Differential location of hemopoietic colonies within liver acini of postnatal and phenylhydrazine-treated adult mice

Crosstalk Between the Hepatic and Hematopoietic Systems During Embryonic Development

Hematopoietic stem cells (HSCs) generated during embryonic development are able to maintain hematopoiesis for the lifetime, producing all mature blood lineages. HSC transplantation is a widely used cell therapy intervention in the treatment of hematologic, autoimmune and genetic disorders. Its use, however, is hampered by the inability to expand HSCs ex vivo, urging for a better understanding of the mechanisms regulating their physiological expansion. In the adult, HSCs reside in the bone marrow, in specific microenvironments that support stem cell maintenance and differentiation. Conversely, while developing, HSCs are transiently present in the fetal liver, the major hematopoietic site in the embryo, where they expand. Deeper insights on the dynamics of fetal liver composition along development, and on how these different cell types impact hematopoiesis, are needed. Both, the hematopoietic and hepatic fetal systems have been extensively studied, albeit independently. This review aims to explore their concurrent establishment and evaluate to what degree they may cross modulate their respective development. As insights on the molecular networks that govern physiological HSC expansion accumulate, it is foreseeable that strategies to enhance HSC proliferation will be improved.

Biliary Epithelial Cells Are Not the Predominant Source of Hepatic CXCL12

Hepatic expression levels of CXCL12, a chemokine important in inflammatory and stem cell recruitment, and its receptor, C-X-C chemokine receptor 4, are increased during all forms of liver injury. CXCL12 is expressed by both parenchymal and nonparenchymal hepatic cells, and on the basis of immunohistochemistry, biliary epithelial cells (BECs) are thought to be a predominant source of hepatic CXCL12, thereby promoting periportal recruitment of C-X-C chemokine receptor 4-expressing lymphocytes. Our study aims to show that BECs may, in fact, not be the predominant source of hepatic CXCL12. We measured CXCL12 secretion and expression from human and murine BECs using enzyme-linked immunosorbent assay and Western blot analysis from cell culture supernatants and whole cell lysates, respectively, whereas CXCL12 expression in murine livers was analyzed in a Cxcl12-Gfp reporter mouse. Cell culture supernatants and whole cell lysates from BECs failed to demonstrate their expression of CXCL12. Furthermore, we confirmed these results with a Cxcl12-Gfp reporter mouse in which green fluorescent protein expression is notably absent from BECs. Interestingly, on the basis of green fluorescent protein expression, we demonstrate a population of CXCL12-expressing cells within the portal tract that are distinct, yet intimately associated with BECs. These findings indicate that BECs are not a predominant source of CXCL12.

Hepatic extramedullary hematopoiesis and macrophages in the adult mouse: histometrical and immunohistochemical studies

Fetal liver hematopoiesis in mice disappears approximately 2 weeks after birth; however, under experimental acute anemia extramedullary hematopoiesis occurs in the livers of adult mice. The hematopoietic foci in the liver during extramedullary hematopoiesis contain erythroblasts and macrophages. In this study, the extent of involvement of macrophages in the development and involutional process of the hematopoietic foci in adult mice livers was clarified by experimentally inducing extramedullary hematopoiesis. Hematopoietic cells appeared in the livers 2 days after phenylhydrazine (PHZ) injections. The number and area of the foci increased rapidly, reaching peak values on the sixth day. F4/80-positive macrophages were observed in the sinusoids as well as the hematopoietic foci, and were tightly surrounded by erythroblasts. Sinusoidal macrophages in normal adult livers were positive for F4/80 but negative for ER-HR3. However, in extramedullary hematopoiesis-induced livers, sinusoidal macrophages became positive for ER-HR3 antibodies. The number of ER-HR3-positive macrophages was 9.2 ± 2.9/mm(2) on the second day after PHZ was administered, and increased to 200.3 ± 4.2/mm(2) on the sixth day. On the seventh day after the PHZ injections, the number decreased and they were no longer detected at 30 days after PHZ was injected. The present study revealed that erythroblasts accumulate around sinusoidal macrophages to form an erythroblastic island with a central macrophage similar to erythropoiesis in the fetal liver. Furthermore, in line with development and regression of extramedullary hematopoiesis, macrophages express ER-HR3, a hematopoiesis related antigen.

Extramedullary erythropoiesis in the adult liver requires BMP-4/Smad5-dependent signaling

OBJECTIVE

In mice, homeostatic erythropoiesis occurs primarily in the bone marrow. However, in response to acute anemia, bone morphogenetic proteins 4 (BMP-4)-dependent stress erythropoiesis occurs in the adult spleen. BMP-4 can also regulate stress erythropoiesis in the fetal liver. In humans, erythropoiesis occurs in the bone marrow. However, in certain pathological conditions, extramedullary erythropoiesis is observed, where it can occur in several organs, including the liver. Given these observations, we propose to investigate whether the BMP-4-dependent stress erythropoiesis pathway can regulate extramedullary erythropoiesis in the livers of splenectomized mice.

MATERIALS AND METHODS

Using splenectomized wild-type and flexed-tail (f) mice, which have a defect in BMP-4 signaling, we compared their recovery from phenylhydrazine-induced hemolytic anemia and characterized the expansion of stress burst-forming unit-erythroid in the livers of these mice during the recovery period.

RESULTS

Our analysis indicates that in the absence of a spleen, stress erythropoiesis occurs in the murine liver. During the recovery, stress burst-forming unit-erythroid are expanded in the livers of splenectomized mice in response to BMP-4 expressed in the liver. f/f mice, which exhibit a defect in splenic stress erythropoiesis do not compensate for this defect by upregulating liver stress erythropoiesis. Furthermore, splenectomized f/f mice exhibit a defect in liver stress erythropoiesis, which demonstrates a role for the BMP-4-dependent stress erythropoiesis pathway in extramedullary erythropoiesis in the adult liver.

CONCLUSIONS

Our data indicate that the BMP-4-dependent stress erythropoiesis pathway regulates extramedullary stress erythropoiesis, which occurs primarily in the murine spleen or in the case of splenectomized mice, in the adult liver.

The prometastatic microenvironment of the liver

The liver is a major metastasis-susceptible site and majority of patients with hepatic metastasis die from the disease in the absence of efficient treatments. The intrahepatic circulation and microvascular arrest of cancer cells trigger a local inflammatory reaction leading to cancer cell apoptosis and cytotoxicity via oxidative stress mediators (mainly nitric oxide and hydrogen peroxide) and hepatic natural killer cells. However, certain cancer cells that resist or even deactivate these anti-tumoral defense mechanisms still can adhere to endothelial cells of the hepatic microvasculature through proinflammatory cytokine-mediated mechanisms. During their temporary residence, some of these cancer cells ignore growth-inhibitory factors while respond to proliferation-stimulating factors released from tumor-activated hepatocytes and sinusoidal cells. This leads to avascular micrometastasis generation in periportal areas of hepatic lobules. Hepatocytes and myofibroblasts derived from portal tracts and activated hepatic stellate cells are next recruited into some of these avascular micrometastases. These create a private microenvironment that supports their development through the specific release of both proangiogenic factors and cancer cell invasion- and proliferation-stimulating factors. Moreover, both soluble factors from tumor-activated hepatocytes and myofibroblasts also contribute to the regulation of metastatic cancer cell genes. Therefore, the liver offers a prometastatic microenvironment to circulating cancer cells that supports metastasis development. The ability to resist anti-tumor hepatic defense and to take advantage of hepatic cell-derived factors are key phenotypic properties of liver-metastasizing cancer cells. Knowledge on hepatic metastasis regulation by microenvironment opens multiple opportunities for metastasis inhibition at both subclinical and advanced stages. In addition, together with metastasis-related gene profiles revealing the existence of liver metastasis potential in primary tumors, new biomarkers on the prometastatic microenvironment of the liver may be helpful for the individual assessment of hepatic metastasis risk in cancer patients.

Roles of spleen and liver in development of the murine hematopoietic system

OBJECTIVE

Hematopoietic stem cells (HSCs) and colony-forming progenitor cells (CFCs) are believed to migrate from liver to bone marrow (BM) around the time of birth, where they remain throughout the animal's life. Although in mice the spleen is also a hematopoietic organ, neither the origin nor the contribution of spleen HSCs to hematopoietic homeostasis has been assessed relative to that of BM HSCs. To investigate these issues we quantitated CFC and HSC activity in the spleen, BM, peripheral blood, and liver of the mouse during ontogeny.

METHODS

CFCs were assessed by clonogenic colony formation, and HSCs by long-term reconstituting ability.

RESULTS

CFCs gradually increased in the BM and decreased in the liver with age. Increased prevalence of CFCs in fetal and pup blood occurred at day (d) 12 postcoitus (pc) and during the period of d16 pc to 4d postbirth, corresponding to the times when hematopoietic cells migrate from the yolk sac and/or aorta-gonad-mesonephros (AGM) to the fetal liver and from the neonatal liver to the BM, respectively. In the spleen, CFCs displayed two peaks of activity at 2d and 14d-15d postbirth. Spleen HSCs also fluctuated during this time period. Neonatal splenectomy did not alter CFC or HSC frequencies in the BM, but CFCs increased in the livers of splenectomized mice.

CONCLUSIONS

These data demonstrate that the liver may act as a site of extramedullary hematopoiesis in the neonate, especially in the absence of the spleen, and imply that the spleen, BM, and liver cooperatively contribute to hematopoietic homeostasis.

Thrombopoietin and Its Receptor, c-mpl, Are Constitutively Expressed by Mouse Liver Endothelial Cells: Evidence of Thrombopoietin as a Growth Factor for Liver Endothelial Cells

Present data suggest that the primary site of thrombopoietin (TPO) mRNA is the liver. Previously, we reported that specific murine liver endothelial cells (LEC-1) located in the hepatic sinusoids support in vitro megakaryocytopoiesis from murine hematopoietic stem cells suggesting that these cells may be a source of TPO. We report here that TPO and its receptor, c-mpl, are coexpressed on cloned LEC-1. Enzyme-linked immunosorbent assay (ELISA), biological assay, and flow cytometry studies confirmed the expression of both TPO and its receptor, respectively, at the protein level. TPO activity was enhanced in supernatants from LEC-1 treated with tumor necrosis factor (TNF)-α and γ-interferon (INF). Our results show that TPO through its receptor stimulated the growth of LEC-1 in vitro. These observations establish LEC-1 as a novel source of TPO in the liver. To our knowledge, this is the first report that liver endothelial cells express both TPO and its receptor, c-mpl, and our findings indicate that this cytokine constitutes a growth factor for liver endothelial cells in vitro.

Hepatic oval cells express the hematopoietic stem cell marker Thy-1 in the rat

Hepatic oval cells (HOC) are a small subpopulation of cells found in the liver when hepatocyte proliferation is inhibited and followed by some type of hepatic injury. HOC can be induced to proliferate using a 2-acetylaminofluorene (2-AAF)/hepatic injury (i.e., CCl4, partial hepatectomy [PHx]) protocol. These cells are believed to be bipotential, i.e., able to differentiate into hepatocytes or bile ductular cells. In the past, isolation of highly enriched populations of these cells has been difficult. Thy-1 is a cell surface marker used in conjunction with CD34 and lineage-specific markers to identify hematopoietic stem cells. Thy-1 antigen is not normally expressed in adult liver, but is expressed in fetal liver, presumably on the hematopoietic cells. We report herein that HOC express high levels of Thy-1. Immunohistochemistry revealed that the cells expressing Thy-1 were indeed oval cells, because they also expressed alpha-fetoprotein (AFP), gamma-glutamyl transpeptidase (GGT), cytokeratin 19 (CK-19), OC.2, and OV-6, all known markers for oval cell identification. In addition, the Thy-1+ cells were negative for desmin, a marker specific for Ito cells. Using Thy-1 antibody as a new marker for the identification of oval cells, a highly enriched population was obtained. Using flow cytometric methods, we isolated a 95% to 97% pure Thy-1+ oval cell population. Our results indicate that cell sorting using Thy-1 could be an attractive tool for future studies, which would facilitate both in vivo and in vitro studies of HOC.

Thrombopoietin and Its Receptor, c-mpl, Are Constitutively Expressed by Mouse Liver Endothelial Cells: Evidence of Thrombopoietin as a Growth Factor for Liver Endothelial Cells

Present data suggest that the primary site of thrombopoietin (TPO) mRNA is the liver. Previously, we reported that specific murine liver endothelial cells (LEC-1) located in the hepatic sinusoids support in vitro megakaryocytopoiesis from murine hematopoietic stem cells suggesting that these cells may be a source of TPO. We report here that TPO and its receptor, c-mpl, are coexpressed on cloned LEC-1. Enzyme-linked immunosorbent assay (ELISA), biological assay, and flow cytometry studies confirmed the expression of both TPO and its receptor, respectively, at the protein level. TPO activity was enhanced in supernatants from LEC-1 treated with tumor necrosis factor (TNF)-α and γ-interferon (INF). Our results show that TPO through its receptor stimulated the growth of LEC-1 in vitro. These observations establish LEC-1 as a novel source of TPO in the liver. To our knowledge, this is the first report that liver endothelial cells express both TPO and its receptor, c-mpl, and our findings indicate that this cytokine constitutes a growth factor for liver endothelial cells in vitro.

Extramedullary hematopoiesis in the adult mouse liver is associated with specific hepatic sinusoidal endothelial cells

In previous work, two anatomically distinct-liver sinusoid endothelial cells (LEC): LEC-1 and LEC-2, have been described. We also reported that extramedullary hepatic hematopoiesis occurs only in close contact with LEC-1, suggesting that these cells may provide the microenvironment necessary for the maintenance and growth of hematopoietic cells. In the present work, we studied the capacity of LEC-1 and LEC-2 to maintain in vitro hematopoiesis. LEC-1 and LEC-2 were isolated and cloned from livers of adult mice. Bone marrow cells (BM) enriched with primitive hematopoietic progenitors were isolated from day-2, post-5-FU-treated mice (5-FUBMC). LEC-1 supported the maintenance and differentiation of hematopoietic progenitors for more than 6 weeks in vitro. In contrast, LEC-2 cells poorly supported the proliferation of hematopoietic cells for only two weeks of the co-culture. LEC-1 and 5-FUBMC cocultures showed cobblestone-area formation and the presence of hematopoietic progenitors that are able to form colonies (CFC) in the adhering fraction after six weeks of coculture. LEC-1 co-cultures treated with a cocktail of cytokines (stem cell factor, interleukin [IL]1alpha, IL-3, and Epo) showed that megakaryocyte (CFU-Mk) and erythrocyte progenitors (BFU-e) were present during the entire period of the culture. Granulocyte-macrophage progenitors (CFU-GM) were present only during the first three weeks of the culture. These results suggest that LEC-1, but not LEC-2, provide an appropriate hematopoietic microenvironment for supporting the proliferation and differentiation of primitive hematopoietic cells. This could explain the anatomical restriction of hematopoietic cells for growing in LEC-1 domains during liver extramedullary hematopoiesis.

Endothelial cell heterogeneity in the normal human liver acinus: in situ immunohistochemical demonstration

While a certain degree of structural and functional intra-lobular heterogeneity of sinusoidal endothelial cells has been observed in rodents, little information is available about the zonal characteristics of sinusoidal endothelial cells in the human liver acinus. We have therefore examined the intra-acinar distribution of a panel of endothelial markers in the normal human liver, including: (a) structural markers of continuous and sinusoidal endothelia (PECAM-1, CD-34 protein, VE-cadherin, 1F10 antigen), (b) functional markers specific for sinusoidal endothelial cells, as previously determined in the laboratory (CD4 protein, the lipopolysaccharide-binding protein receptor (CD 14), aminopeptidase N, ICAM-1, receptors II and III for the Fc fragment of immunoglobulins G), (c) endothelial cell-matrix adhesion proteins and leukocyte-endothelial cell adhesion molecules. We observed a heterogeneous distribution for: (a) the 1F10 antigen, whose distribution in the human liver acinus was restricted to vessels situated along the axis of acinar zone 1, (b) the lipopolysaccharide-binding protein receptor and the receptor III for the Fc fragment of IgG, not expressed or only barely expressed in acinar zone 1. The distribution of the other markers tested did not display significant intra-lobular variation. Our in situ results suggest the existence of a degree of zonal heterogeneity in the structural and functional characteristics of sinusoidal endothelial cells in the human liver acinus. This might contribute to the constitution of distinct microenvironments within the human liver parenchyma.

Isolation and enrichment of two sublobular compartment-specific endothelial cell subpopulations from liver sinusoids

Similar to the well-recognized phenotypical heterogeneity of hepatocytes, in situ sublobular variations have recently been detected in the cell structure, fenestration patterns, filtrating efficiency, surface glycosylation, scavenger function and pathological responses of the sinusoidal lining endothelium. However, unlike other liver cell populations, until now no endothelial cell subpopulations had been isolated or defined with clarity, much less with sublobular/acinar zone-related differential properties. On the basis of our previous studies showing that periportal segments of mouse liver sinusoids express a significantly higher number of wheat germ agglutinin-binding sites than do perivenous ones, we used this differential feature for in vitro labeling of the specific sublobular derivation of isolated sinusoidal lining endothelial cells to correlate their original lobular position with other features determined on flow cytometry, centrifugal elutriation, discontinuous arabinogalactan density gradients and electron microscopy. Our results revealed additional heterogeneous properties whose association with high or low wheat germ agglutinin-binding capacity made it possible to define in vitro two dominant endothelial cell subpopulations that appear similar to the differential features in the periportal and perivenous sinusoidal segments. Type 1 endothelial cells had low forward angle light scatter and high integrated side scatter, low cytoplasmic porosity index (12% +/- 5%) and high wheat germ agglutinin-binding efficiency (160 +/- 35 fluorescence intensity units/cell size); these findings are similar to what was observed in situ in the periportal sinusoidal endothelium. On the other hand, type 2 endothelial cells, with high forward angle light scatter and low integrated side scatter, had a high cytoplasmic porosity index (25% +/- 8%) and low wheat germ agglutinin-binding efficiency (60 +/- 15 fluorescence intensity units/cell size), findings similar to in situ observations of the perivenous sinusoidal lining endothelium. Moreover, these physical and morphological differences entail different cell sedimentation behaviors: type 1 endothelial cell sedimented at high centrifugal elutriation counterflow rates (23 to 37 ml/min) and high arabinogalactan density gradient levels (10% to 15%), whereas type 2 endothelial cell sedimented at low counterflow rates (18 to 23 ml/min) and low density levels (6% to 10%). The combination of these separation procedures made it possible to isolate a 90%-enriched type 1 endothelial cell population in the 12% to 15% interphase of the 23 and 37 ml/min elutriation flow rates and a 75%-enriched type 2 endothelial cell population in the 6% to 10% interphase of the 18 and 23 ml/min flow rates.

Differences in the lectin-binding patterns of the periportal and perivenous endothelial domains in the liver sinusoids

We have studied the distribution patterns of carbohydrate terminals on the endothelial surface of the mouse liver microvasculature. For this purpose, a wide battery of FITC lectins specific to glucose, mannose, galactose, fucose, N-acetyl-neuraminic acid, N-acetyl-galactosamine and N-acetyl-glucosamine residues were incubated on liver cryostat sections or intraportally perfused under physiological conditions. All the resulting hepatic sections were examined under fluorescent microscopy and confocal laser scanning microscopy. With the exception of N-acetyl-galactosamine- and fucose-binding lectins, all the perfused lectins specifically bound to the microvascular wall as confirmed by blocking methods using their corresponding sugars. A wide range of binding was, however, observed among the lectins, and the latter were classified into four groups according to their affinities for the different segments of the hepatic microvasculature: (a) equal affinity for all segments (concanavalin A); (b) different affinities depending on acinar zone (wheat germ agglutinin, Ricinus communis toxin, phytohemagglutinin E, Erythrina cristagalli agglutinin and Pisum sativum agglutinin); (c) preferential binding to the sinusoidal network (Lathyrus odoratus, phytohemagglutinin); and (d) lectins that fail to bind to the hepatic microvasculature (N-acetyl-galactosamine- and fucose-binding lectins). Sinusoidal segment walls in acinar zone 1 expressed a higher concentration of certain lectin-binding carbohydrate residues (N-acetyl-neuraminic acid, N-acetyl-galactosamine, galactose, mannose and glucose) than in acinar zone 3. The labeling patterns obtained through the incubation of liver sections or through in vivo perfusion with the different lectins did not always coincide. Only concanavalin A, wheat germ agglutinin and phytohemagglutinin E lectins proved to be concordant (i.e., they produced identical labeling patterns in both procedures).(ABSTRACT TRUNCATED AT 250 WORDS)

Estimating anatomical-functional position coordinates in liver tissue

Hepatocyte enzyme activity was demonstrated by examining adult C57BL/6 mouse liver cryostat sections under a succinate dehydrogenase (SDH) histochemical reaction, and quantified by microspectrophotometry and microdensitometry. The hepatocyte SDH activity gradient along the path between the portal veins (PV) and efferent terminal hepatic venules (THV) was analyzed by measuring the concentration of the chromophore precipitated in 10 consecutive hepatic parenchymal domains located along imaginary lines drawn across the entire PV-to-THV distance. The profiles of intensity or of normalized relative optical density obtained on a high number of lines were correlated with distance values along the PV-to-THV pathway, enabling us to establish a general mathematical function relating SDH activity (chromophore concentration) to position values on a scale of 0 to 10 corresponding to the theoretical PV-to-THV distance. The equation can be used to interpolate the SDH activity surrounding any intrahepatic object located between the PV and the THV, thus making it possible to calculate the object's anatomical-functional position coordinates in the liver acinus. To demonstrate how this method is used, we have calibrated the intrahepatic position of hemopoietic foci induced in the liver tissue of adult mice treated with phenylhydrazine (PHZ), and show that these foci are located on coordinate 3.31 (maximum range 1.25-4.86) of the sinusoidal domain-that is, on the borderline between Rappaport's acinar zones 1 and 2.

Coincident implantation, growth and interaction sites within the liver of cancer and reactive hematopoietic cells

We have examined the anatomical-functional sites within mouse liver where phenylhydrazine (PHZ)-induced hematopoietic foci, and M5076 reticulum cell sarcoma, B16F10 melanoma and Lewis lung-carcinoma cells specifically develop as colonies after intrasplenic injection. Cancer foci occurred predominantly in the 2.4 to 4.0 segment of the sinusoidal pathway, corresponding to hepatic acinar zone I. No significant differences were detected between different types of tumor, including their different tendencies to spontaneously metastasize liver, or as a result of the different procedures used for obtaining foci or metastases. In addition, PHZ-treatment of mice previously injected with tumor cells, resulted in double colonization of the liver tissue by both hematopoietic and cancer cells, predominantly in zone I. This spatial coincidence indicates that non-cancer-specific mechanisms operate in zone I, either promoting implantation and/or growth of cell colonies or, alternatively, inhibiting these processes in the region surrounding the central vein (Rappaport zone 3). Our observations failed to reveal mutual displacement of cancer or hematopoietic foci by potential competition for development sites in zone I. Enumeration and diameter measurements of cancer foci in PHZ-treated animals showed that the presence of hepatic hematopoietic foci coincided with a significant increase in the hepatic metastasis volume. However, the fact that no significant differences in pulmonary metastases occurred in both the PHZ-treated and control mice given tail-vein injection of cancer cells, and that PHZ reduces cancer cell proliferation in vitro, reveal evidence of local interactions with hematopoietic foci which promote growth of cancer foci in liver.

Functional variations in liver tissue during the implantation process of metastatic tumour cells

We have examined several properties of sinusoidal cells in the unaffected tissue of micrometastasis-containing livers. Tumour cells from either B16 melanoma (B16F10) or Lewis lung carcinoma (LLC) were injected intrasplenically in syngeneic mice and sacrificed on the 7th day. Light and scanning electron microscopy (SEM) showed tumour cells in hepatic veins and sinusoids in close contact with endothelial walls and macrophages. Following quantitative analysis of SEM images from sinusoidal walls it was found that endothelial fenestrae from B16F10 or LLC-colonized livers were diffusely reduced both in size and density/microns 2 throughout the sinusoid wall, although especially affected zone 3 segments. Following the intrasplenic injection of 1 microns fluorescent latex particles 1 h prior to sacrifice of the mice a significant reduction of the latex particle uptake by sinusoidal cells was detected in B16F10-colonized livers (27% of controls) which was in contrast to the significant increase in LLC-colonized mice (180% of controls). Despite the focal character of the tumour cell implantation process, hepatic sinusoidal cells reacted diffusely to metastatic cells. However, over liver acini, endothelial cell changes were mainly expressed in zone 3 while phagocytic properties mainly varied in zone 1 and depending on the tumour type. Although the significance of these sinusoidal changes on metastatic development is unclear, data suggests that "soil" conditions in the liver are different before and after being metastasized by tumour cells.