Portal hypertension highly affects the gut-liver crosstalk and the physiology and pathophysiology of both organs. By using an untargeted peptidomics approach, this study identified novel molecular features to be differentially occurring in the blood of patients undergoing TIPS treatment for the reduction of portal hypertension.
Portal hypertension leads to venous congestion in the intestine, disrupting the intestinal barrier. Since the knowledge of gut-liver crosstalk is highly limited, there is an increasing number of studies focusing on analyzing the role of gut-liver crosstalk in chronic inflammatory diseases32. This includes the investigation of the presence of gut dysbiosis in liver diseases like NAFLD, alcoholic liver disease, liver cirrhosis and hepatic carcinogenesis 33, as well as the development of novel therapeutic strategies that target the gut microbiota in liver diseases, e. g. by employing probiotics to restore the eubiosis of the gut and reduce alcoholic steatohepatitis 20. Analyzing the plasma composition of patients affected by portal hypertension will help in the discovery of novel biomarkers that could be used as diagnostic tools and/or therapeutic candidates and will contribute to clarifying the role of gut and liver crosstalk in hepatic diseases.
The cohort considered in this work was selected with the aim of investigating, specifically, the portal hypertension changes after TIPS and the plasma samples were collected within 7-14 days after surgery, which represents the usual recovery window for this type of intervention 34.
Since peptidomics is a powerful tool to discover and determine new biomarkers 23,24 as well as to identify novel peptides to be employed in prevention and prediction models 35, we developed and applied a workflow to prepare, analyze and interpret peptidomics data by the employment of mass spectrometry and statistics to investigate the plasma of patients suffering from portal hypertension and liver diseases, comparing the molecular profile before and after TIPS treatment to reduce portal hypertension. The peptidomics approach adopted in this study focused on identifying only features that showed to be significantly different between the two groups analysed (before and after TIPS treatment).
The peptidomics data were analyzed by ´limma´ statistics 36, 37. The data were skewed and ´limma´ assumes a normal distribution, however, it is reported in the literature that independent variables can be treated as normally distributed when the sample size is big enough 38. Moreover, several studies proved that a sample size ranging from 15 to 80 removed the effect of skewness on t-statistics39 and a parametric test can be applied to sample sizes smaller than 50 40. Given these premises, the distribution of the molecular features was treated as approximately normal and moderated t-statistics were applied without disregarding eventual assumptions of the ´limma´ statistics. Peptidomics/metabolomics data are characterized by 10% to 50% missing or zero intensities values 41. The molecular features were defined by an RT and m/z and when the signal of the molecular feature was below the limit of detection, this was registered as zero intensity. Given that, when analyzing the differences within the study cohorts, the zero intensities were maintained, as they were most likely underlying diversities in the groups. This led to the identification of features that were mostly depleted after TIPS treatment and one feature that, instead, increased.
Correlation analyses showed that for at least the ∆s of four molecular features (187.8, 863.4, 943.0, 951.4 each m/z) there is a significant correlation with respectively leucocytes amount, ALT, serum creatinine and PHPG. However, due to the smaller size of the dataset and high presence of zero values intensities, these results will need to be validated in an independent wider cohort of patients, but show, anyhow, the potential of the novel biomolecules as useful biomarkers in the contest of the portal hypertension diagnosis and treatment.
The focus of this study was to discover biomolecules influenced by TIPS. All the identified differentially expressed molecular features, except one (154.2 m/z), resulted to be significantly decreased after TIPS intervention. From these 11 biomolecules, four were identified and sequenced as protein fragments (Table 3). Sialomucin core protein 24 (CD164), a small 97 amino acids protein with different isoforms involved in haematopoiesis 42, moreover, in line with its highest expression in the pathological context of portal hypertension before TIPS treatment, it was found to be increasingly expressed in liver tissue in pro-inflammatory, high-cholesterol conditions 43. Meckelin is a protein important for ciliary structure and function 44, and the gene (TMEM67) encoding the meckelin protein has been reported to cause Joubert/Meckel syndrome when mutated, a condition which causes nephronophthisis and hepatic fibrosis 45. Histone-lysine N-methyltransferase (MLL3, KMT2C) regulates metabolic processes and the hepatic circadian control of bile acid homeostasis 46 and was also reported to play a role in fatty liver development 47. Transient receptor potential cation channel subfamily V member 5 (TRPV5) regulates Ca2+ reabsorption in the kidney and intestine 48,49, with Ca2+ reported playing a key in the protection of the liver from NAFLD 50.
The remaining molecular features were metabolites derived from gut and liver metabolism and/or previously related to liver (patho)physiology (Table 3). Dopamine is a neurotransmitter and 50% of total dopamine is produced in the gastrointestinal tract by enteric neurons and intestinal epithelial cells51. Dopamine is involved in cell immunity regulation and has been shown to play an important role in the suppression of autoimmune hepatitis 51. Uric acid has also been investigated concerning liver pathophysiology. It is produced in part by the liver and induces hepatic steatosis through mitochondrial oxidative stress production 52. Also in patients, higher serum levels of uric acid are associated with a higher probability of developing NAFLD53. The metabolite 3-carboxy-4-methyl-5-propyl-2-furanpropionic acid (CMPF) is a known derivate from furan fatty acid metabolism 54. High plasma levels of this metabolite are usually associated with type 2 diabetes mellitus 55 and serum levels of CMPF negatively correlated with lipid metabolism, based on which CMPF was suggested as a potential NAFLD prediction marker56. Mechanistically, CMPF has been reported to promote the production of reactive oxygen species and diabetes development in animal models 55, whereas others found CMPF to prevent and reverse liver steatosis 57,58, thus urging further investigations into a potential pathophysiological role of CMPF. Homoarginine is negatively correlated with mortality risk and the ‘Model for End-Stage Liver Disease’ (MELD) score in patients with liver diseases59, with potential beneficial effects through nitric oxide regulation suggested59. Leucine derivates such as leucylproline levels resulted to be higher in samples from patients affected by NAFLD-cirrhosis 5. Finally, 5-methyluridine, a metabolite found in humans 61 and E. Coli 62, hasn´t been so far linked to liver cirrhosis, NAFLD, NASH or portal hypertension.
In summary, 11 low molecular weight mediators resulted to be differentially expressed in plasma samples derived from peripheral veins before compared to after TIPS. The concentration of ten biomolecules in plasma was significantly decreased and one low molecular mediator (154.2 m/z), which so far remains unidentified, significantly increased after TIPS treatment. These biomolecules now require further investigation in relation to gut-liver crosstalk since they are crucial in the understanding of this organ crosstalk in pathophysiology and might indicate an effect of TIPS on intestinal permeability, the inflammatory process and/or liver pathology. Regarding the sequenced peptides, the parent proteins MLL3 and CD164 were shown to be involved in fatty liver development and NAFLD 47 or were upregulated in the liver in high cholesterol conditions 43, respectively, leading to the hypothesis that in portal hypertensive patients increased levels of peptides from these proteins reflect scarring of the liver. TRPV5, instead, might play a protective role in the liver. No clear conclusions, instead, can be drawn for meckelin. Uric acid is reported as a possible marker for NAFLD, whereas contradictory information or no information is currently available regarding the function of homoarginine, leucylproline, CMPF and 5-methyluridine in relation to liver pathology. In our patient cohort, these metabolites resulted drastically decreased or depleted after the TIPS treatment, suggesting a biomarker potential for portal hypertension and an influence of TIPS on gut released metabolites and through this, potentially, liver pathology and/or gut barrier permeability.