In this issue of Hepatology, Starlinger et al. present a successful example of translational research 7 years following the discovery in mice that platelet-derived serotonin mediates liver regeneration in vivo. Effectively, they provide the first evidence that platelet-derived serotonin plays a critical role in liver regeneration in humans following a hepatectomy and that low preoperative serotonin in platelets is associated with delayed postoperative liver regeneration as well as poorer clinical outcomes.
During the last 20 years, liver surgery has benefited from a number of breakthroughs, which have broadened the indications for liver resection, in order to increase the chances of curing many patients with malignant liver tumors. One of the most significant advances has been a better understanding of the mechanisms of liver regeneration. New knowledge in this area has led to innovative strategies to perform safer extended hepatectomies (more that 70% of the liver volume) often through the manipulation of the portal flow prior to resection to allow efficient hypertrophy of the future liver remnant. In liver transplantation, partial grafts from cadaveric donors (split liver) or from living donors have enabled increasing the pool of available organs for transplantation. In living donor liver transplantation, for example, the donor operation involves a major liver resection (up to 60% of the liver) in a healthy adult, whose remnant liver regenerates to almost its normal size within 4 to 6 weeks. In the recipient, the graft also increases rapidly in volume in order to overcome the metabolic function of the diseased native liver. Major liver resection, however, exposes patients to postoperative liver failure because of insufficient liver mass. Below a certain threshold of the liver remnant volume (about 20-25% of the native liver volume) and depending of the quality of the liver parenchyma (e.g., underlying fibrosis or steatosis), the remnant liver may not regenerate, leading to liver failure, so-called small-for-size syndrome (SFSS). This widely used term in the surgical literature means that a liver remnant is too small for the size of an individual.
From a molecular point of view, liver regeneration encompasses the activation of many intra- and extracellular pathways. The current understanding is that liver regeneration after partial hepatectomy involves a large number of genes organized into three networks: cytokines (e.g., tumor necrosis factor-α [TNF-α] and interleukin 6 [IL-6]), growth factors (e.g., hepatocyte growth factor [HGF], epidermal growth factor [EGF], vascular endothelial growth factor [VEGF], and platelet-derived growth factor [PDGF]), and metabolic, although marked redundancy exists among them. Many factors pertaining to liver regeneration have been extensively studied, among which platelets and platelet-derived serotonin appeared in the forefront of the early activation triggering hepatocyte proliferation. Our interest in the impact of platelets on the liver dates back to the early 1990s when platelets were found to contribute to reperfusion injury after cold preservation of the liver as well as normothermic ischemia.[7, 8] Subsequently, we developed an interest in liver regeneration, which has led us to investigate platelets and factors stored by them in relation to hepatocyte proliferation in vivo. We observed that liver regeneration was significantly impaired following a partial hepatectomy in mice subjected to immune thrombocytopenia. Similar results were documented after applying an inhibitor of platelet aggregation (clopidogrel), suggesting that a factor released by platelets may induce or maintain liver regeneration. We then identified platelet-derived serotonin as a new mediator of regeneration in vivo. Figure 1 illustrates the metabolism of serotonin in humans. Experiments with serotonin antagonists showed that serotonin mediated its effect by way of the receptor HTR2A, and to a lesser degree through HTR2B. In another approach, we used knockout mice of tryptophan hydroxylase 1 (tph-1), which display a normal phenotype, except for a dramatic reduction in serotonin content in platelets. These mice failed to induce liver regeneration after a 70% hepatectomy. Since tph-1−/− mice exhibit normal levels of amino acid decarboxylase, they were then treated with the serotonin precursor 5-hydroxytryptophan (5-htp), which restores normal serotonin content in platelets. This reload of serotonin efficiently rescued liver regeneration, demonstrating the novel role of platelet-derived serotonin in triggering hepatocyte proliferation in vivo.
Other groups confirmed this observation through a variety of experiments. For example, thrombocytotic mice consistently exhibited increased liver regeneration, while thrombocytopenic animals showed impaired regeneration. Infused platelets in the liver by way of the portal vein enhanced liver regeneration after a partial hepatectomy, and evidence also indicates that a direct contact between platelets and hepatocytes is necessary for hepatocyte proliferation. These observations in animals led us to speculate that platelet-hepatocyte interaction initiates mitogenic signaling, chiefly mediated through HGF, VEGF, and insulin-like growth factor-1 (IGF-1). Importantly, clinical observations also supported the theory that platelets might be critical for liver regeneration after a hepatectomy in humans.[12-15] In a series of 216 consecutive patients, who underwent partial liver resection for colorectal liver metastases, low immediate postoperative platelet counts (<100 × 109/L) were an independent predictor of delayed postoperative liver function recovery and a risk factor for postoperative mortality. Similarly, low preoperative platelet counts predicted poorer outcomes after hepatectomy for hepatocellular carcinoma. Following a living donor liver transplantation, platelet transfusion was independently associated with liver graft regeneration. More recently, we reported that platelet counts <60 × 109/L at postoperative day 5 (the 60-5 criterion) after liver transplantation are an independent factor associated with severe complications, and early graft and patient survivals.
Following these results, our group investigated whether the enhancement of serotonin signaling may prevent SFSS following a partial liver transplantation. In a mouse SFSS model (remnant liver: 30% of total liver volume), the use of serotonin agonists substantially improved SFSS liver function and animal survival; an effect that was associated with enhanced micro-perfusion and was mediated through the htr2b receptor.
In other sets of experiments using elderly mice, we observed an impaired capacity of the liver to initiate regeneration after a partial hepatectomy, similar to older humans.[3, 17] During aging, hepatic sinusoids lose their dense fenestration and become “pseudocapillarized,” a process that conceivably limits the exchange between the blood and hepatocytes. Serotonin, by way of htr2 class receptors, was able to rescue the age-associated decline in both regeneration and survival following liver resection through a VEGF-dependent pathway. The application of a serotonin agonist induced the reopening of the sinusoidal fenestrae demonstrating an improvement in sinusoidal micro-perfusion prior to the hepatectomy. Together, these findings suggest that a close communication between blood factors, platelets, and hepatocytes through sinusoidal fenestrae is key to the initiation of liver regeneration.
Additionally, increasing evidence supports an involvement of serotonin in a variety of pathological conditions of the liver including chronic cholestasis, nonalcoholic steatotic hepatitis, chronic inflammation of the liver, and liver fibrosis.[19-21]
Starlinger et al. (this issue) have to be commended for providing the first evidence that platelet-derived serotonin is relevant to liver regeneration in humans. With their expertise in preparing plasma and serum without artificial platelet activation, these authors could accurately monitor the intraplatelet pool of serotonin (IP 5-HT) in 60 patients who underwent liver resection. Thirty-five patients served as exploration and 25 as validation. The IP 5-HT pool decreased immediately after liver resection, irrespective of the extent of tissue loss, but progressively recovered until postoperative day 5. In patients with postoperative liver dysfunction (to differentiate from liver failure, see definition in the article), both preoperative and postoperative IP 5-HT pools were lower than in patients with normal postoperative liver function. These observations do not permit, however, concluding whether this effect is associated with the intrahepatic overconsumption of serotonin or whether the platelets could not be loaded with serotonin. Notably, three patients who received selective serotonin reuptake inhibitors (SSRIs), preventing platelets from reloading with serotonin, developed postoperative liver dysfunction. Interestingly, low preoperative IP 5-HT pool predicted the occurrence of postoperative liver dysfunction and morbidity more accurately than preoperative platelet counts. These findings emphasize the fact that platelet composition is more important than the platelet counts per se. This may also explain why we along with others have failed to document any correlation between preoperative platelet counts and the outcomes of liver resection or transplantation.
Following novel findings in experimental models, this article efficiently illustrates that platelet-derived serotonin represents a potential target to improve liver regeneration and outcomes after hepatectomy and liver transplantation, and thus opens new avenues for interventions in patients undergoing major liver surgery.
Mickaël Lesurtel, M.D., Ph.D.Pierre-Alain Clavien, M.D., Ph.D.
Department of Surgery Swiss Hepato-Pancreato-Biliary (HPB) and Transplantation Center University Hospital Zurich Zurich, Switzerland