Skip to main content
NCBI home page
ACS Bio & Med Chem Au logoLink to ACS Bio & Med Chem Au
. 2022 Jan 20;2(2):161–170. doi: 10.1021/acsbiomedchemau.1c00062

Identification of Natural Product Sulfuretin Derivatives as Inhibitors for the Endoplasmic Reticulum Redox Protein ERO1α

Brennan D Johnson †, Sridhar Reddy Kaulagari ∥, Wei-Chih Chen †, Karen Hayes ∥, Werner J Geldenhuys ‡,§, Lori A Hazlehurst †,‡,*
PMCID: PMC9312093  NIHMSID: NIHMS1819107  PMID: 35892127

Abstract

graphic file with name bg1c00062_0009.jpg

The flavin adenine dinucleotide containing endoplasmic reticulum oxidoreductase-1 α (ERO1α) catalyzes the de novo disulfide bond formation of secretory and transmembrane proteins and contributes toward proper protein folding. Recently, increased ERO1α expression has been shown to contribute to increased tumor growth and metastasis in multiple cancer types. In this report, we sought to define novel chemical space for targeting ERO1α function. Using the previously reported ERO1α inhibitor compound, EN-460, as a benchmark pharmacological tool, we were able to identify a sulfuretin derivative, T151742, which was approximately 2-fold more potent using a recombinant enzyme assay system (IC50 = 8.27 ± 2.33 μM) compared to EN-460 (IC50= 16.46 ± 3.47 μM). Additionally, T151742 (IC50 = 16.04 μM) was slightly more sensitive than EN-460 (IC50= 19.35μM), using an MTT assay as an end point. Utilizing a cellular thermal shift assay (CETSA), we determined that the sulfuretin derivative T151742 demonstrated isozyme specificity for ERO1α as compared to that for ERO1β and showed no detectable binding to the FAD-containing enzyme LSD-1. T151742 retained activity in PC-9 cells in a clonogenicity assay, while EN-460 was devoid of activity. Furthermore, the activity of T151742 inhibition of clonogenicity was dependent on ERO1α expression as CRISPR-edited PC-9 cells were resistant to treatment with T151742. In summary, we identified a new scaffold that shows specificity for ERO1α compared to that for the closely related paralog ERO1β or the FAD-containing enzyme LSD-1 that can be used as a tool compound for the inhibition of ERO1α to allow for pharmacological validation of the role of ERO1α in cancer.

Keywords: redox, PDI, FAD, flavoenzyme, ERO1α, lung cancer

Introduction

Lung cancer remains the most common cancer death worldwide at approximately 1.6 million deaths occurring each year. Non-small lung cancer treatment regimens have had an increasing clinical benefit over the last two decades with the discovery of molecular drivers and development of tyrosine kinase inhibitors and immunotherapies.1 However, the emergence of resistance to targeted and immunotherapies remains a clinical obstacle for increasing overall survival and improving patient outcomes.2−5 Recently, the increased expression of endoplasmic reticulum oxidoreductase-1α (ERO1α) has been identified as a key player in promoting metastatic burden, proliferation, and immune escape, and the increased expression of ERO1α correlates with worse clinical outcomes in multiple tumor indications.6−11 ERO1α belongs to the flavoenzyme family of enzymes. Flavoenzymes are enzymes that meet two criteria: the enzyme (1) must be an oxidoreductase and (2) must contain flavin adenine dinucleotide (FAD) as a prosthetic group.12 Examples of flavoenzymes include ERO1α, ERO1β, LSD-1, MAO-A, and MAO-B. The endoplasmic reticulum (ER) resident enzyme ERO1α is known to form de novo disulfide bridges and aid in the folding of transmembrane and secretory proteins in conjunction with protein disulfide isomerase (PDI).13−16 ERO1α molecular interactions and oxidation of PDI and protein folding have been well established.17,18 While in its reduced form, ERO1α utilizes molecular oxygen, becoming oxidized, while reducing FAD through electron transfer and producing hydrogen peroxide as an intermediate. This process is important for disulfide bond formation that occurs during the protein folding process. Despite the cell being able to deal with minimal peroxide intermediates, it is also tightly regulated to avoid futile oxidative cycles from occurring in the ER and therefore aids in maintaining homeostasis within the ER.19 ERO1α and PDI are the two main players in oxidative reactions occurring within the ER, both of which are highly conserved from yeast to mammals.13−16 Upon homeostatic imbalances within the endoplasmic reticulum, the unfolded protein response is generally activated, resulting in one of two possible outcomes: (1) the unfolded protein response can return endoplasmic stress to homeostatic levels or (2) if homeostasis cannot be established, then apoptosis can be activated.20,21 Thus, the unfolded protein response is a cellular response mechanism for resolving ER stress, which can lead to survival or death depending on the context. Due to the critical enzymatic function, it is not surprising that ERO1α is tightly regulated via post-translational phosphorylation22 and through regulatory disulfide bridging. Cysteine bridging between Cys94 and Cys99 allows for nucleophilic attack on bridged cysteines of Cys394–Cys397, allowing for Cys397 to undergo a nucleophilic attack onto bound FAD.23 This reaction process utilizes molecular oxygen as a terminal electron acceptor to form hydrogen peroxide as an intermediate. Cancer cells typically have increased ER stress compared to normal cells and occur more so in secretory tumors such as lung, breast, multiple myeloma, and pancreatic cancers, but other tumors can exhibit high levels of ER stress following exposure to chemotherapy or hypoxia.24−27 ERO1α plays a crucial role in ER homeostasis, via its interaction with PDI to aid in protein folding, and has emerged as a key player in aiding in the tolerance of cancer cells to ER stress.28 Despite the emerging evidence that ERO1α is an attractive target, few chemical probes are available to allow for further validation of the target. The challenge, in part, is due to the conserved nature of the FAD binding domain across enzymes. Developing specific inhibitors or more potent inhibitors to a specific flavoenzyme is critical for pharmacological validation of the target. Despite this challenge, there are flavoenzyme-targeting agents that are more specific such as the LSD-1 inhibitor IV. LSD-1 inhibitor IV forms a covalent adduct with FAD within LSD-1 and is approximately 6-fold more selective toward LSD-1 compared to MAO-A and MAO-B and was originally described and synthesized by Neelamegam et al.29 The first reported ERO1α inhibitor was identified from a high-throughput screen. The hit molecule called EN-460 was shown to inhibit ERO1α activation and prevent reoxidation.30 However, recently, our laboratory demonstrated that EN-460 inhibits multiple flavoenzyme family members including LSD-1, MAO-A, and MAO-B6 and thus EN-460 is not an ideal compound for pharmacological credentialing of ERO1α. Due to this gap in knowledge, we sought to identify additional chemical space for targeting ERO1α.

The aurone class of compounds with sulfuretin as a primary example is a major group of flavonoid small molecules isolated from the heartwood of Rhus verniciflua and can be used to reduce oxidative stress.31 Chalones are another class of flavonoids that exhibit a close biochemical relationship with the aurones.32 Flavonoid compounds belong to the low-molecular-weight class of phenolic compounds that are widely distributed in the plant kingdom.33 Flavonoid compounds contain specific structural features being the 2-phenyl-benzol (α)pyrane, containing two benzene rings, linked by the heterocyclic pyrane ring.34 The main classes of flavonoids includes the aurones, chalcones, flavones, flavonones, isoflavones, anthrocyanidins, flavan-3-ols, flavans, flavan-3,4-diols, dihydrochalcones, and proanthocyanidins.35,36 Flavones are classified by the double bond at positions 2 and 3 with a ketone at position 4 of the C ring, while most flavones have a hydroxyl group occurring at position 5 of the A ring.33 In this paper, we sought to determine whether the aurone and/or chalcone family demonstrates activity toward ERO1α and to determine the specificity of the most potent analogue.

Results and Discussion

Determination of Inhibition of ERO1α by Sulfuretin Derivatives

Based on previous work that established an aurone chemical class of compounds as FAD-containing enzyme inhibitors,37 we evaluated sulfuretin analogues for ERO1α enzyme inhibition. The Amplex Red assay was used to determine if the sulfuretin derivatives were capable of inhibition of ERO1α in a recombinant assay system previously established by our laboratory.6 As shown in Figure 1, the sulfuretin derivatives inhibited ERO1α in a concentration-dependent manner. The most potent aurone analogue referred to as T151742 was found to be approximately 2-fold more potent compared to EN-460 (IC50 = 8.27 ± 2.33 μM), T151750 (IC50 = 11.34 ± 2.49 μM), T151688 (IC50 = 18.91 ± 5.47 μM), and EN-460 (IC50 = 16.46 ± 3.47 μM).

Figure 1.

Figure 1

Open in a new tab

T151742 is the most potent aurone and is 2-fold more potent compared to EN-460. Purified human ERO1α (0.0625 mg/mL) and purified human PDI (0.0625 mg/mL) were incubated with increasing concentrations of EN-460, T151742, T151688, or T151750. Each data point shown is the average ± S.D (N = 3 independent experiments and performed in triplicates). ****p < 0.0001 by two-way ANOVA.

Based on our initial studies shown in Figure 1, we sought to determine whether the 4-dimethylamino on the benzylidene ring was critical for the inhibition of ERO1α and synthesized a small set of compounds to determine the impact of substitution the benzylidene ring has on enzyme activity. The compound structures are shown in Figure 2 and Table 1 and were tested at 10 μM, which is the approximate IC50 of T151742.

Figure 2.

Figure 2

Open in a new tab

Structures of the compounds evaluated for ERO1α enzyme activity.

Table 1. Inhibition of ERO1α Enzyme Activity, Reported as Percent Inhibition Compared to That of Vehicle Controla,b.

compound % inhibition (10 μM) classification
T151742 55.31 ± 0.98 aurone
T151688 40.46 ± 7.32 aurone
T151750 48.60 ± 4.68 aurone
SR-F-114 28.62 ± 5.57 aurone
SR-F-115 27.48 ± 1.82. aurone
SR-F-125 20.73 ± 4.92 aurone
SR-F-126 38.74 ± 1.56 aurone
SR-F-127 33.40 ± 2.72 aurone
SR-F-128 37.86 ± 2.42 chalcone
SR-F-129 26.75 ± 1.44 chalcone
SR-F-130 27.26 ± 3.16 chalcone
SR-F-131 24.16 ± 0.52 chalcone
Open in a new tab
a

The average of N = 3 ± S.D is shown.

b

%inhibition for each compound at 10 μM was determined by the recombinant enzyme assay as previously described.6

The sulfuretin derivatives were found to inhibit ERO1α, with the Aurone scaffolds demonstrating similar activity to the chalcone compounds (Table 1). With the small changes in activity profile, our findings suggest the 4-dimethylamino-moiety to be important for activity since none of the new aurones or chalcones were more potent than T151742. Additionally, the para-hydroxy on the benzofuran is important for activity, as loss of this functional group led to a decrease in inhibition activity. Comparing the electrostatic surfaces of EN-460 to the compounds tested here, we found that interestingly, T151742 shared a similar surface pattern as EN-460, while other compounds such as SR-F-126, SR-F-128, and T151750 shared some overlap in electrostatics but were less pronounced and concentrated in certain areas, as compared to EN-460 (Figure 3). This, in part, could explain the differences in the activity profile of the compounds compared to T151742 and EN-460.

Figure 3.

Figure 3

Open in a new tab

Electrostatic surfaces of compounds. Surface-electrostatic potential differences are noticeable on the aurone and chalcone derivatives as compared to the ERO1α inhibitor EN-460, with T151742 showing a stronger correlation. Blue shows positive charge areas, and red shows negative charge areas.

Since the newly identified sulfuretin derivative T151742 was the most potent of the compounds tested, we evaluated its activity profile in classical enzyme kinetics assays. Currently, the standard model of FAD-containing enzymes is that the inhibitors available inhibit flavoenzyme interaction directly with FAD substrate pocket and are competitive inhibitors, with a select set of compounds designed as covalent inhibitors.38 To validate the mechanism of inhibition T151742 had on ERO1α, varying PDI concentrations (the substrate for ERO1α) and varying concentrations of T151742 were used and the double reciprocal of this plot was taken to produce the Lineweaver–Burk plot (see Figure 4A,B). As shown in Figure 4A,B, KM is increased, while VMax is reduced with the increasing concentration of T151742. This is observed in the double-reciprocal plot in Figure 4B, as the slope of the lines increases in a concentration-dependent manner, and both x- and y-intercepts change in a concentration-dependent manner as well. These findings suggest that T151742 was acting upon ERO1α as a noncompetitive (mixed) inhibitor and was not binding directly to the active site.

Figure 4.

Figure 4

Open in a new tab

T151742 binds directly to ERO1α and acts as a noncompetitive (mixed) inhibitor. (A) Purified human ERO1α (0.0625 mg/mL), varying concentrations of purified human PDI (0.25 μM, 0.125 μM, 0.0625 μM, 0.03125 μM, and 0.5% DMSO), and varying concentrations of T151742 (100 μM, 75 μM, 50 μM, 25 μM, 10 μM, 0.1 μM, and 0.5% DMSO) were assayed using amplex red assay to quantify enzymatic activity. (B) The Lineweaver–Burk plot produced from taking the double reciprocal (1/[PDI]) vs (1/[T151742]). C) Surface plasma resonance demonstrates the binding of T151742 directly to ERO1α (Kd =31.4 μM).

Surface plasma resonance (SPR) was utilized as an orthogonal assay to confirm direct binding between T151742 and ERO1α. For the SPR assay, recombinant ERO1α was cross-linked to a C5 chip and varying concentrations of ligand T151742 were evaluated for the detection of changes in refractive index due to change in mass on the chip when the ligand is bound to the protein. As shown in Figure 4C, T15742 binds to ERO1α in a concentration-dependent manner with a Kd =31.4 μM. The relatively high Kd value is, in part, due to the fast off-rate. Taken together, our data indicate that the sulfuretin derivative T151742 (i) binds directly to ERO1α in a recombinant system, (ii) is the most potent analogue in this small series at inhibiting ERO1α and 2-fold more potent than EN-460, (iii) acts as a noncompetitive (mixed) inhibitor, and (iv) a remaining potential liability is the fast off-rate of the compound.

Several of the flavoenzyme family of proteins contain a substrate pocket in which specific substrates are metabolized in the presence of FAD, e.g., dopamine with MAO-A.39 This has led to the development of irreversible covalent inhibitors of the cofactor FAD such as tranylcypromine.40 The development of the covalent inhibition led to the inclusion of the active moiety propargylamine, leading to derivatives deprenyl/selegiline and rasagiline, which are used clinically to treat Parkinson’s disease.41 Furthermore, the LSD-1 inhibitor IV, RN-1 was among the first lysine demethylase inhibitors discovered, where LSD-1 is a FAD-containing enzyme.29 Tranylcypromine forms a covalent modification directly with FAD.40 LSD-1 inhibitor IV, RN-1, contains the classical tranylcypromine template that allows for the covalent modification of targeted enzymes containing a FAD moiety. We determined that the LSD-1 inhibitor IV and tranylcypromine both known to covalently cross-link FAD did not inhibit ERO1α activity (see Supporting Information Figure 1 for concentration–response). Conversely, T151742 was shown not to inhibit LSD-1 enzymatic activity (see Supporting Information Figure 2 for concentration–response) These data suggest that the FAD-containing ring within ERO1α is concealed within the enzyme and not easily targeted by small molecules. Based on the crystal structure of ERO1α (3ahq.pdb), the FAD is contained in a pocket close to the redox-active cysteines, fully occupying the pocket space. Taken together, these findings suggest that the putative binding pocket of the ERO1α inhibitors is not in the FAD pocket but is likely allosteric to the FAD binding pocket.

Based on the experimental data, we explored binding pockets allosteric to the FAD binding pocket of ERO1α. The Site Finder algorithm of MOE 2020 was used to identify the putative pocket in proximity to the FAD binding pocket, as well as the reactive cysteines CYS394 and CYS397 (Figure 5A). We used the reduced form of the crystal, based on our enzyme assay, which still produced hydrogen peroxide in the presence of the inhibitors. As shown in Figure 5A, T151742 was predicted to bind by hydrogen bonding to GLN93 and GLY138, with the 3-ketone oriented toward the reactive cysteines CYS394 and CYS397 (<4 A). Figure 5B shows that the loss of the 6-hydroxyl moiety, which yielded T151688, leads to a loss of the GLN93 binding interaction, potentially contributing to differences in activity when comparing T151742 with T151688. Predicted affinity using YASARA Autodock Vina corroborated these findings, with a predicted inhibition of 2.31 μM for T151742 and 13.67 μM for T151688. Lastly, Figure 5C shows that the loss of the 4-dimethylamino in T151750 leads to the loss of the 3-ketone hydrogen bond with GLY138 as compared to T151742. From these docking studies, we propose that the ERO1α inhibitors are interacting with ERO1α via an allosteric binding site located in proximity to the redox-active CYS394–CYS397 pair, preventing the oxidation of the disulfide bond.

Figure 5.

Figure 5

Open in a new tab

Docking studies of ERO1α inhibitors. The crystal structure of the active ERO1α (3ahq.pdb) with the cofactor FAD was used for binding studies. (A) T151742 is shown docked with the optimal hydrogen bonding of the 6-hydroxyl to GLN93 and the 3-ketone to GLY138; (B) loss of the 6-hydroxyl in T151688 led to a decrease in binding; and (C) loss of the 4-dimethylamino leads to reposing with GLY138 bond not forming compared to that of T151742.

Determining the Selectivity of T151742 Using In Vitro Cell Culture Models

The ERO1α inhibitor EN-460 found in a high-throughput screen from natural products was shown to inhibit the reduced form of ERO1α.30 However, this compound has a liability as a tool to pharmacologically validate ERO1α as a target as it has been shown to inhibit other flavoenzyme family members including LSD-1, MAO-A, and MAO-B.6,30 It was previously shown that T151742 showed activity against MAO-A and MAO-B, albeit was not the most potent Aurone member.37 To date, the activity of T151742 toward other flavoenzymes has not been identified at this time. We utilized the cellular thermal shift assay (CETSA) to determine target engagement of T151742 in vitro using the non-small cell lung cancer (NSCLC) cell line PC-9. The EGFR-driven PC-9 cell line was chosen as new treatment strategies are needed in EGFR-driven NSCLC as most patients relapse and become resistant to current targeted therapies with the development of further mutations such as T790M and C797S, as well as the emergence of EGFR-independent tumors.42,43 CETSA utilizes thermodynamic principles that have been well described and upon binding of ligands to a protein, the ligand-bound protein becomes more thermostable with respect to the temperature required to denature the protein or expose hydrophobic regions of the protein leading to aggregation and insolubility of the protein.44 All proteins have an aggregation temperature, which is inherent to their intracellular thermostability. Upon the addition of a ligand binding, the thermostability increases. First, we determined the aggregation temperature (Tagg) of ERO1α, ERO1β, and LSD-1, which are all FAD-dependent enzymes and potential in vitro targets of T151742. Tagg was defined as the point at which 10% of protein or less is remaining when compared to the room-temperature control of the target protein by Western blot analysis. As shown in Figure 6A, CETSA analysis determined the Tagg for the unbound target protein (ERO1α Tagg = 54 °C; ERO1β; Tagg = 54 °C; and LSD-1 Tagg = 52 °C). The values reported are averages from three independent experiments. When T151742 was added to the culture media at a concentration of 100 μM for 1 h at 37 °C, the Tagg for ERO1α shifted approximately ∼6° to >60 °C (see Figure 6A). In contrast, the incubation of cells with 100 μM T151742 did not change the Tagg for LSD-1 or ERO1β. These results indicate that T151742 bound and stabilized ERO1α compared to the flavoenzymes ERO1β and LSD-1. Although ERO1α and ERO1β have highly conserved sequences and functions, it appears their tertiary structures are more diverse than previously thought and specific inhibitors can be developed to distinguish the activity of inhibiting one enzyme using in vitro models. The Tagg for LSD-1 also remained unchanged upon the addition of T151742 and aided in confirming that T151742 shows more specificity toward ERO1α as compared to other members of the flavoenzyme family, with the exception of MAO-A, which was previously reported.37 The results of the CETSA assay combined with the in vitro recombinant assay demonstrated that T151742 does not inhibit the enzymatic activity of LSD-1 (see Supporting Information Figure 2 for concentration–response).

Figure 6.

Figure 6

Open in a new tab

CETSA assay validates target engagement of T151742 toward ERO1α, and proliferation is partially dependent on ERO1α expression by MTT assay in PC-9 cells. (A) PC-9 cells were subjected to the cellular thermal shift assay in the absence or presence of T151742. Upon the addition of T151742, Tagg was quantified for ERO1α, ERO1β, and LSD-1. Quantification of the average of n=3 independent experiments based on Western blot analysis for ERO1α, ERO1β, and LSD-1. The dashed line represents the cutoff point at 10%, which was used to define our aggregation temperature (Tagg) of individual proteins. (B) MTT viability assay comparing potency between ERO1α inhibitors EN-460, T151742, and T151742 derivative SR-F-114, revealing that SR-F-114 is about 2-fold less effective in vitro versus T151742 and EN-460 in the PC-9 non-small cell lung cancer cell line. Combined n = 3 independent experiments performed in quadruplicates ***p < 0.001 by one-way ANOVA. C) Validation of ERO1α knockout in the PC-9 cell line utilizing CRISPR. (D) Two CRISPR-deleted ERO1α clones were subjected to the treatment of T151742 as well as the nontargeted control PC-9 cell line. Cellular viability was determined by MTT assay. n = 3 independent experiments performed in quadruplicates. ****p < 0.0001; by two-way ANOVA.

Evaluating the Efficacy of T151742 in the PC-9 Lung Cancer Cell Line

We utilized an MTT assay to determine the potency of the compounds for inhibiting cell growth. As shown in Figure 6B, we compared T151742, EN-460, and SR-F-114. Consistent with the enzymatic assay, T151742 was the most potent, while SR-F-114 was approximately 2-fold less active (T151742 IC50 = 14.05 ± 3.99 μM, EN-460 IC50= 16.62 ± 4.34 μM, and SR-F-114 IC50= 32.36 ± 6.01). We next asked whether T151742 induced the inhibition of cell growth was dependent on ERO1α expression. As shown in Figure 6C, two ERO1α deleted clones in the PC-9 cell line were isolated. Using an MTT assay to determine the overall cellular viability, a nontargeted control (scramble) and both ERO1α knockout clones were subjected to treatment with T151742 for 72 h (see Figure 6D). PC-9 ERO1α-deleted cells were approximately 2–3-fold more resistant to T151742 treatment compared to the control cell line (PC-9 WT IC50 = 14.13 μM ± 1.15, PC-9 ERO1α KO Clone 1 IC50 = 32.23 ± 1.09 μM and PC-9 ERO1α KO Clone 2 IC50 = 41.54 ± 1.08 μM). We were also able to determine that the IC50 values in myeloma cell lines were 7–9 μM, indicating that myeloma cells demonstrate increased sensitivity to the inhibition of ERO1α. These data suggest that ERO1α may be important to further credential as a target for the treatment of multiple myeloma. Importantly, in the normal lung epithelial cell line Beas-2B, the IC50 value was found to be >100 μM, suggesting that cancer cells are more dependent on ERO1α compared to normal cells (see Supporting Information Table 1 for IC50 values of all cell lines tested). Though only slight differences were observed between T151742 and EN-460, we tested the effects of these compounds in a clonogenic assay, which is a better test for the inhibition of self-renewal. Utilizing soft agar assays, we were able to determine that T151742 was significantly more potent at reducing colony number in a concentration-dependent manner in comparison to EN-460 (Figure 7A,B). PC-9 ERO1α knockout clones were found to be completely resistant to T151742 using a clonogenic assay as an end point (Figure 7C,D). Together, these data suggest that the activity of T151742 is dependent upon ERO1α expression when using clonogenicity as an end point.

Figure 7.

Figure 7

Open in a new tab

Soft agar clonogenic assays demonstrate that T151742 is more potent compared to EN-460 and ERO1α knockout clones are resistant to T151742 treatment. (A, B) 1000 cells/well were plated in a 12-well plate in triplicate with either 0.5% DMSO, T151742, or EN-460 and allowed to form colonies for 14 days. Colonies were subsequently stained with 0.01% Crystal Violet, imaged, and quantified using ImageJ. (C, D) 1500 cells/well were plated in a 12-well in triplicate with either 0.5% DMSO or T151742 and allowed to form colonies for 14 days. Colonies were stained with 0.01% Crystal Violet, imaged, and quantified using ImageJ.

Conclusions

In conclusion, we identified T151742 as the most potent aurone analogue, which demonstrated activity against ERO1α, with T151742 having activity against ERO1α in a recombinant assay system and binds the target using in vitro assays. Aurone and chalcone molecules synthesized showed similar activity, but no derivatives made were found to be more potent than T151742. We were also able to demonstrate that T151742 activity was completely dependent on ERO1α expression using soft agar clonogenic assay as an end point. Additionally, our data indicate that clonogenic assays may be a better end point compared to proliferation or a death assay when screening compounds against ERO1α. The CETSA assay indicates that T151742 has specificity toward ERO1α over other flavoenzymes, and these data indicate that this assay can be used to validate ERO1α target coverage using in vitro assays. Further studies are required to determine whether the CETSA assay can be used to determine the target coverage of ERO1α using in vivo models of cancer. Lastly, we were able to determine through structure–activity relationship built upon here that both the hydroxyl group at position 5 on the A ring and the tertiary amine group on the B ring are important for activity against ERO1α, as all derivatives with changes at these positions lost efficacy toward ERO1α in our recombinant system. Ultimately, our results suggest that the aurone scaffold can be exploited to further develop novel specific inhibitors to pharmacologically credential the target ERO1α using in vitro models of cancer. Future studies are warranted to determine whether T151742 can be used to credential ERO1α using in vivo models of cancer. Studies using a clonogenic assay comparing EN-460 and T151742 suggest that specificity may be more important for targeting self-renewal or clonogenic growth in soft agar. However, further studies will be required to determine whether a promiscuous inhibitor targeting multiple FAD-containing enzymes will be more effective compared to the specific targeting of ERO1α for the treatment of cancer.

Materials and Methods

Compounds used in the assays were obtained from commercial sources. EN-460 (Sigma-Aldrich, CAT# 328501), LSD-1 Inhibitor IV (Sigma-Aldrich, CAT# 489479), and tranylcypromine (Sigma-Aldrich, CAT # 616431) were purchased. T151742 and all derivatives that were used in experiments were made synthetically, and experimental procedures on synthetic strategy, 1H, 13C, 13C–DEPT, and 19F NMR spectra with HPLC traces of compounds, high-resolution mass spectrometry (ESI-TOF), yield, melting points, and purity of compounds tested are attached in the Supporting Information.

Protein Expression, Purification, and Enzymatic Assays

Human ERO1α (a.a. 22-468) containing hyperactivating triple mutation (C104A, C131A, and C166A19) was purified, as previously described.6 Human PDI (a.a. 18-479) was synthesized and purified, as previously described.6 Quantification of the enzymatic activity of ERO1α was performed using the amplex red, as previously described.6

Cell Lines

PC-9 cell line used was obtained from Sigma-Aldrich and grown in RPMI-1640 medium with 10% fetal bovine serum and 1% penicillin/streptomycin. HCC4006, U266, MM1.s, and Beas-2B cell lines were obtained from ATCC and cultured in either RPMI-1640 or DMEM with 10% fetal bovine serum and 1% penicillin/streptomycin. All cell lines were also checked for mycoplasma every 6 months and subjected to short tandem repeat (STR) analysis for cell line validation.

CRISPR Cas9

CRISPR Guide RNA 2 (TGTGAACAAGCTGAACGACT) targeting ERO1α and CRISPR Guide RNA 4 (AATTAAATCTGCGAGCTACA) were purchased from Genscript in the pLentiCRISPR V2 plasmid containing puromycin/ampicillin selection marker. Lentivirus was made from HEK 293T cells by cotransfection with Guide RNA of interest, pVSVG, and pPAX2 plasmids. Lentivirus was collected after 48 h and concentrated using 40% PEG8000. Concentrated virus was used to infect PC-9 cells. Cells were placed under puromycin selection after 72 h and cloned out using limiting dilutions.

Lineweaver–Burk

Varying concentrations of PDI (0–0.25 mg/mL) and T151742 were assayed using the Amplex red kit, as described above. The reaction was incubated for 30 min at 37 °C. The plate was then read on the Biotek Cytation 5 at 530/590 nm.

MTT

Cells were plated at 10,000 cells/well in quadruplets in a 96-well plate and allowed to attach for 24 h. EN-460 or T151742 was added at increasing concentrations (100, 50, 25, 12.5, 6.25, 3.125, 1.5625 μM) with DMSO as control and allowed to incubate at 37 °C and 5% CO2 for an additional 72 h. After 72 h, MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) dye (2 mg/mL) was added to each well and incubated at 37 °C and 5% CO2 for 4 h. Dye and media were aspirated, and DMSO was added to the plate and placed on a rocker at room temperature for 10 min. The plate was then read at 570 nm. Treated wells were all compared to DMSO control wells to calculate percent growth inhibition, and the IC50 was calculated using Graphpad (log (inhibitor) vs growth inhibition).

Soft Agar Clonogenic

A base layer consisting of 1 mL of 1% agar in RPMI-1640 was set and allowed to cool to room temperature. A top layer was then added containing 500 μL of 0.5% agar in RPMI-1640 and cells at the indicated concentration on the figure legend. Once agar had solidified at 37 °C, 500 μL of RPMI-1640 culture media was added to the top. Media was changed every 2-3 days. Plates were incubated for 14 days and were then stained with 0.01% Crystal Violet and imaged. Images were then quantified using ImageJ.

Cellular Thermal Shift Assay

PC-9 cell line (9 million cells) was incubated with either 0.5% DMSO control or T151742 at 100 μM for 1 hr at 37 °C. Cells were aliquoted at 1 million cells per pcr tube in PBS and heated in a Bio-Rad T100 thermal cycler at indicated temperatures (RT, 46, 48, 50, 52, 54, 56, 58, and 60 °C) for 3 min, and aggregation temperature was determined by Western blot analysis. The CETSA protocol was adapted from Jafari et al.45 ERO1α antibody (Santa Cruz Cat# sc-365526), ERO1β (Proteintech Cat #11261-2-AP), and LSD-1 (Cell Signaling Cat#2139S) were obtained. Secondary antibodies used for detection include antirabbit IgG (Jackson Laboratories Cat# 111-035-003) and antimouse IgG (Jackson Laboratories Cat# 115-035-003). The substrate used for Western blot detection was ECL-A/ECL-B (Thermo Fisher CAT# 32209). All Western blot images and analyses were completed using the Amersham imager 680.

Biacore Analysis

Affinity analysis was carried out using a Biacore T200 instrument (GE Healthcare Life Sciences and analysis was performed at Creative Biolabs, Shirley, NY). Briefly, ERO1α protein was directly immobilized on the chip (Serie S-type CM5) using an amine coupling kit (GE Healthcare Life Sciences). The protein was diluted into 50 μg/mL with the immobilization buffer. The CM5 sensor surface was activated using 400 mM EDC and 100 mM NHS, injected at a flow rate of 10 μL/min, with a contact time of 420 s. ERO1α (50 μg/mL) was injected into FC2 at a flow rate of 10 μL/min. The amount of ERO1α immobilized was about 10 000 RU. Then, cholamine was injected for blocking at a flow rate of 10 μL/min, with a contact time of 420 s. T151742 was serially diluted with the running buffer to give a concentration of 100, 50, 25, 12.5, 6.25, 3.125, 1.563, and 0 μM. Data analysis was performed on the Biacore T200 computer and with Biacore T200 evaluation software.

Acknowledgments

This work was supported by the NIH (NCI grants LH RO1 CA95727 and R44-CA22154 LH), Additional funding was obtained from the George M & M Freda B Vance Medical Scholarship Fellowship (WC), the Alexander Bland Osborn Center Endowment, and the Allen Lung Cancer Endowment. Imaging experiments were performed in the WVU Microscope imaging Facilities, which have been supported by the WVU Cancer Institute, The WVU HSC Office of Research and Graduate Education, and NIH Grants P20GM121322, P20GM103434, and U54HM104942. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Foundations.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsbiomedchemau.1c00062.

  • Additional experimental details and methods for the synthesis are included, as well as the 1H and 13C NMR spectra and HPLC chromatograms; additional experimental data presented demonstrates that covalent modifiers of FAD do not inhibit ERO1α enzymatic activity and T157142 does not inhibit LSD-1 enzymatic activity (PDF)

The authors declare the following competing financial interest(s): Lori Hazlehurst is a co-founder of Modulation Therapeutics.

Supplementary Material

bg1c00062_si_001.pdf (21.4MB, pdf)

References

  1. Herbst R. S.; Morgensztern D.; Boshoff C. The biology and management of non-small cell lung cancer. Nature 2018, 553, 446–454. 10.1038/nature25183. [DOI] [PubMed] [Google Scholar]
  2. Abbosh C.; Birkbak N. J.; Swanton C. Early stage NSCLC - challenges to implementing ctDNA-based screening and MRD detection. Nat. Rev. Clin. Oncol. 2018, 15, 577–586. 10.1038/s41571-018-0058-3. [DOI] [PubMed] [Google Scholar]
  3. Chaudhuri A. A.; Chabon J. J.; Lovejoy A. F.; Newman A. M.; Stehr H.; Azad T. D.; Khodadoust M. S.; Esfahani M. S.; Liu C. L.; Zhou L.; Scherer F.; Kurtz D. M.; Say C.; Carter J. N.; Merriott D. J.; Dudley J. C.; Binkley M. S.; Modlin L.; Padda S. K.; Gensheimer M. F.; West R. B.; Shrager J. B.; Neal J. W.; Wakelee H. A.; Loo B. W. Jr.; Alizadeh A. A.; Diehn M. Early Detection of Molecular Residual Disease in Localized Lung Cancer by Circulating Tumor DNA Profiling. Cancer Discovery 2017, 7, 1394–1403. 10.1158/2159-8290.CD-17-0716. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Nair R. R.; Tolentino J.; Hazlehurst L. A. The bone marrow microenvironment as a sanctuary for minimal residual disease in CML. Biochem. Pharmacol. 2010, 80, 602–612. 10.1016/j.bcp.2010.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Ootsuka S.; Asami S.; Sasaki T.; Yoshida Y.; Nemoto N.; Shichino H.; Chin M.; Mugishima H.; Suzuki T. Useful markers for detecting minimal residual disease in cases of neuroblastoma. Biol. Pharm. Bull. 2008, 31, 1071–1074. 10.1248/bpb.31.1071. [DOI] [PubMed] [Google Scholar]
  6. Hayes K. E.; Batsomboon P.; Chen W. C.; Johnson B. D.; Becker A.; Eschrich S.; Yang Y.; Robart A. R.; Dudley G. B.; Geldenhuys W. J.; Hazlehurst L. A. Inhibition of the FAD containing ER oxidoreductin 1 (Ero1) protein by EN-460 as a strategy for treatment of multiple myeloma. Bioorg. Med. Chem. 2019, 27, 1479–1488. 10.1016/j.bmc.2019.02.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Tanaka T.; Kajiwara T.; Torigoe T.; Okamoto Y.; Sato N.; Tamura Y. Cancer-associated oxidoreductase ERO1-alpha drives the production of tumor-promoting myeloid-derived suppressor cells via oxidative protein folding. J. Immunol. 2015, 194, 2004–2010. 10.4049/jimmunol.1402538. [DOI] [PubMed] [Google Scholar]
  8. Tanaka T.; Kutomi G.; Kajiwara T.; Kukita K.; Kochin V.; Kanaseki T.; Tsukahara T.; Hirohashi Y.; Torigoe T.; Okamoto Y.; Hirata K.; Sato N.; Tamura Y. Cancer-associated oxidoreductase ERO1-alpha drives the production of VEGF via oxidative protein folding and regulating the mRNA level. Br. J. Cancer 2016, 114, 1227–34. 10.1038/bjc.2016.105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Tanaka T.; Kutomi G.; Kajiwara T.; Kukita K.; Kochin V.; Kanaseki T.; Tsukahara T.; Hirohashi Y.; Torigoe T.; Okamoto Y.; Hirata K.; Sato N.; Tamura Y. Cancer-associated oxidoreductase ERO1-alpha promotes immune escape through up-regulation of PD-L1 in human breast cancer. Oncotarget 2017, 8, 24706–24718. 10.18632/oncotarget.14960. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Wang Y.; Zhang L.; He Z.; Deng J.; Zhang Z.; Liu L.; Ye W.; Liu S. Tunicamycin induces ER stress and inhibits tumorigenesis of head and neck cancer cells by inhibiting N-glycosylation. Am. J. Transl. Res. 2020, 12, 541–550. [PMC free article] [PubMed] [Google Scholar]
  11. Yang S.; Yang C.; Yu F.; Ding W.; Hu Y.; Cheng F.; Zhang F.; Guan B.; Wang X.; Lu L.; Rao J. Endoplasmic reticulum resident oxidase ERO1-Lalpha promotes hepatocellular carcinoma metastasis and angiogenesis through the S1PR1/STAT3/VEGF-A pathway. Cell Death Dis. 2018, 9, 1105. 10.1038/s41419-018-1134-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Neims A. H.; Hellerman L. Flavoenzyme catalysis. Annu Rev Biochem 1970, 39, 867–88. 10.1146/annurev.bi.39.070170.004251. [DOI] [PubMed] [Google Scholar]
  13. Cabibbo A.; Pagani M.; Fabbri M.; Rocchi M.; Farmery M. R.; Bulleid N. J.; Sitia R. ERO1-L, a human protein that favors disulfide bond formation in the endoplasmic reticulum. J. Biol. Chem. 2000, 275, 4827–4833. 10.1074/jbc.275.7.4827. [DOI] [PubMed] [Google Scholar]
  14. Fagioli C.; Mezghrani A.; Sitia R. Reduction of interchain disulfide bonds precedes the dislocation of Ig-mu chains from the endoplasmic reticulum to the cytosol for proteasomal degradation. J. Biol. Chem. 2001, 276, 40962–40967. 10.1074/jbc.M107456200. [DOI] [PubMed] [Google Scholar]
  15. Frand A. R.; Kaiser C. A. The ERO1 gene of yeast is required for oxidation of protein dithiols in the endoplasmic reticulum. Mol. Cell 1998, 1, 161–170. 10.1016/S1097-2765(00)80017-9. [DOI] [PubMed] [Google Scholar]
  16. Mezghrani A.; Fassio A.; Benham A.; Simmen T.; Braakman I.; Sitia R. Manipulation of oxidative protein folding and PDI redox state in mammalian cells. EMBO J. 2001, 20, 6288–6296. 10.1093/emboj/20.22.6288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Tu B. P.; Weissman J. S. Oxidative protein folding in eukaryotes: mechanisms and consequences. J. Cell Biol. 2004, 164, 341–346. 10.1083/jcb.200311055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Tu B. P.; Weissman J. S. The FAD- and O(2)-dependent reaction cycle of Ero1-mediated oxidative protein folding in the endoplasmic reticulum. Mol. Cell 2002, 10, 983–994. 10.1016/S1097-2765(02)00696-2. [DOI] [PubMed] [Google Scholar]
  19. Inaba K.; Masui S.; Iida H.; Vavassori S.; Sitia R.; Suzuki M. Crystal structures of human Ero1alpha reveal the mechanisms of regulated and targeted oxidation of PDI. EMBO J. 2010, 29, 3330–3343. 10.1038/emboj.2010.222. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Ron D.; Walter P. Signal integration in the endoplasmic reticulum unfolded protein response. Nat. Rev. Mol. Cell Biol. 2007, 8, 519–529. 10.1038/nrm2199. [DOI] [PubMed] [Google Scholar]
  21. Sevier C. S.; Kaiser C. A. Ero1 and redox homeostasis in the endoplasmic reticulum. Biochim Biophys Acta 2008, 1783, 549–556. 10.1016/j.bbamcr.2007.12.011. [DOI] [PubMed] [Google Scholar]
  22. Zhang J.; Zhu Q.; Wang X.; Yu J.; Chen X.; Wang J.; Wang X.; Xiao J.; Wang C. C.; Wang L. Secretory kinase Fam20C tunes endoplasmic reticulum redox state via phosphorylation of Ero1alpha. EMBO J 2018, 37, e98699 10.15252/embj.201798699. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Johnson B. D.; Geldenhuys W. J.; Hazlehurst L. A. The Role of ERO1alpha in Modulating Cancer Progression and Immune Escape. J. Cancer Immunol. 2020, 2, 103 10.33696/cancerimmunol.2.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hanahan D.; Weinberg R. A. Hallmarks of cancer: the next generation. Cell 2011, 144, 646–674. 10.1016/j.cell.2011.02.013. [DOI] [PubMed] [Google Scholar]
  25. Majmundar A. J.; Wong W. J.; Simon M. C. Hypoxia-inducible factors and the response to hypoxic stress. Mol. Cell 2010, 40, 294–309. 10.1016/j.molcel.2010.09.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Li X.; Zhang K.; Li Z. Unfolded protein response in cancer: the physician’s perspective. J. Hematol. Oncol. 2011, 4, 8 10.1186/1756-8722-4-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Sano R.; Reed J. C. ER stress-induced cell death mechanisms. Biochim. Biophys. Acta 2013, 1833, 3460–3470. 10.1016/j.bbamcr.2013.06.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Yadav U. C.; Rani V.; Deep G.; Singh R. K.; Palle K. Oxidative Stress in Metabolic Disorders: Pathogenesis, Prevention, and Therapeutics. Oxid. Med. Cell Longev. 2016, 2016, 9137629 10.1155/2016/9137629. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Neelamegam R.; Ricq E. L.; Malvaez M.; Patnaik D.; Norton S.; Carlin S. M.; Hill I. T.; Wood M. A.; Haggarty S. J.; Hooker J. M. Brain-penetrant LSD1 inhibitors can block memory consolidation. ACS Chem. Neurosci. 2012, 3, 120–128. 10.1021/cn200104y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Blais J. D.; Chin K. T.; Zito E.; Zhang Y.; Heldman N.; Harding H. P.; Fass D.; Thorpe C.; Ron D. A small molecule inhibitor of endoplasmic reticulum oxidation 1 (ERO1) with selectively reversible thiol reactivity. J Biol Chem 2010, 285, 20993–1003. 10.1074/jbc.M110.126599. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Lee J. C.; Lim K. T.; Jang Y. S. Identification of Rhus verniciflua Stokes compounds that exhibit free radical scavenging and anti-apoptotic properties. Biochim. Biophys. Acta 2002, 1570, 181–191. 10.1016/S0304-4165(02)00196-4. [DOI] [PubMed] [Google Scholar]
  32. Bohm B. A.Chalcones, Aurones and Dihydrochalcones. In The Flavonoids, Harborne J. B.; Mabry T. J.; Mabry H., Eds.; Springer US: Boston, MA, 1975; pp 442–504. [Google Scholar]
  33. Panche A. N.; Diwan A. D.; Chandra S. R. Flavonoids: an overview. J. Nutr. Sci. 2016, 5, e47 10.1017/jns.2016.41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Brown J. P. A review of the genetic effects of naturally occurring flavonoids, anthraquinones and related compounds. Mutat. Res. 1980, 75, 243–277. 10.1016/0165-1110(80)90029-9. [DOI] [PubMed] [Google Scholar]
  35. Havsteen B. Flavonoids, a class of natural products of high pharmacological potency. Biochem Pharmacol 1983, 32, 1141–1148. 10.1016/0006-2952(83)90262-9. [DOI] [PubMed] [Google Scholar]
  36. Xu H. X.; Lee S. F. Activity of plant flavonoids against antibiotic-resistant bacteria. Phytother Res 2001, 15, 39–43. . [DOI] [PubMed] [Google Scholar]
  37. Geldenhuys W. J.; Funk M. O.; Van der Schyf C. J.; Carroll R. T. A scaffold hopping approach to identify novel monoamine oxidase B inhibitors. Bioorg. Med. Chem. Lett. 2012, 22, 1380–1383. 10.1016/j.bmcl.2011.12.056. [DOI] [PubMed] [Google Scholar]
  38. Barth J.; Abou-El-Ardat K.; Dalic D.; Kurrle N.; Maier A. M.; Mohr S.; Schutte J.; Vassen L.; Greve G.; Schulz-Fincke J.; Schmitt M.; Tosic M.; Metzger E.; Bug G.; Khandanpour C.; Wagner S. A.; Lubbert M.; Jung M.; Serve H.; Schule R.; Berg T. LSD1 inhibition by tranylcypromine derivatives interferes with GFI1-mediated repression of PU.1 target genes and induces differentiation in AML. Leukemia 2019, 33, 1411–1426. 10.1038/s41375-018-0375-7. [DOI] [PubMed] [Google Scholar]
  39. Di Monte D. A.; DeLanney L. E.; Irwin I.; Royland J. E.; Chan P.; Jakowec M. W.; Langston J. W. Monoamine oxidase-dependent metabolism of dopamine in the striatum and substantia nigra of L-DOPA-treated monkeys. Brain Res. 1996, 738, 53–59. 10.1016/0006-8993(96)00761-5. [DOI] [PubMed] [Google Scholar]
  40. Gaweska H.; Fitzpatrick P. F. Structures and Mechanism of the Monoamine Oxidase Family. Biomol Concepts 2011, 2, 365–377. 10.1515/BMC.2011.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Heijnen W. T.; De Fruyt J.; Wierdsma A. I.; Sienaert P.; Birkenhager T. K. Efficacy of Tranylcypromine in Bipolar Depression: A Systematic Review. J. Clin. Psychopharmacol. 2015, 35, 700–705. 10.1097/JCP.0000000000000409. [DOI] [PubMed] [Google Scholar]
  42. Jia Y.; Yun C. H.; Park E.; Ercan D.; Manuia M.; Juarez J.; Xu C.; Rhee K.; Chen T.; Zhang H.; Palakurthi S.; Jang J.; Lelais G.; DiDonato M.; Bursulaya B.; Michellys P. Y.; Epple R.; Marsilje T. H.; McNeill M.; Lu W.; Harris J.; Bender S.; Wong K. K.; Janne P. A.; Eck M. J. Overcoming EGFR(T790M) and EGFR(C797S) resistance with mutant-selective allosteric inhibitors. Nature 2016, 534, 129–132. 10.1038/nature17960. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Kashima K.; Kawauchi H.; Tanimura H.; Tachibana Y.; Chiba T.; Torizawa T.; Sakamoto H. CH7233163 Overcomes Osimertinib-Resistant EGFR-Del19/T790M/C797S Mutation. Mol. Cancer Ther. 2020, 19, 2288–2297. 10.1158/1535-7163.MCT-20-0229. [DOI] [PubMed] [Google Scholar]
  44. Kopecký V. Jr.; Ettrich R.; Pazderka T.; Hofbauerova K.; Reha D.; Baumruk V. Influence of ligand binding on structure and thermostability of human alpha1-acid glycoprotein. J. Mol. Recognit. 2016, 29, 70–79. 10.1002/jmr.2496. [DOI] [PubMed] [Google Scholar]
  45. Jafari R.; Almqvist H.; Axelsson H.; Ignatushchenko M.; Lundback T.; Nordlund P.; Martinez Molina D. The cellular thermal shift assay for evaluating drug target interactions in cells. Nat. Protoc. 2014, 9, 2100–2122. 10.1038/nprot.2014.138. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

bg1c00062_si_001.pdf (21.4MB, pdf)

Articles from ACS Bio & Med Chem Au are provided here courtesy of American Chemical Society

RESOURCES