The enzyme telomerase maintains the length of specialized repetitive structures called telomeres, which are found at the ends of chromosomes. When they become damaged or shortened, telomeres can stop cells from dividing1. Most cells in adult humans have very low or undetectable levels of telomerase and relatively short telomeres, and therefore have a limited ability to replicate2. However, elevated telomerase levels are seen in various animal and human stem cells that must retain their replicative capacity for self-renewal3. Telomerase defects are associated with tissue scarring (fibrosis) in the livers of both mice and humans4,5, but which cells in the liver express telomerase, and whether they act as stem cells, has been unclear. In a paper in Nature, Lin et al.6 characterize this cell population in mice.
First, the authors identified telomerase-expressing cells in the mouse liver and tracked descendent cells. The group genetically engineered mice to contain a modified version of the gene Tert, which encodes a subunit of telomerase. When the mice are treated with a drug, this alteration causes cells expressing Tert to be indelibly labelled by a fluorescent protein. Once the genetically modified cells are triggered in this way, they and all their descendants produce the fluorescent protein, even if the cells no longer express Tert itself.
Lin et al. found that 3–5% of hepatocytes, the most prevalent type of cell in the liver, fluoresce in response to drug treatment. The authors confirmed, by quantitation of messenger RNA levels, that these cells express Tert. Next, they examined the livers of adult mice one year after drug treatment. The initially labelled cells (dubbed TertHigh) had given rise to clusters of descendants dispersed throughout the liver’s lobes, making up about 30% of the liver’s total mass (Fig. 1). Adult hepatocytes die and are replaced infrequently, so the increase in labelled cells over long periods indicates that the TertHigh hepatocytes contribute to the gradual renewal of the liver under normal conditions.
A key question is whether the TertHigh hepatocytes are a stable, self-renewing population. Alternatively, Tert could be expressed in certain cells for a period of time, then shut off in those hepatocytes and expressed in others. In support of the former case, when Lin et al. triggered fluorescent-protein labelling three times over a ten-week period, they found that the numbers of labelled hepatocytes were comparable to those for a single trigger. Next, they showed that 75% of labelled hepatocytes expressed high levels of Tert mRNA when they were examined a month after a single drug treatment, whereas only 18% did so after a year, indicating that, as the population gradually expands, TertHigh cells not only self-renew but also give rise to progeny that do not express Tert (TertLow). Finally, the researchers demonstrated that TertHigh hepatocytes proliferate more than TertLow cells, whereas TertLow cells exhibit higher expression of genes relating to metabolism and biosynthesis than do TertHigh cells.
Taking these data together, the authors suggest that TertHigh hepatocytes behave like stem cells. But before concluding that the TertHigh cells are bona fide stem cells for the liver, it will be necessary to determine whether the TertHigh population becomes exhausted or remains at similar levels in older mice (because hepatocytes are still renewed in ageing mice), and whether TertLow cells convert to TertHigh over longer periods than those used here (which would indicate that this population is not acting as stem cells). It will also be interesting to determine the processes by which cells transition from TertHigh to TertLow, and how this change relates to homeostatic control of liver mass.
Importantly, stem cells typically reside in a special tissue compartment, or niche, that supports their regenerative capacity. Yet the TertHigh cells are dispersed throughout the liver. This dispersal of TertHigh cells is interesting because hepatocytes reside in different zones in each lobe of the liver, and earlier studies7 implicated one zone or another as being more relevant to liver regeneration. By contrast, Lin et al. provide evidence for a ‘distributed model’ for hepatocyte renewal. The research indicates that, although the TertHigh hepatocytes possess features of stem cells, those features are not of a conventional type.
In the past three years, one regenerative hepatocyte population near the central vein has attracted particular attention. The population responds to venous signals to self-renew during homeostasis, producing progeny that migrate outwards from the central zone8. Lin et al. found a few TertHigh hepatocytes in the central zone in healthy livers, but these cells did not reside close enough to the central vein to respond to its signals. However, when the authors damaged the central-vein zone, TertHigh descendants appeared there and responded to venous signals. Moreover, after damage to the liver tissue in another region, around the portal vein, hepatocytes descended from TertHigh cells appeared abundantly in the periportal and mid-lobular zones, and the researchers found that ablation of TertHigh hepatocytes impaired this regenerative response, leading to liver fibrosis. Taking the above findings together with those of other studies of liver injury, it seems that various types of hepatocyte (as well as cells from the bile duct)9,10,11,12 can regenerate the mouse liver under a range of damage conditions.
In the future, it will be crucial to assess how relevant these findings in mice are to human liver regeneration. The fact that ablation of TertHigh hepatocytes results in fibrosis in the injured mouse liver seems to support relevance for humans, because people who harbour mutations in TERT and genes that encode other telomere-related factors can also exhibit fibrosis and cirrhosis (the latter being a predictor of liver cancer)5. However, TertHigh hepatocytes have not been seen in human livers — although the possibility has not yet been assessed with the sensitivity of the genetic-labelling approach used in mice by Lin and colleagues. An alternative explanation for diseases in humans who have telomerase-related mutations is that excessive telomere shortening in early development might affect many organ progenitors in a nonspecific way.
More-detailed studies in humans will be needed to confirm how telomerase-based regeneration forestalls liver disease, and possibly liver cancer. Nevertheless, Lin and colleagues’ study provides insight into a previously unidentified, dispersed-cell mode of liver regeneration.
Nature 556, 181-182 (2018)