Updates on hepatic homeostasis and the many tiers of hepatobiliary repair

Deciphering the complex circuitry of liver homeostasis and repair is required to improve regenerative therapies for hepatic diseases. Studies in 2018 have identified subsets of hepatic cells that have unique reparative abilities, and clarified the role of biomechanical forces and hepatobiliary reprogramming as sustainable modes of tissue repair.

The liver needs to perform >500 functions while being anatomically located to receive portal blood rich in nutrients, toxins and bacterial products, so its capability of many tiers of repair is not surprising. In homeostasis, the liver has a slow cell turnover and mature hepatocytes can proliferate to replace dying cells. Upon injury, these mature cells can proliferate to restore hepatic mass and function. In chronic forms of injury, hepatocytes and cholangiocytes can acquire reciprocal patterns of gene expression to help in repair, since these cells originate from a common progenitor. Key observations in 2018 have greatly improved our understanding of the homeostatic process and the various modes of hepatobiliary repair (Figure 1).

Figure 1 |. Hepatic homeostasis and tiers of hepatobiliary repair.

Randomly located, rare TERT^High cells have a role in maintaining homeostasis (1) and in repopulation after injury (2). After partial hepatectomy, increased sinusoidal blood flow causes enhanced biomechanical stress on liver sinusoidal endothelial cells, which results in expression of angiocrine factors, such as Wnt and HGF (3). Both TERT^High (4) and other hepatocytes (5) can give rise to cholangiocytes after biliary injury. Conversely, cholangiocytes can give rise to hepatocytes (6).

Adult hepatocytes replenish dying hepatocytes as part of normal attrition. Whether all or a few hepatocytes contribute to homeostasis and/or to repair after injury was unknown. Hepatocytes in the periportal region uniquely expand and repopulate liver after carbon tetrachloride (CCl₄) injury, whereas cells in the pericentral region contribute to homeostasis¹. However, zone 1 and zone 3 are especially prone to common forms of hepatic damage, making these cells vulnerable. Moreover, cells that express the proliferation marker PCNA, are randomly distributed, constitute <0.01% of all hepatocytes representing slow turnover, and are not confined to specific zones. In addition, the location of regenerating hepatocytes after injury is dictated by the kind of injury. Thus, the presence of a randomly dispersed cell population with unique homeostatic and reparative properties would impart a distinct advantage. A study in 2018 discovered such a population. Using a reporter mouse, Lin et al.² identified rare hepatocytes with high endogenous telomerase expression (TERT^High hepatocytes or THH) that are randomly localized throughout the liver lobule. Cell tracking demonstrated repopulation of ≥30% of the liver by 1 year, as these cells could both yield both TERT^High and TERT^Low progeny and were indistinguishable from non-labelled cells. Following CCl₄ challenge, which kills zone 3 hepatocytes, THHs in adjacent zones gave rise to daughter cells that restored zone 3. After a more periportal insult with 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC), THHs expanded substantially and some even acquired a cholangiocyte fate. Selective elimination of THHs aggravated injury and fibrosis after DDC. THHs showed enrichment of cell division and receptor tyrosine kinase gene expression and downregulation of metabolic, ribosomal and mitochondrial function gene expression. The expression of high levels of telomerase does endow these cells with long-term proliferative capacity and chromosome stability. Altogether, THHs seem to be a major contributor to homeostasis and epithelia after damage.

The immense regenerative potential of the liver is well known. Removal of 2/3^rd of the liver is well-tolerated and activates various signalling pathways to ensure timely onset and execution of regeneration¹. Hepatocyte growth factor (HGF) and epidermal growth factor signalling are the most critical and redundant pathways, without which there is a complete failure of regeneration and even fulminant hepatic failure basally, if both pathways are eliminated simultaneously³. Other pathways, such as Wnt–β-catenin and RSPO–LGR4/5 signalling, also contribute to an optimal regenerative response, as their absence results in transient but notable delays in peak hepatocyte proliferation after hepatectomy¹. Liver sinusoidal endothelial cells (LSECs) are a key source of many of these signalling molecules, and commonly referred to as angiocrine signals⁴. The most upstream effectors activating angiocrine signalling to promote hepatic growth have remained obscure. In a study in 2018, Lorenz et al.⁵ used liver development as a model of liver growth in the developing embryo and found increased biomechanical forces caused by enhanced blood flow due to increasing cardiac output, to be the main inducer of angiocrine signalling and hepatocyte proliferation⁵. In embryonic and adult liver cultures, increased perfusion rates through sinusoids increased the shear stress and cyclic stretch on LSECs, activating the mechanosensing receptors β1 integrin and VEGFR3, which in turn induced expression of HGF. Although not shown in the context of liver regeneration, LSECs probably experience similarly increased biomechanical forces after partial hepatectomy, due to increased blood flow through the remnant liver⁴. Indeed HGF, WNT2 and WNT9B expression is induced after hepatectomy^3,4,6, and orbital shear stress induced expression of Wnt genes in cultured primary and immortalized LSECs⁶.

Sustained and profound liver injury is associated with the appearance of hepatocyte markers in cholangiocytes and vice versa. Whether this transformation is transient representing metaplasia, or a stable process of transdifferentiation, has remained unclear. Several recent reports have shown successful mutual transdifferentiation, dependent on the extent of injury and the cell type predominantly affected by that injury. Following an important study published in 2017 that showed generation of functional hepatocytes from cholangiocytes after a hepatocyte-specific injury while hepatocyte proliferation was concomitantly inhibited⁷, two independent groups report in 2018 that hepatocyte-directed injury can lead to transdifferentiation of cholangiocytes into hepatocytes when remaining hepatocytes were senescent or incapable of replication^8,9. Following long-term DDC or thioacetamide treatment in mice, Deng et al.⁸ showed both spontaneous senescence and deficient proliferation of hepatocytes. This led to expression of hepatocyte marker in cholangiocytes without appearance of stem cell markers, and showed proliferative advantage resulting in a remarkable contribution to hepatic parenchyma both during extended ongoing injury and after recovery. Similarly, our group⁹ showed that short-term injury induced by a choline-deficient, ethionine-supplemented diet in mice with hepatocyte knockout of β-catenin, a key mediator of hepatocyte proliferation, led to notably increased injury compared with wild-type mice. Upon switching to normal diet, the damaged parenchyma in knock-out mice was repaired through appreciable and stable repopulation with cholangiocyte-derived hepatocytes. Conversely, a functional conversion of hepatocytes into cholangiocytes was demonstrated by Schaub et al.¹⁰. In mice, genetic disruption of Notch signalling and HNF6 resulted in a lack of peripheral bile ducts at birth reminiscent of Alagille syndrome (AS), but a substantial proportion of mice survived and a functional biliary tree was visible at 4 months. Fate-tracing studies identified hepatocytes as the origin of the newly formed ducts. Hepatocyte transplantation studies complemented these observations. Excitingly, hepatocyte-derived cholangiocytes spontaneously organized themselves to functional ducts and became contiguous with the extrahepatic biliary ducts. Gene expression analysis showed active TGFβ signalling in hepatocyte-derived peripheral cholangiocytes, which can have prognostic or therapeutic implications in diseases such as AS. Tissue from patients with AS showed unique nuclear pSMAD3 localization in peripheral bile ducts indicating ongoing TGFβ signalling and spontaneous transdifferentiation.

Clinical specimens from various pathologies that primarily affect hepatocytes (for example, viral hepatitis and steatohepatitis) show prominent ductular reaction composed of expanding cholangiocytes. By contrast, cholangiopathies are often accompanied by the presence of biliary markers in hepatocytes. The findings published in 2018 show that such clinical observations might not just represent transient metaplasia but denote the plasticity of hepatocytes and cholangiocytes to aid in hepatobiliary repair via transdifferentiation. Careful elucidation of the molecular mechanisms of these processes in relevant preclinical models and validation in clinical specimens should have major implications for hepatic regenerative medicine. Through pharmacological modulation of specific signalling pathways, it might be feasible to direct transient metaplasia to stable transdifferentiation and facilitate hepatobiliary repair in various pathologies to eventually obviate the need for organ transplantation.

Key advances:

• Rare and random TERT^High hepatocytes in all metabolic zones of a hepatic lobule represent a distributed model of hepatic homeostasis and regeneration²

• Mechanosensing of altered blood flow by sinusoidal endothelial cells regulates the release of angiocrine factors that influence hepatocyte proliferation during development and regeneration⁵

• Hepatocytes and cholangiocytes can transdifferentiate into each other to aid in hepatobiliary repair when an injury is chronic or excessive and prohibits survival and proliferation of the default cell type ^8,9,10