Patients Epidemiological And Clinicopathological Characteristics
A total of 31 HBV-ACLF patients expecting liver transplantation and 3 more living donors (NC cases) were include. Their pre-operative assessments are shown in Table 1. No significant differences were noted in the distribution of the demographic factors. Active HBV viral replication was indicated by the HBV-DNA levels, clearly correlating with the grades of HBV-ACLF. So do the levels of liver function. Notably, the coagulation function valued by International Ratio (INR) indicated preferentially deterioration in grade 2 compared with grade 3 samples. Besides, the platelet count provided additional support for this finding.
Table 1
Clinical characteristic of HBV-ACLF patients classified by COSSH grading
Parameter
|
ACLF-grade1 (n = 9)
|
ACLF-grade2(n = 7)
|
ACLF-grade3(n = 15)
|
Age (years)
|
46(38.5,56.5)
|
47(39,56)
|
47(39,55)
|
Gender
|
Male
|
77.8%(7)
|
85.7(6)
|
93.3%(14)
|
Female
|
22.2%(2)
|
14.3%(1)
|
6.7%(1)
|
Laboratory test
|
HBV-DNA level (IU/ml)
|
≤ 103
|
88.9% (8)
|
42.9% (3)***
|
20% (3)***
|
103-106
|
11.1% (1)
|
57.1% (4)**
|
60% (9)***
|
> 106
|
0
|
0
|
20% (3)***
|
Alanine aminotransferase (U/L)
|
40.57 ± 17.61
|
89.78 ± 60.79**
|
237.73 ± 167.28***
|
Aspartate aminotransferase (U/L)
|
56.29 ± 12.93
|
151 ± 50.67**
|
218.27 ± 124.92**
|
Albumin (g/L)
|
38.49 ± 5.65
|
32.07 ± 3.26*
|
33.83 ± 6.10*
|
Total bilirubin (umol/L)
|
434.53 ± 166.66
|
423.33 ± 183.87NS
|
432.75 ± 79.45NS
|
Alkaline phosphatase (U/L)
|
135.89 ± 61.64
|
96.71 ± 18.89*
|
137.60 ± 46.47NS
|
γ-glutamyl transpeptidase (U/L)
|
63.22 ± 35.45
|
42 ± 31.52**
|
78.87 ± 38.57*
|
Creatinine(umol/L)
|
56.44 ± 24.92
|
183.43 ± 23.34**
|
188.87 ± 28.59**
|
Sodium (mmol/L)
|
135.5 ± 5.4
|
135.29 ± 4.39NS
|
135.42 ± 5.50NS
|
White blood cell count (109/L)
|
8.00 ± 6.40
|
5.75 ± 3.09*
|
8.94 ± 4.53NS
|
Hemoglobin (g/L)
|
103.8 ± 18.7
|
96.7 ± 17.9NS
|
93.5 ± 23**
|
Platelet count (109/L)
|
64.3 ± 22.4
|
47.86 ± 18.45**
|
91.73 ± 38.52***
|
INR
|
2.52 ± 0.73
|
5.89 ± 1.12**
|
3.32 ± 0.41*
|
Alpha fetoprotein (ug/L)
|
9.46 ± 9.09
|
9.97 ± 7.71NS
|
106.60 ± 25.14***
|
NH4(umol/L)
|
23.24 ± 9.01
|
52 ± 17.31***
|
60.59 ± 24.68***
|
Renal or liver replacement treatment
|
Yes
|
33.3% (3)
|
42.9% (3)
|
73.3% (11)
|
Extrahepatic organ failure
|
Coagulation
|
33.3% (3)
|
85.7% (6)
|
80.0% (12)
|
Kidney
|
11.1% (1)
|
57.1% (4)
|
93.3% (14)
|
Cerebral
|
0
|
28.5% (2)
|
46.6% (7)
|
Respiration
|
0
|
0
|
20% (3)
|
Circulation
|
0
|
28.5% (2)
|
33.3% (5)
|
Data are presented as the means ± SD, medians (p25, p75) or percentages (numbers of patients).*p < 0.05, **p < 0.01,***p < 0.001 and NA refers to no significant for comparisons between the ACLF-grade2/ACLF-grade3 and ACLF-grade1 groups. |
Extrahepatic organ failure is a key factor determining the grades of HBV-ACLF. Coagulation disorders occur most frequently in all three grades. Kidney is the most frequently disturbed solid organ by liver-renal syndrome along HBV-ACLF development. The majority of the patients in grade 3 received renal or hepatic replacement therapy prior to transplantation owing to severe liver failure combined with kidney failure, significantly more frequent than the other two grades, as shown by our data. Patients with grade 3 HBV-ACLF emerged with multiple organ failure, including those involved in the respiratory and circulatory systems, resulting in disease deterioration without instant liver transplantation.
Transcriptome Divergence Of Pbmcs And Hepatic Immune Cells In Hbv-aclf Patients
The overview of the transcriptome expression was indicated by hierarchical clustering analysis (Fig. 1B). Intra-grade divergence and inter-grade homoplasy of the expression profile was noted in hepatic immune cells (generally represented by CD45 + lymphocytes, named as group C) as well as in PBMCs (named as group P). Differentially expressed genes (DEGs) analysis (Fig. 1C and 1D) revealed a large spectrum of DEGs between normal (named as NC) and HBV-ACLF samples from both groups C and P. Compared with C1 and C3, Group C2 indicated the lowest frequency of DEGs, possibly due to its transitional state in liver. However, compared with PBMCs from the normal donor (NC-P) group, P2 and P3 from HBV-ACLF grade 2 and 3 patients exhibited closing frequencies of DEGs. The key element in distinguishing grade 1 from grade 2 is based on the presence of the extra-hepatic organ damage, whereas the distinction of grade 2 and 3 was based merely on the severity of the extra-hepatic organ damage (mainly the number of involved organs). Therefore, it would be easier to accept the significantly altered state of the systemic response presented by PBMCs gene profiling from grade 1 to grade 2 & 3.
Functional Synergy Analysis Of Pbmcs And Hepatic Immune Cells In Hbv-aclf Patients
Gene ontology (GO) assessment was carried out to depict distribution of DEGs identified by pair-comparison analysis (Fig. 2A, Supplementary Fig. 1–6). The Biological Process (BP) ontology analysis on DEGs in PBMCs suggested an overall activation of the innate immune system and a generally irritated inflammatory response. In comparison between P3 and NC-P, preferential enrichment in coagulation and hemostasis modules were noted. BP analysis in ACLF-C yet indicated a sophisticated enrichment pattern in regulations of the adaptive immune cells. Enrichment assessment consisting of DEGs with mono-direction change was further presented in Supplementary Fig. 7.
We further classifed the enriched BP items in functional synergy analysis into the following 7 categories: Wounding, viral, metabolism, inflammation, immune system, coagulation and apoptosis (Fig. 2B). These categories were pathologically related to the development of ACLF. Compared with NC, the DEGs detected in ACLF-P indicated functional concentration in modules such as immune system and apoptosis, depicting cellular damage induced immune disturbance state. An interaction network was constructed from the enriched BP modules (Fig. 2C). Although the immune-related BP modules exhibited a major proportion in the network of both liver-resident immune cells and PBMCs, the immune modules interacted in a different manner. In PMBCs, the general regulations on the innate immune system including initial immune recognition, energy and cellular preparation and bridging process to adaptive immune took central place in the network. As for liver-resident CD45+ cells, complement activation and mononuclear cell proliferation took major place in innate immune response. The adaptive immune response mediated by T cell chemotaxis (migration) and the regulation of the general lymphocyte population, otherwise, played a major part. Notably, the lipid-metabolism module was found taking prominent position in the network. The enrichment on the immune regulation and metabolism modules may indicate potential interaction where lipid metabolic turbulence in specific immune cells in turn influences their cellular function in the immune system.
Immune Heterogeneity Exists Between Inflammatory Circulation And Hepatic Immune Microenvironments In Hbv-aclf
Gene profiling of PBMCs and liver CD45+ cells can respectively reflect the state of the circulatory and focal immune activity in blood and liver. The Blood Transcriptomic Module (BTM) is a suitable assessment tool for evaluating key functional modules in immune response via GSEA analysis on immune cells from either peripheral blood or solid organs[8]. The BTM revealed significant divergence between systemic and liver-regional immune responses along the development of HBV-ACLF through a thorough assessment of both innnate and adaptive perspective in HBV-ACLF (Fig. 3A and 3B). We further confirmed the contrasting changes of ACLF-derived PBMCs in the innate and adaptive immune response and extended the discovery to all three COSSH grades. Of novelty, we observed that liver-resident CD45+ cells indicated only partially depressed innate immune response when adaptive immunity was generally inhibited. The depressed modules were concentrated on inflammation and dendritic cell function. Dendritic cells work for antigen presentation for subsequent activation of the adaptive immune response. Consistently, the inhibition of the inflammation modules mostly congregated on the pattern recognition receptor (PRR) response. Of note, COSSH grade 2 exhibited an opposite elevation in the inflammation modules most likely to be responsible for extrahepatic damage in this phase. The enhanced inflammatory response resulted in a circulatory inflammatory cytokine storm, leading to organ damage in addition to liver failure. The interferon signaling modules were continuously activated across all three grades, indicating that HBV-relapse could affect the virus-targeted responses. Moreover, the monocyte function modules were also significantly upregulated both in liver and peripheral blood. In conclusion, in the liver microenvironment, the antigen-presenting process and the inflammatory response were partially downregulated, leading to a disruption of the immune regulation and subsequent induction of caspase systemic inflammatory reactions.
Exhausted immune cells accumulate in the HBV-ACLF liver microenvironment.
CIBERSORT was applied to estimate the cellular constitute of the hepatic immune-microenvironment in HBV-ACLF (Fig. 4A and 4B). The results provided consistent findings with those of the BTM assessment regarding both the innate and adaptive immune systems. In the view of innate immune system, the functional polarization of liver-resident macrophages (also known as Kupffer cells) has been widely acknowledged as a hallmark for liver injury and liver failure. In the present study, we observed a clear increasing ratio of the M2-type macrophages, indicating an ongoing M2 polarization along with the HBV-ACLF exacerbation. Moreover, NK cells indicated markedly decreased infiltration in grade 2 and a minor increase in grade 3. With regard to the adaptive immune system, the subtypes of pro-immune T lymphocytes including CD8+ cytotoxic T cells, and CD4+ memory and activated T cells showed sharply decreased levels of infiltration. By contrast, the regulatory T cells, which are mainly responsible for the negative regulation of the immune response, were significantly accumulated in the higher grade.
Consistent results were verified by flow cytometry analysis on the HBV-ACLF patient samples (Fig. 4C and 4E), where cytotoxic CD8+ T cells and NK cells showed a decreasing trend in infiltration, Foxp3+T-reg cells continue to accumulate and polarization of macrophages shift to M2 along ACLF developes. Immune flourence provived consistent phenomenal evidence of immune infiltration (Fig. 4D and 4F) where higher ratio of exhausted state of cytotoxic CD8+ T cells and that of immune modulatory cells including Foxp3+T-reg cells and M2 macrophage. In conclusion, following the elevation of HBV-ACLF grade, the effective cellular components of both the innate and adaptive immune responses gradually represented an exhausted state with a concomitant accumulation of the negatively regulatory counterparts.
The metabolic signature based on three biomarkers depicts cholesterol-centered metabolic phenotypes of immune cells in HBV-ACLF
The role of liver metabolic alterations in influencing immune status and disease severity was further investigated. A total of 91 DEGs were identified that were differentially expressed in all 3 grades by cross-matching DEGs derived from comparisons between each grade and the NC group (Fig. 5A). The interaction network was established based on these DEGs (Fig. 5B). Notably, metabolism-related BP modules were at the central spot. Lipid metabolism activities (lipid localization, lipid transport, lipoprotein particle clearance, etc.) were enriched. Among the 91 DEGs, 3 genes including APOE, APOC1 and SPP1(OPN) showed the highest frequency appearing in the metabolic modules (Fig. 5C). To confirm correlation between selected marker genes and disease severity, qRT-PCR and Western Blot were carried out to detect the mRNA levels and protein concentration of the three biomarkers in HBV-ACLF patients of varied grades (Fig. 5D and 5E). Immunohistological staining further provided evidence that intra-tissue concentration of the three biomarkers was positively related to HBV-ACLF grades (Fig. 5F and 5G). The three biomarkers demonstrated an elevation in both transcription and translation level along the higher grades of the disease, indicating their potentiality as biomarkers depicting consistency with the clinical severity.
Apoe+ Polarized Macrophages With Effluxed Cholesterol Represent Immune Signature Of Hbv-aclf
For exploring which immune cell subsets in HBV-ACLF contribute mostly significant elevation of the three biomarkers, relevant cell subsets (CD8+PD-1+ T cell, CD4+Foxp3+ T-reg cell and macrophage) were abstracted from HBV-ACLF liver tiusses. It turned out that liver-resident macrophages showed most significant expressions of all three biomarkers in both mRNA and protein levels (Fig. 6A and 6B). APOE, APOC1 and SPP1 were all previously reported to actively involve in cholesterol metabolism. Besides, cholesterol transport was enriched at the central part of the interaction network established above (Fig. 5B). Intracellular cholesterol assessment confirmed that macrophages showed intracellular cholesterol consumption in HBV-ACLF compared to NC (Fig. 6C), while other subsets showed no significant fluctuation in cholesterol concentration. As APOE accounts for cholesterol efflux through ABCA1 (ATP Binding Cassette Transporter A1) transporting cholesterol and varied APOE expression could modulate function state of macrophage via inducing polarization according to published papers[9]. We further confirmed that M2 polarized macrohpages in ACLF showed elevated protein levels of APOE, ABCA1 and endogenuous cholesterol production key enzyme HMG-CoA (Fig. 6D), indicating activation of ABCA1 cholsetrol transport pathway. Also, APOE expression showed strong correlation with CD206 (Fig. 6E and 6F). APOE+ M2 macrophages (APOE+ CD206+) increase to infiltrate in liver along higher grades of HBV-ACLF (Fig. 6G), connecting intracellular cholesterol metabolism shifting with functional polarization of macrohpages in HBV-ACLF.
Rosuvastatin Alleviated Aclf Progress Via Inhibiting Cholesterol Efflux-induced Machophage Polarization In Mice Model
In order to confim consistency between samples from ACLF patitients and ACLF mice models[10] so that further intervention therapy could be applied for in vitro experiments. (Fig. 7A). The tissue sections of the mouse livers were obtained for H&E staining and additional histological inspection. The serum concentration levels of specific indices, including ALT and Total Bilirubin (TBL) were detected by ELISA to provide biochemical evidence for acute liver failure in mice with prior cirrhosis (Fig. 7B). IHC sections confirmed elevated intratissue levels of three biomarkers in ACLF mice (Fig. 7C). IF staining revealed similarly increased APOE+ polarized macrophages infiltration (Fig. 7D, 7G and 7H), suggesting that ACLF mice models may perfectly serves as experiment tool mimicking actual disease background of ACLF. We applied Rosuvastatin, an effective lipid-lowing medicine whose effective mechanism include inhibiting HMG-CoA reductase to reduce the synthesis of cholesterol, was applied to revert the cholestrol metabolism shift in polarized machrophage. As the results indicated, Rosuvastatin could significantly reverse cholesterol efflux induced machrophage polarization, meanwhile clearly alleviated the damage and nacrosis in liver tissue, as well as renal dysfunction came along (Fig. 7E, 7F and 7H) .