Acute kidney injury is associated with increased levels of circulating microvesicles in patients with decompensated cirrhosis

Abstract

Background:

Microvesicles (MVs) play a role in inflammation, coagulation, and vascular homeostasis in liver disease.

Aim:

To characterize circulating plasma MVs profile in patients with decompensated cirrhosis and acute kidney injury (AKI).

Methods:

We measured the levels of total, endothelial, platelet, tissue factor (TF) +, leukocyte and hepatocyte MVs by new generation flow-cytometry in a prospective cohort of patients with decompensated cirrhosis with and without AKI.

Results:

Eighty patients with decompensated cirrhosis were recruited (40 each with and without AKI). Patients with cirrhosis with AKI had significantly higher calcein + (total), endothelial, and platelet-MVs. Conversely, TF +, leukocyte, and hepatocyte-MVs were comparable between groups. Resolution of AKI was associated with significantly decreased total and endothelial-MVs that became comparable with those in patients without AKI. Platelet MVs significantly decreased but remained higher compared to patients without AKI. TF + MVs significantly decreased and became lower than patients without AKI. Leukocyte and hepatocyte-MVs remained unchanged. Creatinine (OR 4.3 [95%CI 1.8–10.7]), MELD (OR 1.13 [95%CI 1.02–1.27]), any bleeding (OR 9.07 [95%CI 2.02–40.6]), and hepatocyte-MVs (OR 1.04 [95%CI 1.02–1.07]) were independently associated with 30-day mortality.

Conclusion:

AKI worsened vascular and cellular homeostasis in patients with cirrhosis, particularly by inducing endothelial dysfunction and platelet activation. AKI did not worsen systemic inflammation and hepatocytes activation.

Microvesicles (MVs), also known as ‘microparticles’, are 100–1000 nm diameter-extracellular vesicles (EVs) released by cells into their microenvironment and that eventually enter circulation [1,2].

It has been suggested that, by dispatching a multitude of active biomolecules (i.e. proteins, lipids, DNA/RNA) to selected targets, MVs act as intercellular messengers, both in health and diseases [3]. Specifically, in patients with cirrhosis, plasma circulating MVs demonstrate pro-inflammatory/pro-coagulant properties and potentially contribute to portal hypertension and circulatory dysfunction [4–7].

The surface antigen composition of each type of MVs depends on its cellular origin and on the physiological (or pathophysiological) processes that are responsible for its release. Therefore, by measuring the levels of one specific MVs, one could potentially have diagnostic and prognostic information about disease backgrounds [8].

However, whether MVs are truly mediators of disease progression (pathophysiological role), or potentially useful non-invasive biomarkers (or both) has not yet been thoroughly understood and may perhaps depend on which cell type(s) generate the MVs as well as on their state of activation [9].

Multiple pathophysiological processes that are associated with extracellular release of MVs occur in patients with cirrhosis, including hepatocyte apoptosis, intra-hepatic activation of immune system, and systemic inflammation [10].

Because these processes are associated with disease progression and development of decompensation, it has been proposed to use circulating MVs as non-invasive biomarkers of patients’ outcomes [11–14]. In fact, recent evidence demonstrates that, among different types of MVs, hepatocyte-derived MVs were independent predictors of mortality in patients with cirrhosis [12].

Acute kidney injury (AKI) is a common complication in hospitalized patients with decompensated cirrhosis, occurring in approximately 20% of these patients [15], and it is associated with increased morbidity and mortality [16].

Since AKI both within and outside the context of cirrhosis is associated with profound systemic inflammation and significant alterations of hemostasis [17,18], we hypothesized that the determination of different types of MVs, previously recognized as potential biomarkers for inflammation and hemostasis [10], may have role as prognostic biomarker in this peculiar condition.

Therefore, the goals of this prospective study were [a] to characterize the profile of circulating plasma MVs in patients with decompensated cirrhosis with vs. without AKI, and [b] to investigate the role of MVs as biomarker to identify patients at higher risk of death.

1. Material and methods

1.1. Patient selection

The study population was recently described [18]. Briefly, adult patients with decompensated cirrhosis admitted to the Yale New Haven Hospital between January 1st and September 1st 2019 were prospectively screened to determine eligibility to participate. Decompensation was defined by clinically evident decompensating events (ascites, variceal hemorrhage [VH], and hepatic encephalopathy) [19].

Patients admitted for VH and/or who experienced VH and/or any other major bleeding [20] in the 30 days prior to hospitalization, those with acute on chronic liver failure (ACLF) [21], and those who were admitted to the intensive care units were ineligible.

At screening, medical records, past medical history, and laboratory results were reviewed for the following exclusion criteria: chronic kidney disease (CKD); presence and/or history of venous thrombosis and/or portal vein thrombosis; presence of extra-hepatic tumors or any primary hematologic diseases; recent surgery; HIV-infection, history of any organ transplantation. Patients on full dose anticoagulation and/or antiplatelet therapy and/or anti-fibrinolytic therapy were also excluded.

Upon enrollment, patients were categorized into cases (with AKI) and controls (without AKI).

AKI was defined per the International Club of Ascites criteria [22]. For patients admitted with AKI, baseline serum creatinine was obtained from the medical records within the prior 3 months and the most recent stable value was considered to be baseline; for patients who developed AKI during hospitalization, baseline serum creatinine was the one at admission.

1.2. Study design

This was a prospective, single-center cohort study approved by the Yale Human Investigation Committee (#2,000,024,288). The study was conducted in compliance with the Declaration of Helsinki and all patients gave written informed consent before enrollment.

MVs evaluation was an ancillary project within a main study on the evaluation of hemostasis in patients with decompensated cirrhosis and AKI [18].

Patients with cirrhosis and AKI were recruited within 24 h after the diagnosis of AKI and evaluation of MVs was performed twice: at enrollment, and on the day after AKI resolution (defined by creatinine returning to a level within 0.3 mg/dL from baseline). When the patient was treated with dialysis, the second evaluation was not performed.

Patients with cirrhosis without AKI were recruited at or near the admission and evaluation of MVs was performed once, at enrollment.

All patients with cirrhosis were prospectively followed and 30-day mortality was collected.

1.3. Sample collection and MVs evaluation

1.3.1. Blood sampling

Peripheral blood was collected via venipuncture in citrate-containing vacutainer tubes using 21 g needles and tourniquet. Per our original protocol, all the samples were collected at 6 am to limit circadian variation in platelet function [18]. The first few milliliters were discarded. Platelet-poor plasma (PPP) was prepared within 1 h by double centrifugation (2 × 10 min at 1500 g) at room temperature. One mL of plasma above the buffy coat after first centrifugation and 1 mL of plasma above the bottom of the tube/pellet after second centrifugation were discharged). Aliquots (1 mL) were immediately frozen and then stored at −80 °C until use.

1.3.2. Plasma circulating microvesicles (MVs) analysis

PPP was thawed in a water bath for 5 min at 37 °C and immediately processed for immunolabeling. PPP was analyzed only after a single freeze-thaw cycle. Prior to the staining, the antibody mixture was centrifuged at 20,0 0 0 g for 30 min to remove fluorescent particles. The pre-analytic phase of MVs analysis has previously been reported [23–25]. Flow cytometry analysis was performed using a CytoFLEX flow cytometer (Beckman Coulter), as previously reported [11,26]. For MV size calibration of the flow cytometer, fluorescent polystyrene beads (Megamix FSC & SSC Plus, BioCytex) were used in sizes of 0.1, 0.16, 0.2, 0.24, 0.3, 0.5 and 0.9 μm. Violet side scatter (VSSC) and FL1 channel gain were set to visualize the beads. The side scatter (SSC) from the 405 nm violet laser (VSSC) was used as a trigger signal to discriminate the noise. Megamix bead solution was gated excluding the background noise (because of the solution itself). After turning the set in VSSC and forward scatter (FSC), a rectangular gate was set between the 0.1 mm and 0.9 μm bead populations and defined as MV gate.

Twenty microliters of PPP were stained with 10 μL of calcein-AM (Sigma-Aldrich) + 4 μL PE (phycoerythrin)-labelled anti-CD62E antibody (BioLegend) + 4 μL Alexa Fluor 647-labelled anti-tissue factor (TF –American Diagnostica #4509). A second panel included 20 μL of PPP stained with 10 μL of calcein-AM + 4 μL of PECy5-labelled anti-CD62P antibody (BioLegend) + 4 μL Alexa Fluor 647-labelled anti-TF + 4 μL labelled anti-CD45 antibody (eBioscience). A third panel included 20 μL of PPP stained with 10 μL of calcein-AM + 4 μL of APC (allophycocyanin)-labelled anti-pan-cytokeratin C-11 antibody (GeneTex International) + PE-labelled anti-CD141 (Thrombomodulin-TM) antibody (BD Pharmingen).

All staining processes were performed for 30 min at 37 °C. Parallel incubation was performed with isotype-matched control antibodies. Fluorescence measured with the respective isotype negative control antibody were subtracted in order to avoid unspecific signal. True MV events were defined as double-positive stained for:

• calcein-AM and anti-CD62E (endothelial-MVs)

• calcein-AM and anti-TF (TF + MVs)

• calcein-AM and anti-CD62P (platelet-MVs)

• calcein-AM and antiCD45 (leukocyte-MVs )

• calcein-AM and anti-pan cytokeratin (hepatocyte MVs)

• calcein-AM and anti-CD141 (TM + MVs).

Triple-positive MVs were evaluated in order to ascertain the origin of TF + MVs and namely calcein-AM, antiCD62E and anti-TF (endothelial-TF + MVs); calcein-AM, antiCD62P and anti-TF (platelet-TF + MVs); calcein-AM, antiCD45 and anti-TF (leukocyte-TF + MVs).

Stained PPP was then diluted by adding 140 mL of sterile filtered PBS. MVs were expressed as events/μL with the volume measurement of the CytoFLEX. Files were exported and data were evaluated by CytExpert (Software Version 1.2, Beckman Coulter).

1.4. Data collection

Data collected from the medical records included causes for admission, patient demographics, presence or absence of infection at time of AKI, laboratory data, and 30-day mortality.

1.5. Data analysis

1.5.1. Study objective

The primary objective of this study was to compare the levels of plasma circulating MVs of different origin in patients with decompensated cirrhosis with vs. without AKI.

The secondary objective of this study was to correlate the profile of MVs with patient’s outcomes, specifically mortality at 30 days.

1.5.2. Statistical analysis

The number of patients included in this ancillary study was based on the sample size calculation for the main protocol [18].

Qualitative data are described using frequency and percentage. Quantitative data are described using median with 25% and 75% quartile ranges. Comparison between two independent groups were performed using the Mann Whitney U test and t-test for continuous variables, and Chi-square test of Fisher’s exact test for categorical variables. Comparison within the same group (baseline vs. resolution of AKI in patients with cirrhosis) was performed using the Wilcoxon signed rank test.

Univariate logistic regression analyses were performed to evaluate parameters significantly associated with the clinical outcome (30 day-mortality). Only variables highly significant associated (p ≤ 0.01) were included in multivariate analysis. Since the relatively small sample size and the low number of clinical outcome, only two variables at time were entered in the multivariate analysis. Odds Ratio (ORs) with 95% confidence intervals (CI) for association were calculated. Statistical significance was set at p ≤ 0.05. All analyses were completed using SPSS version 26.

2. Results

2.1. Demographics

Of 136 patients with decompensated cirrhosis screened for recruitment, 80 were included (40 each with and without AKI). Median time between hospital admission and baseline samples collection was 1.5 days (IQR 1–4) and 2 days (IQR 1–3) in patients with and without AKI, respectively (p = 0.9). Reasons for exclusion were as follows: admission for VH (n = 6); ACLF (n = 5); admission to the intensive care units (n = 10); CKD (n = 11); portal vein thrombosis (n = 3); pulmonary embolism (n = 1); recent surgery (n = 2); liver or liver/kidney transplantation (n = 4); therapeutic anticoagulation/anti-thrombotic therapy (n = 2); blood product transfusion in the 3 days before screening (n = 8); refused to participate (n = 2).

Demographics, reasons for hospitalization, and severity of cirrhosis comparable between patients with and without AKI (Table 1). Overall, alcohol, hepatitis C virus infection, and non-alcoholic steatohepatitis were the etiology of cirrhosis in 55%, 19%, and 14% of the patients, respectively. MELD score was significantly higher in patients with AKI (25 vs. 18 in AKI vs. no-AKI, respectively), due solely to a significant difference in serum creatinine (1.8 mg/dL vs. 0.8 mg/dL in AKI vs. no-AKI, respectively). In contrast, serum bilirubin and INR were comparable between groups (Table 1). Approximately 45% of patients in both groups received antithrombotic prophylaxis with low molecular weight heparin (LMWH), with no statistically significant difference between the groups (Table 1).

Table 1.

Baseline characteristics in patients with decompensated cirrhosis.

	AKI (n = 40)	No AKI (n = 40)	P values
Age (years)	56 (52–65)	57 (52–64)	ns
Male gender (%)	60	73	ns
Child class B/C (n)	17/19	23/21	ns
Pugh score ^	10 (7–13)	10 (7–12)	ns
Ascites (%)	83	85	ns
History of decompensation^^ (%)	90	85	ns
MELD score	25 (20–29)	18 (11–26)	<0.001
Reason for admission (%)			–
Abdominal pain/suspected	20	25
infection	23	30
Ascites	20	30
AMS or HE	12	0
AKI	15	0
Trauma	10	15
Other
Bacterial infection (%)	48	23	0.02
Non-selective beta-blockers (%)	35	43	ns
Statins (%)	8	13	ns
VTE prophylaxis (%)	47.5	45	ns
Etiology of AKI (%)	60/20/20	–	–
Prerenal/HRS/ATN
AKI stage (%)	30/50/20	–	–
1/2/3
Total bilirubin, mg/dL	2.9 (1.9–4.7)	2.6 (1.5–5.8)	ns
Albumin, g/dL	3.2 (2.6–3.5)	2.9 (2.4–3.3)	ns
Hemoglobin, g/dL	8 (7.3–9.2)	9.2 (8–11)	0.01
INR	1.7 (1.4–1.8)	1.5 (1.3–1.8)	ns
Platelet count, 109/L	79 (61–127)	66 (48–96)	ns
Creatinine, mg/dL	1.8 (1.6–2.5)	0.8 (0.7–0.9)	<0.001
Sodium, mmol/L	134 (132–138)	136 (130–138)	ns
Potassium, mmol/L	4.2 (3.9–4.6)	3.9 (3.6–4.3)	0.01
AST, U/L	39 (31–61)	52 (35–65)	ns
ALT, U/L	20 (15–33)	31 (23–45)	0.004

Median values reported with 25th and 75th percentiles in parenthesis.

^{^}Median (range).

^{^^}Including ascites, variceal hemorrhage, and hepatic encephalopathy.

MELD: Model for End-Stage Liver Disease; AMS: altered mental status; HE: hepatic encephalopathy; AKI: acute kidney injury; VTE: venous thromboembolism; ns: non-significant; HRS: hepatorenal syndrome; ATN: acute tubular necrosis; INR: international normalized ratio; AST: aspartate aminotransferase; ALT: alanine aminotransferase.

In patients with AKI, etiology of renal dysfunction was as follows: pre-renal azotemia (60%), hepatorenal syndrome (20%), and acute tubular necrosis (20%). AKI was present at time of admission in 80% of patients, while the remaining developed AKI during hospitalization with a median time from admission to patient’s recruitment of 5 days.

2.2. Microvesicles in patients with decompensated cirrhosis with vs. without AKI

As shown in Figs. 1 and 2, at baseline, patients with cirrhosis and AKI had significantly higher levels of calcein + (total), endothelial, platelet, and platelet-TF + MVs. Conversely, levels of TF +, CD45 +, TM +, hepatocyte, endothelial-TF +, and leukocyte-TF + MVs were comparable between groups (Figs. 1 and 2). Numerical values of circulating MVs in patients with vs. without AKI are reported in Table 2.

Fig. 1.

Circulating MVs in patients with decompensated cirrhosis with vs. without AKI. AKI is associated with increased levels of total, endothelial, and platelet-MVs.

MVs: microvesicles; AKI: acute kidney injury; TF: tissue factor; TM: thrombomodulin.

Fig. 2.

Hepatocyte and triple-positive MVs in patients with decompensated cirrhosis with vs. without AKI.

MVs: microvesicles; AKI: acute kidney injury; TF: tissue factor.

Table 2.

MVs in patients with cirrhosis with vs. without AKI at baseline.

MVs/uL	AKI (n = 40)	No AKI (n = 40)	P value
Calcein + MVs (total)	2108 (1652–2977)	1523 (947–1931)	0.002
ESelectin + MVs (endothelial)	1055 (677–1513)	563 (343–755)	<0.0001
PSelectin + MVs (platelet)	777 (535–1369)	425 (213–535)	<0.0001
TF + MVs	54 (34–140)	50 (31–80)	0.3
CD45 + MVs (leukocyte)	52 (32–141)	50 (23–76)	0.05
TM + MVs	728 (423–956)	665 (429–800)	0.3
Hepatocyte-MVs	27 (13–44)	32 (20–42)	0.6
ESelectin + TF + (endothelial-TF + )	85 (61–189)	75 (44–102)	0.1
PSelectin + TF + (platelet-TF + )	46 (22–108)	36 (10–60)	0.006
CD45 + TF + (leukocyte-TF +)	26 (10–62)	25 (7–40)	0.1

Median values reported with 25th and 75th percentiles in parenthesis.

MV: microvesicles; TF: Tissue factor; TM: thrombomodulin; Hep: hepatocyte; AKI: acute kidney injury.

Resolution of AKI occurred in 23 patients (53%) and was associated with profound changes in MVs levels (Figs. 3 and 4). Calcein +, endothelial, and platelet-TF + MVs significantly decreased and became comparable with those in patients without AKI. Platelet MVs significantly decreased but remained higher compared to patients without AKI. TF + and endothelial-TF + MVs significantly decreased and became lower than patients without AKI. TM + MVs significantly decreased but remained comparable with patients without AKI. CD45 +, CD45 + TF +, and hepatocyte-MVs remained unchanged. Numerical values of circulating MVs in patients with AKI at baseline vs. at AKI resolution are reported in Supplementary Table 1.

Fig. 3.

Circulating MVs in patients with decompensated cirrhosis with AKI: baseline vs. resolution of AKI. MVs: microvesicles; AKI: acute kidney injury; TF: tissue factor; TM: thrombomodulin.

Fig. 4.

Hepatocyte and triple-positive MVs in patients with decompensated cirrhosis with AKI: baseline vs. resolution of AKI. MVs: microvesicles; AKI: acute kidney injury; TF: tissue factor.

In patients with cirrhosis with AKI, no significant correlation was found between levels of plasma circulating platelet-derived MVs and platelet aggregation and secretion (markers of platelet function) both at baseline and after AKI resolution (Supplementary Table 4).

2.3. Correlation of MVs with 30-day mortality

Eleven (27.5%) patients with cirrhosis with AKI died vs. 4 (10%) patients without AKI (p = 0.05). Variables associated with this outcome at univariate analysis are shown in Supplementary Table 2. In a two-variables at time multivariate analysis model, creatinine (OR 4.3 [95%CI 1.8–10.7]; p = 0.0008), MELD (OR 1.13 [95%CI 1.02–1.27]; p = 0.025), any bleeding (OR 9.07 [95%CI 2.02–40.6]; p = 0.004), post-procedural bleeding (OR 6.75 [95%CI 1.08–42]; p = 0.036), and levels of hepatocyte-MVs (OR 1.04 [95%CI 1.02–1.07]; p = 0.0011) remained significantly associated with 30-day mortality (Supplementary Table 3).

3. Discussion

Cell response to a number of stressors and inflammatory mediators is key to maintenance of tissue homeostasis. Early manifestations of a cell activation process are the formation of membrane blebs and the shedding of nanoscale membrane fragments known and designated as microvesicles (0.1 to 1 μm in diameter) [27]. Circulating MVs have various (patho-)physiological functions, including the transport of membrane components from the parent cell to other cells and the direct or indirect activation of inflammation and coagulation [1,24,28–30].

Although several aspects of MV functions in patients with chronic liver disease are still unclear, it has been shown that MVs play an important role in inflammation, coagulation, and vascular homoeostasis in these patients [10]. Because inflammation and alterations of vascular homoeostasis (i.e.: development of clinically significant portal hypertension) are main drivers of cirrhosis progression and development of decompensation [31], MVs have been proposed as potential biomarkers to stratify risk and predict patient outcome [10].

AKI is a relatively common complication of decompensated cirrhosis, occurring in approximately 20% of hospitalized patients [15], and is associated with an increased risk of death [16]. However, which is the effect of AKI on the profile of circulating MVs in patients with cirrhosis and whether MVs could serve as biomarker to identify patients with cirrhosis and AKI at greater risk of complications remains to be evaluated.

To our knowledge, this is the first study to extensively characterize levels of several MVs subtypes in patients with decompensated cirrhosis with vs. without AKI as well as to address the potential role of MVs as biomarkers in this clinical setting.

We found that patients with cirrhosis with AKI had a significantly higher level of total MVs compared to patients with cirrhosis without AKI but comparable severity of underlying liver disease. The significantly higher levels of endothelial-derived, platelet-derived, and platelet-derived TF + MVs in patients with vs. without AKI along with the comparable level of leucocyte and hepatocyte MVs between the two groups would suggest that the cells most sensitive to the AKI-driven cellular activation are endothelial cells and platelets.

It has been previously shown that patients with cirrhosis have higher levels of circulating MVs compared to healthy controls [4,11], and that levels of circulating MVs increase in parallel with severity of liver disease [12]. Our new findings show that, in hospitalized patients with decompensated cirrhosis, AKI further worsens cellular homeostasis and leads to a further increased cellular vesiculation. This is in line with previous data by Agarwal et al. [32] who demonstrated that MV-activity was significantly increased in patients with acute liver failure with vs. without AKI as well as with other studies that found AKI and CKD (in patients without liver disease) to be associated with increased levels of circulating MVs [33,34]. Another condition associated with endothelial and microvascular dysfunction is sepsis, in which the loss of antithrombotic properties by damaged endothelium and parenchymal micro-thrombosis have been implicated in disease progression and development of sepsis-related complication, including AKI [9,35]. It has been shown that, in patients with sepsis, MVs may have a role in determining the risk of sepsis-related complications [33]. Interestingly, patients with sepsis with AKI and patients with cirrhosis with AKI demonstrate comparable alterations of circulating MVs [36], which suggests a common occurrence of endothelial dysfunction by the superimposed acute kidney injury.

The significant reduction of total MVs with AKI resolution, which became comparable to that observed in patients with cirrhosis without AKI, is further proof that AKI is responsible for the increased levels of MVs observed in patients with cirrhosis. However, while endothelial, platelet-TF + MVs and TM + MVs significantly decreased and became comparable with those in patients without AKI, platelet MVs significantly decreased but remained relatively higher compared to non-AKI. This would suggest that the endothelium perturbation decreases rapidly after AKI resolution while on the other hand the effect of platelet activation persists longer after normalization of serum creatinine. Additionally, TF + and endothelial-TF + MVs significantly decreased and became lower than patients without AKI. This result would confirm that normalization of renal function is associated with resolution of endothelial dysfunction, as TF is usually expressed in damaged and activated endothelium [37].

As for platelet function, our previous findings showed that patients with cirrhosis and AKI had platelet aggregation and secretion significantly lower than patients without AKI with a significant improvement in both platelet aggregation and secretion after AKI resolution [18]. Here we demonstrate that, despite reduced aggregation and secretion (markers of platelet function), AKI patients have increased levels of platelet-derived MVs that persisted after AKI resolution, which suggests increased platelet activation. Since we did not find any correlation between increased levels of platelet-derived MVs and platelet aggregation and secretion, the increased platelet activation appears to be mostly a reflection of AKI-driven endothelial dysfunction [35] rather than a true compensating event. This hypothesis is also confirmed by the fact that also platelet-TF + MVs were significantly increased in AKI patients, denoting the cross talk between activated endothelial cells exposing TF and platelets.

Finally, CD45 +, CD45 + TF +, and hepatocyte-MVs remained unchanged after AKI resolution. This suggests that inflammation and hepatocytes proliferation are not affected by superimposed AKI but are more a reflection of cirrhosis severity itself.

AKI-driven endothelial dysfunction and increased platelet activation in patients with decompensated cirrhosis may have implications in both cirrhosis progression [38,39] and risk of subsequent CKD [35,40], which further denotes the well-known importance of early proactive treatment of AKI in cirrhosis. This also suggests, if our results will be validated, that dysfunctional endothelium and activated platelets may become new potential therapeutic targets to prevent, or at least mitigate, the detrimental sequelae associated with AKI in cirrhosis. Finally, it has been shown that patients with cirrhosis and AKI have increased risk of bleeding [41,42] due to platelet dysfunction, low level of fibrin-stabilizing factor XIII, and perhaps enhanced fibrinolysis [18,43]. As our MVs-based data appears to suggest, endothelial dysfunction is also implicated in the AKI-driven coagulopathy and it should be further evaluated whether it may contribute to the purported increased bleeding tendency of these patients.

As far as the possible role of MVs as biomarkers, we documented significant but weak association between hepatocyte-MVs (anti-pan cytokeratin-MVs) and 30-day mortality. Cytokeratins are a class of intermediate filaments that are involved in differentiation of epithelia [44]. Specific keratins are expressed in different epithelia, and hepatocytes express only one pair of keratins (K8 and K18). Therefore, by measuring cytokeratin 18-expressing MVs, one can detect the MVs released by hepatocytes and thereby indirectly assess hepatocytes vesiculation capacity.

In patients with cirrhosis, previous data demonstrate that plasmatic levels of hepatocyte MVs increase in parallel with histological liver necro-inflammatory activity, which suggests that hepatocyte chronic injury is the main driver for hepatocyte release of MVs [12]. Additionally, in the seminal study by Payance et al. including 242 patients with decompensated cirrhosis in the absence of an acute complication, hepatocyte MVs (but not endothelial nor leukocyte MVs) were independent predictor of patient’s survival [12]. Our study confirms these findings and show that, despite superimposed AKI being associated with worsening vascular homeostasis, as reflected by increased levels of endothelial MVs, hepatocyte MVs remain the only independent predictor of mortality in decompensated cirrhosis [18].

The strengths of the study lie in the longitudinal design, the use of a standardized pre-analytical protocol to detect MVs, the evaluation of several types of MVs involved in coagulation, inflammation, endothelial dysfunction and hepatocytes proliferation using a new generation flow cytometry. On the other hand, the main limitations of this study pertain to i) the small sample size, notwithstanding the very homogeneously enrolled and longitudinally followed population; ii) the lack of standardized method to analyze circulating MVs. Particularly, the centrifugation protocol deviated somewhat from the proposed methods [43] because the centrifugation speed was lower than that recommended. However, during pre-analytic phase, we took specific precautions to prevent platelet contamination (1 mL of plasma above the buffy coat after first centrifugation and 1 mL of plasma above the bottom of the tube/pellet after second centrifugation were discharged) and we prevented contamination by particles ≥ 1 μm by using by using a specific MVs gate at cytofluorimetry. Additionally, there is controversy on TF + MVs detection by flow-cytometry because of the limited number of TF epitopes per MV, quality of available antibodies, blockade of TF with factor VII and tissue factor pathway inhibitor, and the lack of sensitivity of current flow cytometers to detect low fluorescent MVs [45]. However, in this study we used a new generation flow-cytometry with high sensitivity to low fluorescence and a mouse monoclonal anti-human TF antibody that reacts with and neutralizes the purified apoprotein of human TF and recognizes TF on the surface of tumor cells and lipopolysaccharides-stimulated monocytes [46]; iii) some patients with AKI were on heparin prophylaxis and because reduced renal clearance may be associated with increased plasmatic concentration of the drug, this might have somehow influenced the MVs analysis in this group of patients; iv) given the small number of patients with the clinical outcome (30-day mortality), the multivariate regression analysis could be impaired.

In conclusion, our study showed that AKI did worsen vascular and cellular homeostasis in patients with cirrhosis by inducing endothelial dysfunction and platelet activation. On the other hand, AKI was not associated with worse systemic inflammation and hepatocytes activation, as leukocyte-MVs and hepatocyte-MVs remained unchanged before vs. after AKI resolution. We also showed that, among these multiple subpopulations of plasma circulating MVs, only hepatocyte-derived MVs were independently correlated with short-term survival. Larger studies are needed to further evaluate the potential role of MVs, particularly hepatocyte-derived MVs, as clinical biomarkers in hospitalized patients with decompensated cirrhosis with and without AKI.

Abbreviations:

MVs

microvesicles

EVs

extra-cellular vesicles

AKI

acute kidney injury

MELD

model for end stage liver disease

VH

variceal hemorrhage

ACLF

acute on chronic liver failure

CKD

chronic kidney disease

PPP

platelet poor plasma

VSSC

violet side scatter SSC

PE

phycoerythrin

ORs

odds-ratios

CI

confidence interval