Good afternoon. I am Adrian DiBascelli. I'm the moderator for this session. It's a pleasure for me to introduce today's speaker for our Hans Popper basic science state of the art lecture. Perhaps for some of our younger members, Dr. Hans Popper was a founding father of hepatology and former president of this association. Our speaker today is Dr. Stephen Artandi. He's the director of the Stanford Cancer Institute, the Jerome and Daisy Lowe Gilbert professor and professor of biochemistry at Stanford University. His research has focused on the role played by telomerase in cancer, aging, and in stem cell function. But today, he's going to tell us about where new hepatocytes come from. It's a pleasure to have you, Dr. Artandi. Well, thank you, Adrian. It's a pleasure to be here and a distinct honor to be giving the Hans Popper lecture. And I'm going to tell you a little bit about our work on hepatocytes and how new hepatocytes are born and the role of telomerase and telomeres in liver biology. I have nothing to disclose. Telomeres are the ends of chromosomes. They represent repeats at chromosome ends in humans and mammals and in all vertebrates. And as shown here, they are these specific dots at every chromosome end. This is a technique called telomere fish, in which you can actually hybridize a probe to detect telomeres. Now, telomeres are not just this simple repeat, TTA, GGG, repeated for up to 10 to 15 kilobases in humans, but they also serve as a platform for binding a protein complex that's called shelter and it's a six-member complex. And what this complex does, together with these repeats, is that they suppress DNA damage responses that would otherwise occur at chromosome ends because they truly represent an end to the chromosome and look like a break. They also will suppress replicative senescence and cell death in a process called crisis. Now, there's a problem here at the end because normal polymerase can't fully replicate the end during cell division. And so, nature has evolved an enzyme called telomerase to actually add, in a de novo fashion, these repeats to chromosome ends. And this is a cryo-EM structure from Kathy Collins' lab that was published last year. Telomerase is a large, complicated enzyme. It's an RNP, a ribonucleoprotein, so it's unlike other simpler enzymes that we tend to think of. It is many proteins assembling on an RNA scaffold shown here in blue. And this RNA serves as a scaffold for telomerase, but it also encodes the complementary sequence to the telomere so that, together, the telomerase RNA and TERT, the telomerase reverse transcriptase, can execute this reverse transcription reaction to elongate telomeres. Now, I'm going to be talking to you a lot about TERT during this seminar because TERT is the subunit that is transcriptionally controlled, and its levels are turned on or off in different cells at different times in humans and in other mammals. Now, there are patients walking around, some of you may have seen them in the clinic, that have inactivating mutations in telomerase-related genes. And those patients, sometimes called diskeratosis congenita, can develop a whole host of tissue failure phenotypes, and I'll just highlight a few of those. One is that they can develop liver cirrhosis. A very common manifestation is pulmonary fibrosis, and another common manifestation is aplastic anemia. So one seems to see that there is a common feature of tissue failure in all of these different phenotypes. Telomerase has also been famously linked to cancer and to immortalization, and here's a paper from many years ago that showed this enzymatic assay and that telomerase activity was associated with nearly all human cancers. And what we learned also many years ago is that human cells have a limited capacity to divide in culture if one tries to grow them in culture. They often will senesce after a very long lag period, many months, and this is the work of Leonard Hayflick from the 1960s, and it turns out that the basis for that replicative senescence is telomere shortening, and that can be opposed or rescued by simply adding TERT into these cells. That's enough to reactivate telomerase, elongate telomeres, and now the cells are immortal. So clearly TERT has a very crucial role to play in cancer more broadly. And for many years we did not know how this process happened, how is TERT actually upregulated in cancer, and just a few years ago we learned from these two papers that there were a set of very fascinating mutations occurring, first in melanoma samples, these are somatic mutations, that occur in the promoter region of TERT. So normally in cancer we're used to thinking about lots of genomic changes, including changes in the coding sequences of genes, but this is a mutation that's a non-coding mutation. It's the most common non-coding mutation in cancer, and it was rapidly found after those papers that these mutations are highly recurrent in other cancer types, including in hepatocellular carcinoma where it's the most common mutation, in melanoma, glioma, and bladder cancer, and actually in all of these cancers it's the most common mutation in each cancer type. There are other cancer types affected, including thyroid carcinoma and some others. So there's been a lot of interest in understanding how these mutations work. What they are, they're one of two mutations in the very proximal promoter region of TERT that create a new consensus sequence, GGAA, for an ETS family. ETS is a very large family of transcription factors, and these mutant sequences recruit one of these ETS factors, called GABP, to the promoter to superactivate it. So a fascinating and very unusual mechanism of cancer oncogene activation. Many of you may be familiar with the pattern of pathways altered in hepatocellular carcinoma. This is from a recent paper, and you can see that the TERT promoters are the most common ones, and after that is p53, which is also common, beta-catenin, which is also common, and then there are less represented pathway mutations after that. The other thing that's fascinating about these mutations in HCC is that they're not only in the invasive HCC compartment, they've also been detected in low-grade dysplastic nodules and in high-grade dysplastic nodules. So they seem to be required for even what we would think of as an early stage of HCC development, and the range in the literature has been anywhere from 43 to 65 percent. There have been evidence for TERT amplification, and another fascinating mechanism is that in HCC that derive from livers that are chronically infected with hepatitis B, there's evidence for hepatitis B virus integration in the promoter sequence, presumably co-opting that promoter and activating it. So it seems to be a really common and critical hallmark of HCC. Now given this pattern of tissue failure phenotypes in people, this led us to propose that perhaps there are TERT-positive stem cells in many of these tissues, and that perhaps they had not yet been identified. And we set out to test that idea using the mouse as a model system. So we knocked in, first we knocked in a TD tomato, which is an RFP protein into the TERT locus so that we can read out TERT promoter activity, and we did this in ES cells, and the ES cells were red because embryonic stem cells have high telomerase. When we differentiated these embryonic stem cells to adipocytes, over time, over this three-week period, you could see that the red cells then became non-red as the reporter was silenced with differentiation to the adipocyte fate. So this is just a confirmatory experiment to indeed show that the TERT promoter is reacting to cell state, and it's active in the pluripotent or the stem cell state, and it goes off when cells differentiate. And we've used this to study the testis, the spermatogonial stem cell compartment, and found that stem cells in this organ have high TERT promoter activity and high telomerase, so that's a feature of the stem cell compartment, and we're working on many different tissues. One of our favorites over the past many years has been to study the liver, and we wondered, how are hepatocytes renewed during life? And of course, a lot of attention has been paid over the years to the diversity of hepatocytes from this portal axis to the central vein, and hepatocytes have been divided into zones based on compartmentalization of the biochemical processes that are enriched in those zones, and also the markers in those zones. And many groups have been trying to understand, where do new hepatocytes come from? How are new hepatocytes born? And a few years ago, the Stanger Lab, the Willenbring Lab, and the Grompy Lab all published papers testing the idea of whether one of the dominant models at the time could be demonstrated through a technique called lineage tracing, and so they used lineage tracing, which I'll introduce in a moment, from the bile duct epithelium, and they found that instead of the bile duct epithelium giving rise to hepatocytes as what had been shown through a lot of beautiful experiments using lineage tracing, they found that essentially there really weren't many or any new hepatocytes that derived from the bile duct epithelium, and instead, they found that new hepatocytes were actually coming from the hepatocyte compartment. And then this raised the question of, well, if that's true, are new hepatocytes coming from all hepatocytes, or are they coming from a subset of hepatocytes? And related to this question is, where does HCC come from, and where does cholangiocarcinoma come from? And I think those are each very important questions that still need to be addressed and are particularly difficult to understand in humans. So we knocked in another reporter into this TERT locus. So in this case, it's CRE-ER. CRE-ER is the CRE recombinase fused to the estrogen receptor, and what that does is it confers estrogen regulatability onto this CRE recombinase. So we knocked this into ES cells made mice, and then we add another transgene into this mouse, which is a strong promoter, LOX-STOP-LOX-TD-TOMATO, again, this RFP variant. So the way that this system works is that there are no red cells in the animal, but if the animal is treated with tamoxifen, which activates CRE-ER to recombine out these cognate LOX-P elements and delete the stop element, now any TERT-positive cell is red, and the important part of this is that it's a permanent labeling, so that if that cell divides, any progeny cell inherit that red label. And the power of this approach is twofold. One can do what's called a short trace experiment, give tamoxifen and analyze the animal just a few days later, and that will tell us, okay, what cells in the tissue are TERT expressing. Or in parallel, we can give tamoxifen and wait a long time, and if these are special cells, then you'll see that these cells give rise to daughter cells, which form a clone. So what we've done is we've done extensive analysis in the liver, and this is the work of an amazing postdoc, Shengdao Lin, who really pioneered all this work and is on the job market, so if anyone's interested, he is looking for an independent position. This is a stitched-together, zoomed-out view of the liver after a three-day trace, only three days after giving a single dose of tamoxifen, and one sees these red cells sprinkled throughout the liver. So we wondered, what are those red cells? And through double immunostaining, what he found is that these red cells are always hepatocytes. They stain for the hepatocyte marker HNF4A, and they're never bile-duct epithelium, nor are they any of the other cell types in the liver. So then Shengdao developed a means for actually purifying these cells from the liver using AAV-expressing GFP in hepatocytes, and he found that by fax, these red cells were about 3% of total hepatocytes, so they're a rare subpopulation of the total pool of hepatocytes in adults. And then we could do biochemical characterization of these cells. So we purified the high and the low cells and did a telomerase enzymatic assay, and we found that telomerase activity is about five times higher than in the low cells. And by qPCR, TERT messenger RNA is about tenfold higher than in the bulk cells. And because we're really obsessed over this, we purified the cells and spun them onto glass slides to perform RNA in situ hybridization, and in this case, you can see the abundant spots of TERT mRNA in the TERT high cells. So indeed, there's a rare subset of cells that have high TERT expression in the liver. Now if these are special cells to repopulate the liver, one will see their progeny expand over time, and that's indeed what we found. So we labeled animals for three days or one month or two months or three months, six months or a year, and you see this really dramatic expansion of the red cell population during this time course. And it's very, very linear over the course of one year, such that about 30 or 35 percent of the cells now of the hepatocytes are red. And these cells are always hepatocytes after a year. These cells don't give rise to bile duct cells or any other cells. Now we also wondered, within the hepatic lobule, where are these rare TERT high progenitor cells? And so we stained sections with antibodies to glutamine synthetase, shown here in green, to light up the pericentral cells and gain insight into the anatomy of the hepatic lobule. And we found that these Tert-High cells were distributed throughout the lobule. They did not follow this rubric of being periportal or pericentral. They were throughout. And at six months, you can see now that they are repopulating the liver in all zones. The zone three, or the pericentral zone, is an interesting one. It's one that's under the influence of Wnt signaling. Our cells are a little bit underrepresented in that zone. But as the animals age, you can see that an increasing percentage of cells in zone three are comprised of progeny from the Tert-High progenitors. And we can use a different approach, which is called rare labeling, to give just a little bit of tamoxifen and make sure that we're labeling only single cells. And when we do that, we can see that, and then we wait six months, we can see that these single cells become clones of two, four, or eight cells from this confocal imaging and 3D image reconstruction. We can make nice movies of these clones and visualize them in three-dimensional space. And then we can also quantitate them. We found that these single cells now are called into action at three months or six months, such that they become clones of increasing size. And we think this is the basis for how the liver is renewed during homeostasis. Importantly, these clones don't diminish over time. So there's no dropout of clones as there is in some other stem cell system where there's clonal competition. And therefore, some clones will lose out. The clone number is maintained. And we also did this low-dose labeling to analyze the location of clones. And we found, at six or 12 months, that the clones are largely in the mid-lobular zones and zones one and two. They can sometimes abut zone three, which, again, is shown here in this green labeling. We have rare clones that will cross this barrier, so be partially in zone two and in zone three. But the bottom line is that the clones are largely mid-lobular and in zones one and two. Because this has been such an area of longstanding interest in the field, we did a very simple experiment, which is just to stain adult liver with antibodies to KEY67. And KEY67 is a marker of cellular proliferation. And then we just simply asked, where are these brown KEY67-positive hepatocytes relative to the portal vein and the central vein? And we graphed those data as a position index shown here. And we found that proliferating cells were also distributed throughout the hepatic lobule, which really fits the data that I just showed you for the behavior of TERT cells. We next wanted to interrogate the cellular mechanism by which the TERT-high cells give rise to clones. And there are two potential models. One is a self-renewal and differentiation model. And according to this model, a high telomerase cell would divide and give rise, in an asymmetric manner, give rise to another high telomerase cell and essentially self-renew, while also giving rise to a more committed hepatocyte that may not have that capacity. And the alternative is that the high telomerase cell may just simply self-duplicate, and you would get an increased number of high TERT cells over time. And so the way that we addressed this was to go back to lineage tracing, give Tamoxifen and analyze at three days, one month, or a year. In that case, we would isolate the RFP-positive cells, and we would assess the percentage of TERT-high cells in that growing clone using RNA fish. And here's the data. What we found is that at three days, almost all the red cells that were labeled in the TERT-ER lineage tracing experiment had very high numbers of these TERT mRNA foci, and that's quantified here. If we look at one month of lineage tracing, not much happens in the liver during the course of just simply waiting homeostatically for one month, and so the results are very similar. The red cells are comprised almost exclusively of cells with high numbers of TERT mRNA molecules. But at one year, after these clones have expanded, as I've shown you, the clone now is comprised largely of TERT-low cells while retaining a subpopulation of TERT-high cells. So what these data show is that when one considers these two alternative models for how the TERT-high cells may renew, this is the model that actually fits the data, that the TERT-high cells divide in an asymmetric manner to renew the TERT-high population and keep that so that the clone can continue to grow while giving rise to more committed or differentiated TERT-low hepatocytes. Now, the liver is an organ that exhibits polyploidy in the hepatocyte population during postnatal life, and we wanted to understand to what extent there might be a difference between the TERT-high and the TERT-low cells in terms of ploidy. So we analyzed ploidy in adult animals using this Herx stain, and we found that there was no significant difference between the TERT-low and the TERT-high cells. They had very similar patterns of cells that had 2C DNA content, 4C DNA content, and 8C DNA content. So ploidy does not explain the difference in the behavior of these cells. So that prompted us to ask, what is the difference in these two populations? And we turned to RNA sequencing to understand how they may be different. And this is what's called a volcano plot, where every dot represents a gene that's differentially expressed between the TERT-low cells on one hand and the TERT-high cells on one hand. And you can see that there are many genes that are significantly differently expressed between the two populations. These are the hepatocyte progenitor cells, and these are the bulk, 97% of hepatocytes. And what we found was very surprising and interesting to me. First, proliferation genes were elevated in the TERT-high population. And this makes sense, because I've shown you that these TERT-high cells are more proliferative than the bulk cells. But what was very surprising is that genes associated with the ribosome were exclusively upregulated in the TERT-low population, as were genes in the electron transport chain. So what this suggested is that there are two types of hepatocytes, based on this analysis. There is the rare subpopulation of hepatocytes, and those are the ones that are acting as progenitors and are more proliferative. And then there are these bulk hepatocytes, and they are the ones doing all the heavy metabolic work of the liver, of the hepatocyte population. They have enriched activity for all the translation activity, all the electron transport chain gene activity that are generating molecules, macromolecules, as well as ATP. And the way I think about it in a simple way is that these are the worker B hepatocytes. And it makes sense for nature to have divided up the work in this way, because of the reactive oxygen species that may be generated during all this metabolic work and the negative consequences of those reactive oxygen species on DNA integrity. So here, by using a subpopulation of cells that express high TERT and lower metabolic activity, this would be a way of keeping the genome of hepatocytes more pristine, and also enabling long-term renewal of the liver, while suppressing the possibility of hepatocellular carcinoma. We also tested directly whether these cells are more proliferative using a seven-day EDU drinking water protocol. We found that they had about a six-fold higher rate of proliferation compared to non-labeled BULK hepatocytes. Now, to understand the role of these TERT-high hepatocyte progenitors during injury responses, we treated animals with a DDC diet. DDC is a chemical that induces widespread damage in the liver. What we found, and this is a normal diet for one month. You see this distributed pattern of the hepatocytes. But after one month of this diet, you see this marked expansion of the red cell lineage. This is really equivalent to one year of renewal under homeostatic conditions, but it's been accelerated by the use of this DDC drug. So the TERT-high cells are activated and renewing the liver during drug treatment. And we tried another injury protocol, which is carbon tetrachloride. And we used this drug in a way such that we gave a single dose of carbon tetrachloride. And when that happens, it induces a zone 3 specific injury because of the enzymes that are located here to metabolize carbon tetrachloride. And this is at three days. You see this necrosis around every central vein. But that necrosis is healed within seven days. And here's an experience to show that these red cells are a little bit underrepresented in that zone 3, but they're nearby in this experiment. But if we injure and analyze at seven days, you see that the newly repaired zone 3 hepatocytes are now represented much more commonly by red cells. And that's quantified here. So this is evidence that the nearby TERT-high progenitors are contributing to healing that chemical wound in zone 3 within seven days. Now, we wanted to answer the question of whether these rare red TERT-high hepatocyte progenitors were actually required for hepatocyte regeneration. And this is a difficult experiment to achieve just for technical reasons. And the way that one wants to do this is to eliminate the red cells from the adult liver and then ask what happens to the liver. And the way that we achieved this was to engineer an AAV, an adeno-associated virus, that has a hepatocyte-specific promoter driving expression of a toxic gene called the diphtheria toxin gene, DTA. But that expression is blocked by a lock, stop, locks element. So if we infect animals with this virus, the hepatocytes get very efficiently incorporated into the hepatocytes. And all the cells are green. Sorry, the cells aren't green at this point because this is stopped. But we treat the animals with tamoxifen. We can selectively delete the stop element and express DTA in the hepatocyte population. So after many months of work, we achieved about an 80% reduction in this population, as shown here. And then we could ask, well, what are these cells really required for? Because we could challenge the animals by putting them back on a DDC diet. And this is what we found. So by staining for serious red, we saw this marked increase in fibrosis in the animals. This is after only one month of challenge with a DDC diet. And commensurate with that fibrotic response, we saw an increase in activation of stellate cells, as shown here, by staining with smooth muscle actin antibodies. So this really shows that the TERT hypergenitors are needed to maintain a functional and healthy liver, and that once they're genetically ablated, the liver becomes fibrotic. So this has led us to propose what we call a distributed model for liver renewal. That's driven by a subset of hepatocytes expressing high TERT. We think that these are hepatocyte progenitors. The way that these cells behave at the single cell level is that they form local clones. And those clones are increasingly comprised of TERT low cells, or cells that would be more metabolically active. During homeostasis, these cells are distributed throughout the lobule, where they're poised to respond to nearby loss of hepatocytes that are lost to injury. And there's this steady and slow expansion of hepatocytes as clones are made. If we injure the liver, chemically or otherwise, this actually accelerates this process, because many, many hepatocytes are killed off. And this further activates these cells to more briskly form clones. And finally, if we ablate the cells, as I showed you genetically, and then challenge the liver to renew, there's impeded regeneration and fibrosis. And this may be something that happens in cirrhosis as there are repeated cycles of damage caused by many different injury types in the human liver. Now, we also are pursuing the idea of whether or not these cells, as they are renewing the liver, they could be what we call the cell of origin for hepatocellular carcinoma. Because as they're doing this heavy, replicative work in the liver, they may sustain mutations that initiate the tumorigenic process. Now, these cells have telomerase. But it's likely that telomerase level is enough for normal renewal. But it's not enough to sustain the high renewal rates of hepatocellular carcinoma. And that may explain why the TERT promoter mutations are such a prominent feature of the HCC transformation process in humans. So with that, I want to thank, again, the scientists who did the work in my lab, Shengda Lin, Liu Chen, Elizabeth Nazimento, and Chandresh Gajara. And I want to thank you very much for your attention. Thank you.

Stanford Cancer Institute

Hans Popper lecture

hepatocytes

telomerase

telomeres

liver biology

TERT gene