Thrombospondin 1 enhances systemic inflammation and disease severity in acute-on-chronic liver failure

Standing as distinct disease entity, acute-on-chronic liver failure (ACLF) develops from chronic liver disease subsequent to various insults and is characterized by severe disease, rapid development, poor prognosis and high mortality [1, 2]. ACLF depicted by regional phenotype specificity caused by different etiologies and precipitants. In Europe and the United States, the main etiology of ACLF is alcoholic cirrhosis, and the primary precipitants including bacterial infection, sepsis and systemic inflammation; while in China, the most common precipitant is hepatitis b virus (HBV), and HBV reactivation in infected patients represents the main trigger of ACLF [3].

Understanding the molecular mechanism of ACLF pathogenesis is essential for the development of effective diagnostic and therapeutic strategies. In this context, prominent dysregulation of immune and metabolic processes has been recognized as the key mechanism underlying ACLF pathogenesis, which was previously reported by the two major comprehensive studies worldwide, the Chronic Liver Failure (CLIF) Consortium Acute-On-Chronic Liver Failure in Cirrhosis (CANONIC) study and our recent large prospective multicenter Chinese Group on the Study of Severe Hepatitis B (COSSH) study on HBV-related ACLF (HBV-ACLF). Both studies highlighted either the intense systemic inflammatory response or an excessive immune response triggered by HBV exacerbation initiated immune-metabolic disorder, the key contribution factors that provokes multiorgan failure and worsen ACLF prognosis [4‐7]. Furthermore, various studies have highlighted the potential roles of autoimmune hepatitis (AIH) and drug-induced liver injury (DILI) in the pathophysiology of ACLF. Typical AIH patients may develop ACLF following a second insult, possibly exacerbated by long-term immunosuppression [8]. In contrast, antitubercular agents and complementary and alternative medications are recognized as classical triggers of DILI leading to ACLF [9].

Recently, we emphasized the role of several potential biomarkers associated with the immune-metabolism dysregulation linked to HBV-ACLF prognosis, among which thrombospondin 1 (THBS1) was recognized within the top significantly differentially expressed key molecules [7]. THBS1, a member of the thrombospondin family, is a matricellular glycoprotein produced by various cells, including platelets, hepatocytes and other cells, where it can interact with multiple targets and receptors, such as CD36, integrins, CD47 and TGF-β, and participates in multiple functions, including inflammation, angiogenesis regulation, apoptosis and other cellular fate determinants [10, 11]. THBS1 is secreted in response to inflammation [12], and has been shown to be involved in tissue injury, inflammatory diseases, liver fibrosis and hepatic cancer [13], but there is no clear evidence regarding the involvement of THBS1 in ACLF pathogenesis and outcome.

On this background, we hypothesized that THBS1 expression is positively correlated with the typical disease features of ACLF, namely, inflammation and hepatocellular apoptosis. Consequently, in this work, we first analyzed the expression profile of THBS1 as a diagnostic/prognostic biomarker, predictor of ACLF severity and disease outcome in a total of 568 subjects with HBV-related etiologies and 110 healthy normal controls, as well as in our previously established ACLF rat model. Notably, we developed a hepatocyte-specific THBS1 knockout (THBS1^KO) mouse model to validate the hepatoprotective capabilities following intoxication, and the overall study design is summarized in Fig. 1A.

ACLF has been a topic of interest for researchers due to its severe incidence rate, rapid disease progression and high mortality [30]. The recognition of ACLF has become more perceptive in recent years and achieved a status of universal acceptance [31‐34]. Single molecules and combination of cytokines or chemokines were studied as potential ACLF biomarkers and showed promising results in predicting outcomes [35‐37]; however, they were identified based on conventional technologies, which unfortunately could not be sufficiently translated to modern clinical applications. At present, many studies on ACLF are based on alcoholic liver disease and hepatitis C in Western countries, but in Asia-Pacific region, the main etiology is HBV-ACLF [15, 17], in which most of the HBV-ACLF biomarker researches and mechanisms comes from small samples and/or single centers, which limits their future applications. At the same time, there is still a relative lack of research on prognostic molecular targets and their specific pathophysiological mechanisms.

Being the eminent ACLF pathological mechanism, the identification of reliable targets related to immune-metabolism disorder is critical for facilitating timely intensive clinical interventions [38]. Previously, we reported serum levels of certain molecules could predict the onset and severity of HBV-ACLF [39, 40]. Although some of the analyzed biomarkers linked to HBV-ACLF pathogenesis in plasma could have stemmed from other affected organs, the detection of liver-specific biomarkers is somewhat limited by the fact that liver tissues are scarce because of the risk of massive bleeding in HBV-ACLF patients [41]. Previously reported biomarker candidates of HBV-ACLF emerged as a single comparison without optimal prediction of patients’ prognosis [42, 43]. In this study, which was based on the COSSH multicenter cohort, we investigated THBS1 as a potential biomarker and predictor of HBV-ACLF pathogenesis, disease severity and outcome. Through comprehensive analyses with various methods, external cohort validation and development of two liver failure animal models, we finally characterized THBS1 as an important target and potential novel biomarker involved in ACLF pathogenesis. Interestingly, the robust performance of THBS1 was equivalent to the classic scores for predicting ACLF prognosis, whereas the plasma THBS1 cut-off value that stratified patients with low- and high-risk might facilitate the clinical transition of THBS1 as non-invasive tool for ACLF disease progression.

Inflammation and apoptosis are decisive events in the incidence and progression of ACLF [44, 45]. Hepatic inflammasome-induced interleukins (especially IL-1α and IL-1β) are viewed as an important driving force for the development of fatal ACLF [46]. We discovered that THBS1 was significantly correlated with inflammation-related and apoptosis-related processes. THBS1 is recognized as a crucial factor in immune activation and inflammatory response. Overexpression of THBS1 has been documented in conditions associated with tissue damage and inflammation. In the progression of ACLF, elevated THBS1 levels could contribute to liver dysfunction, intensifying hepatic damage and cellular death through the activation of key pathways, such as TGF-β1 and NF-κB, thereby precipitating liver failure. Furthermore, THBS1 can stimulate the up-regulation of IL-1β, IL-6, and TNF-α in immune cells, potentially regulating cell differentiation and apoptosis by activating inflammatory cells [47]. Researchers were able to verify the involvement of THBS1 in various tissue injuries and inflammatory disorders [48‐51], mainly through interactions with multiple targets and receptors, such as TGF-β, CD36, integrins and CD47 [11, 52]. In line with this evidence, Starlinger et al. highlighted the effects of THBS1 on liver resection and regeneration abilities, as THBS1 is considered a valuable and clinically relevant predictor of postoperative liver dysfunction and poor clinical outcome after liver resection [53]. Moreover, our results highlighted the significant positive correlations between plasma THBS1 levels and the biochemical indices ALT and γ-GT, the clinical manifestations of liver failure. Both ALT and γ-GT have long been considered as biomarkers reflecting liver disease severity and associated with increased risk of metabolic syndrome and mortality [54]. These findings further support the role of immune-metabolism disorder in the HBV-ACLF pathogenesis previously reported by our group [7, 22]. Overall, THBS1 may further aggravate liver failure by promoting hepatic inflammatory responses and inducing hepatocyte apoptosis. With these multiple lines of evidence, THBS1 has emerged as a new biomarker for predicting ACLF short-term mortality, indicated by a significantly higher AUROC, emphasizing the possible combination of THBS1 with the existing scores for predicting severity could improve the predictive accuracy of ACLF short-term mortality.

In the absence of specific HBV-CLF animal models [55], we applied our previously established ACLF rat model [22], to further assess the potential key role of THBS1 in ACLF pathogenesis. The THBS1 biomarker identified from the human PBMC transcriptome in the COSSH multicenter cohort was externally validated in the serum and liver tissues of ACLF rats, which further confirmed the functional association between THBS1, the immune response and cellular apoptosis. Previously, THBS1 was shown to regulate immune activation, enhance the inflammatory response and accelerate fibrosis [56, 57]. Generally, thrombospondins family are induced in sites of tissue damage or active remodeling, and positively involved in the regulation of cellular responses to injury [58, 59]. THBS1 can modulate inflammation through interaction with the major anti-inflammatory cytokines [60], as blocking THBS1 expression not only prevented inflammatory exacerbation but also was linked to augmentation of anti-inflammatory responses in THBS1^KO mice. Similar results were also found in high-fat diet-fed THBS1-deficient mice [61], as well as ureteric obstruction model, which showed blockade of inflammatory lesions and lower renal interstitial inflammation [62]. Interestingly, Min-DeBartolo et al. revealed that THBS1 null mice exhibited a decrease in serum lipid and inflammatory marker levels and hepatic fibrosis in nonalcoholic steatohepatitis [63], but the detailed molecular mechanism of THBS1 regulation of inflammatory response remains to be elucidated.

Many studies have linked THBS1 to the activation of numerous processes that impact cellular apoptosis [64‐66]. Our current findings showed upregulation of hepatic THBS1 expression, which was positively associated with hepatocyte apoptosis, denoting that THBS1 promotes cellular death through the induction of apoptotic pathways. In line with this evidence, a significant reduction in hepatic apoptosis, associated with less injury and liver tissue damage, was observed in THBS1^KO mice following intoxication. Moreover, in vitro/in vivo experiments revealed that THBS1 selectively triggers cellular apoptosis via sequential provocation of the caspase family [67, 68]. However, in-depth verification is still required to reveal the outcomes of THBS1^KO in the dynamic disease course of ACLF. Based on our findings, given the significant correlations of THBS1 with several inflammation- and apoptosis-related genes, we proposed a hypothetical mechanism of THBS1 involvement in ACLF pathogenesis (Fig. 7), which could provide a theoretical foundation and technical support for prospective studies.

Nevertheless, our study is associated with shortcomings, including the study population is based on single precipitant (HBV); however, some prognostic features and disease outcome could vary in ACLF patients with other etiologies as alcoholic cirrhosis, AIH and DILI. Comprehensive investigation of the role of potential confounders that could interfere with THBS1 in predicting the HBV-ACLF outcome also needed to be explored. Second, current lack of universally accepted HBV-based ACLF animal model that represent the disease manifestations, despite the successfulness of our ACLF rat model to resemble the clinical spectrum of the disease. Regardless the shared clinical symptoms between ALF and ACLF [69], which was previously confirmed even at the monocyte transcriptome level [70], we cannot exclude the possibility of distinct differences in the disease entity between ACLF and ALF in our animal models, as mice were insensitive to porcine serum in establishing an immune-based ACLF [71]. If such model is established, the genuine protective capacity of THBS1 knockout could be examined. Third, in light with above findings, a comprehensive exploration of the THBS1 intricate mechanism in the ACLF pathogenesis is still necessary to supplement, and in-depth validations will be conducted in near future with suitable investigations, although it is unlikely to change the overall conclusions of this study.

In summary, THBS1 could be used as a new biomarker of ACLF pathogenesis, prediction of disease severity and short-term outcome. This study characterized THBS1 as robust ACLF disease-related biomarker, which has great importance in providing clinicians with new targets for improving diagnosis and treatment strategies and reducing the high mortality rate of ACLF. Fully uncovering the comprehensive mechanisms of THBS1 in ACLF pathogenesis is still of utmost importance, particularly in a non-HBV-ACLF population, dynamic longitudinal cohort and in suitable ACLF animal models.