Plasma metabolism biomarker combination for predicting and intervening propofol-induced respiratory depression, detection method and application

Abstract

The invention discloses a plasma metabolism biomarker combination for predicting and intervening propofol-induced respiratory depression, a detection method and application, and relates to the technical field of biological science. According to the method, non-target metabonomics detection is conducted on preoperative plasma of 298 subjects through UPLC-MS, 14-metabolite panels are obtained through machine learning supervised feature screening, a diagnosis model is established through verification of discovery and verification queues, and the prediction AUC of severe respiratory depression reaches 0.86-0.91 (discovery queue) and 0.86-0.91 (verification queue). Mouse in-vivo function verification proves that the key metabolite alpha-linolenic acid obviously improves the respiratory rate, the tidal volume and the minute ventilation volume under two administration paradigms of pretreatment and treatment of a mouse propofol-RD model, shortens the recovery time of righting reflex, has a good safety foundation, and has a good application prospect. Prompt can be converted into a foresight supplement or rapid treatment medication thought in a perioperative period, and a'prediction and intervention 'integrated clinical value is formed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of biological science technology, and in particular to a blood plasma metabolite biomarker combination for predicting and intervening propofol-induced respiratory depression, a detection method and application. BACKGROUND

[0002] Propofol, as one of the most commonly used intravenous anesthetics in clinic, can cause dose-dependent respiratory depression (RD), manifested as tidal volume reduction, respiratory rate decrease, and even respiratory arrest. Due to the lack of specific antagonists, clinical treatment mainly relies on airway support and mechanical ventilation, so how to identify high-risk patients before anesthesia and intervene in advance has important clinical significance. At present, the means for assessing the risk of respiratory depression (such as age, BMI, anesthetic dose, blood oxygen / C Monitoring, etc.) are mostly empirical or post-knowledge, which is difficult to explain individual sensitivity differences and cannot accurately stratify before anesthesia. The core problem lies in the lack of specific prediction tools based on plasma molecules, which cannot reveal the nature of individual sensitivity differences to propofol respiratory depression from the metabolic mechanism level, resulting in lack of targeted clinical intervention.

[0003] In recent years, metabolomics has been regarded as a key technology connecting drug response and phenotypic results. Unlike genomics or proteomics, metabolites directly reflect cell energy status, lipid homeostasis, redox level, and inflammatory response, which are highly related to drug sensitivity and physiological changes. Ultra-performance liquid chromatography-mass spectrometry (UPLC-MS) is the most mature metabolomics detection platform in current clinical research, with the advantages of high sensitivity, wide dynamic range, and simultaneous quantification of hundreds of endogenous metabolites, which can systematically capture the overall metabolic state of patients before anesthesia, and mine potential markers and metabolic pathways related to propofol respiratory depression. In addition, combined with machine learning algorithms, stable multi-molecular combinations can be screened from high-dimensional metabolic data to establish prediction models.

[0004] However, there are few existing metabolomics studies on propofol-related respiratory depression, lacking: 1) large sample preoperative blood metabolite data; 2) repeatable detection procedures and feature selection methods; 3) lack of functional verification of key metabolites - i.e. whether the candidate metabolites can actively improve respiratory depression, rather than just a simple correlation, so that there is still a lack of functional molecules that can be safely used to prevent or treat this adverse reaction in clinic. These technical shortcomings collectively result in the current inability to provide a clinically translatable precise prediction model, and the formation of a complete management system of "preoperative risk assessment - active intervention", which cannot reduce the risk and harm of respiratory depression from the source.

[0005] Therefore, those skilled in the art are dedicated to developing a method for predicting respiratory depression based on plasma metabolomics and proposing candidate metabolites that can be used for intervention, which can not only be used for anesthesia risk assessment, but also provide a new strategy for preoperative nutritional supplementation or drug regulation. Summary of the Invention

[0006] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is to develop a combination of plasma metabolic biomarkers for predicting and intervening in propofol-induced respiratory depression, a detection method thereof, and their applications.

[0007] To achieve the above objectives, the present invention provides a combination of plasma metabolic biomarkers for predicting and intervening in propofol-induced respiratory depression, a detection method, and their applications.

[0008] Furthermore, a plasma metabolic biomarker set for predicting and intervening in propofol-induced respiratory depression consists of the following 14 metabolites: 4-allylphenol sulfate, 13-hydroxyoctadecadienoic acid, α-linolenic acid, hexadecanoic acid, ibostane, corticosterone, aspirin, valdiate, histidine, 3-op-coumarylquinic acid, P-type 18:0 / 22:6(4Z,8Z,10Z,13Z,16Z,19Z)-OH(7) phosphatidylcholine, 20:4(5Z,8Z,11Z,14Z)-OH(16R) / 22:0 type phosphatidylethanolamine, P-type 18:0 / 20:3(8Z,11Z,14Z) phosphatidylethanolamine, and 20:0 / 20:3(6,8,11)-OH(5) type phosphatidylethanolamine.

[0009] Furthermore, the 14 metabolites are divided into 6 downregulated metabolites and 8 upregulated metabolites; the downregulated metabolites include 4-allylphenol sulfate, 13-hydroxyoctadecadienoic acid, α-linolenic acid, hexadecanoic acid, ibostane, and corticosterone; the upregulated metabolites include acetosiline, valdiate, histidine-lysine dipeptide, 3-Op-coumarylquinic acid, P-type 18:0 / 22:6(4Z,8Z,10Z,13Z,16Z,19Z)-OH(7) phosphatidylcholine, 20:4(5Z,8Z,11Z,14Z)-OH(16R) / 22:0 type phosphatidylethanolamine, P-type 18:0 / 20:3(8Z,11Z,14Z) phosphatidylethanolamine, and 20:0 / 20:3(6,8,11)-OH(5) type phosphatidylethanolamine.

[0010] Furthermore, among the 14 metabolites, α-linolenic acid had a p-value of 2.87E-05, which was the smallest p-value among the downregulated metabolites, and its expression level in the severe respiratory depression group was significantly lower than that in the non-respiratory depression group.

[0011] Furthermore, the detection method for the plasma metabolic biomarker combination includes the following steps: Step 1: Patients undergoing hysteroscopic surgery who received propofol anesthesia were prospectively enrolled and divided into four groups: no RD, mild, moderate, and severe RD. EDTA anticoagulated plasma was collected before anesthesia and frozen at -80°C. Step 2: Take 100 μL of plasma, add 400 μL of methanol / acetonitrile solution with a volume ratio of 1:1, extract the protein, centrifuge and take the supernatant, vacuum dry, and then reconstitute with 50% acetonitrile aqueous solution. Step 3: Inject the reconstituted supernatant into the UPLC-MS system, and perform full scan and secondary mass spectrometry acquisition using optimized chromatographic gradient and mass spectrometry parameters in positive and negative ion mode, m / z 60-1200. Step 4: Peak extraction and alignment were performed using Progenesis QI software, and metabolite annotation was performed based on the HMDB database, resulting in a total of 740 metabolic features. Step 5: Screening of 740 metabolic features. Based on the conditions of LASSO score not less than 0.04, intergroup comparison p value less than 0.05, and single marker AUC > 0.65, the plasma metabolic biomarker combination described in claim 1 is obtained.

[0012] Furthermore, the method for verifying the performance of the plasma metabolic biomarker combination includes: dividing the 298 subject samples into a discovery cohort (199 cases) and a validation cohort (99 cases) at a 2:1 ratio, and using the method described in claim 4 to detect the expression levels of 14 metabolites in the two groups of samples, and using the area under the ROC curve (AUC), sensitivity, and specificity as evaluation indicators to verify the predictive performance of the biomarker combination.

[0013] Furthermore, the diagnostic performance of the 14 metabolites was evaluated using five typical machine learning algorithms: Support Vector Machine (SVM), Lasso Regression (LR), Neural Network (NN), Elastic Network (EN), and Ridge Regression (RR).

[0014] Furthermore, the procedure includes in vivo functional verification steps for α-linolenic acid in the combination, as detailed below: S1: Animal model preparation: A respiratory depression model was established by intraperitoneal injection of 50 mg / kg propofol using male C57BL / 6 mice and a whole-body volume plethysmography system. S2: Pretreatment experiment: 15 minutes before propofol injection, 100 mg / kg α-linolenic acid was injected via the tail vein, and the respiratory rate, tidal volume and anesthesia recovery time of the mice were detected and recorded. S3: Treatment Experiment: 100 mg / kg α-linolenic acid was injected via the tail vein 5 minutes after propofol injection, and the recovery of respiratory function in mice was dynamically monitored.

[0015] Furthermore, the pretreatment experiment results showed that α-linolenic acid significantly reduced propofol-induced respiratory depression (p<0.001) and shortened the anesthesia recovery time (p<0.05); the treatment experiment results showed that α-linolenic acid could rapidly improve the respiratory function of mice (p<0.01).

[0016] Furthermore, the application of the plasma metabolic biomarker combination in the preparation of drugs to intervene in propofol-induced respiratory depression, wherein the plasma metabolic biomarker is α-linolenic acid.

[0017] Furthermore, the application of the aforementioned plasma metabolic biomarker combination in predicting propofol-induced respiratory depression.

[0018] In a preferred embodiment 1 of the present invention, the process of plasma metabolite detection and data acquisition is described in detail.

[0019] In another preferred embodiment 2 of the present invention, the screening and performance verification process of 14-metabolite biomarker combinations is described in detail.

[0020] In another preferred embodiment 3 of the present invention, the in vivo functional verification process of the key metabolite (α-linolenic acid) is described in detail.

[0021] Technical effects: 1. This invention uses UPLC-MS to perform target-free metabolomics detection on preoperative plasma of 298 subjects, and combines machine learning with supervised feature screening to obtain a panel of 14 metabolites. A diagnostic model was established through discovery and validation cohorts. The predictive AUC for severe respiratory depression reached 0.86-0.91 (discovery cohort) and 0.86-0.91 (validation cohort), which can realize preoperative risk stratification and perioperative management optimization, filling the gap in accurate prediction at the molecular level of plasma.

[0022] 2. This invention has demonstrated through in vivo functional verification in mice that the key metabolite α-linolenic acid significantly improved respiratory rate, tidal volume, and minute ventilation under both pretreatment and rescue administration paradigms in a mouse propofol-RD model, and shortened the righting reflex recovery time, demonstrating a good safety profile. This suggests that it can be transformed into a perioperative prospective supplemental or rapid rescue medication approach, forming an integrated clinical value of "prediction + intervention".

[0023] 3. This invention uses mature UPLC-MS technology, which only requires 0.25 mL of blood sample to complete the detection. No additional invasive operations are required. The detection process, quality control methods and model calculations have been standardized and can be quickly deployed in hospital laboratories and third-party medical testing laboratories.

[0024] The following will further explain the concept, specific structure, and technical effects of the present invention in conjunction with the accompanying drawings, so as to fully understand the purpose, features, and effects of the present invention. Attached Figure Description

[0025] Figure 1 This is a diagram illustrating the LASSO score threshold optimization process for feature selection in a preferred embodiment 2 of the present invention.

[0026] Figure 2 This is a time-effect graph of different treatments on the respiratory function of mice with propofol-induced respiratory depression in the pretreatment experiment of a preferred embodiment 3 of the present invention. Figure 3 This is a time-effect graph of different treatments on the respiratory function of mice with propofol-induced respiratory depression in a treatment experiment in a preferred embodiment 3 of the present invention. Detailed Implementation

[0027] The following description, with reference to the accompanying drawings, illustrates several preferred embodiments of the present invention to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms, and the scope of protection of the present invention is not limited to the embodiments mentioned herein.

[0028] Example 1: Detection and Data Acquisition of Plasma Metabolites

[0029] Step 1: Instrument and reagent preparation: Ultra-high performance liquid chromatography-mass spectrometry (UPLC-MS) system (SCIEX ExionLC tandem Zeno TOF 7600), chromatographic column (Waters HSS T3), HPLC grade reagents such as acetonitrile, methanol, and formic acid.

[0030] Step 2: Cohort Sample Preparation: 298 patients scheduled for hysteroscopic surgery under propofol anesthesia were prospectively enrolled and divided into no-RD, mild-RD, moderate-RD, and severe-RD groups according to the degree of respiratory depression (RD) during anesthesia. EDTA anticoagulated plasma was collected before anesthesia and frozen at -80℃.

[0031] Step 3: Plasma sample pretreatment: Add 400 μL of a 1:1 methanol-acetonitrile mixed extraction solution, vortex to mix, and precipitate proteins; centrifuge at 12000 r / min for 15 min at 4℃, collect the supernatant and vacuum dry; reconstitute with 50% acetonitrile aqueous solution, filter through a 0.22 μm organic phase filter membrane, and then proceed with analysis.

[0032] Step 4: UPLC-MS data acquisition: Inject the reconstituted supernatant into the UPLC-MS system and perform full scan and secondary mass spectrometry acquisition using optimized chromatographic gradient and mass spectrometry parameters (positive and negative ion mode, m / z = 60-1200).

[0033] Step 5: Data preprocessing and identification: Peak extraction and alignment were performed using Progenesis QI software, and metabolite annotation was performed based on the HMDB database, resulting in a total of 740 metabolic features.

[0034] Example 2: Screening and Performance Validation of 14-Metabolite Biomarker Combinations

[0035] Step 1: Feature Screening: Metabolic data from 114 samples (n=61) in the severe renal impairment (RD) group and the non-RD group (n=53) were analyzed using a multi-step screening strategy (LASSO score ≥ 0.04, p < 0.05 and AUC > 0.65, see...). Figure 1 ).

[0036] Step 2: After multiple screening steps, a 14-metabolite biomarker combination consisting of 6 downregulated metabolites and 8 upregulated metabolites was finally obtained (see Table 1 for specific metabolite information).

[0037] Step 3: Build a prediction model: The 114 samples were randomly divided into a discovery queue (n=76) and a validation queue (n=38) in a 2:1 ratio. Five machine learning algorithms, namely Support Vector Machine (SVM), Lasso Regression (LR), Neural Network (NN), Elastic Network (EN), and Ridge Regression (RR), were used to build a prediction model.

[0038] Step 4: Validate model performance: The 14-marker combination showed high predictive performance in both the discovery and validation cohorts (see Table 2), with ROC-AUC of 0.86-0.91 in both the discovery and validation cohorts, and this performance was independent of patient age.

[0039] Step 5: Screening of key metabolites: Correlation analysis of 14 biomarkers from 298 patients revealed that α-linolenic acid and 13-hydroxyoctadecadienoic acid were significantly associated with respiratory depression (p < 0.05). Therefore, α-linolenic acid and its upstream metabolite linoleic acid were selected as key metabolites for subsequent functional validation.

[0040] Table 1. Information on 14 metabolic biomarkers

[0041] Table 2. Diagnostic performance of 14 metabolic biomarkers in five typical machine learning algorithms.

[0042] Example 3: In vivo functional validation of the key metabolite (α-linolenic acid)

[0043] Step 1: Animal model preparation: A propofol (50 mg / kg) induced respiratory depression model was established using male C57BL / 6 mice and a whole-body volume plethysmography system, with untreated mice and mice treated with corresponding concentrations of DMSO solution as controls.

[0044] Step 2: Pretreatment Experiment: 15 minutes before propofol injection, α-linolenic acid (100 mg / kg), linoleic acid (100 ug / kg), and 13-hydroxyoctadecadienoic acid (100 ug / kg) were administered via tail vein injection. Results showed that only α-linolenic acid significantly improved respiratory function at different time points, including increases in respiratory rate, tidal volume, and minute ventilation. Figure 2 It also shortened the anesthesia recovery time (p<0.05).

[0045] Step 3: Treatment Experiment: Five minutes after propofol injection, alpha-linolenic acid (100 mg / kg), linoleic acid (100 ug / kg), and 13-hydroxyoctadecadienoic acid (100 ug / kg) were administered via tail vein. Results showed that alpha-linolenic acid treatment significantly improved respiratory function, including increases in respiratory rate, tidal volume, and minute ventilation. Figure 3 It also shortened the anesthesia recovery time (p<0.05).

[0046] The technical solution of this invention has significant practicality and industrialization prospects, mainly reflected in the following aspects: 1. Significant clinical application value The plasma metabolic biomarker combination obtained by this invention can accurately predict the risk of propofol-induced severe respiratory depression before anesthesia. In clinical validation, the AUC reached 0.86–0.91, and the predictive efficacy was stable and reliable. At the same time, this invention also found that α-linolenic acid has a functional effect of improving respiratory depression, which enables it not only to achieve preoperative risk identification, but also to be used for perioperative intervention, and has the clinical value of "prediction + treatment".

[0047] 2. The testing platform is mature and its operation is replicable.

[0048] This invention employs ultra-high performance liquid chromatography-mass spectrometry (UPLC-MS) for the detection of preoperative plasma metabolites. The method is mature, highly sensitive, and requires only 0.25 mL of blood sample for detection, without any additional invasive procedures. The detection process, quality control methods, and model calculations have all been standardized, allowing for rapid deployment in hospital laboratories and third-party medical testing laboratories, demonstrating excellent feasibility.

[0049] 3. Possesses a clear transformation path and industrialization feasibility.

[0050] The technical solution of this invention has been validated in clinical cohorts, and the metabolite panel has been locked and the algorithm model has been fixed, making it ready to be transformed into a "mass spectrometry detection kit + intelligent analysis software" or medical software. The key metabolite, α-linolenic acid, as a safe fatty acid, also has the potential to be developed into a preoperative nutritional supplement or adjuvant drug. The overall technology chain is complete, conforms to the registration path for medical devices / clinical testing projects, and has a lower transformation threshold than drug development.

[0051] 4. Clear market demand and wide range of application scenarios

[0052] Propofol, the most commonly used intravenous anesthetic globally, is widely used in surgical anesthesia, painless endoscopy, day surgery, and ICU sedation, with over 200 million uses worldwide annually. The incidence of respiratory depression is 5-15%, but preoperative assessment tools and effective preventative measures are still lacking. The technical solution of this invention can be widely applied in surgical anesthesia, painless endoscopy, day surgery centers, and ICU sedation management, helping to screen high-risk individuals, guide anesthetic dosage, and reduce postoperative respiratory complications. It has clear market demand and potential for widespread application.

[0053] In summary, this invention has significant advantages in terms of clinical application value, technological advancement, and industrialization foundation, and possesses good industrialization prospects and market potential.

[0054] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A combination of plasma metabolic biomarkers for predicting and intervening in propofol-induced respiratory depression, characterized in that, The plasma metabolic biomarker combination consists of the following 14 metabolites: 4 - Allylphenol sulfate, 13-hydroxyoctadecadienoic acid, α-linolenic acid, hexadecanoic acid, ibostane, corticosterone, aspirin, Valdiate, hist-lysine, 3-Op-coumarylquinic acid, P-type 18:0 / 22:6(4Z,8Z,10Z,13Z,16Z,19Z)-OH(7) phosphatidylcholine, 20:4(5Z,8Z,11Z,14Z)-OH(16R) / 22:0 type phosphatidylethanolamine, P-type 18:0 / 20:3(8Z,11Z,14Z) phosphatidylethanolamine, 20:0 / 20:3(6,8,11)-OH(5) type phosphatidylethanolamine.

2. The plasma metabolic biomarker combination as described in claim 1, characterized in that, The 14 metabolites are divided into 6 downregulated metabolites and 8 upregulated metabolites. The downregulated metabolites include 4-allylphenol sulfate, 13-hydroxyoctadecadienoic acid, α-linolenic acid, hexadecanoic acid, ibostane, and corticosterone. The upregulated metabolites include acetosiline, valdiate, histidine-lysine dipeptide, 3-Op-coumarylquinic acid, P-type 18:0 / 22:6(4Z,8Z,10Z,13Z,16Z,19Z)-OH(7) phosphatidylcholine, 20:4(5Z,8Z,11Z,14Z)-OH(16R) / 22:0 type phosphatidylethanolamine, P-type 18:0 / 20:3(8Z,11Z,14Z) phosphatidylethanolamine, and 20:0 / 20:3(6,8,11)-OH(5) type phosphatidylethanolamine.

3. The plasma metabolic biomarker combination as described in claim 1, characterized in that, Among the 14 metabolites, α-linolenic acid had the smallest p-value among the downregulated metabolites, with an intergroup comparison p-value of 2.87E-05. Its expression level in the severe respiratory depression group was significantly lower than that in the non-respiratory depression group.

4. The detection method of the plasma metabolic biomarker combination according to claim 1, characterized in that, Includes the following steps: Step 1: Patients undergoing hysteroscopic surgery who received propofol anesthesia were prospectively enrolled and divided into four groups: no RD, mild, moderate, and severe RD. EDTA anticoagulated plasma was collected before anesthesia and frozen at -80°C. Step 2: Take 100 μL of plasma, add 400 μL of methanol / acetonitrile solution with a volume ratio of 1:1, extract the protein, centrifuge and take the supernatant, vacuum dry, and then reconstitute with 50% acetonitrile aqueous solution. Step 3: Inject the reconstituted supernatant into the UPLC-MS system, and perform full scan and secondary mass spectrometry acquisition using optimized chromatographic gradient and mass spectrometry parameters in positive and negative ion mode, m / z 60-1200. Step 4: Peak extraction and alignment were performed using Progenesis QI software, and metabolite annotation was performed based on the HMDB database, resulting in a total of 740 metabolic features. Step 5: Screening of 740 metabolic features. Based on the conditions of LASSO score not less than 0.04, intergroup comparison p value less than 0.05, and single marker AUC > 0.65, the plasma metabolic biomarker combination described in claim 1 is obtained.

5. The detection method for the combination of plasma metabolic biomarkers as described in claim 4, characterized in that, The method for verifying the performance of the plasma metabolic biomarker combination includes: dividing 298 subject samples into a discovery cohort and a validation cohort at a ratio of 2:1, and using the method described in claim 4 to detect the expression levels of 14 metabolites in the two groups of samples, and using the area under the ROC curve, sensitivity, and specificity as evaluation indicators to verify the predictive performance of the biomarker combination.

6. The detection method for the combination of plasma metabolic biomarkers as described in claim 4, characterized in that, The diagnostic performance of 14 metabolites was evaluated using five typical machine learning algorithms: support vector machine, logistic regression, neural network, elastic network, and ridge regression.

7. The detection method for the combination of plasma metabolic biomarkers as described in claim 4, characterized in that, The in vivo functional verification steps for α-linolenic acid in the aforementioned combination are as follows: S1: Animal model preparation: A respiratory depression model was established by intraperitoneal injection of 50 mg / kg propofol using male C57BL / 6 mice and a whole-body volume plethysmography system; S2: Pretreatment experiment: 15 minutes before propofol injection, 100 mg / kg α-linolenic acid was injected via the tail vein, and the respiratory rate, tidal volume and anesthesia recovery time of the mice were detected and recorded. S3: Treatment Experiment: 100 mg / kg α-linolenic acid was injected via the tail vein 5 minutes after propofol injection, and the recovery of respiratory function in mice was dynamically monitored.

8. The method for detecting the combination of plasma metabolic biomarkers as described in claim 7, characterized in that, Pretreatment experiments showed that α-linolenic acid significantly reduced propofol-induced respiratory depression (p<0.001) and shortened anesthesia recovery time (p<0.05); treatment experiments showed that α-linolenic acid rapidly improved respiratory function in mice (p<0.01).

9. The use of the plasma metabolic biomarker combination of claim 1 in predicting propofol-induced respiratory depression.

10. The use of the plasma metabolic biomarker combination of claim 1 in the preparation of a drug to intervene in propofol-induced respiratory depression, characterized in that, The plasma metabolic biomarker is α-linolenic acid.

Images