The gut-liver axis: A busy two-way street

See Article on Page 1976 The intimate anatomical and functional relationship between the digestive tract and the liver extends into the field of immunology, and a number of associations have been demonstrated. The term “gut-liver axis” has been applied to the immunological phenomenon linking the two in health and disease.1 Such associations include transfer of molecules associated with the gut microbiome to the liver. Many microbiome-associated and immunologically active molecules such as lipopolysaccharide (LPS) enter the portal circulation during health. In addition changes in intestinal permeability and microbiome composition occur in clinically relevant situations such as nonalcoholic steatohepatitis (NASH), ASH, and cirrhosis. The topic of this editorial is programming of T cells upon encountering antigen in one organ such that they subsequently localize to specific sites. This has been best demonstrated for naive T cells, which upon interacting with a specific antigen on a population of dendritic cells in the gut-associated lymphoid tissue (GALT), or on microfold (M) cells in payers patches, acquire high levels of the integrin α4β7 and the chemokine receptor CCR9 which provide the molecular signals allowing subsequent localization to the small intestine.2 Acquisition by T cells of the ability to localize to the site of origin of an antigen seems intuitively necessary for the effector arm of a cellular immune system. What was less obvious was the existence of T-cell homing to the liver after priming by GALT-derived dendritic cells.3 This has been demonstrated for the healthy liver, which possesses selective molecules such as VAP1, and particularly during hepatic inflammation when molecules such as CCL25 and MADCAM1 are up-regulated on liver sinusoidal endothelial cells (LSECs).4 This gut-liver circulation is thought to be important for the development of T-cell-mediated hepatic diseases associated with gut inflammation, but its role in health is not clear.5 GALT, gut-associated lymphoid tissue; iTreg, induced regulatory T cells; LSEC, liver sinusoidal endothelial cell; M, microfold; RA, retinoic acid; RD, retinaldehyde dehydrogenase. The study by Neumann and colleagues6 builds on their earlier work in which they demonstrated that CD4 T cells activated by LSECs (TLSEC) acquire the capacity to home to the liver, which is reminiscent of the ability of GALT-primed T cells to home to the intestine. In addition to hepatic homing there was also significant homing to the small intestine. In the current article, Neumann et al.7 demonstrate that priming of CD4+ T cells by LSEC resulted in a T-cell phenotype that promoted homing to the intestine and the GALT, and this was largely dependent on very specific molecular events (Fig. 1). They initially demonstrated that T cells primed by LSEC (TLSEC) express the gut homing molecules α4β7 and CCR9, but not skin homing molecules.7 Interestingly, the expression of α4β7 remained stable after restimulation by LSEC or splenic cells, but CCR9 expression was lost after restimulation by splenic cells. They subsequently showed that, as predicted by the expression profile of α4β7 and CCR9, priming by LSEC resulted in homing to the liver and mesenteric lymph nodes, but not peripheral lymph nodes. There is a requirement for vitamin A-derived retinoic acid (RA) for the induction of gut-homing molecules in GALT, and this was tested for gut-homing of TLSEC.8 Purified LSEC expressed messenger RNA (mRNA) for two of the four isoforms of the enzyme retinaldehyde dehydrogenase (RD), which converts vitamin A to RA, and also possess enzymatic activity expected from RD. To further confirm the role of this pathway there was significantly less α4β7 expression on TLSEC when they were primed by LSEC from vitamin A-deficient mice, and this could be recovered by exogenously added vitamin A. Priming of CD4+ T cells by LSEC is associated with production of RA by the LSEC. This activates nuclear RA receptors in the T cell and induces up-regulation of the intestinal homing integrins α4β7 and the chemokine receptor CCR9. Additional consequences of signaling to T cells by RA are Th17 and iTreg development, as well as enhancing early T-cell development. The components demonstrated in this article are shown in solid lines. Additional pathways that are known to be driven by RA and may overlap with intestinal homing are shown as dashed lines. There are a number of interesting and significant aspects to the above findings. First, it further cements the importance of RA in the biology of T-cell activation and homing in the gastrointestinal system.9 LSEC now join CD103+ GALT dendritic cells (DCs) as cells that can metabolize vitamin A to produce RA, and regulate T-cell homing to the intestine. RA, however, regulates a much broader range of CD4+ T-cell functions on priming. These include the requirement of RA as a cofactor in the development of induced regulatory T cells (iTreg).10, 11 Oral tolerance is the active suppression of inflammatory responses to orally ingested antigens, and is critically dependent on the iTreg cells.12 As expected, the generation of iTreg cells in response to antigen feeding is abrogated in animals deficient in vitamin A.13 This is very relevant, as oral tolerance is significantly reduced if blood from the intestines bypasses the liver, and hepatic production of iTreg cells by way of RA-dependent LSEC priming may be an important mechanism for this.6 In addition to having a role in the generation of iTreg cells, which limit immune responses to food antigens, RA is also important in the generation of T-helper (Th)17 cells that produce interleukin (IL)-17, IL-21, and IL-22 and are important in control of bacterial and fungal infections.14 This can result in apparently paradoxical effects of RA deficiency, reduced oral tolerance to food antigens, and also reduced immune responses against pathogens. For example, there is a loss in the ability to clear infection with Toxoplasma gondii, and to mount cellular responses to vaccines in the absence of RA.13 Finally, RA is also important in early T-cell activation events, and this may be an issue in states of severe vitamin A deficiency.15 The above known consequences of RA manipulation on T-cell activation and subtype differentiation now conceptually overlap with aspects of liver immunology. The first of these has already been touched upon and relates to the tolerogenic ability of antigens delivered to the liver. A number of mechanisms have been proposed for this ability, and the role of RA adds another valuable mechanism. A very important and poorly understood corollary to the phenomenon of hepatic tolerance is the question of how and when hepatic tolerance is switched off, such that an effective immune response can be mounted.16 In relation to this, it is well known that hepatic stellate cells lose their store of vitamin A upon activation after liver injury. It will be interesting to determine if this loss of vitamin A results in eventual loss of RA, and a reduction in LSEC-mediated production of regulatory CD4+ T cells. This would predict that in the injured and possibly fibrotic liver there may be reduced production of regulatory cells and a more active immune response. There are also a number of liver-specific immune diseases under the umbrella heading of autoimmune hepatitis, and TLSEC have been shown to suppress hepatic inflammation, opening up the possibility that derangements in RA-based signaling has a role in autoimmune hepatitis. Finally, we should be prepared to accept this liver-gut trafficking as a new and unexpected aspect of the better-established gut-liver axis, which is clearly a two-way street.