Shreya Pandey (Centre of Biomedical Research, India)
LinkedIn: Shreya Pandey; X: @shreyapandey171
Abstract: The landscape of chronic liver disease has changed significantly, with metabolic dysfunction-associated steatotic liver disease(MASLD) now emerging as the most widespread form worldwide. In Asia, particularly in India, the prevalence of MASLD is increasing, largely driven by poor dietary habits and a sedentary way of life. MASLD spans from fat deposition to inflammation and fibrosis. Fibrosis stands out as the most critical indicator of liver-related complications and overall risk of death in MASLD. Early identification of fibrosis is critical, but current tests are often invasive or unreliable. While studies have explored metabolic changes in MASLD, few have focused on distinguishing early-stage fibrosis from steatosis.
In this study, we used NMR-based metabolomics to analyse serum samples from n = 103 MASLD patients, divided into fibrosis (n = 44) and non-fibrosis (n = 59) groups based on standard non-invasive scoring systems. We identified seven metabolites—arginine, glycerol, aspartate, glucose, phenylalanine, histidine, and citrate—that significantly differed between the two groups and showed good diagnostic potential (AUROC> 0.70). Pathway analysis revealed disruptions in arginine and nitrogen metabolism, associated with liver scarring processes, and in energy and lipid metabolism, pointing to mitochondrial dysfunction and lipotoxic stress. Reduced aspartate levels also suggested loss of natural protection against fibrosis.
This is the first study of MASLD cohort to differentiate early-stage fibrosis from steatosis using metabolomics. Our findings highlight the potential of a simple NMR based blood test to aid early diagnosis, guide treatment decisions, and personalize care—offering a non-invasive alternative to improve MASLD management.
-
Thanks for the nice presentation. I have following questions related to the presentation:
1. As per the study enhanced level of arginine leads to increased proline synthesis. Is this enhanced proline level reflected in metabolic profile?
2. How is the lipid profile? Are there specific lipids which are changing?
3. How does enhanced level of collagen synthesis lead to fibrosis?
4. What portion of the result was used in machine learning?Thanks again
-
Thank you sir.
1- Yes sir, even the proline level was enhanced in the NMR profiling , however it didnot match the criteria to be considered as significant metabolite ( despite having p value 1 , the AUC value was 0.65 so we had to exclude it).
2- The few lipids that we obtained using diffusion edited pulse program also profile had significant difference. -CO-CH2-CH2- (corresponding to cholesterol and FA{TAG and Phospholipids}) was found to be increased in fibrotic cohort, similarly PUFA was found to be decreased in fibrotic cohort. As we did using NMR we have limited data corresponding to lipids. Once we use LC we might get broader insights which we will be starting soon
3- When there is continuous injury or inflammation to liver , Hepatic stellate cells gets activated due to cascade of events. These HSCs are major contributers for collagen. When the level of collagen increases , it starts accumulating in the liver , distorting the normal structure and function and liver and eventually forming scarred tissue. This condition is called fibrosis.
4- Only the data from bins of significant metabolites was used. We excluded the water region and the regions that were not significant
-
-
Thanks for a nice presentation. I have following questions:
1. As shown in the presentation the enhanced level of arginine leads to increased collagen production via increased proline level. Is increased proline reflected in the NMR profiling?
2. How is does lipid profile look? Any specific lipids which are enhanced?
3. How does increased collagen lead to liver fibrosis?
4. What exact data was used in machine learning?Thanks again
-
Thank you sir.
1- Yes sir, even the proline level was enhanced in the NMR profiling , however it didnot match the criteria to be considered as significant metabolite ( despite having p value 1 , the AUC value was 0.65 so we had to exclude it).
2- The few lipids that we obtained using diffusion edited pulse program also profile had significant difference. -CO-CH2-CH2- (corresponding to cholesterol and FA{TAG and Phospholipids}) was found to be increased in fibrotic cohort, similarly PUFA was found to be decreased in fibrotic cohort. As we did using NMR we have limited data corresponding to lipids. Once we use LC we might get broader insights which we will be starting soon
3- When there is continuous injury or inflammation to liver , Hepatic stellate cells gets activated due to cascade of events. These HSCs are major contributers for collagen. When the level of collagen increases , it starts accumulating in the liver , distorting the normal structure and function and liver and eventually forming scarred tissue. This condition is called fibrosis.
4- Sir, the data from the binned sheet that we obtained from chenomx was used in machine learning learning.
-
-
Interesting work! Could you tell the pulse program used? What data did you use for building the model ? Is NMETA available online
-
Thank you Daniel, we have used CPMG pulse program which is basically used to suppress large molecules. We have used the binned sheet generated using chenomx for creating model. As far as NMETA is concerned, it is not available online we are still working on it.
-
-
The presentation looks very informative but can you answer me the following question: What is NMeta? What all information does it require?
-
Thank you, Karthik,
NMETA is a web-based application we are currently working on as part of our effort to develop a non-invasive alternative to liver biopsy.
The process is simple: perform a 1D NMR experiment on a serum sample and upload the resulting spectrum to our webpage. The tool will then provide the probability of the individual having liver fibrosis.
-
-
Hello Shreya!
Interesting work, I was wondering how did you handle the large lipo protein signals arising from the blood samples. Did you filter them out? Is there any evidence of these proteins to be a biomarker of the disease?-
Thank you, Dr. Schiavina,
We have not filtered the serum as we had to perform diffusion edited experiment on the same serum sample. Instead, we have used the CPMG pulse sequence to suppress signals from large molecules, particularly lipoproteins and lipid fragments, thereby minimizing their interference.
While previous studies have compared MASLD (formerly NAFLD) with hepatocellular carcinoma (reference: ref:-https://www.thelancet.com/journals/ebiom/article/PIIS2352-3964(21)00455-2/fulltext), none have specifically examined non-fibrotic MASH versus early fibrotic MASH within the MASLD cohort. We are currently investigating this comparison and expect to share promising results soon.
-
Leave a Reply