Portal venous endothelium in developing human liver contains haematopoietic and epithelial progenitor cells

Abstract

Future treatments for chronic liver disease are likely to involve manipulation of liver progenitor cells (LPCs). In the human, data characterising the regenerative response is limited and the origin of adult LPCs is unknown. However, these remain critical factors in the design of cell-based liver therapies. The developing human liver provides an ideal model to study cell lineage derivation from progenitors and to understand how foetal haematopoiesis and liver development might explain the nature of the adult LPC population. In 1st trimester human liver, portal venous endothelium (PVE) expressed adult LPC markers and markers of haematopoietic progenitor cells (HPCs) shared with haemogenic endothelium found in the embryonic dorsal aorta. Sorted PVE cells were able to generate hepatoblast-like cells co-expressing CK18 and CK19 in addition to Dlk/pref-1, E-cadherin, albumin and fibrinogen in vitro. Furthermore, PVE cells could initiate haematopoiesis. These data suggest that PVE shares phenotypical and functional similarities both with adult LPCs and embryonic haemogenic endothelium. This indicates that a temporal relationship might exist between progenitor cells in foetal liver development and adult liver regeneration, which may involve progeny of PVE.

Introduction

Organ transplantation remains the only cure for chronic liver disease, a condition characterised by severe impairment of hepatocytic function coupled with progressive liver fibrosis. As a result of insufficient donor organ availability, research effort is ongoing to find alternatives to solid organ transplantation. Recently, there has been an increased understanding of the role played by liver progenitor cells (LPCs) in hepatic regeneration after injury [1], [2], [3], [4], [5], [6], [7]. Therefore, characterising the regenerative capacity of the liver and identifying progenitors capable of producing cells with mature hepatocytic function may enable therapeutic manipulation of LPCs either through direct targeting in vivo or through cell therapy.

Liver regeneration is normally provided by division of mature hepatocytes [4]. However, following chronic and/or severe hepatic insult, quiescent LPCs can become activated and contribute to organ regeneration [2], [7], [8]. LPCs, termed oval cells in the rat, are bi-potential and capable of generating progeny with either a hepatocytic or biliary fate [3], [4], [9], [10], [11]. In the human, LPCs have been identified in a number of liver diseases and are likely to originate from the terminal branches of the biliary tree—the Canals of Hering [5], [6], [12], [13], [14], [15], [16], [17]. LPCs are small in number and most extensively characterised in the adult rodent, although phenotypic disparities exist between species [7] and liver injury models [18]. Whilst this has hindered comparative human studies, recent work has identified putative human LPC phenotypes suitable for analysis, although their precise relationship with rodent LPCs requires further clarification [19], [20]. However, in contrast to the situation in adults, the developing human liver is abundant in progenitor cell populations allowing ready investigation [21], [22], [23], [24], [25], [26].

During liver development, ventral endoderm-derived hepatoblasts express the liver epithelial-specific markers Dlk/pref-1, E-cadherin, albumin and α-fetoprotein, in addition to the hepatocytic and cholangiocytic markers cytokeratin CK18 and CK19, conferring an ability to adopt either a hepatocytic or biliary fate [21], [23], [27], [28], [29], [30], [31]. In this respect, hepatoblasts share several phenotypic and functional similarities with adult LPCs [2]. In response to as yet poorly defined signalling events, subsequent hepatoblast differentiation eventually populates the liver parenchyma with functional hepatocytes and bile duct epithelium [32], [33].

As well as generating cells with functional hepatic maturity in an ordered architectural structure, the developing human liver also provides a niche for definitive haematopoiesis. Prior to liver migration and colonization by haematopoietic progenitor cells (HPCs), haematopoietic clusters arise in the embryonic yolk sac and dorsal aorta [34], [35], [36]. These cells originate immediately adjacent to vascular endothelial cells, resulting in the parallel development of haematopoiesis and the primitive foetal circulation. It is established that, whilst yolk sac haematopoiesis generates principally primitive erythroid cells, definitive foetal haematopoiesis is attributable to HPCs arising from haemogenic vascular endothelium in the region of the dorsal aorta [37]. Following transition of haematopoiesis to the developing liver, haematopoietic cells are clustered within portal and sinusoidal endothelial structures [30], indicating that a similar mechanism may contribute to foetal liver haematopoiesis.

In addition to hepatoblast marker expression [2], human and rodent LPCs have been shown to share markers with HPCs including CD133, CD34, c-kit, Sca-1 and mRNA for flt3 [12], [13], [38], [39], [40]. Although these data led some authors to suggest that bone marrow-derived progenitor cells could differentiate into hepatocyte-like cells in vitro [41], [42] and in vivo [43], [44], [45], [46], [47] more recent studies have refuted these findings [48], [49], [50], [51]. Nevertheless, a possible explanation for haematopoietic and epithelial marker co-expression on LPCs might relate to an association between liver development and regeneration, where HPCs would contribute to the epithelial compartment during organogenesis and share phenotypic and functional similarities with their later LPC counterparts [52]. Indeed, in the developing liver, haematopoietic and epithelial lineages exist in immediate anatomical proximity [30].

We have previously investigated this haematopoietic–epithelial lineage relationship during liver development using CD34-expressing cells from 2nd trimester human foetal liver, demonstrating a bi-potential progenitor capability with respect to the hepatocytic and cholangiocytic lineages [22]. Moreover, in later experiments with 1st trimester human liver, a population of CD34-positive cells co-expressing the haematopoietic marker Thy-1, the endothelial marker CD31 and mesenchymal marker vimentin, was localised to the developing portal venous endothelium (PVE) [30]. As a result of these findings, we sought to explore the possibility that PVE might possess an epithelial progenitor capacity during human liver organogenesis, whilst retaining a haematopoietic progenitor capability akin to dorsal aortic endothelium. This would support the notion that progeny of developing PVE might reside quiescently as the presumptive facultative LPC population activated following appropriate injury in the adult liver.

Therefore, since improved understanding of the functional capability of PVE might provide insight into the origin of LPCs, this study aimed to characterise PVE in 1st trimester human liver and investigate the potential of PVE cells for epithelial, haematopoietic and endothelial differentiation.