Search

The fenestrae of liver sinusoidal endothelial cells (LSECs) allow passive transport of solutes, macromolecules, and particulate material between the sinusoidal lumen and the liver parenchymal cells. Until recently, fenestrae and fenestrae-associated structures were mainly investigated using electron microscopy on chemically fixed LSECs. Hence, the knowledge about their dynamic properties has remained to date largely elusive. Recent progress in atomic force microscopy (AFM) has allowed the study of live cells in three dimensions (X, Y, and Z) over a prolonged time (t) and this at unprecedented speeds and resolving power. Hence, we employed the latest advances in AFM imaging on living LSECs. As a result, we were able to monitor the position, size, and number of fenestrae and sieve plates using four-dimensional AFM (X, Y, Z, and t) on intact LSECs in vitro. During these time-lapse experiments, dynamic data were collected on the origin and morphofunctional properties of the filtration apparatus of LSECs. We present structural evidence on single laying and grouped fenestrae, thereby elucidating their dynamic nature from formation to disappearance. We also collected data on the life span of fenestrae. More especially, the formation and closing of entire sieve plates were observed, and how the continuous rearrangement of sieve plates affects the structure of fenestrae within them was recorded. We observed also the dawn and rise of fenestrae-forming centers and defenestration centers in LSECs under different experimental conditions. Conclusion: Utilizing a multimodal biomedical high-resolution imaging technique we collected fine structural information on the life span, formation, and disappearance of LSEC fenestrae; by doing so, we also gathered evidence on three different pathways implemented in the loss of fenestrae that result in defenestrated LSECs.

Non-alcoholic fatty liver disease (NAFLD) is a widespread liver disease in Western society, but its multifactorial pathogenesis is not yet fully understood. Ultrastructural analysis of liver sinusoidal endothelial cells (LSECs) in animal models and in vitro studies shows defenestration early in the course of NAFLD, promoting steatosis. LSECs and fenestrae are important in the transport of lipids across the sinusoids. However, human ultrastructural data, especially on LSECs and fenestrae, are scarce. This study aimed to explore the ultrastructural changes in perfusion type fixed liver biopsies of NAFLD patients with and without non-alcoholic steatohepatitis (NASH), with a special focus on LSECs and their fenestration. Liver biopsies from patients with NAFLD were fixed using two perfusion techniques, jet and injection fixation, for needle and wedge biopsies, respectively. Ultrastructural changes were studied using transmission electron microscopy. NASH was diagnosed by bright-field microscopy using the SAF score (steatosis, activity, fibrosis). Thirty-seven patients were included, of which 12 (32.4%) had NASH. Significantly less defenestration was found in NASH compared to noNASH samples (p=0.002). Other features, i.e., giant mitochondria and fenestrae size did not differ between groups. Furthermore, we described new structures, i.e., single cell steatonecrosis and inflammatory fat follicles (IFF) that were observed in both groups. Concluding, defenestration was more common in noNASH compared to NASH in human liver samples. Defenestration was not related to the degree of steatosis or fibrosis. We speculate that defenestration can be a protective mechanism in simple steatosis which is lacking in NASH.

The liver represents a frontline immune organ that is constantly exposed to a variety of gut-derived antigens as a result of its unique location and blood supply. With a predominant role in innate immunity, the liver is enriched with various innate immune cells, among which natural killer (NK) cells play important roles in host defense and in maintaining immune balance. Hepatic NK cells were first described as 'pit cells' in the rat liver in the 1970s. Recent studies of NK cells in mouse and human livers have shown that two distinct NK cell subsets, liver-resident NK cells and conventional NK (cNK) cells, are present in this organ. Here, we review liver NK cell subsets in different species, revisiting rat hepatic pit cells and highlighting recent progress related to resident NK cells in mouse and human livers, and also discuss the dual roles of NK cells in liver immunity.

To-date serial block-face scanning electron microscopy (SBF-SEM) dominates as the premier technique for generating three-dimensional (3-D) data of resin-embedded biological samples at an unprecedented depth volume. Given the infancy of the technique, limited literature is currently available regarding the applicability of SBF-SEM for the ultrastructural investigation of tissues. Herein, we provide a comprehensive and rigorous appraisal of five different SBF-SEM sample preparation protocols for the large-volume exploration of the hepatic microarchitecture at an unparalleled X, Y and Z resolution. In so doing, we qualitatively and quantitatively validate the use of a comprehensive SBF-SEM sample preparation protocol, based on the application of heavy metal fixatives, stains and mordanting agents. Employing the best-tested SBF-SEM approach, enabled us to assess large-volume morphometric data on murine parenchymal cells, sinusoids and bile canaliculi. Finally, we integrated the validated SBF-SEM protocol with a correlative light and electron microscopy (CLEM) approach. The combination of confocal scanning laser microscopy and SBF-SEM provided a novel way to picture subcellular detail. We appreciate that this multidimensional approach will aid the subsequent research of liver tissue under relevant experimental and disease conditions.

Background:Oxaliplatin is used as a neo-adjuvant therapy in hepatic colorectal carcinoma metastasis. This treatment has significant side effects, as oxaliplatin is toxic to the sinusoidal endothelial cells and can induce sinusoidal obstruction syndrome (SOS), which is related to decreased overall survival. Glutathione has an important role in the defence system, catalysed by glutathione S-transferase (GST), including two non-enzyme producing polymorphisms (GSTM1-null and GSTT1-null). We hypothesise that patients with a non-enzyme producing polymorphism have a higher risk of developing toxic injury owing to oxaliplatin.Methods:In the nontumour-bearing liver, the presence of SOS was studied histopathologically. The genotype was determined by a semi-nested PCR.Results:Thirty-two of the 55 (58%) patients showed SOS lesions, consisting of 27% mild, 22% moderate and 9% severe lesions. The GSTM1-null genotype was present in 25 of the 55 (46%). Multivariate analysis showed that the GSTM1-null genotype significantly correlated with the presence of (moderate-severe) SOS (P=0.026).Conclusion:The GSTM1-null genotype is an independent risk factor for SOS. This finding allows us, in association with other risk factors, to conceive a potential risk profile predicting whether the patient is at risk of developing SOS, before starting oxaliplatin, and subsequently might result in adjustment of treatment.British Journal of Cancer advance online publication, 3 January 2013; doi:10.1038/bjc.2012.590 www.bjcancer.com.

We previously produced mice with human hepatocyte (h-hep) chimeric livers by transplanting h-heps into albumin enhancer/promoter-driven urokinase-type plasminogen activator-transgenic severe combined immunodeficient (SCID) mice with liver disease. The chimeric livers were constructed with h-heps, mouse hepatocytes, and mouse hepatic sinusoidal cells (m-HSCs). Here, we investigated the morphological features of the chimeric livers and the h-hep gene expression profiles in the xenogeneic animal body. To do so, we performed immunohistochemistry, morphometric analyses, and electron microscopic observations on chimeric mouse livers, and used microarray analyses to compare gene expression patterns in hepatocytes derived from chimeric mouse hepatocytes (c-heps) and h-heps. Morphometric analysis revealed that the ratio of hepatocytes to m-HSCs in the chimeric mouse livers were twofold higher than those in the SCID mouse livers, corresponding to twin-cell plates in the chimeric mouse liver. The h-heps in the chimeric mouse did not show hypoxia even in the twin-cell plate structure, probably because of low oxygen consumption by the h-heps relative to the mouse hepatocytes (m-heps). Immunohistochemical and electron microscopic examinations revealed that the sinusoids in the chimeric mouse livers were normally constructed with h-heps and m-HSCs. However, a number of microvilli projected into the intercellular clefts on the lateral aspects of the hepatocytes, features typical of a growth phase. Microarray profiles indicated that similar to 82% of 16 605 probes were within a twofold range difference between h-heps and c-heps. Cluster and principal component analyses showed that the gene expression patterns of c-heps were extremely similar to those of h-heps. In conclusion, the chimeric mouse livers were normally reconstructed with h-heps and m-HSCs, and expressed most human genes at levels similar to those in human livers, although the chimeric livers showed morphological characteristics typical of growth. Laboratory Investigation (2013) 93, 54-71; doi:10.1038/labinvest.2012.158; published online 12 November 2012

Aims: Oxaliplatin is an important chemotherapeutic agent, used in the treatment of hepatic colorectal metastases, and known to induce the sinusoidal obstruction syndrome (SOS). Pathophysiological knowledge concerning SOS is based on a rat model. Therefore, the aim was to perform a comprehensive study of the features of human SOS, using both light microscopy (LM) and electron microscopy (EM). Methods and results: Included were all patients of whom wedge liver biopsies were collected during a partial hepatectomy for colorectal liver metastases, in a 4-year period. The wedge biopsy were perfusion fixated and processed for LM and EM. The SOS lesions were selected by LM and details were studied using EM. Material was available of 30 patients, of whom 28 patients received neo-adjuvant oxaliplatin. Eighteen (64%) of the 28 patients showed SOS lesions, based on microscopy. The lesions consisted of sinusoidal endothelial cell detachment from the space of Disse on EM. In the enlarged space of Disse a variable amount of erythrocytes were located. Conclusion: Sinusoidal endothelial cell detachment was present in human SOS, accompanied by enlargement of the space of Disse and erythrocytes in this area. These findings, originally described in a rat model, were now for the first time confirmed in human livers under clinically relevant settings. (C) 2016 Elsevier Ltd. All rights reserved.

Hepatocytes are a key target for gene therapy of inborn errors of metabolism as well as of acquired diseases such as liver cancer and hepatitis. Gene transfer efficiency into hepatocytes is significantly determined by histological and functional aspects of liver sinusoidal cells. On the one hand, uptake of vectors by Kupffer cells and liver sinusoidal endothelial cells may limit hepatocyte transduction. On the other hand, the presence of fenestrae in liver sinusoidal endothelial cells provides direct access to the space of Disse and allows vectors to bind to receptors on the microvillous surface of hepatocytes. Nevertheless, the diameter of fenestrae may restrict the passage of vectors according to their size. On the basis of lege artis measurements of the diameter of fenestrae in different species, we show that the diameter of fenestrae affects the distribution of transgene DNA between sinusoidal and parenchymal liver cells after adenoviral transfer. The small diameter of fenestrae in humans may underlie low efficiency of adenoviral transfer into hepatocytes in men. The disappearance of the unique morphological features of liver sinusoidal endothelial cells in pathological conditions like liver cirrhosis and liver cancer may further affect gene transfer efficiency. Preclinical gene transfer studies should consider species differences in the structure and function of liver sinusoidal cells as important determinants of gene transfer efficiency into hepatocytes.

Background/Aims Investigation of apoptosis is pivotal in searching for mechanisms that eliminate colon cancer cells getting trapped in liver sinusoids at the time of surgical removal of the primary tumor. This study focuses on nitric oxide (NO), Fas/FasL and the involvement of interferon-gamma (IFNγ) in liver sinusoidal endothelial cells (LSECs) and in the colon carcinoma cell line CC531s. Methods Apoptosis was quantified and visualized in vitro by specific DNA fragmentation, specific staining and electron microscopy. In vivo experiments were also conducted. Results In co-cultures of LSECs with CC531s, apoptosis of CC531s was observed only when they were pre-treated with IFNγ, and was unaffected by blocking the Fas/FasL pathway. However, LSECs continuously produced NO, and apoptosis was inhibited by NO-inhibitors (NMMA and dexamethasone). When IFNγ-sensitized CC531s were injected into rats, liver weight was lower, in contrast to control conditions where liver weight was higher. Conclusions (i) LSECs induce apoptosis in IFNγ-sensitized CC531s in vitro; (ii) LSECs express FasL; (iii) Fas on CC531s becomes active after IFNγ-treatment; however, (iv) blocking the Fas/FasL pathway had no effect; (v) apoptosis was inhibited by NO-inhibitors; (vi) the immune system uses this IFNγ-activated pathway to support LSECs in killing tumor cells.

Immunocytochemistry has been widely used to localize molecules involved in apoptosis. In this short report, we describe with the aid of confocal laser scanning microscopy the immunolocalization of Fas and FasL on liver sinusoidal endothelial cells, and show how the localization of these two molecules differ when the cells are fixed with different fixation protocols. Methanol fixation shows diffuse staining of Fas and FasL in the cytoplasm, as well as in the nucleus. In contrast, paraformaldehyde fixation reveals the presence of Fas and FasL polarized at one side of the cell and only in the cytoplasm. After fixation with a combination of paraformaldehyde and glutaraldehyde the polarization is still present although the fluorescence is concentrated and located as bright dots in the cytoplasm. In conclusion, paraformaldehyde preserves the (nuclear) membrane-associated structures better then methanol and results in a more accurate localization of Fas and FasL. Understanding the different outcome of these common used fixation protocols will assist investigators to select the most suitable method for visualizing membrane-bound Fas and FasL.

Atomic force microscopy (AFM) has been used to image a wide variety of cells and has proven to be successful in cellular imaging, by comparing results obtained by AFM with SEM or TEM. The aim of the present study was to investigate further the conditions for AFM imaging of living cells and compare the results with those obtained by SEM. We chose to image skin fibroblast and liver sinusoidal endothelial cells of two different sources, because these cells have been well described and characterized in earlier studies. AFM imaging of living cells mainly reveals submembranous structures, which could not be observed by SEM. This concerns the visualization of the overall cytoskeletal architecture and organelles, without the necessity of any preparative steps. The AFM study of living cells allows a time lapse study of dynamic changes of the actin cytoskeleton under the influence of the cytoskeleton-disturbing drug cytochalasin B in cells that can be followed individually during the process. However, softer samples, such as the fenestrated parts of living rat liver sinusoidal endothelial cells in culture could not be visualized. Apparently, these cell parts are disrupted due to tip-sample interaction in contact mode. To avoid the lateral forces and smearing artefacts of contact mode AFM, non-contact imaging was applied, resulting in images of higher quality. Still, endothelial fenestrae could not be visualized. In contrast, contact imaging of immortomouse liver sinusoidal endothelial cells, which are devoid of fenestrae, could easily be performed and revealed a detailed filamentous cytoskeleton.

Backgroundl Aims: Previous studies have shown that hepatic natural killer (NK) cells, also called pit cells, have a higher cytotoxicity against certain tumor cells and have a higher expression of the cell adhesion molecule CD11a as compared with blood NK cells. We further investigated the involvement of the adhesion molecules, reported to be involved in target cell killing by blood NK cells, in pit cell-mediated colon carcinoma cell killing. Methods: 51 Cr-release and DNA fragmentation were used to quantify target cell lysis and apoptosis, respectively. Adhesion of pit cells to CC531s monolayers was quantitated. Results: Flow cytometric analysis showed that pit cells expressed CD2, CD11a, CD18 and CD54. CC531s cells expressed only CD54. Treatment of freshly isolated pit cells with monoclonal antibodies (mAbs) to CD11a and CD18 inhibited not only the pit cell-mediated CC531s cytolysis but also the pit cell-induced apoptosis of CC531s cells. The combination of mAbs to CD11a, CD18 and CD54 further increased the inhibition of pit cell-mediated CC531s cytolysis and apoptosis. Anti-CD2 mAb did not affect these processes. The binding of pit cells to CC531s cells was also inhibited by anti-CD11a, and CD18 mAbs, but not by anti-CD2 mAb. Anti-CD54 mAb reduced the target cell killing and the binding only slightly. Conclusions: These results indicate that CD11a/CD18 (LFA-1) present on pit cells plays an important role in pit cell-mediated target cell adhesion, lysis and apoptosis. This finding might explain why pit cells, which have a higher expression of LFA-1 as compared to blood NK cells, are more cytotoxic against tumor cells as compared to blood NK cells.

The high mortality of colon cancer is to a large extent caused by the frequent occurrence of liver metastasis. This is remarkable, because the liver harbors two distinct cell populations that can eliminate invading cancer cells, namely hepatic natural killer (NK) cells and Kupffer cells. These hepatic NK cells, known as pit cells, are the most cytotoxic cells of the naturally occurring NK cells. However, the mechanism by which pit cells eliminate tumor cells is largely unknown. Because we recently found an indication that apoptosis is involved, we tried to assess the role of this mode of cell death using an in vitro system with isolated pure pit cells (>90%) and CC531s cells, a rat colon carcinoma (CC) cell line. Pit cells induced apoptosis in CC531s cells as shown by quantitative DNA fragmentation, agarose gel electrophoresis, and different modes of microscopy. When extracellular Ca2+ was chelated by ethylene glycol-bis(beta-aminoethyl ether)-N,N,-tetraacetic acid (EGTA) during coincubation or when the pit cells were preincubated with the granzyme inhibitor 3,4-dichloroisocoumarin (DCI), the induction of apoptosis was abolished. These results show that pit cells use the Ca2+-dependent perforin/granzyme pathway to induce apoptosis in the CC531s cells, and not the alternative Ca2+-independent Fas pathway. To further exclude the possibility of the involvement of the Fas pathway, we treated CC531s cells with recombinant Fas ligand. This treatment did not result in the induction of apoptosis, indicating that CC531s cells are resistant to Fas-mediated apoptosis. We conclude therefore that pit cells induce apoptosis in CC cells in vitro by the perforin/granzyme pathway.

Pit cells, or hepatic natural killer (NK) cells, present in rat liver sinusoids, represent an organ-associated NK cell population, with a higher level of activation and a different morphology when compared with peripheral blood NK cells. These cells are the result of an influx of peripheral blood NK cells in the liver microenvironment, followed by an activation or differentiation process toward the highly activated phenotype. In this work we investigated the role of Kupffer cells in this differentiation process of NK cells in the liver sinusoids. In vivo elimination of Kupffer cells with the macrophage cytotoxic drug dichloromethylene diphosphonate induced a decrease in number of hepatic NK cells that paralleled that of Kupffer cells. This effect was further investigated in vitro. Kupffer cell-conditioned medium appeared to enhance the viability, tumor-cytotoxic activity, and adherence of hepatic NK cells to liver endothelial cells in vitro. We conclude therefore that Kupffer cells, present in the microenvironment of the liver sinusoids, play an essential role in the differentiation process of peripheral blood NK cells to the highly activated hepatic NK cell population.

Vreuls C P H, Van Den Broek M A, Winstanley A, Koek G H, Wisse E, Dejong C H, Olde Damink S W M, Bosman F T & Driessen A ?(2012) Histopathology similar to 61, 314318 Hepatic sinusoidal obstruction syndrome (SOS) reduces the effect of oxaliplatin in colorectal liver metastases Aims: Oxaliplatin is an important chemotherapeutic agent used to reduce hepatic colorectal metastases, resulting in tumour reduction and permitting surgical resection. This treatment has significant side effects, as oxaliplatin can induce sinusoidal obstruction syndrome (SOS) in the non-tumour-bearing liver, resulting in increased morbidity. We hypothesized that SOS might impede hepatic perfusion, thereby interfering with the tumour environment and attenuate the response to the chemotherapy. Methods and results: From the prospective database of the Maastricht University Medical Centre we collected 50 patients with hepatic colorectal carcinoma metastases. All patients received neo-adjuvant oxaliplatin followed by partial hepatectomy. Metastases and non-tumour-bearing liver were studied histopathologically. Thirty-two of 50 (64%) patients showed SOS lesions, classified as mild (26%) and moderatesevere (38%). The response to treatment, as expressed in the tumour regression grade (TRG), was grade 1 (10%); grade 2 (14%); grade 3 (28%); grade 4 (32%) and grade 5 (16%). Statistical analysis showed that a higher grade of SOS was associated with a higher grade of TRG (P = 0.016). Conclusion: Developing SOS is associated with a lower tumour response to neo-adjuvant oxaliplatin treatment. Hepatic hypoperfusion due to sinusoidal obstruction syndrome might induce hepatic hypoxia, diminishing the response to chemotherapy.

For an electron microscopic study of the liver, expertise and complicated, time-consuming processing of hepatic tissues and cells is needed. The interpretation of electron microscopy (EM) images requires knowledge of the liver fine structure and experience with the numerous artifacts in fixation, embedding, sectioning, contrast staining and microscopic imaging. Hence, the aim of this paper is to present a detailed summary of different methods for the preparation of hepatic cells and tissue, for the purpose of preserving long-standing expertise and to encourage new investigators and clinicians to include EM studies of liver cells and tissue in their projects.

The distribution and the cellular source of the novel extracellular matrix glycoprotein tenascin were studied in normal and fibrotic rat liver. Cryostat sections of normal rat livers, livers of rats treated with intraperitoneal injections of CCl4 and 4-day-old and 8-day-old primary fat-storing cell cultures were stained for tenascin and desmin using an immunoperoxidase procedure or a double-label immunofluorescence technique. Fat-storing cell cultures were metabolically labeled with H-proline. Radiolabeled proteins were immunoprecipitated from the supernatant with antitenascin antiserum and subjected to polyacrylamide gel electrophoresis. In normal rat livers, tenascin was detected discontinuously along the sinusoids, whereas portal tracts were devoid of staining. In fibrotic rat livers, tenascin was preferentially expressed in areas of cell damage, in slender septa or at connective tissue–parenchymal interfaces. The middle region of broad septa was negative. Desmin-positive fat-storing cells accumulated in areas strongly immunoreactive for tenascin, and double-label immunofluorescence showed cells positive for both tenascin and desmin. In fat-storing cell cultures, both intracellular positivity for tenascin and staining of extracellular fibers were seen. Gel electrophoresis of immunoprecipitated proteins revealed two major and three minor bands with molecular weights consistent with tenascin. We conclude that tenascin is a component of the extracellular matrix of both normal and fibrotic rat livers. The strong expression of tenascin in areas of cell damage, in “early” septa or at septal-parenchymal interfaces, in contrast to its absence from the middle region of mature septa, suggests a role in early matrix organization. Fat-storing cells synthesize and secrete tenascin. (HEPATOLOGY 1992;15:909–916).

In this study, we have investigated the cell population kinetics of fat-storing cells in livers of rats intoxicated with CCl4. Fat-storing cells were identified in cryostat sections by immunoperoxidase staining of desmin. The peroxidase label was visualized using diaminobenzidine/hydrogen peroxide containing Ni2+ and Co2+-ions (Nico/diaminobenzidine method). In normal rats, we found 12.8 fat-storing cells/0.1 mm2 in periportal areas vs. 9.4 in pericentral fields. After one injection of CCl4, the number of pericentral cells increased gradually to reach a maximum of 39.4 cells/0.1 mm2 96 hr after injection. The desmin staining intensity of the pericentral fat-storing cells increased from 48 hr onward. At 72 to 120 hr, strongly stained cells were observed in pericentral areas and in bands of tissue between adjacent central veins, reminiscent of the connective tissue septa in fibrotic livers. In the periportal areas the number of fat-storing cells was not altered. After a second and third injection of CCl4, the number of cells increased further in the pericentral areas. When more than three injections were given, the pericentral fat-storing cell population reached a new steady state with the cell number being seven times higher than in control animals. Proliferation of fat-storing cells at different stages of CCl4 intoxication was studied by intravenous administration of 3H-thymidine, followed by combined desmin staining and autoradiography. Autoradiographical labeling of fat-storing cells was nearly absent in control animals and at 24 hr after a single CCl4 injection. At 48 to 96 hr, labeling indices of pericentral fat-storing cells were significantly higher than in control animals, with a maximum at 72 hr when 22.9% of the cells were labeled. After multiple injections of CCl4, labeling indices between 4.9% and 8.4% were found. We conclude that fibrogenesis is preceded by a strong expansion of the fat-storing cell population in the pericentral areas of the liver lobules and in bands of tissue between adjacent central veins. Local proliferation is an important mechanism underlying the expansion of this cell population. (HEPATOLOGY 1991;13:1193–1202.)

Liver sinusoidal endothelial cell (LSEC) fenestrae are membrane-bound pores that are grouped in sieve plates and act as a bidirectional guardian in regulating transendothelial liver transport. The high permeability of the endothelial lining is explained by the presence of fenestrae and by various membrane-bound transport vesicles. The question as to whether fenestrae relate to other transport compartments remains unclear and has been debated since their discovery almost 40 years ago.In this study, novel insights concerning the three-dimensional (3D) organization of the fenestrated cytoplasm were built on transmission electron tomographical observations on isolated and cultured whole-mount LSECs. Classical transmission electron microscopy and atomic force microscopy imaging was performed to accumulate cross-correlative structural evidence.The data presented here indicate that different arrangements of fenestrae have to be considered: i.e. open fenestrae that lack any structural obstruction mainly located in the thin peripheral cytoplasm and complexes of multifolded fenestrae organized as labyrinth-like structures that are found in the proximity of the perinuclear area. Fenestrae in labyrinths constitute about one-third of the total LSEC porosity. The 3D reconstructions also revealed that coated pits and small membrane-bound vesicles are exclusively interspersed in the non-fenestrated cytoplasmic arms.

Fenestrae allow the passage of gene transfer vectors from the sinusoidal lumen to the surface of hepatocytes. We have previously shown that the diameter of fenestrae correlates with species and strain differences of transgene expression following intravenous adenoviral transfer. In the current study, we demonstrate that the diameter of fenestrae in humans without liver pathology is 107+/-1.5 nm. This is similar to the previously reported diameter in New Zealand White (NZW) rabbits (103+/-1.3 nm) and is significantly smaller than in C57BL/6 mice (141+/-5.4 nm) and Sprague-Dawley rats (161+/-2.7 nm). We show that the diameter of fenestrae in one male NZW rabbit and its offspring characterized by a more than 50-fold increase of transgene expression after adenoviral gene transfer is significantly (113+/-1.5 nm; P

Collagen type I and procollagen type III were localized at the ultrastructural level on ultrathin frozen sections of rat liver by the protein A–gold technique using affinity-purified primary antibodies. Both collagen type I and procollagen type III were localized on nearly all solitary and bundled fibrils in the space of Disse. Simultaneous localization of collagen type I and procollagen type III by a double-labeling procedure using protein A–gold probes of different sizes unequivocally demonstrated the presence of both collagens in the same fibrils. Measurement of the diameter of large numbers of collagen fibrils in the space of Disse of the rat liver showed a unimodal distribution of the fibril diameters around an average value of 62.4 nm (S.D. = 12.8 nm), and 91% of the collagen bundles contained less than 30 fibrils. Additional measurements on epoxy resin–embedded material of five biopsy specimens of normal human liver showed a comparable unimodal distribution of the fibril diameters around an average value of 57.2 nm (S.D. = 9.6 nm), and 74% of the bundles contained less than 60 fibrils. The latter observation demonstrates that human liver contains broader interstitial collagen bundles than rat liver. From these results, we conclude that the space of Disse of normal rat and human liver contains a uniform population of striated interstitial collagen fibrils. In the rat liver, these fibrils contain both collagen type I and procollagen type III. Therefore the concept that procollagen type III is predominantly localized in small diameter fibrils or bundles, whereas collagen type I is preferentially localized in thick ones, does not hold. (HEPATOLOGY 1990;12:233–241).

The present paper reviews the literature on the ultrastructure and function of sinusoidal fat-storing cells and pit cells in the mammalian liver. Ultrastructurally, fat-storing cells are characterized by the presence of cytoplasmic fat droplets, well developed rough endoplasmic reticulum; a Golgi complex; multivesicular bodies; one or two centrioles; and few, rather small, lysosomes. These lysosomes are sometimes associated with fat droplets. Fat-storing cells may bear a cilium and project characteristic cytoplasmic processes into the space of Disse. These processes contain microtubules and filaments. Fat-storing cells are the main storage site of retinol esters in the mammalian body. Moreover, these cells have the potential of synthesizing several connective tissue components including the collagens type I, III, and IV; fibronectin; laminin; heparan sulfate; chondroitin sulfate; and dermatan sulfate. Pit cells are polarized cells, with most organelles localized at one site of the nucleus near the cytocentre. They are characterized electron microscopically by the presence of dense cytoplasmic granules with a specific ultrastructure, by rod-cored vesicles, and by multivesicular bodies. It has recently been shown that pit cells have natural killer activity to certain tumor cells and have many features in common with large granular lymphocytes. They therefore may act in the liver as a first line of defense against neoplasia, metastasis, and viral infections.

Colorectal cancer (CRC) is a common malignant disease and the severe nature of cases in men and women who develop colorectal cancer makes this an important socio-economic health issue. Major challenges such as understanding and modeling colorectal cancer pathways rely on our understanding of simple models such as outlined in this paper. We discuss that the development of novel standardized approaches of multidimensional (correlative) biomolecular microscopy methods facilitates the collection of (sub) cellular tissue information in the early onset of colorectal liver metastasis and that this approach will be crucial in designing new effective strategies for CRC treatment. The application of X-ray micro-computed tomography and its potential in correlative imaging of the liver vasculature will be discussed.

Hepatic NK cells are more cytotoxic than blood NK cells against tumor cells. To understand the basis of this difference in cytotoxicity we analyzed RNA derived from freshly isolated rat blood and hepatic NK cells [high-density (HD) and low-density subpopulations] by high-density oligonucleotide arrays (Affymetrix), containing about 9,000 genes and expressed sequence tags. IL-2-treated blood NK (A-NK) cells and IL-2-treated hepatic HD cells were used as a reference of NK cell activation. About 150 genes and expressed sequence tags were differentially expressed between hepatic and blood NK cells. Surprisingly, more than half of the increased expressed genes in hepatic NK cells were not increased in A-NK cells. Differentially expressed genes like the stem cell factor receptor c-kit and the chemokine receptor CCR5 can contribute to the homing and differentiation of hepatic NK cells in the liver sinusoids. Several of the differentially expressed genes can possibly contribute to the enhanced cytotoxic activity of hepatic NK cells: cell membrane receptors like NKG2D, NKG2C, CD94, ecto-ATPase; signaling molecules like phosphatidylinositol 3-kinase; granule-associated effector molecules like granzymes and defensin NP3. Moreover, it appears that redirection of cytotoxic granules and increase in intracellular Ca2+ are convergence points of several of these genes.

Liver sinusoidal endothelial cells (LSECs) can optimally be imaged by whole mount transmission electron microscopy (TEM). However, TEM allows only investigation of vacuum-resistant specimens and this usually implies the study of chemically fixed and dried specimens. Cryo-electron microscopy (cryo-EM) can be used as a good alternative for imaging samples as whole mounts. Cryo-EM offers the opportunity to study intact, living cells while avoiding fixation, dehydration and drying, at the same time preserving all solubles and water as vitrified ice. Therefore, we compared the different results obtained when LSECs were vitrified using different vitrification conditions. We collected evidence that manual blotting at ambient conditions and vitrification by the guided drop method results in the production of artefacts in LSECs, such as the loss of fenestrae, formation of gaps and lack of structural details in the cytoplasm. We attribute these artefacts to temperature and osmotic effects during sample preparation just prior to vitrification. By contrast, by using an environmentally controlled glove box and a vitrification robot (37 degrees C and 100% relative humidity), these specific structural artefacts were nearly absent, illustrating the importance of controlled sample preparation. Moreover, data on glutaraldehyde-fixed cells and obtained by using different vitrification methods suggested that chemical prefixation is not essential when vitrification is performed under controlled conditions. Conditioned vitrification therefore equals chemical fixation in preserving and imaging cellular fine structure. Unfixed, vitrified LSECs show fenestrae and fenestrae-associated cytoskeleton rings, indicating that these structures are not artefacts resulting from chemical fixation.

Liver sinusoidal endothelial cells (LSECs) possess fenestrae arranged in sieve plates. Hepatic endothelial fenestrae are open pores approximately 100 to 200 nm in diameter. Alterations in their number or diameter by hormones, xenobiotics, and diseases have important implications for hepatic microcirculation and function. Numerous reports of hepatotoxin-induced defenestration suggest that the cytoskeleton and the energy status of hepatic endothelial cells play a key role in the regulation of fenestrae. Therefore, we investigated the effect of antimycin A, an inhibitor of mitochondrial energy production, on the number of fenestrae in cultured LSECs using high-resolution microscopy and immunocytochemistry. Prolonged incubation (greater than 30 min) with antimycin A resulted in defenestrated cells and coincided with the appearance of F-actin dots, whereas the distribution of G-actin remained unchanged. Adenosine triphosphate (ATP) was depleted dramatically to less than 5% within 30 minutes within the LSECs. After treatment with antimycin A, unusual elevated fenestrated complexes were apparent, organized as a meshwork of anastomosing fenestrae at the center of and above the sieve plates. The position and appearance of these novel structures and their association with defenestration suggest that they are implicated in the process of defenestration. In conclusion, the results of experiments with antimycin A suggest that ATP is needed to maintain fenestrae and the underlying fenestrae-associated cytoskeleton rings that maintain fenestrae patency. Antimycin A-induced defenestration of LSECs is associated with the development of a structure in the sieve plate that appears to be intrinsically involved in defenestration.

The atomic force microscope (AFM) is a powerful tool to investigate surface and submembranous structures of living cells under physiological conditions at high resolution. These properties enabled us to study the interaction between live hepatic natural killer (NK) cells, also called pit cells, and colon carcinoma cells in vitro by AFM. In addition, the staining for filamentous actin and DNA was performed and served as a reference, because actin and nuclear observations at the light microscopic level during the cytotoxic interaction between these two cell types have been presented earlier. In this study, we collected evidence that conjugation of hepatic NK cells with CC531s colon carcinoma cells results in a decreased binding of CC531s cells to the substratum as probed with the AFM in contact mode as early as 10 min after cell contact (n = 11). To avoid the lateral forces and smearing artefacts of contact mode AFM, non-contact imaging was performed on hepatic NK/CC531s cell conjugates, resulting in identical observations (n = 3). In contrast, the first cytotoxic signs, as determined with the nuclear staining dye Hoechst 33342, could be observed 3 h after the start of the co-culture. This study illustrates that the AFM can be used to probe early cytotoxic effects of effector to target cell contact in nearby physiological conditions. Other routine cytotoxicity tests detect the first cytotoxic effects after 1.5-3 h co-incubation at the earliest.

Hepatic natural killer (NK) cells are located in the liver sinusoids adherent to the endothelium. Human and rat hepatic NK cells induce cytolysis in tumor cells that are resistant to splenic or blood NK cells. To investigate the mechanism of cell death, we examined the capacity of isolated, pure (90%) rat hepatic NK cells to kill the splenic/blood NK-resistant mastocytoma cell line P815. Cell death was observed and quantified by fluorescence and transmission electron microscopy, DNA fragmentation, and 51Cr release. RNA and protein expression were determined by real time reverse transcription-polymerase chain reaction and Western blotting. Compared with splenic NK cells, hepatic NK cells expressed higher levels of perforin and granzyme B and readily induced apoptosis in P815 cells. Although P815 cells succumbed to recombinant Fas ligand (FasL) or isolated perforin/granzyme B, hepatic NK cells used only the granule pathway to kill this target. In addition, hepatic NK cells and sinusoidal endothelial cells strongly expressed the granzyme B inhibitor, protease inhibitor 9 (PI-9)/serine PI-6 (SPI-6), and P815 cells and hepatocytes were negative. Transfection of target cells with this inhibitor resulted in complete resistance to hepatic NK cell-induced apoptosis. In conclusion, hepatic NK cells kill splenic/blood NK-resistant/FasL-sensitive tumor cells exclusively by the perforin/granzyme pathway. Serine protease inhibitor PI-9/SPI-6 expression in liver sinusoidal endothelial cells may protect the liver microenvironment from this highly active perforin/granzyme pathway used to kill metastasizing cancer cells.

21-Aminosteroids are potent anti-inflammatory and antioxidant drugs that provide remarkable endothelial protection in different models of tissue ischemia-reperfusion and inflammation. The effects of 21-aminosteroids in sepsis, a highly inflammatory condition leading to panendothelial activation and injury, are largely uninvestigated. We therefore explored the effects of the 21-aminosteroid U74386G on hepatic blood flow, endothelial cell function, and sinusoidal structure in a canine model of fluid-resuscitated, hyperdynamic endotoxic shock.Following invasive hemodynamic monitoring and placement of ultrasonic flow probes around the common hepatic artery and the portal vein, 12 anesthetized dogs received 2 mg/kg iv of Escherichia coli endotoxin, followed by generous saline infusion, before randomization into two groups. One group (N = 6) received U74389G as an iv bolus of 80 microg/kg, followed by a continuous infusion of 10 microg/kg. min. The other group (N = 6) received an equivalent volume of vehicle. Hyaluronic acid was measured in plasma for in vivo evaluation of endothelial cell function. Liver biopsies were taken after 4 h of endotoxic shock and prepared for light and electron microscopic examination.Compared with the vehicle-treated controls, U74389G maintained a higher blood flow in the hepatic artery and in the portal vein, without markedly influencing the systemic hemodynamic response. The endotoxin-induced increase in plasma hyaluronic acid levels was significantly attenuated following U74389G treatment (70 +/- 14 vs 188 +/- 24 ng/mL after 3 h of endotoxic shock; P0.05). Morphological studies showed that the U74389G-treated group had less sinusoidal endothelial cell damage together with a dramatic reduction of neutrophil infiltration into the liver tissue.U74389G can preserve the functional and structural integrity of endothelial cells in the hepatic sinusoid during hyperdynamic endotoxic shock. This endothelial-protective effect was associated with a better maintained hepatic blood flow and a significant attenuation of inflammatory liver injury.

Pit cells are liver-specific natural killer (NK) cells and belong to the group of sinusoidal cells, together with Kupffer, endothelial, and fat-storing cells. Pit cells are lymphoid cells containing specific granules, classifying them also as large granular lymphocytes (LGL). They probably originate from the bone marrow, circulate in the blood, and marginate in the liver, where they develop into pit cells by lowering their density and increasing the number of granules, which decrease in size. Pit cells remain in the liver about 2 weeks and are dependent on Kupffer cells. Pit cells also proliferate locally, when stimulated with interleukin-2, biological response modifiers, or other agents. Pit cells adhere to tumor target cells during killing. They possess a high level of natural cytotoxicity against a variety of tumor cell lines, which is comparable to the cytotoxicity level of lymphokine-activated killer (LAK) cells. Tumor cell killing is synergistically enhanced when pit cells attack tumor cells together with Kupffer cells. Further investigations are needed to clarify the mechanisms of pit cell cytotoxicity and the role of these cells in killing virus-infected cells, such as during viral hepatitis.

Previous studies showed that blood large granular lymphocytes (LGL), which possess natural killer (NK) activity, develop within rat liver sinusoids into high-density (HD) and subsequently into low-density (LD) pit cells which show an increasing level and spectrum of tumor cytotoxicity. In this study, we investigated the role of adhesion molecules, such as CD2, CD11a, CD18, and CD54 in the recruitment of pit cells to the liver. Immunostaining for electron microscopy, and two color flow cytometry showed that most pit cells expressed CD2, CD11a, CD18, and CD54. After intravenous injections into rats with anti-CD2, anti-CD11a, and anti-CD18 antibodies, the number of pit cells per square millimeter in frozen sections of liver tissue decreased. Treatment of rats with zymosan increased the number of pit cells fivefold, whereas subsequent treatment with anti-adhesion-molecule antibodies resulted in approximately 60% lower number of pit cells. Anti-CD54, supposed to block CD54 expression on sinusoidal endothelial cells, also decreased the number of pit cells. The number of blood LGL was, however, not affected by these antibodies. These results indicate that blocking of CD2, CD11a, CD18, and CD54 antigens on blood LGL and/or liver endothelium decreased the number of pit cells in the liver. These adhesion molecules therefore play an important role in the recruitment of pit cells in the liver.

Transforming growth factor β (TGF-β), a potent fibrogenic cytokine, is secreted in latent form. We examined which cell type in both normal and carbon tetrachloride (CC1 4 )-induced fibrotic rat liver bears surface type II IGF/mannose 6-phosphate (IGF-II/M6P) receptor, known to facilitate activation of TGF-β. In addition, the role of the IGF-II/M6P receptor in activation of latent TGF-β was investigated in a coculture system with sinusoidal endothelial cells. Northern hybridization analysis for IGF-II/M6P receptor messenger RNA (mRNA) was performed on total RNA of different isolated and purified liver cell types. In normal liver, cells expressed little IGF-II/M6P receptor mRNA. In fibrotic liver, we found significant expression only in fat-storing cells. The presence of IGF-II/M6P receptors was established by [ 125 I]IGF-II binding assays on freshly isolated fat-storing cells from normal and CC1 4 -exposed rat livers. We found specific binding of [ 125 I]IGF-II only on CC1 4 exposed fat-storing cells. As determined by polyacrylamide gel electrophoresis after affinity labeling, the specific binding involved 220 kD type II IGF receptors. Scatchard analysis revealed the presence of 3,043 ± 1,378 IGF-II/M6P high-affinity receptors/fat-storing cell, with a Kd of 387 ± 165 pmol/L. With a mink lung epithelial cell (Mv 1 Lu) proliferation inhibition assay, inhibition of proliferation (a measure of active TGF-β function) was determined using conditioned media of activated fat-storing cells, cocultures of fat-storing cells, and endothelial cells and pure endothelial cell cultures. We found that conditioned media of cocultures of fat-storing cells and endothelial cells inhibits the growth of Mv 1 Lu cells more strongly than conditioned media of homotypic cultures. Addition of neutralizing anti-TGF-β antibodies neutralizes this effect. The contribution of IGF-II/M6P receptors on the cell membrane of activated fat-storing cells to the activation of latent TGF-β was demonstrated by incubating the cells with M6P, or antibodies directed to the IGFII/M6P receptor, both of which diminishes activation of TGF-β. We conclude that IGF-II/M6P receptors are present on activated fat-storing cells and that the presence of this receptor at the cell surface is necessary, but not sufficient, for activation of latent TGF-β. Other factors, derived from endothelial cells, are involved.

The aim of the present study was to identify the cell types that express collagen alpha 1(I), alpha 1(III) and alpha 1(IV), fibronectin and laminin B1 genes in normal rat liver. Parenchymal, sinusoidal endothelial, Kupffer and fat-storing (Ito) cells were isolated and purified. Total RNA of the freshly isolated cells was subjected to Northern hybridization analysis. We also compared the steady state levels of specific mRNAs in freshly isolated fat-storing cells to the levels in myofibroblast-like cells obtained from purified fat-storing cells cultured for two passages. The average purity of each cell preparation, and the percentage of contaminating cells, were determined by transmission electron microscopy and by examining the presence of vitamin A-autofluorescent cells. Fibronectin and collagen alpha 1(III) mRNAs were detected in total RNA of purified parenchymal cells. In poly(A)+ enriched RNA, small amounts of collagen alpha 1(I) mRNA were also present. In total RNA of freshly isolated fat-storing cells, collagen alpha 1(III), alpha 1(IV), and laminin B1 transcripts were found, whereas collagen alpha 1(I) and fibronectin mRNAs were not detected. Cultured fat-storing cells, however, did contain high levels of collagen alpha 1(I) and fibronectin mRNAs. The molecular size of the latter transcript was larger than the fibronectin transcript found in parenchymal cells and the whole liver. Endothelial cells contained small amounts of alpha 1(IV) mRNA. Kupffer cells did not contain the investigated transcripts. We conclude that normal parenchymal, fat-storing and endothelial cells each express a typical pattern of connective tissue molecules. When fat-storing cells are allowed to differentiate into myofibroblast-like cells, they express high levels of collagen alpha 1(I) and fibronectin mRNAs, in addition to collagen alpha 1(III) and alpha 1(IV), and laminin B1 chain mRNAs.

The motility and deformability of lymphokine-activated killer cells, purified by their adherence to plastic (A-LAK cells), was investigated in vivo and in vitro. In vitro, A-LAK cells were observed as intrinsically motile cells. They continuously changed their shape while forming protopods or pseudopods and crawled over the culture surface. A-LAK cells were able to migrate across micropores of 3 μm diameter, which was three times smaller than the average diameter of an A-LAK cell. Even in the absence of serum factors and of interleukin-2 (IL-2), more than half of the inoculated cells migrated across such micropore membranes within 18 h. Electron microscopic examination of these micropore membranes showed that the A-LAK cells were highly deformable. A-LAK cells also migrated across a confluent monolayer of endothelium-like 10T1/2 cells. After 24 h incubation on the monolayers, about 20% of the A-LAK cells were found underneath the monolayer. There, they actively moved in the narrow space between the monolayer and the bottom of the culture dish. In vivo IL-2-activated cells showing the large granular lymphocyte morphology accumulated in the hepatic microvasculature. Electron microscopic observations indicate that these cells did not accumulate by mechanical entrapment due to rigidity and size but rather by adherence to the endothelial lining of the sinusoidal capillaries. This also appeared to be the case for adoptively transferred A-LAK cells, which were seen to be attached to the capillary wall. These observations stress the importance of adhesive entrapment to explain the accumulation of intravenously transferred LAK cells in the vascular beds.

There is now sufficient evidence to consider the pit cells as a hepatic, intrasinusoidal population of large granular lymphocytes. They should be considered as resident sinusoidal cells of the liver, even if they depend on a continuous influx of precursor cells from the bone marrow, since they differ in several phenotypic and functional respects from their circulating (precursor) counterparts. There is growing evidence that a comparable population of activated or further differentiated, liver-specific pit cells (or LGL) is not only present in the rat but also in the human liver. Pit cells exert high spontaneous cytotoxic activity against tumor cell lines and may act as a primary defense barrier to metastasizing tumor cells and to virus infections. The total number of pit cells in the liver can be greatly increased by the administration of interleukin-2 or by a variety of biological response modifiers, which may represent a mechanism of defense against disease. Their involvement in liver pathologies deserves more attention.

In an attempt to elucidate the role of fat-storing cells (FSCs) in liver fibrosis, we investigated the collagen synthesis by FSCs freshly isolated from rats treated with CCl4, with vitamin A, and from untreated rats. FSCs from CCl4-treated rats contained a small number of lipid droplets and an abundant rough endoplasmic reticulum (RER), while those from vitamin A-treated rats showed numerous large lipid droplets and scanty RER. The population doubling times of FSCs isolated from normal, CCl4-treated, and vitamin A-treated rats were 38 +/- 4.3, 24 +/- 2.5, and 48 +/- 6.3 hr, respectively. The rate of collagen synthesis by FSCs from CCl4-treated rats was four- to sixfold enhanced, while collagen synthesis by FSCs from vitamin A-treated rats was suppressed. The ratio of collagen type I to type III produced by FSCs from CCl4 rats was enhanced as compared with control rats (94.7:5.3 vs 87.6:12.4). Therefore, FSCs can be considered to play an important role in the pathogenesis of liver fibrosis.

Fat-storing cells were isolated and purified from livers of normal adult rats and maintained in primary culture. By light and electron microscopy it was established that they underwent phenotypic changes into cells with the ultrastructural characteristics of myofibroblasts, between the third and sixth day in culture. These morphological changes were accompanied by a 2-fold increase of l-[ 3 H]proline incorporation into secretory proteins and an 11-fold increase into secreted collagenase-sensitive proteins. In contrast, incorporation into ceil layer-associated proteins and into cell layer-associated collagenase-sensitive proteins was not significantly elevated. Sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in combination with fluorography, demonstrated that the main collagen type secreted by the myofibroblast-like cells was collagen type I. Collagen types III and IV, and fibronectin were present in lesser amounts. The similarity between the well known in vivo alterations of fat-storing cells under pathological conditions and the spontaneous in vitro differentiation described in this study, makes primary cultures of fat-storing cells a valuable tool for studying their role in chronic liver disease.

The distribution of adrenergic nerves in guinea pig and rat liver was studied by the immunolocalization of fibers containing tyrosine hydroxylase and dopamine beta-hydroxylase, enzymes involved in the synthesis of catecholamines. In both species, adrenergic fibers were identified within portal tracts, often in close proximity to hepatic artery branches. In guinea pig liver, but not rat liver, abundant intraacinar fibers were identified; fibers were also seen within the walls of terminal hepatic vein radicles and larger hepatic veins. The presence of peptidergic nerves containing the regulatory peptide neuropeptide tyrosine and the C-flanking peptide CPON was investigated by indirect immunofluorescence. The distribution of these nerves was similar to that of tyrosine hydroxylase- and dopamine beta-hydroxylase-positive nerves and showed the same species difference. The effector sympathetic nature of tyrosine hydroxylase- and neuropeptide tyrosine-positive fibers in rat liver was confirmed by chemical denervation studies using 6-hydroxydopamine.

The in vitro cytotoxic activity of two types of hepatic sinusoidal cells, i.e., natural killer (NK) cells and macrophages (Kupffer cells), was tested against a syngeneic rat colon adenocarcinoma cell line (DHD-K12). Purified hepatic NK cells (85% cells with large granular lymphocyte morphology) were spontaneously cytolytic, whereas Kupffer cells (90% pure) were not able to kill the DHD-K12 cells. This carcinoma cell line was found to be resistant to the action of mouse recombinant tumor necrosis factor which is considered as the major cytolytic molecule secreted by macrophages. However, colon carcinoma cells were readily lysed by soluble factors present in the culture supernatant of NK cells. It is postulated that hepatic NK cells, which are strategically located within the lumen of the sinusoids, may form a first line of defense to metastasizing colon carcinoma cells.

The interaction between fat-storing cells (FSCs) and Kupffer cells (KCs) in vitro has been studied in an attempt to clarify certain aspects of the pathogenesis of fibrotic process in the liver. FSCs and KCs were isolated from the livers of rats either treated with CCl4 for 6 weeks, or with vitamin A for 6 weeks or from untreated rats by the pronase-collagenase digestion method. FSCs were further purified by centrifugation over a double layered metrizamide gradient, and KCs were separated from other sinusoidal cells by the dish adherence technique. FSCs from CCl4-treated rats divided rapidly, while those from vitamin A-treated rats divided slowly, as compared with untreated rats. Furthermore, the proliferation of FSCs was enhanced in the presence of KCs from CCl4-treated rats, but was slightly suppressed by KCs from normal and vitamin A-treated rats. This enhancement was mediated by a non-dialyzable, soluble factor present in the conditioned medium of KCs from CCl4-treated rats, but was not detected in the conditioned medium of KCs from normal or vitamin A-treated rats. From the present study, a growth factor secreted by KCs from CCl4-treated rats may play an important role in controlling the proliferation of FSCs during the pathogenesis of liver fibrosis.

In normal rat liver, Kupffer cells were unequivocally identified using peroxidase cytochemistry by light microscopy in semithin plastic sections. The Kupffer cell population was found to constitute 31% of the sinusoidal cells and by morphometry and serial sectioning, a mean absolute number of 14 to 20 X 10(6) Kupffer cells per g liver was calculated. The mean distribution of Kupffer cells in the liver lobules was 43% in the periportal, 28% in the midzonal and 29% in the central area of the lobule. Administration of latex particles labeled only 64% of all Kupffer cells, and in particular centrally located cells, showed a lower activity of latex uptake, even at overloading doses. Furthermore, the latter cells were of smaller size than periportal Kupffer cell profiles. The mean number and distribution of latex-labeled Kupffer cells did not change over a period of 3 months, indicating a long lifetime for these resident macrophages. This slow population turnover was supported by the observed small mitotic index, 0.06% after a 6 hr arrest by vinblastine, and by the small [3H]thymidine labeling index which did not change over a period of 3 weeks after administration of the label. It is proposed that the Kupffer cell population, under physiologic conditions, is a long-living and self-renewing population, the kinetics of which substantially differ from those of other sinusoidal cell types.

The numbers of Kupffer cells (macrophages) and pit cells (large granular lymphocytes) were counted by light and electron microscopy in perfusion-fixed liver sinusoids. After a single intravenous injection of the biological response modifiers zymosan, Propionibacterium acnes and OK-432, a 4- to 6-fold increase in the number of pit cells and a 2- to 4-fold increase in the number of Kupffer cells were observed within a period of 4 to 7 days. The mechanisms of the pit cell accumulation were further studied by the metaphase arrest method and by selective irradiation of the liver, or of the rest of the body, with a dose of 8.5 Gy. Stimulated pit cells showed mitotic activity in the liver and to a lesser extent in peripheral blood. The zymosan-induced increase in pit cell number was inhibited by irradiation of the liver but not by irradiation of the rest of the body (with shielded liver). It is concluded that the hepatic pit cell population, which has been shown to have natural tumoricidal activity, can be induced to expand by biological response modifiers, and local division of preexisting cells contributes significantly to this expansion.

The rat liver contains a population of natural killer cells consisting of two morphologically and functionally different subsets, a low-density and a high-density fraction. In this work we describe the influence of low-dose radiation on hepatic natural killer activity. The effect on the cytotoxicity against YAC-1 lymphoma and CC531 colon adenocarcinoma tumor cells was measured in chromium-51 assays, and morphological changes were analyzed by means of electron microscopy. The low-density natural killer fraction showed increased cytotoxicity against YAC-1 which was associated with an increased binding of natural killer cells to the YAC-1 tumor cells shortly after irradiation. These phenomena were paralleled by an increased number of multivesicular bodies and cytoplasmic granules with an electron-lucent halo. In contrast, the other hepatic natural killer cell fraction, the high-density natural killer cells, did not show increased cytotoxicity, binding or morphological alterations. The radiation-stimulated lysis of YAC-1 cells was also observed when in vivo irradiated cells were isolated and tested immediately for in vitro lysis of YAC-1 cells. Sixteen hours after in vitro or in vivo irradiation, the cytotoxicity of hepatic natural killer cells against YAC-1 was no longer enhanced. The cytolysis of and binding of hepatic natural killer cells to CC531 colon adenocarcinoma cells was not stimulated by in vitro irradiation. From these experiments, we conclude that low-dose radiation stimulates the cytotoxicity of hepatic low-density natural killer cells against YAC-1 lymphoma cells immediately after irradiation as the result of enhanced binding of the cells to the tumor cells, in addition to the synthesis of new cytotoxic granules. (HEPATOLOGY 1994;19:1453–1458.)

Liver sinusoids contain a population of large granular lymphocytes or natural killer cells, originally termed pit cells. After isolation and purification, these cells were separated into a low-density and a high-density fraction. The liver low-density fraction differs significantly in morphology and function from cells of the blood, whereas the liver high-density fraction shows intermediate properties. In this study we demonstrate that this morphological and functional heterogeneity is based on subsequent steps of differentiation of the large granular lymphocytes within the liver. When cell proliferation was suppressed by sublethal total body irradiation, the life span of the hepatic large granular lymphocytes could be determined: high-density and low-density populations were totally depleted within 1 and 2 wk after irradiation, respectively. By using intravenous asialo-GM1 anti-serum to deplete animals of asialo-GM1-positive cells, we found that the depletion of the asialo-GM1-positive cells preceded the depletion of asialo-GM1-negative hepatic low-density large granular lymphocytes by approximately 1 wk. Direct evidence that the asialo-GM1-positive high-density large granular lymphocytes are precursors of the low-density large granular lymphocytes was given by adoptive transfer experiments with fluorescent-labeled high-density cells. Three days after their injection, labeled large granular lymphocytes were found in the hepatic low-density fraction of the recipient rat, and these cells had developed morphological characteristics of low-density large granular lymphocytes. It is concluded therefore that marginating blood large granular lymphocytes differentiate through high-density large granular lymphocytes into the typical liver specific low-density large granular lymphocytes or pit cells.

Pit cells are liver-specific natural killer (NK) cells that can be divided into high- (HD) and low-density (LD) subpopulations. The characteristics of pit cells were further investigated in this report. LD and HD pit cells express the specific NK-activation markers gp42, CD25, and ANK44 antigen. LD cells and IL-2-activated NK cells have a high mRNA expression of perforin, granzymes, interferon-gamma, and tumor necrosis factor-alpha. LD pit cells, unlike spleen NK cells, have a weak response to IL-2 with regard to proliferation, cytotoxicity, and production of NK-related molecules. The characteristics of HD cells are intermediate between LD and spleen NK cells. These results show that pit cells, especially LD cells, possess characteristics similar to IL-2-activated NK cells. This is the first evidence on a molecular level that pit cells could be considered in vivo activated NK cells.