed by 6-phosphogluconolactonase (6PGL) to form 6-phosphogluconate, which is converted by 6-phosphogluconate dehydrogenase (6PGD), resulting in ribulose 5-phosphate and NADPH.¹ The oxidative PPP thus yields NADPH essential to maintain the redox equilibrium² and adequate ribose-5-phosphate levels crucial for nucleotide biosynthesis in proliferating cells.³

Several studies suggest an important role of G6PD in the pathology of various human diseases such as heart failure, type 2 diabetes, hypertension,⁴ and cancer.^3,5

In the context of cancer research, glucose metabolism has been of interest for more than 80 years because of initial observations that cancer cells take up high amounts of glucose and produce high amounts of lactate.⁶ Because lactate production is pronounced even under aerobic conditions, a high aerobic glycolytic rate was associated with cancer cell metabolism, which is known as the Warburg effect.⁷ At present, the increased glucose uptake by cancer cells is commonly used as a diagnostic tool for tumors.⁸ However, the assumption that the increased glucose metabolism involves only glycolysis has been questioned.⁷ Boros et al.^3,9 observed that inhibition of the nonoxidative and/or oxidative PPP reduces cell proliferation of pancreatic carcinoma cells in vitro, suggesting an important role of the PPP in cancer cells. These findings are supported by the fact that mammary carcinoma cells present an increased glucose flux through the PPP in vitro.¹⁰ Moreover, G6PD activity is increased in tumor tissues compared with normal cells,^5,11 and silencing G6PD in tumor cells results in decreased proliferation and enhanced apoptosis in vitro.¹² Very recently, it was shown that the tumor suppressor p53 regulates the PPP activity by binding to and suppressing G6PD, highlighting the impact of G6PD in cancer cells.¹³

Inhibitors with in vivo activity against human G6PD are 6-aminonicotinamide (6AN) and 17-ketosteroides such as dehydroepiandrosterone (DHEA). 6AN, a competitive G6PD and 6PGD inhibitor,^14,15 was used in the past for chemotherapy of various tumors.⁴ However, 6AN treatment had severe side effects, such as nerve damage and vitamin B deficiency. Disadvantages of the uncompetitive G6PD inhibitor DHEA are high oral doses and the generation of active androgens.⁴ The recent finding that antiproliferative effects of DHEA are most likely not due to G6PD inhibition but rather to changes in mitochondrial gene expression¹⁶ highlights the need for novel selective G6PD inhibitors with few side effects in order to investigate the impact of G6PD on human diseases.

Here we describe the discovery of novel G6PD inhibitors using a high-throughput screening (HTS) approach. Selected G6PD inhibitors were characterized and investigated in the MCF10 model for mammary carcinoma.¹⁷ Our results could provide a basis for research on several human diseases and potentially for the development of novel antitumor therapies.

Materials and Methods

G6PD Overexpression and Purification

hG6PD was overexpressed and purified according to our previously described protocol.¹⁸ In brief, a clone (ID 282264) from the National Institutes of Health (NIH) Mammlian Gene Collection was purchased from Invitrogen (Carlsbad, CA). The hG6PD sequence (GenBank accession number NP_001035810) was cloned into the pET24a expression vector (Novagen) and overexpressed with a C-terminal His tag in Escherichia coli BL21 cells (Invitrogen) containing pRAREII. Overexpression was performed at 23 °C in 2xYT medium (16 g tryptone, 10 g yeast, 5 g NaCl per liter medium) containing 50 µg/mL kanamycin and 12.5 µg/mL chloramphenicol. Protein expression was induced with 0.1 mM IPTG at an optical density at 600 nm of 0.1. Cells were harvested after 24 h, lysed, and purified by metal affinity chromatography using Ni-NTA. Buffer used for hG6PD lysis and purification was 50 mM Tris/HCl, 300 mM NaCl, 0.1 mM NADP⁺, pH 8.0. hG6PD was eluted from the Ni-NTA column with 150 and 300 mM imidazole and stored with 1.8 M ammonium sulfate + 0.1 mM NADP⁺ at 4 °C.

Assay Reagents

NADP⁺ (Amresco), G6P (Sigma-Aldrich, St. Louis, MO), diaphorase (Sigma-Aldrich), and resazurin (Sigma-Aldrich) were prepared as concentrated stock solutions, which were stored in aliquots at −80 °C. Stability of the assay reagents was checked by monitoring K_m values and enzyme titration patterns of freshly purified hG6PD from time to time. K_m and V_max determinations for one hG6PD batch over several weeks revealed information about enzyme stability. DHEA (Sigma-Aldrich) and 6AN (Sigma-Aldrich) were prepared at 30 and 90 mM, respectively and stored at −80 °C.

Assay Development

The HTS assay development, including determination of K_m values, appropriate hG6PD, and resazurin concentrations, DMSO tolerance, signal-to-background (S/B) ratio, and Z′-factor, was performed according to our previously described HTS assay development protocols for G6PD from Plasmodium falciparum.¹⁹

Compound Libraries for HTS

The LOPAC library (Sigma) consists of 1280 druglike molecules in the fields of cell signaling and neuroscience, which were present at a concentration of 10 mM. The Spectrum collection (Microsource Discovery Systems, Gaylordsville, CT) includes 2320 biologically active and structurally diverse compounds, experimental bioactives, and pure natural products, which were also present at a concentration of 10 mM. The DIVERSet (ChemBridge, San Diego, CA) consist of ~50 000 druglike compounds with maximum pharmacophore diversity at a concentration of 0.5 mg/mL. LOPAC and Spectrum libraries were stored in a desiccator cabinet at 20 °C, whereas the DIVERSet was stored at −20 °C.

Primary and Secondary Screening

The LOPAC and Spectrum compound libraries were screened following the HTS procedure described previously.¹⁹ In brief, 9 µL of a substrate mix (Tris [pH 7.5], MgCl₂, G6P, diaphorase, resazurin) were transferred into 384-well plates (Greiner, Kremsmuenster, Austria) using the Multidrop Combi reagent dispenser (Thermo Scientific, Waltham, MA). A total of 36 nL of the compounds was then added with the Echo 555 liquid handler (Labcyte, Sunnyvale, CA) to columns 3 to 22. Columns 1 and 2 served as positive controls (100% inhibition) and contained the substrate mix without G6P. Columns 23 and 24 were negative controls (0% inhibition) without compounds. A total of 36 nL DMSO was added to positive and negative controls. Afterward, 9 µL of an enzyme mix (NADP⁺, Tween 20, hG6PD) was added, and the plates were incubated in the dark for 40 min. Finally, the plates were measured at ex 530/em 590 using the multiwell plate reader Analyst HT (LJL Biosystems), and the data were analyzed with CBIS (Chemical and Biological Information Systems, http://www.cheminnovation.com). The final assay concentrations were 0.05 M Tris (pH 7.5), 3.3 mM MgCl₂, 96 µM G6P, 1 U/mL diaphorase, 0.1 mM resazurin, 11 µM NADP⁺, 0.005% Tween 20, and 0.51 nM hG6PD. The compounds were screened at 20 µM with an incubation time of 40 min. For screening of the DIVERSet, the substrate mix contained 122 µM G6P and 130 µM resazurin and the enzyme mix 8 µM NADP⁺ due to slight K_m variations of different protein batches. Compounds of the DIVERSet were screened at a final concentration of 5 µg/mL. In addition, the incubation time was increased to 60 min. Linearity of the reaction was achieved over 60 min using an enzyme concentration of 0.51 nM. Compounds inhibiting hG6PD ≥50% were retested in dose response in a secondary screening.

Orthogonal Assay

Selected hit compounds were purchased from ChemBridge in powder form, dissolved in DMSO, and included in IC₅₀ determinations in an orthogonal assay, which did not include resazurin and diaphorase. The orthogonal assay was performed according to a previously described protocol.¹⁹ In brief, 6 µL of a substrate mix containing 0.15 M Tris (pH 7.5), 9.9 mM MgCl₂, 78 µM NADP⁺, and 354 µM G6P were transferred into wells of a 384-well plate (Greiner). Six microliters of varying compound concentrations (0–15 µg/mL) were added. DMSO was added instead of compounds as negative control (0% inhibition); the substrate mix without G6P was used as positive control (100% inhibition). Afterward, 6 µL of enzyme mix composed of 0.015% Tween 20 and 1.1 nM hG6PD were added to start the reaction (enzyme concentrations were adjusted if necessary). The fluorescence of NADPH was monitored over 20 to 30 min at ex 340/em 460 nm using the multiwell plate reader Infinite M2000 combined with the Magellan 6 software (Tecan). IC₅₀ values were calculated using the GraphPad Prism software based on the endpoint of linear reaction curves. Dose-response tests for DHEA and 6AN included final concentrations of 0 to 1 mM DHEA or 0 to 5 mM 6AN. Fluorescence of NADPH was detected for these compounds with the SpectraMax M5 (Molecular Devices, Sunnyvale, CA) multiwell plate reader.

Mechanistic Characterization

The mechanism of inhibition of selected compounds was investigated by titrating the compounds at varying substrate and co-substrate concentrations as described in the following. Six microliters of G6P (0–2400 µM) or NADP⁺ (0–264 µM) in 9.9 mM MgCl₂, 3 mM TCEP, 0.15 M Tris (pH 7.5), and 33 µM NADP⁺ for G6P titration or 300 µM G6P for NADP⁺ titration were transferred to 384-well plates (Greiner). Six microliters of varying compound concentrations (final about 0–10× IC₅₀) were added, after which 6 µL of an enzyme mix consisting of 0.015% Tween 20 and 0.41 nM hG6PD were added to start the reaction. Fluorescence of NADPH was measured at ex 340/em 460 nm using the SpectraMax M5 (Molecular Devices) multiwell plate reader for several minutes, and the initial slope of the reaction was determined with the SoftMax Pro 5.2 software (Molecular Devices). Based on these data, K_m and V_max values were determined with the GraphPad Prism software, and α from the following rate equation (equation 1) used for general description of inhibitors was calculated with SigmaPlot (Systat Software).

\begin{array}{l} v = \frac{V_{max}^{app} [S]}{K_{m}^{app} + [S]}, \\ where V_{max}^{app} = \frac{V_{max}}{(1 + [I] / α K_{i})} and K_{m}^{app} = K_{m} \frac{{1 + [I] / K}_{i}}{(1 + [I] / α K_{i})} \end{array}

Determination of the mechanism of action was based on the following assumptions²⁰: competitive inhibitors present increasing K_m^app values with increasing compound concentrations, whereas V_max^app values remain constant, α = ∞; noncompetitive inhibitors show decreasing V_max^app values with increasing compound concentrations, whereas K_m^app values stay constant, α = 1; uncompetitive inhibitors present decreasing K_m^app and V_max^app values with increasing compound concentrations, α << 1; mixed-type inhibitors are represented by parameters between those of competitive and noncompetitive inhibitors.

Time Dependency and Reversibility

To investigate whether compounds bind to hG6PD in a time-dependent manner, enzyme and inhibitor were preincubated at high concentrations in order to shift the binding equilibrium and promote time-dependent processes. The experiments were performed as described previously.¹⁹ In brief, 50 µL of varying compound concentrations (0–0.15 mg/mL compound [=37× final concentration or max 3% DMSO]) were prepared in the first three columns of a 96-well plate (Greiner). At time point 0, 50 µL of an enzyme mix consisting of 0.1 M Tris (pH 7.5), 2 mg/mL BSA, and 0.12 µM hG6PD (=37.5× final concentration) was added to column 1. After 1 h, 8 µL of the enzyme/inhibitor mix of column 1 was added to 92 µL of buffer, which consisted of 0.05 M Tris (pH 7.5), 1 mg/mL BSA, and 0.0082% Tween 20 (=1.635× final concentration). At the same time, 50 µL enzyme mix was added to column 2. After another hour, 8 µL of enzyme/inhibitor mix of column 2 was added to 92 µL of buffer. Then, 50 µL enzyme mix was added to column 3, of which 8 µL was directly added to 92 µL of buffer. Finally, 12 µL of the enzyme/inhibitor/buffer mixes were transferred to 384-well plates, and 6 µL of a substrate mix (9.9 mM MgCl₂, 30 µM NADP⁺ [=3× final concentration], 0.05 M Tris, 1 mg/mL BSA, 390 µM G6P [3× final concentration]) was added to start the reaction. The substrate mix without G6P was used as a control representing 100% inhibition. Fluorescence of NADPH was detected at ex 340/em 460 with the SpectraMax M5 (Molecular Devices) multiwell plate reader, and the IC₅₀^app values were calculated with GraphPad prism based on the initial slope of the reaction. In this way, 1 h preincubation and 1 h postdilution (1:1), 1 h preincubation and 0 h postdilution (1:0), and 0 h preincubation and 0 h postdilution (0:0) were determined. Condition 0:0 was used as the control. A decrease in IC₅₀^app values following the incubation suggests time-dependent binding and dissociation. If the IC₅₀^app after 1:0 h incubation was lower than the control, the compound was assigned to inhibit time dependently. If the IC₅₀^app after 1:1 h was lower than the control, the compound inhibited irreversibly; if the IC₅₀^app after 1:1 h was equal or higher than the control, the compound inhibited reversibly.

Cell Viability Assay for MCF10-A and MCF10-AT1 Cells

MCF10-A (ATCC, Manassas, VA) and MCF10-AT1 (The Barbara Ann Karmanos Cancer Institute, Detroit, MI) cell lines were cultured in DMEM medium (CellGro; 10% FBS, 1% penicillin-streptomycin, 100 ng/mL EGF) in an atmosphere of 37 °C and 5% CO₂. A total of 15 000 to 30 000 MCF10-A cells and 10 000 to 15 000 MCF10-AT1 cells were seeded in the wells of a 96-well plate. After 24 h, the cells were ~50% confluent, and compounds were added in varying concentrations (0–100 µM) to the wells (final 0.98% DMSO). DMSO was used as negative control; the known G6PD inhibitor 6AN (max 325 µM) was used as a positive control. The plates were incubated for 24 h. Afterward, the cells were washed twice with phosphate-buffered saline (PBS), and the cells were fixed by adding 50 µL paraformaldehyde (PFA; 4%) for 10 min. PFA was removed, and the cells were incubated with 50 µL 0.1% methylene blue for 1 h. After washing with PBS, the cells were incubated with 0.1 M HCl. After 1 h, the absorbance was measured at 660 nm using a Synergy x Multiwell reader (Biotek, Winooski, VD), and IC₅₀ values were calculated using GraphPad Prism. The background (absorbance of 0.1 M HCl) was subtracted from all values. To test growth inhibition at varying confluences, MCF10-A and MCF10-AT1 cells were seeded at varying concentrations.

Solubility and Permeability Analysis

Solubility analysis was performed using a direct UV kinetic solubility method in a 96-well format. All liquid dispense and transfer steps were performed with the Freedom Evo automated liquid handler (Tecan US). Solubility measurements were performed in an aqueous buffer solution (pION) at pH 5.0, 6.2, and 7.4 in duplicate. Samples were incubated at room temperature for 18 h to achieve equilibrium, then filtered to remove any precipitate. The concentration of the compound was then measured by UV absorbance.

Permeability was assessed using the Parallel Artificial Membrane Permeability Assay (PAMPA) in a 96-well format. Permeability measurements were performed in an aqueous buffer solution (pION) at pH 5.0, 6.2, and 7.4, in duplicate. A sandwich plate consisting of a donor bottom plate and an acceptor filter plate was used. The donor wells contained the compound in 180 µL system solution and magnetic stir bars. The filter on the bottom of each acceptor well was coated with GIT-0 liquid and filled with 200 µL of Acceptor Sink Buffer, pH 7.4, containing a surfactant to mimic the function of serum proteins. The permeation time was 30 min, and moderate stirring was applied. After the permeation time, the sandwich was disassembled and compound present in both the donor and acceptor wells was measured by UV absorbance.

Results

Assay Development

The HTS assay used in this study was developed based on a resazurin/diaphorase–coupled G6PD assay, the principle of which was described previously.¹⁹ In brief, G6P is converted to 6-phosphoglucono-δ-lactone by the enzyme G6PD.¹ The reaction is accompanied by production of NADPH, which is used in a coupled reaction to reduce resazurin by means of the enzyme diaphorase (Clostridium kluyveri) to form the highly fluorescent resorufin.²¹ The final amount of generated resorufin is proportional to the converted G6P and thus reflects G6PD activity.

We overexpressed and purified mg quantities of hG6PD according to our previously developed protocol¹⁸ and stored the enzyme with 1.8 M ammonium sulfate and 0.1 mM NADP⁺ at 4 °C. Under these conditions, the enzyme was stable for several weeks. Only slight changes in K_m and V_max values occurred after longer storage periods, and all parameters were confirmed before using the enzyme. Assay reagents such as NADP⁺, G6P, resazurin, and diaphorase were stable over months, as judged by K_m values and enzyme titration patterns comparable to freshly purified enzyme (data not shown).

K_m values were determined for all protein batches that were used during the study by titrating the substrate and co-substrate. hG6PD presented a K_m value of 95 to 120 µM for the substrate G6P ( Fig. 1A ) and about 10 µM for the co-substrate NADP⁺ ( Fig. 1B ).

Figure 1.

High-throughput screening assay development. Michaelis-Menten curves for G6P (A) and NADP⁺ (B) for hG6PD. (C) hG6PD titration shown as initial slope versus hG6PD concentration. (D) Dose-response curve for dehydroepiandrosterone (DHEA). The signal of an endpoint reaction was plotted against the logarithmic DHEA concentration. Assay without substrate was used as positive control (100% inhibition). Assay without any compound represents the negative control (0% inhibition). RFU = relative fluorescence units.

Previous development of a resazurin-coupled HTS assay for P. falciparum G6PD revealed that the fluorescence signal of resorufin was quenched at higher resazurin concentration,¹⁹ consistent with the strong inner filter effect of resazurin. A similar effect of resazurin concentration was observed in the hG6PD assay (Fig. S1). Based on these findings, a resazurin concentration slightly higher than the G6P concentration was included, sufficient for not limiting the reaction.

hG6PD titration showed that the signal was linear for up to 60 min, including 0.51 nM hG6PD. Furthermore, the ratio of signal-to-enzyme concentration was linear up to a concentration of at least 2 nM ( Fig. 1C ), confirming that the included enzyme concentration was in an appropriate range. Based on these initial experiments, the screening assay was tested using 0.05 M Tris (pH 7.5), 3.3 mM MgCl₂, 96 µM G6P, 1 U/mL diaphorase, 0.1 mM resazurin, 11 µM NADP⁺, 0.005% Tween 20, and 0.51 nM hG6PD. The test run presented an S/B ratio of 10 and a Z-factor of 0.84 (Fig. S2). Furthermore, DMSO titration revealed that hG6PD tolerates DMSO up to a concentration of at least 4% (Fig. S3). The assay was verified by testing the known hG6PD inhibitor DHEA²² in varying concentrations. It was possible to detect inhibitory effects of DHEA and to determine an IC₅₀ value based on the developed HTS assay ( Fig. 1D ).

Screening of Compound Libraries

The HTS assay was used to screen three small-molecule compound libraries. First, LOPAC and Spectrum libraries, which contain a smaller number of compounds (<3000), were screened for assay verification. Compounds of these libraries were included in the assay at a final concentration of 20 µM. By screening of the LOPAC library (1280 compounds), we identified seven compounds inhibiting hG6PD ≥50%, which corresponds to a hit rate of 0.55%. The average S/B ratio and Z-factor were 5.5 and 0.82, respectively ( Table 1 ).

Table 1.

Screening Results of Primary High-Throughput Screen of LOPAC (20 µM), Spectrum (20 µM), and DIVERSet Compound Libraries (5 µg/mL)

Library (Comp. No.)	Hits (≥50% Inhibition)	Hit Rate (%)	S/B Ratio	Z-Factor
LOPAC (1280)	7	0.5	5.5 ± 0.1	0.82 ± 0.01
Spectrum (1969)	25	1.3	7.3 ± 0.3	0.80 ± 0.04
DIVERSet (49 14;971)	107	0.2	8.1 ± 2.7	0.85 ± 0.03

Twenty-five hG6PD inhibitors were found by screening the Spectrum library (1969 compounds), presenting a hit rate of 1.3% (≥50% inhibition), an S/B ratio of 7.3, and a Z-factor of 0.80 ( Table 1 ). Compounds showing ≥50% inhibition of hG6PD were later included in a secondary screening for hit confirmation and IC₅₀ determination.

The DIVERSet (~50 000 compounds) was screened at a concentration of 5 µg/mL. Primary screening resulted in 107 hit compounds (≥50% inhibition), a hit rate of 0.21%, an S/B ratio of 8.1, and a Z-factor of 0.85 ( Table 1 ). Hit compounds were included in a secondary screening, resulting in a hit confirmation of 66%. The secondary screening was performed at the same day of the primary screening to prevent several thawing cycles of the DIVERSet library.

Orthogonal Assay

Four of the hit compounds, namely, CB63, CB70, CB72, and CB104, inhibited hG6PD with an IC₅₀ value <4 µM in the orthogonal assay, the kinetic assay based on NADPH fluorescence intensity. Structures of these compounds and CB83, an hG6PD inhibitor identified in a previous study,¹⁹ are shown in Figure 2 . Dose-response curves of CB63 (IC₅₀ 0.6 ± 0.0 µM), CB70 (IC₅₀ 2.6 ± 1.1 µM), CB72 (IC₅₀ 3.1 ± 0.8 µM), CB104 (IC₅₀ 3.0 ± 1.2 µM), and the previously identified inhibitor CB83²² are given in Figure 3 . For the known G6PD inhibitors DHEA and 6AN, we found IC₅₀s of 483 ± 202 µM and >3 mM, respectively (uncoupled G6PD assay, data not shown). More accurate IC₅₀ determinations for DHEA and 6AN were not possible because of solubility issues of DHEA at concentrations >350 mM and DMSO limitations greater than 5 µM of 6AN.

Figure 2.

Structures of selected hG6PD inhibitors. (a) CB83 was identified as an hG6PD inhibitor in a previous study.¹⁹

Figure 3.

Dose-response curves for selected hG6PD inhibitors. Dose-response curves for CB63 (A), CB70 (B), CB72 (C), CB104 (D), and CB83 (E). Data for CB83 were obtained in a previous study.¹⁹ The percentage activity inhibition was plotted against the logarithmic compound concentration. No included substrate was used as positive control (100% inhibition); no included compound represents the negative control (0% inhibition). For each compound, one representative IC₅₀ curve out of at least three independent determinations is shown.

Mechanistic Characterization, Time Dependency, and Reversibility

CB63, CB70, CB72, and CB104 were furthermore mechanistically characterized. All four compounds showed competitive inhibition to the substrate G6P ( Table 2 ). CB63 showed noncompetitive or mixed-type inhibition against NADP⁺, CB70 and CB104 showed mixed-type inhibition against the cofactor, and CB72 demonstrated competitive inhibition with respect to NADP⁺. Figure 4 shows representative data distributions for CB63 versus G6P (A) and NADP⁺ (B), as well as K_m and V_max changes with increasing inhibitor concentrations, α, and K_i, based on which the mechanism of inhibition was determined.

Table 2.

Time Dependency and Reversibility of hG6PD Inhibitors^a

			IC₅₀^app (µM), Hours Preincubation:Postdilution
Compound	MOI vs. G6P	MOI vs. NADP⁺	0:0	1:0	1:1
CB63	Competitive	Noncompetitive or mixed type	1.36 ± 0.68	0.32 ± 0.11	0.83 ± 0.19
CB70	Competitive	Mixed type	4.59 ± 1.35	5.80 ± 4.20	19.39 ± 0.5
CB72	Competitive	Competitive	—	—	—
CB83^b	—	—	>30	0.37 ± 0.03	0.21 ± 0.03
CB104	Competitive	Mixed type	5.08 ± 2.43	1.91 ± 0.79	3.43 ± 0.64

IC₅₀^app values are given as average with standard deviation of two to three independent experiments, each performed in duplicate.

Values determined in a previous study.

Figure 4.

Mechanistic characterization of CB63 for hG6PD. Three-dimensional view of data distribution of G6P and CB63 titration (A) and NADP⁺ and CB63 titration (B). The average of duplicates of one representative data set out of at least three independent experiments is shown. RFU = relative fluorescence units.

Data for CB70, CB72, and CB104 are given in Figure S4. CB63 and CB104 demonstrated time-dependent components in inhibition of hG6PD as judged by shifts in the compound concentration inhibiting hG6PD activity by 50% under the specified conditions (IC₅₀^app). The IC₅₀^app values decreased from 1.4 to 0.3 µM for CB63 and from 5.1 to 1.9 µM for CB104 after 1 h preincubation of hG6PD and inhibitor, both present at 16× assay concentration. Interestingly, additional incubation after 10-fold dilution resulted in recovery of the original IC₅₀ values observed in the mix-and-measure (control) assay ( Table 2 ). This suggests that both binding and dissociation of CB63 and CB104 are reversible yet slow and require extended time for reaching equilibrium. This behavior is in stark contrast to the previously observed irreversible inhibition with CB83 that demonstrated >100-fold lower IC₅₀^app after preincubation¹⁹; 1 h incubation after the dilution failed to reverse this effect ( Table 2 ). CB70 inhibited hG6PD reversibly. IC₅₀^app determination for CB72 was not possible under the assay conditions chosen. Without preincubation, there was no inhibition detectable, whereas G6PD activity increased with increasing CB72 concentrations with preincubation/postincubation ( Table 2 ).

Effect of hG6PD Inhibitors on Cancer Cells

In our study, we used the MCF10 model of mammary carcinoma¹⁷ to investigate effects of hG6PD inhibitors on breast cancer cells. The spontaneously immortalized MCF10-A cells were used as a model for normal cells,²³ whereas the tumorigenic MCF10-AT1 cells served as a model for breast cancer cells.²⁴ It has been previously reported that the flux of G6PD to ribose is approximately four times higher in the MCF10-AT1 cells than in the MCF10-A cells.¹⁰ We tested the effects of 0.1 to 100 µM of the most potent hG6PD inhibitors identified in our previous HTS approach on both cell lines. Of the tested compounds, CB83 showed cell growth inhibition that was stronger for MCF10-AT1 (IC₅₀ around 25 µM) compared to MFC10-A cells (IC₅₀ >50 µM; Fig. 5 ). CB63, CB72, and CB104 did not inhibit cell growth at concentrations ≤50 µM (data not shown).

Figure 5.

Effects of CB83 on MCF10-A (A) and MCF10-AT1 (B) cell viability. A representative data set is shown. Experiments were performed in duplicate, and values are given as average with standard deviation. Studies including varying confluences were performed once, and data were confirmed by three to four further independent experiments including confluences between 40% and 70%.

Solubility studies for CB83 revealed an aqueous solubility in pION’s buffer of 13.2 µg/mL (pH 5.0), 6.4 µg/mL (pH 6.2), and 3.4 µg/mL (pH 7.2) and of 8.2 µg/mL in 1× PBS (pH 7.4). PAMPA assays presented at an acceptor pH of 7.4 values of 579/647/573 × 10⁻⁶ cm/s at a donor pH of 5.0/6.2/7.4.

Discussion

Assay Development and Screening

Several human diseases, including cancer,^3,5 are associated with increased G6PD activity. However, only a few hG6PD inhibitors are available, and their use is limited because of adverse side effects.⁴ In this study, we developed an HTS assay to identify new inhibitors for hG6PD. The assay development was based on previously established HTS assays for Leuconostoc mesenteroides (PubChem AID 1020) and P. falciparum G6PD,¹⁹ using a resazurin/diaphorase–coupled detection system.²¹ Detection of reduced resazurin (ex 530/em 590) instead of NADPH (ex 340/em 460) provides higher assay sensitivity and reduces the risk of detecting optical artifacts compared to the standard G6PD assay.²⁵ Determination of reagent and enzyme stability, K_m values, optimal enzyme and resazurin concentrations, DMSO tolerance, dose response of a known inhibitor, and of the Z-factor built the assay development process.

Our HTS assay was set up to identify competitive, noncompetitive, and uncompetitive inhibitors and thus was run at G6P and NADP⁺ concentrations at their respective K_m values. Based on our findings, hG6PD presented K_m values of ~100 µM for G6P ( Fig. 1A ) and ~10 µM for NADP⁺ ( Fig. 1B ), which are similar to the ones determined in previous studies.¹⁸ Slightly lower values were reported by other groups,^26,27 which is most likely due to methodical differences (e.g., variations in co-substrate concentrations). A resazurin titration confirmed our previous finding¹⁹ that high resazurin concentrations decrease the assay sensitivity by quenching the signal (Fig. S1). Therefore, we included resazurin at a low concentration that was assured not to limit the reaction. An enzyme concentration of 0.03 µg/mL, which was in the linear range of signal-to-enzyme ratio ( Fig. 1C ), allowed for a linear reaction over 60 min. With a Z-factor of 0.84 and an S/B ratio of 10 (Fig. S2), our assay presented excellent HTS statistics^28,29 and was further verified by a dose-response test of the known hG6PD inhibitor DHEA ( Fig. 1D ).³⁰ With our assay system, we determined an IC₅₀ of ~330 µM for DHEA, which was confirmed to be in a similar range in the orthogonal assay. However, the accuracy of the IC₅₀ calculation for this specific compound was limited because of insolubility of DHEA >350 µM. Earlier studies from the 1960s reported lower IC₅₀ values (~10–80 µM),^31,32 which may have been caused by the use of enzyme purified from mammalian cells or blood lysates, saturating substrate and co-substrate concentrations, or the detection of NADPH absorbance. Recently published values using recombinant hG6PD (~11 µM)³³ might be due to variations in the enzyme assay used for IC₅₀ determination.

Moreover, the commercially available LOPAC and Spectrum libraries (<3000 compounds) were screened at a concentration of 20 µM to verify the established HTS assay. At a 50% cutoff, hit rates were with 0.5% for LOPAC and 1.3% for the Spectrum library in a normal range, and Z-factor (0.8) and S/B ratio (>5) were excellent ( Table 1 ). Subsequent screening of the DIVERSet library (49 971 compounds) presented similar HTS statistics and presented a hit rate of 0.2% ( Table 1 ), which is in the lower range for HTS assays. Because the compound composition of the libraries was different and consisted of different scaffolds and chemical classes, variations in hit rates are normal and usually expected to be about 0.5% to 1%. Furthermore, compared to screenings of the LOPAC and Spectrum collections, the final compound concentration was with 5 µg/mL (MW of compounds <500) slightly lower for screening of the DIVERSet. In a secondary screening, we were able to confirm 66% of the DIVERSet hits as hG6PD inhibitors.

Orthogonal Assay and Characterization of Inhibitors

Several hit compounds were retested in a G6PD assay that did not contain the resazurin/diaphorase–coupled system for identification of false-positives that were due to compound interference with the coupled system. We found four compounds (CB63, CB70, CB72, CB104) that inhibited recombinant hG6PD with IC₅₀ values <4 µM ( Fig. 3 ), which are 100- to 1000-fold lower compared to those of the known hG6PD inhibitors DHEA and 6AN. Mechanistic characterization studies revealed that all four compounds inhibited hG6PD competitively versus G6P. For the co-substrate, CB70 and CB104 showed mixed-type inhibition, whereas CB63 inhibited noncompetitively and CB72 competitively versus NADP⁺ ( Table 2 , Fig. 4 , Fig. S4). Compared to the uncompetitive hG6PD inhibitor DHEA³² and the NADP⁺-competitive inhibitor 6AN,¹⁴ the compounds identified in this study follow different overall mechanisms of inhibition.

Experiments to determine the mechanism of inhibition were limited for CB83. As discussed previously, adding BSA is necessary for assays run over several hours in order to stabilize hG6PD. However, adding BSA resulted in increased IC₅₀^app values, which was most likely due to the compound binding to BSA. Inclusion of 10xIC₅₀ concentrations needed for mechanistic characterization was not possible because of autofluorescence and insolubility of the compound at these high concentrations.¹⁹

CB63 and CB104 demonstrated time-dependent but fully reversible binding and dissociation to hG6PD. CB70 was found to be a reversible hG6PD inhibitor ( Table 2 ).

Based on our findings, CB72 caused improved hG6PD activity with increasing concentrations in the reversibility studies, because of which IC₅₀ values could not be determined under the assay conditions. A possible explanation for this finding is that CB72 mimics NADP⁺ that serves as a structural component³⁴ and that binding of CB72 stabilizes hG6PD and prevents loss of enzyme activity over time.

In a previous study, CB63, CB70, and CB104 were found to inhibit P. falciparum G6PD (PfGluPho).¹⁹ The sequence of the G6PD part of PfGluPho is very similar to hG6PD, except for a 62-amino-acid insertion,³⁵ which may explain this finding. However, all three compounds inhibit hG6PD two- to fourfold more potently than PfGluPho and present differences in the mechanism of inhibition. These variations may be caused by a combination of G6PD and 6PGL in the bifunctional PfGluPho, which most likely results in a slightly different folding compared to the isofunctional hG6PD.

Effects of hG6PD Inhibitors on MCF-10 Cells

Besides characterization of hG6PD inhibitors using the recombinant enzyme, we were further interested in effects of the compounds on human cells. Because several forms of cancer are associated with increased G6PD activity,^5,11 we investigated the effects of hG6PD inhibitors on cell viability of mammary carcinoma cells. We used the MCF10 model,¹⁷ including the spontaneously immortalized MCF10-A cells²³ as control (“normal” cells) and the tumorigenic MCF10-AT1²⁴ as cancer cells, which have the same genetic background. CB63, CB72, and CB104 presented no or very slight inhibition of MCF10-A and AT1 cell growth after a 24 h incubation with 0 to 100 µM of the compounds. Possible explanations for this finding are low membrane permeability or stability of the compounds as well as potential binding to other molecules (such as BSA, discussed above). These aspects will be addressed in structure-activity relationship studies in the future.

CB70 was not available at the time of the cell-based experiments. However, in a previous study, we observed inhibitory effects of another compound, CB83, against hG6PD.¹⁹ This compound showed time-dependent effects with an IC₅₀ of 0.37 µM after a 1 h incubation. Because CB83 showed antiplasmodial activity,¹⁹ we assumed that it was membrane permeable and thus included it in our studies. CB83 indeed showed an effect on survival of MCF10 cells, with IC₅₀ values of ~25 µM for the 10AT1 ( Fig. 5B ) and >50 µM for the 10A cell line ( Fig. 5A ). Permeability and solubility studies showed that CB83 has moderate membrane permeability, which might also be responsible for the relatively high concentrations needed and poor solubility. Solubility and permeability could potentially be improved by chemical alteration of functional side groups. Furthermore, detailed studies of CB83 on various metabolic pathways as well as on the underlying mechanisms that lead to reduced cell viablity should be addressed in the future.

Increased hG6PD activity is suggested to be involved in the pathogenesis of several human diseases.⁴ Only a few hG6PD inhibitors are known,^14,33 the use of which, however, is limited.⁴ In this study, we developed an HTS assay for recombinant hG6PD and screened three small-molecule compound libraries. Overall, we were able to identify four new hG6PD inhibitors with IC₅₀ values <4 µM. These new inhibitors are much more potent and follow different mechanisms of inhibition compared to previously known hG6PD inhibitors. Furthermore, one of the inhibitors reduced cell viability of mammary carcinoma cells. Thus, our study provides novel hG6PD inhibitors that could be used for research in several human diseases and potentially as a starting point for the development of new antitumor therapies.

Footnotes

Acknowledgements

The authors would like to thank Beate Hecker and Carolin Marx for their excellent technical assistance. We would also like to thank Layton Smith and Danielle McAnally of the pharmacology core at Sanford-Burnham Medical Research Institute for determining the solubility and permeability for CB83.

Supplementary material for this article is available on the Journal of Biomolecular Screening Web site at .

Abbreviations: 6AN, 6-aminonicotinamide; 6PGD, 6-phosphogluconate dehydrogenase; 6PGδL, 6-phosphoglucono-δ-lactone; 6PGL, 6-phosphogluconolactonase; DHEA, dehydroepiandrosterone; G6P, glucose-6-phosphate; G6PD, glucose-6-phosphate dehydrogenase; HTS, high-throughput screening; PAMPA, parallel artificial membrane permeability assay; P., Plasmodium; PPP, pentose phosphate pathway; S/B: signal-to-background.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was supported by the National Institutes of Health (1R21AI082434-01 to L.B.), NIH Roadmap Initiative grant U54 HG005033 (A.P., M.H., E.S.), and the Deutsche Forschungsgemeinschaft (Be1540/18-1 to K.B., S.R.).

References

Wamelink

M. M.

Struys

E. A.

Jakobs

The Biochemistry, Metabolism and Inherited Defects of the Pentose Phosphate Pathway: A Review. J. Inherit. Metab. Dis. 2008, 31, 703–717.

Kuo

W. Y.

Tang

T. K.

Effects of G6PD Overexpression in NIH3T3 Cells Treated with Tert-butyl Hydroperoxide or Paraquat. Free Radic. Biol. Med. 1998, 24, 1130–1138.

Boros

L. G.

Puigjaner

Cascante

Lee

W. N.

Brandes

J. L.

Bassilian

Yusuf

F. I.

Williams

R. D.

Muscarella

Melvin

W. S.

Schirmer

W. J.

Oxythiamine and Dehydroepiandrosterone Inhibit the Nonoxidative Synthesis of Ribose and Tumor Cell Proliferation. Cancer. Res. 1997, 57, 4242–4248.

Gupte

S. A.

Glucose-6-phosphate Dehydrogenase: A Novel Therapeutic Target in Cardiovascular Diseases. Curr. Opin. Investig. Drugs 2008, 9, 993–1000.

Wang

Yuan

Chen

Hou

Overexpression of G6PD Is Associated with Poor Clinical Outcome in Gastric Cancer. Tumour. Biol. 2012, 33, 95–101.

Warburg

Wind

Negelein

The Metabolism of Tumors in the Body. J. Gen. Physiol. 1927, 8, 519–530.

X. L.

Guppy

Cancer Metabolism: Facts, Fantasy, and Fiction. Biochem. Biophys. Res. Commun. 2004, 313, 459–465.

Burt

B. M.

Humm

J. L.

Kooby

D. A.

Squire

O. D.

Mastorides

Larson

S. M.

Fong

Using Positron Emission Tomography with [(18)F]FDG to Predict Tumor Behavior in Experimental Colorectal Cancer. Neoplasia 2001, 3, 189–195.

Boros

L. G.

Bassilian

Lim

Lee

W. N.

Genistein Inhibits Nonoxidative Ribose Synthesis in MIA Pancreatic Adenocarcinoma Cells: A New Mechanism of Controlling Tumor Growth. Pancreas 2001, 22, 1–7.

10.

Richardson

A. D.

Yang

Osterman

Smith

J. W.

Central Carbon Metabolism in the Progression of Mammary Carcinoma. Breast Cancer Res. Treat. 2008, 110, 297–307.

11.

Bokun

Bakotin

Milasinovic

Semiquantitative Cytochemical Estimation of Glucose-6-phosphate Dehydrogenase Activity in Benign Diseases and Carcinoma of the Breast. Acta Cytol. 1987, 31, 249–252.

12.

Zhu

Tang

Yang

Tong

A New G6PD Knockdown Tumor-Cell Line with Reduced Proliferation and Increased Susceptibility to Oxidative Stress. Cancer Biother. Radiopharm. 2009, 24, 81–90.

13.

Jiang

Wang

Mancuso

Gao

Yang

p53 Regulates Biosynthesis through Direct Inactivation of Glucose-6-phosphate Dehydrogenase. Nat. Cell Biol. 2011, 13, 310–316.

14.

Kohler

Barrach

Neubert

Inhibition of NADP Dependent Oxidoreductases by the 6-Aminonicotinamide Analogue of NADP. FEBS Lett. 1970, 6, 225–228.

15.

Lange

Proft

E. R.

Inhibition of the 6-Phosphogluconate Dehydrogenase in the Rat Kidney by 6-Aminonicotinamide. Naunyn Schmiedebergs Arch. Pharmakol. 1970, 267, 177–180.

16.

H. Y.

Cheng

M. L.

Chiu

H. Y.

Weng

S. F.

Chiu

D. T.

Dehydroepiandrosterone Induces Growth Arrest of Hepatoma Cells via Alteration of Mitochondrial Gene Expression and Function. Int. J. Oncol. 2008, 33, 969–977.

17.

Miller

F. R.

Xenograft Models of Premalignant Breast Disease. J. Mammary Gland Biol. Neoplasia 2000, 5, 379–391.

18.

Jortzik

Mailu

B. M.

Preuss

Fischer

Bode

Rahlfs

Becker

Glucose-6-Phosphate Dehydrogenase-6-Phosphogluconolactonase: A Unique Bifunctional Enzyme from

Plasmodium falciparum. Biochem. J. 2011, 436, 641–650.

19.

Preuss

Hedrick

Sergienko

Pinkerton

Mangravita-Novo

Smith

Marx

Fischer

Jortzik

Rahlfs

Becker

Bode

High-Throughput Screening for Small-Molecule Inhibitors of Plasmodium falciparum Glucose-6-phosphate Dehydrogenase 6-Phosphogluconolactonase. J. Biomol. Screen. 2012, 17, 738–751.

20.

Cornish-Bowden

: Reversible inhibition and activation. In : Fundamentals of Enzyme Kinetics. London: Wiley-Blackwell, 2004.

21.

De Jong

D. W.

Woodlief

W. G.

Fluorimetric Assay of Tobacco Leaf Dehydrogenases with Resazurin. Biochim. Biophys. Acta 1977, 484, 249–259.

22.

Ziboh

V. A.

Dreize

M. A.

Hsia

S. L.

Inhibition of Lipid Synthesis and Glucose-6-phosphate Dehydrogenase in Rat Skin by Dehydroepiandrosterone. J. Lipid Res. 1970, 11, 346–354.

23.

Soule

H. D.

Maloney

T. M.

Wolman

S. R.

Peterson

W. D.

Brenz

McGrath

C. M.

Russo

Pauley

R. J.

Jones

R. F.

Brooks

S. C.

Isolation and Characterization of a Spontaneously Immortalized Human Breast Epithelial Cell Line, MCF-10. Cancer Res. 1990, 50, 6075–6086.

24.

Dawson

P. J.

Wolman

S. R.

Tait

Heppner

G. H.

Miller

F. R.

MCF10AT: A Model for the Evolution of Cancer from Proliferative Breast Disease. Am. J. Pathol. 1996, 148, 313–319.

25.

Beutler

Red Cell Metabolism: A Manual of Biochemical Methods. New York: Grune and Stratton; 1984.

26.

Birke

Kim

H. W.

Periclou

Schorsch

Grouse

Craney

Kinetics of Human Erythrocyte Glucose-6-Phosphate Dehydrogenase Dimers. Biochim. Biophys. Acta 1989, 999, 243–247.

27.

Wang

X. T.

S. W.

Lam

V. M.

Engel

P. C.

Recombinant Human Glucose-6-Phosphate Dehydrogenase.:Evidence for a Rapid-Equilibrium Random-Order Mechanism. Eur. J. Biochem. 2002, 269, 3417–3424.

28.

Iversen

P. W.

Eastwood

B. J.

Sittampalam

G. S.

Cox

K. L.

A Comparison of Assay Performance Measures in Screening Assays: Signal Window, Z′ Factor, and Assay Variability Ratio. J. Biomol. Screen. 2006, 11, 247–252.

29.

Inglese

Johnson

R. L.

Simeonov

Xia

Zheng

Austin

C. P.

Auld

D. S.

High-Throughput Screening Assays for the Identification of Chemical Probes. Nat. Chem. Biol. 2007, 3, 466–479.

30.

Tsutsui

E. A.

Marks

P. A.

Reich

Effect of Dehydroepiandrosterone on Glucose 6-Phosphate Dehydrogenase Activity and Reduced Triphosphopyridine Nucleotide Formation in Adrenal Tissue. J. Biol. Chem. 1962, 237, 3009–3013.

31.

Levy

H. R.

The Interaction of Mammary Glucose 6-Phosphate Dehydrogenase with Pyridine Nucleotides and 3beta-hydroxyandrost-5-en-17-one. J. Biol. Chem. 1963, 238, 775–784.

32.

Oertel

G. W.

Rebelein

Effects of Dehydroepiandrosterone and Its Conjugates upon the Activity of Glucose-6-Phosphate Dehydrogenase in Human Erythrocytes. Biochim. Biophys. Acta 1969, 184, 459–460.

33.

Hamilton

N. M.

Dawson

Fairweather

E. E.

Hamilton

N. S.

Hitchin

J. R.

James

D. I.

Jones

S. D.

Jordan

A. M.

Lyons

A. J.

Small

H. F.

Thomson

G. J.

Waddell

I. D.

Ogilvie

D. J.

Novel Steroid Inhibitors of Glucose-6-Phosphate Dehydrogenase. J. Med. Chem. 2012, 55, 4431–4445.

34.

S. W.

Gover

Lam

V. M.

Adams

M. J.

Human Glucose-6-Phosphate Dehydrogenase: The Crystal Structure Reveals a Structural NADP(+) Molecule and Provides Insights into Enzyme Deficiency. Structure 2000, 8, 293–303.

35.

Clarke

J. L.

Sodeinde

Mason

P. J.

A Unique Insertion in Plasmodium berghei Glucose-6-Phosphate Dehydrogenase-6-Phosphogluconolactonase: Evolutionary and Functional Studies. Mol. Biochem. Parasitol. 2003, 127, 1–8.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.48 MB

Identification and Characterization of Novel Human Glucose-6-Phosphate Dehydrogenase Inhibitors

Abstract

Keywords

Introduction

Materials and Methods

G6PD Overexpression and Purification

Assay Reagents

Assay Development

Compound Libraries for HTS

Primary and Secondary Screening

Orthogonal Assay

Mechanistic Characterization

Time Dependency and Reversibility

Cell Viability Assay for MCF10-A and MCF10-AT1 Cells

Solubility and Permeability Analysis

Results

Assay Development

Screening of Compound Libraries

Orthogonal Assay

Mechanistic Characterization, Time Dependency, and Reversibility

Effect of hG6PD Inhibitors on Cancer Cells

Discussion

Assay Development and Screening

Orthogonal Assay and Characterization of Inhibitors

Effects of hG6PD Inhibitors on MCF-10 Cells

Footnotes

Acknowledgements

Declaration of Conflicting Interests

Funding

References

Supplementary Material