Up to this point in time, the microanatomical location of the hepatic stem/progenitor cell (HSPC) niche has been described based on circumstantial evidence alone. The first observable cells that enter the cell cycle following certain chemical injury regimens have been considered to demarcate the HSPC compartment. Proposed anatomical niches include the junction of the canal of Hering and the liver cord, within the canal of Hering, periductal, and intraductal. Circumstantial supporting evidence for these niches is presented in a recent review article by Sell.1 The current work by Kuwahara and colleagues2 reported in this issue of HEPATOLOGY appears to largely substantiate what has been suspected for decades. Kuwahara has used label retention assays, classically applied to tissues with high turnover rates, to label asymmetrically dividing hepatic cells. To circumvent the problem of general quiescence within the liver, Kuwahara induces moderate hepatic injury via sublethal acetaminophen (APAP) intoxication. Previous studies conducted within the laboratory of Dr. Neil Theise (senior author of the current work) indicate that this particular injury activates the HSPC compartment.3 Incorporation of the thymidine analog bromodeoxyuridine into newly synthesized DNA labels the proliferating cells, and a subsequent dose of APAP is used to dilute the label within the transient amplifying cell population. The resulting label-retaining cells should pinpoint the locations of asymmetric cell division—the HSPC niche.
By this method, Kuwahara has identified four distinct anatomical locations containing cells that divide asymmetrically in response to APAP treatment (Fig. 1). The first location lies at the junction of the hepatocyte bile canaliculi and the proximal biliary ductules at the canals of Hering. This represents the areas where the liver parenchyma and biliary tree make contact, and many investigators have reasonably suspected that a bipotential HSPC may reside within this structure. The second location is within the intralobular bile ducts. Because this location does not make direct contact with the hepatic parenchyma, Dr. Kuwahara questions whether or not these cells may give rise to hepatocytes and cautions that label retention by these cells may simply be due to insufficient washout of the label. The third location is peripheral to the bile ducts and is represented by a cell that does not express classical HSPC cell markers. Cells matching this description have been identified following treatment with agents that induce portal zone necrosis, particularly allyl alcohol.4 The final HSPC niche identified by Kuwahara's studies lies within the hepatic parenchyma and harbors a cell that resembles a small, transitional hepatocyte. Grisham and coworkers5 have published several articles on the retrorsine/partial hepatectomy model that seem to induce proliferation of this particular cell.
Why would direct evidence of the HSPC niche warrant an editorial? On the surface, the data may seem rather esoteric. From an academic perspective, Kuwahara's data help explain discrepancies in reported growth patterns of expanding progenitor cell populations within the liver. Early observations of HSPC proliferation induced by ethionine described a periductular response comprised of cells that did not generally arrange into ducts.6 In light of the current study, it seems likely that this particular model favors induction of the periductular HSPC compartment. More recently, many laboratories, including our own, have favored some variation of the Solt-Farber model of HSPC activation. In these models, mature hepatocyte proliferation is chemically inhibited prior to chemical or physical liver injury. The regenerative response to this treatment is largely HSPC-mediated.7, 8 These cells do form ducts, which probably derive from the intraductular HSPC compartments, including the canal of Hering. There has been a recent trend in the literature to equate HSPC proliferation in animals and humans with a ductular reaction. This is probably a reflection of the current predominance of modified Solt-Farber models in HSPC studies. From a functional standpoint, these two injury models are identical—that is, they both result in restoration of the hepatic parenchyma and biliary tree. However, two distinct cell populations are clearly responsible for repair. This notion has important ramifications with regard to the comparison of studies that use different injury models, because the distinct HSPC compartments are obviously responding to different, though overlapping, environmental cues.
It is interesting to speculate whether or not the asymmetrically dividing cells that populate these niches represent some form of lineage hierarchy within the HSPC population.1 It is quite appealing to imagine the periductular “null” cell as giving rise to the cytokeratin-positive cell of the canal of Hering, which in turn gives rise to the intraductular cell and periductular hepatocyte-like cell. Kuwahara has demonstrated through analysis of serial sections that the periductular, small hepatocyte-like cell probably derives from the canal of Hering cell population. If this is the case, the HSPC can, indeed, change phenotype while retaining the property of asymmetric division. However, double injury studies conducted within our own laboratory have shown that chemical liver injury administered during the HSPC response to 2-acetylaminofluorene/partial hepatectomy results in the proliferation of non–duct-forming cells from within the population of duct-forming HSPCs (unpublished data). Transition from a duct-forming to a non–duct-forming phenotype would seem to run counter to the expected progression.
It is well known that the body has multiple mechanisms of repair, and perhaps during certain repair processes, cells from all of these niches are invoked. However, one cell type/niche ultimately becomes selected as the most efficient and thus becomes the avenue of repair. Occam's razor states that the explanation of any phenomenon should make as few assumptions as possible and favor the least complicated theory. The rule of parsimony does not, however, supersede the scientific method or intuitive logic. When new evidence arises to support a more complicated explanation, positions must be re-evaluated. (Although the law of parsimony is widely credited to William Ockham [c. 1285-1349], the term Ockham's razor first appeared in 1852 in the works of Sir William Rowan Hamilton [1805-1865]; the origins of the principle can be traced back as far as the fourth century BC.9) Many have attributed discrepancies in HSPC data to differing behavior of “the” HSPC under various experimental conditions. In light of the current work by Kuwahara, it seems more likely that experimental conditions are dictating the responding HSPC niche, resulting in the expansion of subtly different cell types.
The data and technique presented by Kuwahara will, undoubtedly, spur research toward the identification of discriminate surface markers for these distinct HSPCs. Armed with a proven technique to label the asymmetric dividing cell population, dual labeling experiments are now possible. This opens up the exciting prospect of isolating HSPCs from the uninjured liver. Studies on isolated oval cells, which derive from the HSPC population, are in actuality working with a transient amplifying population. These cells maintain bipotentiality but have obviously taken the first steps toward commitance. The true fruits of Kuwahara's work will become apparent in the coming years, and it will be exciting to determine whether or not the cells from these niches have different potentials in terms of liver repopulation following cell transplantation.