Microbiota transplantation in portal hypertension: promises and pitfalls

Sinusoidal portal hypertension (PH) is the major cause of morbidity and mortality in cirrhosis. While the pathophysiology through which PH develops is complex, changes in portal vein pressure can be expressed as a function of blood flow and resistance in accordance with Ohm’s law [1]. The sinusoidal form of PH begins with increased intrahepatic vascular resistance (IHVR), which is a consequence of structural and functional changes in the liver sinusoids [2]. While structural components of IHVR tend to be permanent (as they mostly represent excessive fibrosis and parenchymal/vascular remodeling of the liver architecture), the dynamic components of IHVR represent potentially reversible abnormalities, such as sinusoidal endothelial cell dysfunction and vascular deregulation [3,4]. As liver disease progresses, a sustained elevation in portal vein pressure promotes vasodilation in the splanchnic arteries (increasing portal venous inflow) and systemic arteries (decreasing systemic vascular resistance) and triggers the development of spontaneous portosystemic shunts (which allow blood to bypass the cirrhotic liver in its return to the systemic circulation) [5]. As a result of these hemodynamic and angiogenic events in cirrhosis, increased inflow into the splanchnic vessels is balanced by the sum of outflow into the liver and into portosystemic collaterals (Figure 1A).

Currently, reduction in PH can be achieved by correcting splanchnic and systemic vasoregulatory changes through the use of non-selective β-blockers, somatostatin analogs, and vasopressin analogs. However, similar therapeutic agents to improve sinusoidal microcirculation and reduce IHVR are lacking [4,6]. One such opportunity may lie in modulating the gut–liver axis, which refers to the bidirectional relationship between gut microbiota and the liver. Cirrhosis negatively affects host–microbiome interactions; in return, disruption of gut microbiota (dysbiosis) appears to contribute to the progression of chronic liver disease [7]. Efforts to restore the composition and function of gut microbiota in cirrhosis have been shown to reduce the severity of hepatic encephalopathy, which is a key decompensating event in liver disease [8,9]. Several molecular and cellular mechanisms can be invoked to explain these observations (Figure 1B). Dysbiosis alters enterohepatic bile acid metabolism, reduces intestinal short-chain fatty acid levels, and impairs the intestinal barrier, allowing the translocation of microbes and their derivatives such as endotoxin (lipopolysaccharide, LPS) into the portal and systemic circulation [10]. Dysbiosis may change the balance of vasoactive transmitters such as nitric oxide (NO) and the release of angiogenic mediators such as vascular endothelial growth factor (VEGF), both in the liver and in the splanchnic circulation. While not necessarily a ‘game changer’ on its own, correction of dysbiosis may work in combination with other factors to favorably alter the development of PH and mitigate adverse clinical outcomes [11].

Despite these advances, the impact of gut dysbiosis on the structural and functional changes that contribute to PH in cirrhosis of different etiologies remain relatively unexplored. This relationship has been investigated in a recent work published in Clinical Science [12]. Huang et al. utilized the bile duct ligation (BDL) model of experimentally induced PH to analyze how microbiota transplantation may ameliorate PH [12]. The authors evaluated changes in hemodynamics, vascular responsiveness, blood chemistries, and the histology and molecular biology of liver and mesenterial tissue in response to two different sources of microbiota transplant obtained from donor (untreated) rats: gut material transplantation (GMT), in which the transplanted material was collected from the terminal ileum and fecal material transplantation (FMT), in which the transplanted material was obtained via collection of fecal pellets. In the first experimental protocol, authors performed GMT or FMT by daily oral gavage for 5 consecutive days starting on day 7 after BDL or sham operation with hemodynamic and molecular biology analysis performed on day 28. In a second group of animals, FMT was administered starting on day 21 with analytical assessment performed on day 35 to assess the therapeutic effects of this intervention later in the course of liver disease [12].

As expected, BDL resulted in significant liver injury and PH, evidenced by higher plasma levels of tumor necrosis factor-α, higher portal vein pressure, higher IHVR, and higher amounts of liver fibrosis as compared with sham rats; liver tissue expression of phosphorylated endothelial nitric oxide synthase (eNOS) was significantly lower, reflecting endothelial dysfunction [12]. BDL rats also showed evidence of splanchnic vasodilation and pathologic mesenteric angiogenesis, indicated by significantly lower superior mesenteric artery (SMA) resistance, higher SMA blood flow, higher density of mesenteric vessels, and higher mesenterial expression of angiogenic proteins (including eNOS, inducible NOS, VEGF, platelet-derived growth factor and hypoxia-inducible factor). Hyperdynamic circulation was also evident in BDL rats based on significantly lower systemic vascular resistance and significantly higher cardiac index. Moreover, increased portosystemic collateral shunting in response to BDL was confirmed by both splenorenal shunt flow rates and the microsphere distribution method [12].

Treatment of BDL rats with early microbial transplantation (days 7–11 after surgery) resulted in lower portal vein pressure at day 28, but this improvement could not be attributed to lower IHVR or improved vascular responsiveness in the liver [12]. Instead, improvement was associated with changes in the splanchnic circulation, evidenced by decreased SMA blood flow and improved vascular responsiveness (but not by improved SMA resistance). Importantly, microbiota transplantation resulted in reduced vascular density and phospho-eNOS levels (but had no effect on the expression of other angiogenic proteins) in the mesentery of BDL rats. In addition, the degree of portosystemic shunting significantly decreased and the vascular contractility of collateral vessels significantly increased in BDL rats after microbiota transplantation. Microbiota transplantation had no impact on BDL-associated changes in systemic vascular resistance and cardiac index. Similar observations were made in response to delayed microbial transplantation when BDL-induced pathological changes were more advanced. Finally, Huang et al. found that bacterial composition of feces collected from BDL rats was significantly different from sham rats. BDL was associated with diminished abundance in Firmicutes including a deficit in Lachnospiraceae. While these deficits persisted despite treatment with FMT or GMT, microbiota transplantation did result in increased abundance of Bifidobacterium in the feces. Moreover, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis indicated significantly decreased prominence of VEGF-associated signaling following microbiota transplantation [12].

The study of Huang et al. provides new evidence for the amelioration of PH hemodynamics in response to microbiota transplantation in the rat model of BDL-induced cirrhosis [12]. Specifically, microbiota transplantation alleviated splanchnic hyperdynamic circulation and reduced blood flow in spontaneous portosystemic shunts, which is directly relevant to PH complications. These changes were accompanied by reduced mesenteric angiogenesis and down-regulation of eNOS in the mesenteric tissue. While there are obvious differences between observations in animal models and the human clinical experience, the impact of microbiota transplantation on hemodynamic parameters and associated molecular changes in splanchnic vessels including portosystemic collaterals provides a mechanistic explanation for the beneficial impact of this intervention on clinical events related to portosystemic collateral formation such as hepatic encephalopathy [9,13].

Neo-angiogenesis has been implicated as a pathological factor in key components of PH such as sinusoidal remodeling, splanchnic hyperemia and collateral formation [6,14]. The work of Huang et al. now shows that attenuation of dysbiosis may favorably affect this process in the mesenterial vasculature. While cellular and molecular details of how gut microbiota drive mesenteric angiogenesis remain incompletely understood, recent research has offered one intriguing piece of the puzzle [15]. Depletion of Paneth cells, which have been known to synthesize and secrete antimicrobial peptides to prevent dysbiosis, significantly attenuated the development of PH and portosystemic collateral formation in both the partial portal vein ligation (PPVL) and BDL models of cirrhosis. Paneth cells grown in intestinal organoid cultures induced a robust angiogenic response when stimulated with microbiota-derived signals, indicating that Paneth cells are directly relevant to mesenteric angiogenesis and subsequent changes in splanchnic circulation [15].

Intriguingly, Huang et al. found that microbiota transplantation had no impact on the extent of liver fibrosis, the degree of sinusoidal vascular responsiveness, the level of hepatic eNOS expression, or the overall IHVR in BDL-induced cirrhosis [12]. Since early stage of PH in cirrhosis mainly reflects increased IHVR, this lack of effect suggests that restorative changes in gut microbiota were not sufficient to affect intrahepatic components of PH in this experimental model. Perhaps BDL was associated with dense fibrosis by the time of sacrifice, masking any beneficial impact of microbiota transplantation on sinusoidal microcirculation. Moreover, as the authors pointed out, microbial transplantation resulted in beneficial changes in, but not the restitution of, original gut microbiota in BDL rats. It is possible that partially improved dysbiosis was not able to correct ongoing lack (or surplus) of microbial metabolites affecting intrahepatic vasoregulation and architectural changes. On this note, however, it was recently reported from another laboratory that ultrasound-based hemodynamic alterations and liver dysfunction in rats with BDL-induced cirrhosis associated with marked dysbiosis were partially corrected by oral administration of Bifidobacterium pseudocatenulatum CECT7765. In that study, eNOS gene expression levels were significantly increased in the liver of BDL rats treated with the bifidobacterial strain, indicating that even simple interventions such as the administration of a single probiotic strain may ameliorate endothelial dysfunction (and the resultant increase in IHVR) in the BDL model [16].

Currently available preclinical models of PH have several limitations and show substantial differences in the pathophysiology [17]. Portosystemic collateral formation is significant and liver function is mostly maintained in the PPVL model, while the use of thioacetamide, carbon tetrachloride or BDL results in advanced chronic liver disease but low degree of collateralization [4,18]. Thus, different mechanisms of host–microbiota interactions may be involved in the development of PH depending on the mode and extent of liver injury [19]. For instance, plasma endotoxin levels are higher when ascites is present, suggesting that the intestinal barrier is increasingly compromised in decompensated cirrhosis and correction of dysbiosis may have a greater impact on disease pathophysiology [18]. In addition, cirrhosis and PH developing in response to mechanical obstruction vs. parenchymal liver disease may be associated with gut microbiota changes of different types and severity.

These considerations are particularly important amidst current epidemiological trends with nonalcoholic fatty liver disease (NAFLD) as we aim to find new approaches for the alleviation of PH [4,20]. NAFLD is closely linked to caloric excess and other dietary changes with direct impact on the gut microbiota. There is now evidence that the intestinal barrier is disrupted and microbial derivatives may contribute to sinusoidal microcirculatory dysfunction in NAFLD before cirrhosis is established [21,22], while restoration of healthy gut microbiota may attenuate early elevations in portal vein pressure associated with increased IHVR [23]. These observations suggest that modulating the inner workings of the gut–liver axis may also become a valid therapeutic goal in the management of pre-cirrhotic liver disease.

In summary, the work of Huang et al. provides new insights into the effects of microbiota transplantation in BDL-induced PH and suggests that simultaneous and successful targeting of IHVR, splanchnic hyperemia, and portosystemic collaterals by this approach may depend on the cause and severity of liver disease as well as on the type and timing of intervention [12]. Further research is needed to confirm and extend these promising observations.

This Commentary piece contains no new data subject to Data Availability Statement.

The authors declare that there are no competing interests associated with the manuscript.

The authors declare that there are no sources of funding to be acknowledged.

Emilie K. Mitten: Writing—original draft. Writing—review & editing. György Baffy: Conceptualization. Writing—original draft. Writing—review & editing.

 
• BDL

  bile duct ligation

 
• eNOS

  endothelial nitric oxide synthase

 
• FMT

  fecal material transplantation

 
• GMT

  gut material transplantation

 
• IHVR

  intrahepatic vascular resistance

 
• NAFLD

  nonalcoholic fatty liver disease

 
• NO

  nitric oxide

 
• PH

  portal hypertension

 
• PPVL

  partial portal vein ligation

 
• SMA

  superior mesenteric artery

 
• VEGF

  vascular endothelial growth factor