Mechanistic Insights into Salidroside's Mitochondrial Protection via AMPK/Sirt1/HIF-1α Pathway in Hypoxic HT22 Cells

The present study identified a unique mechanism by which salidroside exerts mitochondrial protective effects on hypoxic HT22 cells, partly through the AMPK/Sirt1/HIF-1α pathway.

Salidroside (Sal), an active ingredient of Rhodiola crenulata (Hook. f. et Thoms.) H. Ohba has been found to exert mitochondrial protective effects by improving metabolism and enhancing the energy supply of brain cells under hypoxic conditions. However, its mechanism of action has not been fully clarified. In the present study, high-performance liquid chromatography was first employed to analyze the effects of Sal on nucleotide (ATP, ADP, and AMP) levels. The cellular thermal shift assay (CETSA), a widely used molecular interaction method for validating and quantifying drug target engagement in cells and tissues across different species, was then chosen to confirm the affinity of Sal for AMPK/Sirt1/HIF-1α pathway-related proteins. The results revealed that Sal increased ATP and ADP levels in hypoxic HT22 cells while reducing AMP levels. Moreover, Sal exhibited stable binding to AMPKα, p-AMPKα, Sirt1, and HIF-1α proteins. In conclusion, Sal may exert mitochondrial protective effects by modulating the AMPK/Sirt1/HIF-1α pathway to regulate nucleotide content. This study provides a methodological reference for nucleotide content analysis in cell samples and contributes to the identification and discovery of targets for compounds derived from traditional Chinese medicine.

The brain is highly sensitive to oxygen due to its high metabolic demands, limited glycolytic capacity, and dependence on oxidative phosphorylation. As a result, exposure to a low-oxygen environment at high altitudes can readily lead to hypobaric hypoxic brain injury (HHBI)^1,2. Epidemiological studies indicate that when individuals unacclimated to high altitudes ascend rapidly to high-altitude regions, the incidence of acute mountain sickness can reach up to 75%, with a fatality rate of approximately 1% for severe cases. Furthermore, in the absence of medical care, mortality rates for high-altitude cerebral or pulmonary edema can be as high as 40%^3,4.

HHBI presents with a broad spectrum of clinical symptoms. Mild to moderate cases may include headaches, dizziness, and memory loss⁵, while severe cases can result in cognitive impairment, altered consciousness, and potentially fatal outcomes⁵. The prevention and treatment of HHBI in high-altitude regions have become a key focus of medical research. Preventative strategies primarily involve adaptive training in high-altitude environments, including adequate rest, a well-balanced diet, proper nutrition, and appropriate physical exercise^6,7. Additionally, pharmacological interventions aimed at protecting brain cells and alleviating cerebral hypoxia remain central to current HHBI research.

Mitochondria serve as the primary energy production centers within cells, synthesizing adenosine triphosphate (ATP) to meet cellular energy demands. Under hypoxic conditions, mitochondrial energy production declines, leading to reduced ATP levels and impaired cell function⁸. The hypoxic injury also disrupts mitochondrial regulation of Ca²⁺ and pH homeostasis, triggering apoptosis and necrosis^9,10. There is a mutually reinforcing relationship between mitochondrial dysfunction and hypoxic brain injury. On the one hand, hypoxia-induced mitochondrial impairment exacerbates oxygen deficiency by further reducing cellular energy metabolism, creating a vicious cycle. On the other hand, mitochondrial dysfunction elevates intracellular Ca²⁺ levels, activating apoptotic cascades and leading to cell death¹¹. Although the mechanisms underlying hypoxic brain injury remain complex and not fully understood, multiple studies have identified impaired neuronal mitochondrial energy metabolism as a critical factor in its pathogenesis^12,13. Therefore, further exploration of mitochondrial function may provide valuable insights into potential therapeutic targets for hypoxic brain injury.

Salidroside (Sal) is an active ingredient extracted from the plateau plant Rhodiola crenulata (Hook. f. et Thoms.) H. Ohba and is widely used in food health products and pharmaceuticals¹⁴. The molecular formula of Sal is C₁₄H₂₀O₇, and it is also known as 2-(4-hydroxyphenyl)-ethyl β-D-glucopyranoside. It possesses diverse pharmacological properties, including anti-hypoxia, antioxidant, anti-fatigue, anti-tumor, immunomodulatory, anti-inflammatory, and cardiovascular and cerebrovascular protective effects^15,16,17. Among these, its anti-hypoxia effect is one of the most well-documented. Recent studies have increasingly highlighted the significant mitochondrial protective effects of Sal as a potential mechanism for its preventive and therapeutic effects on plateau-induced brain injury in mice^14,18. However, the precise molecular mechanisms by which Sal influences ATP, ADP, and AMP levels remain poorly understood.

AMP-activated protein kinase (AMPK) acts as a key energy sensor that helps maintain cellular energy homeostasis. Activation of AMPK stimulates Sirtuin 1 (Sirt1), leading to an increase in intracellular NAD⁺ levels¹⁹. Studies have shown that Sirt1 can regulate hypoxia-inducible factor 1-alpha (HIF-1α) to coordinate the cellular response to hypoxia²⁰. Previous research has demonstrated that Sal inhibits the opening of the neuronal mitochondrial permeability transition pore, regulates HIF-1α-mediated mitochondrial energy processes, attenuates neuronal apoptosis, and maintains blood-brain barrier integrity, thereby protecting rats from plateau-induced brain injury^14,21. However, the effect of Sal on ATP and its metabolites, ADP and AMP, remains uncertain.

To investigate this, high-performance liquid chromatography (HPLC) was first employed to quantify the levels of these three nucleotides. Additionally, the cellular thermal shift assay (CETSA), a widely used biophysical technique introduced in 2013 to study ligand-protein interactions in intact cells²², was utilized. This method is commonly applied to validate and quantify drug-target engagement in cells and tissues across different species. Specifically, after co-incubating target cell lysates with the drug at varying temperatures for a set duration, the drug-bound protein exhibits increased thermal stability, making it less prone to denaturation and precipitation. The precipitated unbound proteins are then removed via centrifugation, and drug-target interactions are subsequently identified through western blot analysis of the supernatant²². To identify potential molecular targets of Sal, CETSA was selected to assess its binding affinity with AMPK/Sirt1/HIF-1α pathway-related proteins.

The commercial details of the reagents and the equipment used in this study are provided in the Table of Materials.

1. Solution preparation

1. Prepare complete Dulbecco's modified Eagle medium (DMEM) by adding 10% fetal bovine serum and 1% penicillin-streptomycin.
2. Prepare a 20 mM Sal solution by dissolving 3 mg of Sal in 500 µL of phosphate-buffered saline (PBS)²¹.
3. Filter 500 mL of chromatographically pure acetonitrile (mobile phase A) using a 0.45 µm organic membrane.
4. Prepare 1000 mL of mobile phase B by adding 3.5 mL of reagent I to 1 L of deionized water. Adjust the pH to 6.15 with reagent II, following the instructions of the nucleotide (ATP, ADP, and AMP) content assay kit (see Table of Materials).
   NOTE: Reagent I and reagent II are included in the nucleotide (ATP, ADP, and AMP) content assay kit.
5. Filter mobile phase B using a 0.22 µm aqueous membrane.
   NOTE: Mobile phase B should be used immediately after preparation. Mobile phases A and B should be sonicated for 30 min before use to remove air bubbles.
6. Dilute a 1 µM ATP, ADP, and AMP standard stock solution with deionized water to obtain 0.5 µM, 0.1 µM, 0.05 µM, 0.01 µM, and 0.005 µM ATP, ADP, and AMP series standard solutions.
7. Filter the series of standard solutions using a 0.22 µm needle-type microporous filter membrane.
   NOTE: Brown injection bottles should be used to store the serial standard solutions. The prepared series of standard solutions should be used as soon as possible.

2. Cell culture

NOTE: HT22 cell culture and the CoCl₂-stimulated hypoxia model were established according to a previous report²¹.

1. Culture HT22 cells at 37 °C with 5% CO₂ until 80% confluence is reached in the Petri dish. Digest the cells with 0.25% trypsin for 1 min.
2. Seed 2 × 10⁵ cells from step 2.1 into 6-well plates with complete DMEM (control group with no drug treatment) or complete DMEM containing the drug (250 µM CoCl₂, 250 µM CoCl₂ + 10 µM Dor, 250 µM CoCl₂ + 20 µM Sal, and 250 µM CoCl₂ + 10 µM Dor + 20 µM Sal). Incubate for 24 h at 37 °C with 5% CO₂.
   NOTE: Three parallel samples were set up in each group.

3. Nucleotide (ATP, ADP, and AMP) content assay

1. Add 1 mL of extraction reagent I to each well in the 6-well plates from step 2.2.
   NOTE: Extraction reagent I is included in the nucleotide (ATP, ADP, and AMP) content assay kit (see Table of Materials).
2. Lyse the cells by sonicating on an ice bath using an ultrasonic cell disruption apparatus (power: 300 W, ultrasonic waves for 3 s, 7 s interval, total time: 3 min).
3. Centrifuge the lysed cells at 10,304 × g for 10 min at 4 °C using a high-speed refrigerated centrifuge.
4. Transfer 0.75 mL of supernatant from step 3.3 into a 2.0 mL tube. Add 0.75 mL of extraction reagent II, shake and mix using a vortex mixer, and centrifuge again at 10,304 × g for 10 min at 4 °C.
   NOTE: Extraction reagent II is included in the nucleotide (ATP, ADP, and AMP) content assay kit.
5. Pipette the supernatant from step 3.4 and filter it using a 0.22 µm needle-type microporous filter membrane into a brown vial.
   NOTE: ATP, ADP, and AMP in the extracted sample are not stable. All experimental procedures should be performed at low temperatures or on ice. Samples should be processed and tested as soon as possible.
6. Open the chromatography software and set the following parameters: injection volume: 10 µL; column temperature: 27 °C; flow rate: 0.8 mL/min; detection wavelength: 254 nm; detection duration: 70 min. Set the gradient elution procedure according to Table 1.
7. Click the execute button to automatically inject the sample and detect the content of ATP, ADP, and AMP in serial standard solutions (step 1.7) and sample solutions (step 3.5) according to the program set in step 3.6.
   NOTE: At the end of the experiment, wash the chromatographic column with 98% mobile phase B to prevent clogging.
8. Plot the standard curves of ATP, ADP, and AMP using graphing and statistical analysis software, with serial standard concentrations on the x-axis and peak areas on the y-axis.
9. Calculate the ATP, ADP, and AMP content in the samples using the standard curve equation from step 3.8.

4. Cellular thermal shift assay (CETSA)

1. Lyse normal HT22 cells for 20 min in a Petri dish using protease inhibitor-containing lysate (lysate: protease inhibitor = 100:1) by repeatedly pipetting the cells.
   NOTE: Perform this step on an ice bath to prevent protein degradation.
2. Collect the cell lysate from step 4.1 and sonicate on an ice bath using an ultrasonic cell disruption apparatus (power: 300 W, ultrasonic waves for 3 s, 7 s interval, total time: 3 min). Centrifuge at 10,304 × g, 4 °C for 20 min and collect the supernatant.
   NOTE: Divide the supernatant into 14 portions: 7 groups for coincubation with Sal and 7 groups for coincubation with PBS as the control. For the Sal-treated group, add 100 µL of supernatant and 0.1 µL of 20 µM Sal per sample. For the control group, add 100 µL of supernatant and 0.1 µL of PBS per sample.
3. Incubate each sample at room temperature for 30 min, then further incubate at 37 °C, 42 °C, 47 °C, 52 °C, 57 °C, 62 °C, and 67 °C for 3 min using a metal heating temperature control instrument.
4. Centrifuge the samples from step 4.3 at 10,304 × g for 20 min at 4 °C and collect the supernatant using a pipette.
5. Measure the total protein concentration of the supernatant from step 4.4 using a BCA protein concentration assay kit, following the manufacturer's instructions (see Table of Materials).
6. Prepare 10% separation gel using a PAGE gel rapid preparation kit, following the manufacturer's instructions (see Table of Materials).
7. Load 3 µL of pre-stained color protein marker and 14 µL of samples into the wells of the separation gel (final protein amount per well: 20 µg).
8. Run the gel using electrophoresis buffer for 1 h at 80-100 mV to obtain separated proteins.
9. Transfer the proteins from step 4.8 onto a PVDF membrane using rapid membrane transfer solution and a sandwich structure (cotton-filter paper-PVDF membrane-protein gel-filter paper-cotton) at 400 mA for 30 min .
10. Block the PVDF membranes from step 4.9 with 5% bovine serum albumin (BSA) solution for 1-2 h at room temperature.
    NOTE: Prepare 5% BSA solution by dissolving 5 mg of BSA powder in 100 mL of TBST solution containing 100 mL of TBS buffer and 0.05% Tween 20 .
11. Incubate the PVDF membranes from step 4.10 with 5 mL diluted primary antibody at 4 °C overnight.
    NOTE: Dilute the primary antibody in BSA solution (1:1000 ratio).
12. Wash the PVDF membranes three times with TBST for 5 min each to remove unbound primary antibodies.
13. Incubate the membranes from step 4.12 with diluted horseradish peroxidase (HRP)-coupled secondary antibody for 2 h at room temperature on a decolorization shaker.
    NOTE: Dilute the HRP-coupled secondary antibody in BSA solution (1:10,000 ratio) .
14. Wash the membranes three times with TBST for 5 min each to remove residual secondary antibody.
15. Cover the membranes from step 4.14 with ECL chemiluminescence developer, then image the target protein signal using an imaging system.
    NOTE: Perform this step in the dark to prevent fluorescence quenching.
16. Analyze and quantify the gray value of target proteins using ImageJ software (see T able of Materials).

The standard curves for ATP, ADP, and AMP detected by HPLC were Y = 7006.5X - 222.99, Y = 5217.3X - 17.796, and Y = 9280.1X + 22.749, respectively (Figure 1A-C). The nucleotide contents measured in each group by HPLC were calculated using the standard curves (Figure 1D-I). It was found that CoCl₂ significantly reduced ATP and ADP levels in HT22 cells compared to the control group (P < 0.01, Figure 1J,K). Furthermore, administration of dorsomorphin (Dor), an AMPK inhibitor, led to a marked reduction in ATP and ADP levels compared to the CoCl₂-treated group (P < 0.01, Figure 1J,K). However, treatment with Sal significantly increased ATP and ADP levels relative to the CoCl₂-treated group (P < 0.05, Figure 1J,K).

Further analysis revealed that AMP levels increased significantly in the CoCl₂-treated group, whereas Sal treatment effectively limited this surge (P < 0.01, Figure 1L). Proper mitochondrial function is essential for maintaining intracellular ATP, ADP, and AMP homeostasis, as well as normal cellular physiological functions. The functional state of mitochondria directly influences the conversion rate and balance of ATP, ADP, and AMP. Mitochondrial damage or dysfunction can impair ATP synthesis and reduce the conversion rates of ADP and AMP, thereby disrupting cellular energy supply and metabolic balance.

These findings suggest that Sal may alleviate hypoxic injury in HT22 cells by modulating ATP, ADP, and AMP conversion. Additionally, CETSA analysis of Sal binding to mitochondria-associated proteins demonstrated an increase in the thermal stability of AMPKα, p-AMPKα, Sirt1, and HIF-1α following Sal treatment. This suggests that Sal may interact with these proteins (Figure 2), consistent with findings from previous studies^21,23,24.

DATA AVAILABILITY:
All raw data are provided (Supplementary File 1 and Supplementary File 2).

Figure 1: Effect of Sal on nucleotide (ATP, ADP, and AMP) content. (A-C) Standard curves for ATP (A), ADP (B), and AMP (C). (D-I) Representative HPLC curves for the determination of ATP, ADP, and AMP content across different treatment groups. (J-L) Statistical analysis of ATP, ADP, and AMP content in different treatment groups (control, 250 µM CoCl₂, 250 µM CoCl₂ + 10 µM Dor, 250 µM CoCl₂ + 20 µM Sal, and 250 µM CoCl₂ + 10 µM Dor + 20 µM Sal). (n = 3, *P < 0.05, **P < 0.01). Please click here to view a larger version of this figure.

Figure 2: Evaluation of Sal binding properties to AMPK/Sirt1/HIF-1α pathway-associated proteins using the CETSA method. β-Actin (A), AMPKα (B), p-AMPKα (C), Sirt1 (D), and HIF-1α (E) protein bands and expression trends in HT22 cells were assessed by CETSA (n = 3). The binding properties of Sal with different proteins were analyzed using statistical and graphing software across varying temperatures. Please click here to view a larger version of this figure.

Table 1: Mobile phase elution procedure.

Supplementary File 1: Raw data for detecting ATP, ADP and AMP by HPLC. Please click here to download this File.

Supplementary File 2: CETSA protein bands. Please click here to download this File.

Mitochondria are key organelles involved in the therapeutic prevention of HHBI^23,24,25. Previous studies by the group have confirmed that Sal regulates AMPK, Sirt1, and HIF-1α protein expression, enhancing neuronal mitochondrial function and protecting against HHBI^21,24. However, the direct effect of Sal on nucleotides in hypoxic cells requires further investigation. Additionally, assessing the binding potential of compounds and proteins under varying temperature conditions is essential for understanding their interactions.

To address these aspects, the present study examined the effect of Sal on mitochondrial ATP, ADP, and AMP levels in HT22 cells using HPLC, providing direct insight into its role in mitochondrial energy metabolism. Based on a previous study¹⁸, 250 µM of CoCl₂ was selected to induce hypoxia in HT22 cells. The results demonstrated that Sal significantly increased ATP and ADP levels while reducing AMP content in hypoxic HT22 cells (Figure 1). Combined with prior findings from the group, these results further confirm that Sal exerts a protective effect on mitochondria in hypoxic neurons.

AMPK is a highly conserved protein kinase in eukaryotes that regulates energy metabolism. It senses changes in cellular ATP, ADP, and AMP levels to coordinate glucose metabolism, fatty acid synthesis, and mitochondrial biogenesis by phosphorylating various substrates²⁶. Activation of AMPK promotes energy production and suppresses energy expenditure, particularly in response to reduced intracellular ATP levels and elevated AMP levels²⁷. Additionally, AMPK enhances ATP reserves by phosphorylating glucoamylase, altering enzymatic catalysis, and rapidly generating ATP from the phosphoarginine-arginine pool²⁸.

In the present study, Dor, a selective phosphorylated AMPK inhibitor, was used to investigate the effects of Sal on nucleotide levels²⁹. HPLC was employed to quantify ATP, ADP, and AMP levels in CoCl₂-treated HT22 cells exposed to Dor alone or in combination with Sal. The findings indicate that Sal modulates the ATP/ADP/AMP ratio, thereby exerting neuroprotective effects against hypoxia (Figure 1). The established HPLC method provides an intuitive, simple, and efficient approach for directly assessing the effects of bioactive compounds on nucleotides at the cellular level.

HIF-1α is a transcription factor activated in response to hypoxic stimulation. Under hypoxic conditions, the stability of HIF-1α increases significantly, allowing it to translocate from the cytoplasm to the nucleus, where it forms a heterodimer with HIF-1β³⁰. This process promotes the expression of genes involved in angiogenesis, glucose metabolism, and glycolysis^31,32,33. Additionally, HIF-1α stability has been shown to be highly correlated with mitochondrial function and intracellular O₂ levels during oxidative phosphorylation³¹. However, hypoxia-induced mitochondrial dysfunction leads to an insufficient intracellular energy supply, which in turn activates the HIF-1α pathway to regulate cellular adaptation to hypoxia^31,32.

Sirt1, a member of the Sirtuin family, is an NAD⁺-dependent deacetylase that reduces HIF-1α protein stability by deacetylation, thereby limiting its nuclear accumulation³². Under hypoxic conditions, decreased intracellular ATP levels lead to increased AMPK activity, which influences the intracellular transport, stability, and transcriptional activity of HIF-1α and its associated proteins through phosphorylation^31,32,33. Additionally, Sirt1 and AMPK modulate the interaction between HIF-1α and coactivators, thereby regulating its transcriptional activity and further facilitating cellular adaptation to hypoxia³³. In hypoxia-injured brain cells, AMPK, Sirt1, and HIF-1α interact to maintain intracellular energy homeostasis and regulate metabolic processes.

Simple and efficient methods for evaluating the thermostability and binding potential of compound-protein complexes in vitro include techniques that assess the stability of proteins upon ligand binding. It is well established that compound-protein interactions increase the thermostability of proteins, making them less sensitive to temperature-induced denaturation and resulting in a relatively stable protein yield. The cellular thermal shift assay (CETSA) combined with western blot analysis, while not revealing the exact binding site, provides a low-cost and efficient approach to studying drug-target interactions and potential off-target effects^22,34. CETSA is widely used to validate and quantify drug-target engagement in cells and tissues across different species.

In this study, CETSA was employed to assess whether Sal interacts with AMPKα, p-AMPKα, Sirt1, and HIF-1α at the cellular level. Following heat treatment, Sal increased the expression of these proteins, suggesting the formation of stable Sal-AMPKα, Sal-p-AMPKα, Sal-Sirt1, and Sal-HIF-1α complexes (Figure 2). These findings align with previous experimental results^21,24. CETSA-western blot thus serves as a robust and efficient approach for large-scale screening of compound-protein interactions and thermostability in vitro.

Despite the insights gained into the protective mechanisms of Sal on mitochondrial energy metabolism using HPLC and CETSA techniques, certain limitations remain. CETSA does not provide information at the amino acid level regarding how compounds influence protein thermostability. A potential advancement would be the integration of CETSA with imaging technologies to visually identify compound-binding sites within proteins, offering a clearer understanding of how compound-protein complexes resist heat-induced denaturation. Achieving this requires further scientific innovation and refined experimental design. Additionally, complementing CETSA with other molecular interaction techniques, such as localized surface plasmon resonance and differential scanning fluorimetry, would provide a more comprehensive analysis of compound-protein interactions³⁵.

The authors have nothing to disclose.

This work was supported by the National Natural Science Foundation of China (82274207 and 82474185), the Science & Technology Department of Sichuan Province (2024NSFSC1845), the Science Foundation for Youths of Science & Technology Department of Sichuan Province (2023NSFSC1776), the Key Research and Development Program of Ningxia (2023BEG02012), Youth Talent Support Project of the China Association of Chinese Medicine for 2024-2026 (2024-QNRC2-B07) and the Xinglin Scholar Research Promotion Project of Chengdu University of TCM (XKTD2022013 and QJJJ2024027).

AUTHOR CONTRIBUTION:
Xiaobo Wang, Yating Zhang, Ya Hou, Rui Li and Xianli Meng conceived this project. Yating Zhang, Ya Hou, and Tingting Kuang performed the experiments and analyzed the data. Yating Zhang and Hong Jiang wrote the manuscript. Xiaobo Wang and Xianli Meng revised the manuscript. All of the authors have read and approved the final manuscript.