Presepsin as a biomarker of bacterial translocation and an indicator for the prescription of probiotics in cirrhosis

Abstract

BACKGROUND

The gut–liver axis and bacterial translocation are important in cirrhosis, but there is no available universal biomarker of cellular bacterial translocation, for which presepsin may be a candidate.

AIM

To evaluate the relationship of the blood presepsin levels with the state of the gut microbiota in cirrhosis in the absence of obvious infection.

METHODS

This study included 48 patients with Child–Pugh cirrhosis classes B and C and 15 healthy controls. The fecal microbiome was assessed using 16S rRNA gene sequencing. Plasma levels of presepsin were measured. A total of 22 patients received a probiotic (Saccharomyces boulardii) for 3 months.

RESULTS

Presepsin levels were higher in patients with cirrhosis than in healthy individuals [342 (91-2875) vs 120 (102-141) pg/mL; P = 0.048]. Patients with elevated presepsin levels accounted for 56.3% of all included patients. They had lower levels of serum albumin and higher levels of serum total bilirubin and overall severity of cirrhosis as assessed using the Child–Pugh scale. Patients with elevated presepsin levels had an increased abundance of the main taxa responsible for bacterial translocation, namely Bacilli and Proteobacteria (including the main class Gammaproteobacteria and the minor taxa Xanthobacteraceae and Stenotrophomonas), and a low abundance of bacteria from the family Lachnospiraceae (including the minor genus Fusicatenibacter), which produce short-chain fatty acids that have a positive effect on intestinal barrier function. The presepsin level directly correlated with the relative abundance of Bacilli, Proteobacteria, and inversely correlated with the abundance of Lachnospiraceae and Propionibacteriaceae. After 3 months of taking the probiotic, the severity of cirrhosis on the Child–Pugh scale decreased significantly only in the group with elevated presepsin levels [from 9 (8-11) to 7 (6-9); P = 0.004], while there were no significant changes in the group with normal presepsin levels [from 8 (7-8) to 7 (6-8); P = 0.123]. A high level of presepsin before the prescription of the probiotic was an independent predictor of a greater decrease in Child–Pugh scores (P = 0.046), as well as a higher level of the Child–Pugh scale (P = 0.042), but not the C-reactive protein level (P = 0.679) according to multivariate linear regression analysis.

CONCLUSION

The level of presepsin directly correlates with the abundance in the gut microbiota of the main taxa that are substrates of bacterial translocation in cirrhosis. This biomarker, in the absence of obvious infection, seems important for assessing the state of the gut–liver axis in cirrhosis and deciding on therapy targeted at the gut microbiota in this disease.

Key Words: Dysbiosis; Gut; Intestinal permeability; Leaky gut; Gut-liver axis; Liver; Microbiota

Core Tip: The gut–liver axis and bacterial translocation are important in cirrhosis; however, there is no available universal biomarker for cellular bacterial translocation, although presepsin may be a candidate. The level of presepsin directly correlated with the abundance in the gut microbiota of the main taxa that are substrates of bacterial translocation in cirrhosis. This biomarker, in the absence of obvious infection, seems important for assessing the state of the gut–liver axis in cirrhosis to decide on therapy targeted at the gut microbiota in this disease.

INTRODUCTION

Cirrhosis is the final stage of chronic liver disease, and its development makes the prognosis poor[1-4]. Currently, the gut–liver axis is of great importance in the pathogenesis of cirrhosis[5-9]. An increase in intestinal permeability and changes in the composition of the intestinal microbiota (gut dysbiosis) occur in cirrhosis. As a manifestation of gut dysbiosis, the number of facultative anaerobes, i.e., bacteria that can survive in oxygenated living human tissues, increases. The facultative anaerobes of the gut microbiota include most representatives of the class Bacilli and the phylum Proteobacteria that are minor taxa in the normal gut microbiota but become much more abundant in cirrhosis[5-9]. These changes lead to bacterial translocation represented by the penetration of bacteria and their components from the intestinal contents into the intestinal wall, lymph nodes, ascitic fluid, and portal and systemic blood flow[5,10-13]. Bacterial translocation triggers a pathological cascade, which results in the worsening of portal hypertension and decreased liver function[5,10-13]. Despite this important role of bacterial translocation, there are currently no adequate biomarkers to evaluate it in clinical practice[14]. Traditionally, lipopolysaccharide (LPS), which is an endotoxin of Gram-negative bacteria, was thought to play this role[14]. Its advantage is that its presence in the blood is a direct manifestation of bacterial translocation. However, its disadvantage is that it reflects only the translocation of Gram-negative bacteria, since Gram-positive bacteria do not contain LPS. In addition, the detection of LPS does not indicate the penetration of bacteria into the body, since this molecule can enter the bloodstream through damaged tight junctions of the intestinal epithelium separately from bacterial cells after their death and destruction. Other candidates for the role of biomarkers of bacterial translocation (LPS-binding protein, procalcitonin, CD14, and others), as a rule, represent various manifestations of the systemic inflammatory response, which causes by bacterial translocation[14]. Presepsin is interesting in this regard. This protein is a fragment of CD14, a molecule that is involved in the recognition of conserved molecular patterns of both Gram-positive and Gram-negative bacteria. CD14 participates in the phagocytosis of recognized bacterial cells, plunging with them into the phagosome. After fusion with lysosomes, their enzymes cleave off the CD14 fragment, which is ejected from the cell outwards. This ejected fragment is presepsin. Experiments have shown that the formation of presepsin is stimulated by living and dead cells of both Gram-positive and Gram-negative bacteria, while their noncellular components, such as LPS, do not have such properties. Therefore, presepsin can be considered a biomarker of phagocytosis of whole bacterial cells, both Gram-negative and Gram-positive[15-19]. Presepsin is successfully used in clinical practice as a biomarker of bacterial infection and sepsis[15-19], including in patients with cirrhosis[20-24]. Its blood level is higher in decompensated cirrhosis than in compensated cirrhosis[24-26]. Its high blood level determines the poor prognosis of patients with cirrhosis[23-27]. We suggest that in the absence of obvious infection, presepsin can also be used as a universal biomarker of cellular bacterial translocation in cirrhosis. The aim of this study was to evaluate the relationship of the blood presepsin levels with the state of the gut microbiota in cirrhosis in the absence of obvious infection.

MATERIALS AND METHODS

Patients

Patients with cirrhosis classes B and C according to their Child–Pugh score who presented to our clinic for routine examination were screened for participation in this study. The study procedures were explained to the potential participants, and written informed consent was obtained before enrollment. The study was approved by the Ethics Committee of Sechenov University and performed in accordance with the Declaration of Helsinki.

The inclusion criteria were the presence of cirrhosis, the diagnosis of which was made based on histology or a combination of physical examination, laboratory and instrumental data, signed written informed consent, and an age of between 18 years and 70 years. The exclusion criteria were as follows: Use of drugs that could affect the composition of the gut microbiota (lactulose, lactitol, or other prebiotics, probiotics, antibiotics, and metformin) in the preceding six weeks; alcohol consumption in the preceding six weeks; current bacterial infection; inflammatory bowel disease, cancer, renal failure, or any other serious disease.

The patient was considered free of current bacterial infection if he had a normal body temperature, no open wounds, no complaints of respiratory (cough, shortness of breath), intestinal (diarrhea, abdominal pain) or urinary (lower back or abdominal pain, urinary disorder) infection, normal chest x-ray and routine urinalysis.

The control group consisted of 15 healthy individuals who visited the clinic for routine health examinations.

Investigations

The day after the initial medical examination, fasting blood was collected from patients and immediately centrifuged. The plasma was separated, divided into several aliquots, and frozen. Once all patients were recruited, aliquots were thawed and the levels of presepsin were assessed by enzyme immunoassay. Assays were performed according to the manufacturers’ instructions. The blood plasma of healthy controls was examined in the same way.

All patients underwent a standard examination that included physical and neurological examination, abdominal ultrasound, complete blood count, blood chemistry, coagulation tests, and a number connection test for the diagnosis of covert hepatic encephalopathy.

Gut microbiota analysis

On the same day as the blood collection, after voluntary defecation by patients, stool samples were collected in a special disposable sterile container and immediately frozen at -80 ℃ for 16S rRNA gene sequencing. Gut microbiota analysis was performed as previously described (Supplementary material)[28-32].

Statistical analysis

Statistical analysis was performed with Statistica 10 (StatSoft Inc., Tulsa, United States) software. The data are presented as medians (interquartile ranges). The Mann-Whitney method was used to assess the differences between continuous variables. The differences between categorical variables were assessed with Fisher’s exact test. Spearman’s rank test was used to assess the correlations between variables. The assessment of changes in the levels of variables was performed with the Wilcoxon test. Comparison of the composition of the gut microbiota between groups was carried out by linear discriminant analysis effect size (LEfSe) using the online resource. P value ≤ 0.05 were considered as statistically significant.

RESULTS

Fifty patients were included in the study. Fecal 16S rRNA gene sequencing was not possible in two patients due to technical problems; these patients were excluded from the study. In total, 48 cirrhosis patients and 15 healthy individuals were included (Figure 1). Patients and controls did not differ in age [50 (43–58) years vs 46 (39–54) years; P=0.194], body mass index [25.7 (23.1-28.9) kg/m²vs 25.0 (23.7–25.8) kg/m²; P = 0.294], and sex distribution (male/female: 19/29 vs 7/8; P = 0.423).

Figure 1  Flow diagram.

Presepsin levels were higher in patients with cirrhosis than in healthy individuals [342 (91-2875) pg/mL vs 120 (102-141) pg/mL; P = 0.048]. The mean and σ for presepsin in healthy individuals was 118 pg/mL and 27 pg/mL, which allowed us to use the m+-2σ rule to estimate the normal interval for presepsin in our study from 64 pg/mL to 172 pg/mL; this included all the results of healthy individuals, with the exception of one that was slightly below (61 pg/mL) this lower limit.

Based on this, cirrhotic patients were divided into a group with increased presepsin (> 172 pg/mL) and a group with normal presepsin (< 172 pg/L). Patients with elevated presepsin levels accounted for 56.3% of all included patients. They had lower levels of serum albumin and higher levels of serum total bilirubin and overall severity of cirrhosis as assessed using the Child–Pugh scale. Moreover, the C-reactive protein level, despite the obvious tendency to be higher in patients with increased presepsin levels, did not differ significantly between groups (Table 1).

Table 1 Main characteristics of cirrhosis patients with increased and normal presepsin levels, n (%).

	Patients with increased presepsin level	Patients with normal presepsin level	P value
Age, yr	49 (43-59)	52 (42-56)	0.868
Body mass index, kg/m2	25.9 (24.7-30.1)	25.3 (21.6-28.4)	0.275
Male/female	12/15	7/14	0.316
Etiology of cirrhosis: Alcohol	13 (48.1)	10 (47.6)	> 0.050
HBV	3 (11.1)	-
HCV	3 (11.1)	3 (14.3)
Metabolic dysfunction-associated steatotic liver disease	2 (7.4)	1 (4.8)
Mixed	4 (14.8)	3 (14.3)
Cryptogenic	2 (7.4)	4 (18.0)
Child–Pugh score	9 (8-10)	8 (7-9)	0.009
Esophageal varices (Grade 1)	9 (33.3)	9 (42.9)	0.353
Esophageal varices (Grade 2-3)	15 (55.5)	10 (47.6)	0.399
Minimal hepatic encephalopathy	19 (70.4)	14 (66.7)	0.588
Overt hepatic encephalopathy	1 (3.7)	3 (14.3)	0.215
Ascites	21 (77.7)	16 (76.2)	0.582
Ascites (Grade 1)	11 (40.7)	12 (57.1)	0.201
Ascites (Grade 2-3)	10 (37.0)	4 (19.0)	0.149
Serum total protein, g/L	73 (65-76)	70 (61-74)	0.344
Serum albumin, g/L	32 (30-34)	35 (33-37)	0.049
Serum total bilirubin, μmol/L	60 (36-78)	37 (26-57)	0.022
International normalized ratio	1.63 (1.39-1.75)	1.48 (1.36-1.61)	0.151
Red blood cells, 1012 cell/L	3.6 (3.1-4.2)	3.6 (3.3-4.2)	0.934
White blood cells, 109 cell/L	3.7 (2.6-5.2)	3.3 (2.6-5.9)	0.942
Platelets, 109 cell/L	94 (52-104)	94 (66-105)	0.64
Serum creatinine, μmol/L	77 (67-91)	76 (66-107)	0.625
Serum sodium, mmol/L	141 (139-142)	141 (140-142)	0.983
Alanine aminotransferase, U/L	26 (18-50)	28 (20-45)	0.723
Aspartate aminotransferase, U/L	52 (41-77)	48 (28-51)	0.151
Gamma glutamyl transferase, U/L	72 (44-114)	68 (37-269)	0.575
Alkaline phosphatase, U/L	261 (194-330)	242 (215-388)	0.506
C-reactive protein, mg/L	9 (5-15)	7 (3-12)	0.399
Splenic length, cm	15.0 (13.0-17.0)	15.0 (13.2-18.4)	0.708

HBV: Hepatitis B virus; HCV: Hepatitis C virus.

Presepsin levels were directly correlated with Child–Pugh scores (r = 0.365, P = 0.011) and C-reactive protein levels (r = 0.483, P = 0.025 for the range of presepsin levels from 1 pg/mL to 1000 pg/mL) and inversely correlated with serum albumin levels (r=-0.340, P = 0.018).

LEfSe showed that in the group of cirrhosis patients with elevated presepsin there was an increased abundance of the main taxa responsible for bacterial translocation, namely Bacilli and Proteobacteria (including the main class Gammaproteobacteria and the minor taxa Xanthobacteraceae and Stenotrophomonas), and a low abundance of bacteria from the family Lachnospiraceae (including the minor genus Fusicatenibacter), which produce short-chain fatty acids that have a positive effect on intestinal barrier function (Figure 2A). In addition, high levels of presepsin in cirrhosis were associated with higher levels of minor taxa Erysipelatoclostridiaceae and Solobacterium and lower levels of minor taxa Holdemania and Alistipes.

Figure 2 Comparison of the gut microbiota of cirrhotic patients. A: Comparison of the gut microbiota of cirrhotic patients using LefSe; B: Comparison of the gut microbiota of cirrhotic patients at the level of major classes. LefSe: Linear discriminant analysis effect size; LDA: Linear discriminant analysis.

At the level of gut microbiota classes, cirrhosis patients with high presepsin levels had a higher proportion of Bacilli and Gammaproteobacteria (the main class of the Proteobacteria phylum in the human gut microbiota) and a lower proportion of the Clostridia class (containing the family Lachnospiraceae) (Figure 2B).

The presepsin level directly correlated with the relative abundance of Bacilli, Proteobacteria, taxa included in this phylum, and some minor taxa, and inversely correlated with the abundance of Lachnospiraceae, Propionibacteriaceae, and some minor taxa (Table 2).

Table 2 Significant correlations between blood presepsin levels and gut microbiota taxa.

Direct correlations			Inverse correlations
Taxon rank	Taxon	R, P value	Taxon rank	Taxon	R, P value
Phylum	Proteobacteria	0.316, 0.029	Family	Lachnospiraceae	0.447, 0.0421
Class	Bacilli	0.470, 0.0321		Propionibacteriaceae	-0.302, 0.038
	Gammaproteobacteria	0.351, 0.014	Genus	Fusicatenibacter	-0.297, 0.040
Family	Erysipelatoclostridiaceae	0.384, 0.007		Holdemanella	-0.315, 0.029
	Eubacteriaceae	0.340, 0.018
	Vibrionaceae	0.366, 0.010
	Xanthobacteraceae	0.363, 0.011
Genus	Escherichia-Shigella	0.338, 0.019
	Stenotrophomonas	0.365, 0.011
	Vibrio	0.346, 0.016

¹For the range of presepsin levels from 1 to 1000 pg/mL.

Some patients further participated in a study on the effect of the probiotic (Saccharomyces boulardii) on the course of cirrhosis[33]. We chose this probiotic because it has pleiotropic effects: It reestablishes the gut microbiome after dysbiosis[34], strengthens the intestinal immune barrier[35], improves the trophic function of gut microbiota[36], restores the impaired gut barrier, and protects against bacterial translocation[37] in experimental models and in patients with gut diseases. S. boulardii administration in an experimental mouse model of cirrhosis led to the correction of gut dysbiosis and decreased intestinal permeability, as well as the reduced severity of liver inflammation and fibrosis[38]. That study included 40 patients from this study, but 2 patients included in that study were excluded from this study because the gut microbiota analysis could not be completed due to technical problems, as described at the beginning of this section. We additionally enrolled 10 patients for this study after the closure of that study. Details of that study are described in the published article[33]. Among the patients receiving the probiotic, 13 patients had elevated presepsin levels, and 9 patients had normal levels. After 3 months of taking the probiotic, the severity of cirrhosis on the Child–Pugh scale decreased significantly only in the group with elevated presepsin levels, while there were no significant changes in the group with normal presepsin levels (Table 3). A significant reduction in the severity of cirrhosis (by more than 1 point on the Child–Pugh scale) after taking the probiotic was observed in 9 (69.2%) patients with elevated presepsin levels and only in 2 (22.2%) patients with normal presepsin levels (P = 0.040). In the placebo group (n = 16), such a significant reduction in the severity of cirrhosis was observed in 3 (18.8%) patients, which was not significantly different from the probiotic subgroup with normal presepsin levels (P = 0.610) and was below the probiotic subgroup with elevated presepsin levels (P = 0.008). A high level of presepsin before the prescription of the probiotic was an independent predictor of a greater decrease in Child–Pugh scores (P = 0.046), as well as a higher baseline level of the Child–Pugh scale (P = 0.042), but not the baseline C-reactive protein level (P = 0.679) according to multivariate linear regression analysis. In addition, in the group with elevated presepsin levels, there was a significant decrease in the levels of total bilirubin, aspartate aminotransferase, and alkaline phosphatase, as well as the incidence of ascites and minimal hepatic encephalopathy, as a result of taking the probiotic, in contrast to the group with normal presepsin levels (Table 3).

Table 3 Change in the values of the main cirrhosis indicators as a result of the probiotic (Saccharomyces boulardii) taking, depending on the baseline presepsin level, n (%).

	Elevated baseline presensin level (n = 13)			Normal baseline presensin level (n = 9)
	At the beginning of the study	After the probiotic course	P value	At the beginning of the study	After the probiotic course	P value
Child–Pugh score	9 (8-11)	7 (6-9)	0.004	8 (7-8)	7 (6-8)	0.124
Child–Pugh class, A/B+C	0/13	5/8	0.020	0/9	2/7	0.235
No hepatic encephalopathy	2 (15.4)	8 (61.5)	0.021	2 (22.2)	3 (33.3)	0.500
Ascites	12 (92.3)	5 (38.4)	0.006	8 (88.9)	5 (55.5)	0.147
Serum albumin, g/L	32 (30-34)	33 (32-39)	0.033	35 (33-36)	38 (37-39)	0.008
Serum total bilirubin, μmol/L	52 (36-120)	29 (26-40)	0.016	34 (26-44)	25 (22-35)	0.066
Alanine aminotransferase, U/L	26 (18-47)	26 (18-40)	0.272	30 (21-41)	33 (25-39)	0.813
Aspartate aminotransferase, U/L	60 (37-75)	39 (27-51)	0.012	45 (28-51)	36 (36-42)	0.314
Gamma glutamyl transferase, U/L	92 (44-207)	56 (40-89)	0.196	37 (31-317)	62 (30-115)	0.193
Alkaline phosphatase, U/L	268 (245-355)	210 (187-245)	0.013	261 (227-446)	215 (188-275)	0.110

DISCUSSION

In this study, we found that the presepsin level directly correlated with the relative abundance of Bacilli and Proteobacteria. These bacteria are represented in the gut microbiota by facultative anaerobes, which are not afraid of oxygen and live normally in oxygenated human tissues[34]. An increase in their abundance is a manifestation of cirrhosis-specific gut dysbiosis[39-43]. They were substrates of bacterial translocation in animal models of cirrhosis[44,45], and their DNA dominated among bacterial DNA in the blood of patients with cirrhosis[46,47].

Moreover, we found that the presepsin level inversely correlated with the abundance of bacteria from the family Lachnospiraceae. These bacteria are strict anaerobes and do not contain LPS. They die at a minimum concentration of oxygen in the environment, and thus, their living cells are not able to penetrate living oxygenated tissues; that is, they are not capable of either cellular or molecular bacterial translocation[39]. The Lachnospiraceae abundance in the gut microbiota decreases in cirrhotic gut dysbiosis[42,43].

Although presepsin in cirrhosis has been studied in many publications, none of them have examined its blood level in combination with the composition of the gut microbiota[19-23], which is the novelty and strength of our study. We were able to confirm our hypothesis that in the stable course of cirrhosis without bacterial infection, the blood level of presepsin directly correlates with the abundance of substrates for bacterial translocation in the gut microbiota and inversely correlates with the abundance of bacteria not involved in it. Therefore, we believe that presepsin can be used as a universal biomarker of cellular bacterial translocation.

Our study, in accordance with previously published studies[24-26], has confirmed that the level of presepsin in the blood increases as cirrhosis worsens according to the Child–Pugh scale. Gut dysbiosis[40,48] and disruption of the intestinal barrier[49-53] also worsen as cirrhosis becomes more severe, which may explain this pattern.

We also studied how the level of presepsin before the therapy targeted at the gut microbiota affects its effectiveness. We recently published a study showing that a probiotic has a positive effect on the severity of cirrhosis[33]. In this study, we examined how this effect depended on presepsin levels before the start of the therapy. A total of 59% of patients in the probiotics group had elevated presepsin levels, suggesting they had significant bacterial translocation. The use of the probiotic in these patients had a significant positive effect, as a result of which the value of the Child–Pugh cirrhosis severity scale in them decreased by an average of 2 points. However, the effect of the same probiotic in patients with normal presepsin levels, which suggests minimal bacterial translocation, was insignificant. Multivariate regression analysis confirmed that the positive effect of the probiotic was determined by a higher level of presepsin, regardless of the severity of cirrhosis on the Child–Pugh scale and the level of C-reactive protein. Moreover, the level of the latter, unlike the level of presepsin, did not determine the effectiveness of this therapy.

A limitation of our study is that we did not have the technology to determine the bacterial DNA level in the blood and conduct a correlation analysis between this direct marker of bacterial translocation and presepsin. This is a task for future research. The second limitation of our study was the small number of participants; however, this did not prevent us from obtaining significant results. Larger studies are needed to confirm our findings.

New studies using other probiotics, antibiotics, prebiotics, etc., are also needed to further test our hypothesis that presepsin levels in cirrhosis can be used to guide decisions about gut microbiota-targeted therapies.

CONCLUSION

The level of presepsin, a proposed universal biomarker of cellular bacterial translocation, directly correlates with the abundance in the gut microbiota of the main taxa that are substrates of bacterial translocation in cirrhosis. This biomarker, in the absence of obvious infection, seems important for assessing the state of the gut–liver axis in cirrhosis and deciding on therapy targeted at the gut microbiota in this disease.

ACKNOWLEDGEMENTS

The authors are grateful to the staff of the Department of Hepatology: Alexei Lapshin, Shauki Ondos, Petr Tkachenko, Igor Tikhonov and others.