A new ferroptosis inhibitor, isolated from Ajuga nipponensis, protects neuronal cells via activating NRF2-antioxidant response elements (AREs) pathway

Abstract

Ferroptosis is a new form of cell death, and inhibition of ferroptosis is a promising strategy to treat neurological diseases. In this work, sixteen compounds were isolated from Ajuga nipponensis and assayed for anti-ferroptosis activity in HT22 mouse hippocampal neuronal cells. Ajudecunoid C (1, ADC), a new neoclerodane diterpenoid, showed significant inhibitory activity against erastin and RSL3-induced ferroptosis with EC50 values of 4.1 ± 1.0 and 3.6 ± 0.3 μM, respectively. Experimental results demonstrated that ADC effectively prevented ferroptosis through scavenging free radical and activating NRF2-antioxidant response elements (AREs) pathway. This study reveals that ADC, as a new ferroptosis inhibitor, is a promising lead compound for the development of drugs against ferroptosis-related neurological diseases.

1. Introduction

Ferroptosis is an iron-dependent form of regulated cell death defined in 2012 [1]. Its morphological, biochemical, and genetic characteristics are distinct from other forms of cell death, such as apoptosis and nec- roptosis. The essence of ferroptosis is excessive lipid peroxidation [1]. Glutathione peroxidase 4 (GPX4), using glutathione (GSH) as co-factor, catalyzes toxic lipid hydroperoxides to non-toxic lipid alcohols, thereby suppressing lipid peroxidation and ferroptosis [2]. Cystine/glutamate antiporter (systemX—) imports cystine into cells for GSH synthesis [1]. Inhibition of GPX4 or systemX— can trigger ferroptosis. Accumulating studies indicate that ferroptosis is implicated in many neurological diseases, such as neurodegenerative diseases and stroke [3]. Many neurological diseases are caused by the dysfunction and cell death in central nervous system, and ferroptosis may be a driving factor [4]. For example, Dixon et al. [1] found glutamate-induced neurotoxicity in rat organotypic hippocampal slice culture models shared common core lethal mechanism with ferroptosis. In animal models of neurological dis- eases, some ferroptosis inhibitors have achieved good results, showing their great potential to treat these diseases. For instance, the ferroptosis inhibitor ferrostatin-1 (Fer-1) can inhibit hemoglobin-induced neuronal death in an intracerebral hemorrhage stroke model [5].

Nuclear factor erythroid 2-related factor 2 (NRF2) is a transcription factor that plays a central role in protecting against oxidative damage [6]. It can bind to antioxidant response elements (AREs) in the nucleus, initiating transcription of ARE genes [7]. The downstream ARE genes include heme oxygenase-1 (HO-1) [8], NAD(P)H quinone oxidoreduc- tase 1 (NQO-1) [9], and other antioxidant genes [6]. Increasing evidence reveals that NRF2-AREs pathway is involved in the regulation of fer- roptosis [10–12]. It was reported that NRF2 induction prevents ferrop- tosis and redox imbalance in Friedreich’s ataxia [13]. However, few compounds were reported to suppress ferroptosis by this pathway. Additionally, most of reported ferroptosis inhibitors are synthetic com- pounds but few are natural products.
Ajuga nipponensis, a plant of the genus Ajuga (Labiatae), is a folk edible medicine for treating inflammation and traumatic injury [14]. Previous phytochemical studies of A. nipponensis have led to the iden- tification of phytoecdysteroids [15,16], ecdysterones [14], neoclerodane diterpenoids [17], and flavonoids [14,18]. Some of these compounds show hypoglycaemic [18], antioxidant [14], and hep- atoprotective [14] activities. However, the anti-ferroptosis activity of constituents from A. nipponensis has not been reported yet.

In this study, sixteen compounds, including twelve diterpenoids (1–12), two cyclic eight-membered α, β-unsaturated ketones (13–14), and two ecdysterones (15–16) were isolated from A. nipponensis (Fig. 1). Among them, compounds 1, 2, and 13 are new ones. Their protection against erastin-induced ferroptosis was evaluated in HT22 mouse hip- pocampal neuronal cells. And ajudecunoid C (1, ADC) showed the most potent protective effect. Subsequent experiments demonstrated that ADC is a new ferroptosis inhibitor. Also, the mechanism of action of ADC inhibiting ferroptosis is described.

2. Results and discussion

2.1. Structural elucidation

Compound 1 showed a molecular formula of C31H48O10 based on its HRESIMS ion peak at m/z 603.3138 [M+Na] + (calcd for C31H48O10Na, 603.3140), indicating eight degrees of unsaturation. Its 1H NMR data HMBC correlations from H-15 (δH 5.27) to the carbon with chemical shift at 63.9 (CH2O-) and 108.7 (C-16) confirmed the ethoxy group connected to C-15. The 6-OAc and 19-OAc were also verified by HMBC correlation signals (Fig. 2). 15α-OEt was assigned by comparing the chemical shift of H-15 with that of the known compounds clerodinin C and D [19]. The NOESY correlations of H-3/H2-19 and H2-19/H3-20, H- 6/H-18, H-6/H-8, and H-8/H-10 revealed CH2OAc-19 and CH3-20 were α-oriented and H-10, H-8, H2-18, and H-6 were β-oriented. The H-13 and H-16 are cofacial based on NOESY correlation of H-13/H-16 and the 1H NMR signal of H-16 (δH 5.27, d, J = 5.4 Hz). It is noted that many of neoclerodane diterpenoids with a hexahydrofurofuran moiety have been reported in Ajuga species [20]. However, the stereochemistry of C-11, C-13, and C-16 has not been fully discussed. In the literature, H-11 was defined as α-oriented, and H-13 and H-16 was assigned as β-orientation. Biogenetically, the configurations of C-11, C-13, and C-16 in compound 1 should be consistent with those reported in the literature. Thus, the relative configurations of 1 were identified as shown in Fig. 1 and compound 1 was named as ajudecunoid C (ADC).

2.2. Compound 1 (ADC) is an effective ferroptosis inhibitor

We used the classic ferroptosis inducer erastin [1] to induce ferrop- tosis in HT22 cells, and evaluated whether these sixteen natural prod- ucts from A. nipponensis can inhibit ferroptosis. As shown in Fig. 3A, erastin significantly decreased cell viability to 9.6 ± 0.4%, while the specific ferroptosis inhibitor with 10 μM ADC or 1 μM Fer-1 for 36 h. Data shown represent mean ± SEM from three independent experiments. ##P < 0.01 compared with the control of siNC, **P < 0.01 compared with the control of siGPX4.

For the neoclerodane diterpenoids (1–7), com- pounds with a hexahydrofurofuran moiety (1–2) showed significant anti-ferroptosis activity. Moreover, the substituents MeBuO at C-3 and ethoxy group at C-15 were necessary for this activity. Four (9–12) of the five abietane diterpenoids with phenolic hydroxy groups exhibited moderate activity. The two cyclic eight-membered α, β-unsaturated ketones (13–14) and two ecdysterones (15–16) didn’t show inhibitory activity against ferroptosis. Notably, compounds 1 and 2 at 10 μМ exhibited the most significant inhibitory activity with cell viability of 91.1 ± 3.9 and 76.8 ± 3.2%, respectively. Their EC50 values for erastin lethality suppression were 4.1 ± 1.0 and 8.2 ± 2.8 μM, respectively. The most potent compound 1 (ADC) (Fig. 3C) was chosen for further anti-ferroptosis evaluation. RSL3, another classic ferroptosis inducer, can trigger ferroptosis by directly inactivating GPX4 [1]. With the assay of RSL3-induced ferroptosis in HT22 cells, the ferroptosis inhibitor Fer-1 significantly prevented cell death. ADC inhibited this cell death as well (EC50 = 3.6 ± 0.3 μM). Besides, similar to Fer-1, ADC also significantly suppressed ferroptosis induced by knockdown of GPX4 in HT-1080 human fibro- sarcoma cells (another ferroptosis-sensitive cell line).

To further confirm the inhibition of ADC against ferroptosis, we observed the morphological changes of HT22 cells upon different treatments. As shown in Fig. 4A, RSL3 treatment made HT22 cells shrink and round, while ADC co-treatment significantly reversed this change in a dose-dependent manner. The increase of lipid reactive oxygen species (ROS) and cytosolic ROS are characteristic biomarkers of ferroptosis [1]. We detected the levels of lipid ROS by BODIPY 581/591 C11 (BODIPY- C11), a lipid peroxidation-sensitive dye, using confocal laser scanning microscopy. As shown in Fig. 4B, RSL3 significantly induced lipid ROS accumulation in HT22 cells, as indicated by the remarkable increase in the level of the green fluorescence. When cells were co-incubated with ADC or Fer-1, the fluorescence intensities were significantly reduced. Flow cytometry results also demonstrated that ADC inhibited lipid ROS accumulation in a dose-dependent manner (Fig. 4C). In addition, we detected the levels of cytosolic ROS by the fluorescent probe 5-(and-6)-carboxy-2',7' di-chloro-dihydro-fluorescein diacetate (carboxy-H2DCFDA), using flow cytometry. As shown in Fig. 4D, ADC also dose- dependently suppressed RSL3-induced cytosolic ROS accumulation in HT22 cells. Upregulation of prostaglandin-endoperoxide synthase 2 (PTGS2) mRNA expression is another hallmark of ferroptosis [2]. As shown in Fig. 4E, RSL3 led to more than three-fold increase in PTGS2 mRNA expression, while ADC co-treatment inhibited this upregulation in a dose-dependent manner.

Besides, we found that ADC had no ability to inhibit staurosporine (STS)-induced apoptosis and hydrogen peroxide (H2O2)-induced necrosis. Together, ADC is an effective and specific ferroptosis inhibitor. Next, we explored the underlying mechanism of action of ADC against ferroptosis.

2.3. ADC has mild antioxidant activity and doesn’t reduce intracellular iron levels

The classic ferroptosis inhibitors Fer-1 and liproxstatin-1 are both radical-trapping antioxidants, which directly scavenge free radical to suppress ferroptosis [31]. To evaluate the antioxidant activity of ADC, we performed the 1,1-diphenyl-2-picryhydrazyl (DPPH) assay in cell- free system using Fer-1 and vitamin C as the positive control. As shown in Fig. 5A, ADC can scavenge about 30% free radical DPPH at 100 μM, suggesting that ADC had mild antioxidant activity. As this antioxidant activity is weak, there may be other mechanisms to play a role in the anti-ferroptosis activity of ADC at the same time.

Since iron chelators such as deferoxamine (DFO) are also a class of classic ferroptosis inhibitors [1], we explored the iron chelating ability of ADC using UV–Vis spectroscopy. As shown in Fig. 5B, DFO showed a characteristic absorption peak at ~430 nm in the presence of Fe3+ or Fe2+. However, no new peak was monitored when adding Fe3+ or Fe2+ to the solution of ADC, indicating that ADC did not chelate with iron.

Considering the important role of ferrous iron in lipid peroxidation and ferroptosis, we detected Fe2+ levels in live cells by Fe2+-selective fluo- rescent probe FerroOrange using confocal laser scanning microscopy and flow cytometry. As shown in Fig. 5C and S25C, ADC didn’t change intracellular Fe2+ content. These results suggested that ADC inhibited ferroptosis not by reducing intracellular iron levels in HT22 cells.

2.4. ADC does not affect GSH/GPX4 pathway

The GSH/GPX4 pathway is a main ferroptosis suppressing system [2]. To determine whether ADC acts on this pathway to inhibit ferrop- tosis, we first measured the intracellular GSH level following ADC treatment. As shown in Fig. 6A, ADC didn’t increase GSH level in HT22 cells. Then we detected the levels of mRNA and protein expression of GPX4 using qRT-PCR and Western blot analysis, respectively. Also, ADC didn’t change GPX4 mRNA and protein expression even at a high con- centration (Fig. 6B–D). As aforementioned, ADC can inhibit ferroptosis induced by knockdown of GPX4 (Fig. 3D). Taken together, these results demonstrated that ADC didn’t act on GSH/GPX4 pathway to inhibit ferroptosis.

Ferroptosis suppressor protein 1 (FSP1), a newly identified antioxi- dant regulator in ferroptosis, can inhibit ferroptosis in a GPX4- independent manner [32,33]. Acyl-CoA synthetase long-chain family member 4 (ACSL4), an enzyme involved in fatty acid metabolism, is required for ferroptosis. Inhibition of ACSL4 can suppress ferroptosis

2.5. ADC activates NRF2-AREs pathway

As aforementioned, NRF2-AREs pathway plays an important role in antioxidant defense. And it has been reported that NRF2 is one of the key ferroptotic regulators [36]. To investigate whether the ferroptosis inhibitory activity of ADC is mediated by NRF2-AREs signaling pathway, we first detected the levels of NRF2 using Western blot analysis. As shown in Fig. 7A and B, upon ADC treatment, the protein level of NRF2 was increased in a dose-dependent manner. Then, we examined the mRNA levels of NRF2 and its downstream target genes HO-1 and NQO-1 using qRT-PCR analysis. As shown in Fig. 7C, ADC didn’t affect NRF2 mRNA expression, but significantly increased the mRNA expressions of HO-1 and NQO-1 in a dose-dependent way. Additionally, the Western blot analysis showed that upon ADC treatment, the protein expressions of HO-1 and NQO-1 were also upregulated dose-dependently (Fig. 7A and B). To sum up, these results demonstrated that ADC inhibited fer- roptosis via activating NRF2-AREs pathway.

2.6. ADC may interfere with Keap1-NRF2 interaction

Given ADC increased the protein level of NRF2 without changing its gene expression, we suspected that this compound inhibited NRF2 degradation. Kelch-like ECH-associated protein 1 (Keap1) is the prin- cipal negative regulator of NRF2. In cells, Keap1 interacts with NRF2 and mediates its ubiquitination and proteasomal degradation [37]. When Keap1 expression is reduced or Keap1-NRF2 interaction is dis- rupted, NRF2 degradation will be inhibited and its protein level will be increased.

To verify our conjecture, we first detected the level of Keap1. As shown in Fig. 8A and B, ADC didn’t change the level of Keap1. There- fore, the upregulation of NRF2 is most likely a result of the disruption of Keap1-NRF2 interaction by ADC. Then we performed molecular docking to predict the binding mode of ADC with Keap1 (PDB code: 5FNR). As shown in Fig. 8C and D, ADC was located in the NRF2-binding pocket of Keap1, and the binding mode was similar to the previously reported Keap1-NRF2 interaction inhibitors [38]. The oxygen atom of acetoxy group at C-19 formed a hydrogen bond with residue Arg415. A water molecule mediated the interaction of the oxygen atom of acetoxy group at C-6 and Ser602. Together, ADC may disrupt Keap1-NRF2 interaction through binding to the NRF2-binding pocket of Keap1, thereby increasing the protein stability of NRF2 and activating NRF2-AREs pathway.

3. Conclusions

In this study, we isolated sixteen compounds from A. nipponensis, including three new (1, 2 and 13) and thirteen known ones (3–12 and 14–16). Among them, compound 1 (ADC), a new natural diterpenoid showed the most effective anti-ferroptosis activity. ADC blocked fer- roptosis induced by ferroptosis inducers (erastin and RSL3) and knock- down of GPX4. The anti-ferroptosis activity of ADC was further confirmed by its inhibition against the lipid ROS and cytosolic ROS accumulation and upregulation of PTGS2 mRNA expression. We also explored the mechanism of action of ADC against ferroptosis. ADC had mild antioxidant activity as demonstrated by DPPH assay. Western blot analysis showed that ADC upregulated NRF2 protein expression in a dose-dependent manner. Besides, ADC treatment dose-dependently increased the levels of mRNA and protein expression of HO-1 and NQO-1, the downstream AREs in the NRF2 signaling pathway. There- fore, we concluded that ADC activated NRF2-AREs pathway in HT22 cells. Furthermore, we found that ADC neither changed intracellular iron levels nor acted on the classic GSH/GPX4 pathway to inhibit ferroptosis. In combination of experimental results and computational analysis, we speculated that ADC may interfere with Keap1-NRF2 interaction, thereby activating NRF2-AREs pathway to inhibit ferroptosis.

In conclusion, this work revealed that the natural product ADC, a new diterpenoid from A. nipponensis, is a potent ferroptosis inhibitor. It inhibits ferroptosis mainly via activating NRF2-AREs signaling pathway. The intrinsic antioxidant ability of ADC also contributes to its anti- ferroptosis activity. Although ferroptosis attracts more and more scientists, Erastin the types of existing ferroptosis inhibitors are relatively limited, and the mechanisms of ferroptosis are not fully elucidated. Our results demonstrate that ADC may be a small molecular probe for studying the mechanism of ferroptosis, and also serve as a lead compound for the treatment of ferroptosis-related neurological diseases.