Increased expression of glomerular von Willebrand factor after irradiation of the mouse kidney.

Ionizing irradiation has been shown to induce an increased release of von Willebrand factor (vWF) in human endothelial cells in vitro. The present study was undertaken to investigate whether an increase in expression of vWF also occurs in glomerular endothelial cells in vivo after irradiation of the kidney. Increased expression of vWF may initiate prothrombotic changes, and the resultant vascular damage could cause renal failure. The amount of adherent leukocytes in the renal cortex after irradiation was also quantified, since this may contribute to the histological changes that occur after irradiation. Changes in expression of glomerular vWF and in the amount of leukocytes were related to the development of impairment of renal function, as assessed with the [51Cr]EDTA retention assay. Mice were given bilateral irradiation (single dose of 16 Gy) or were sham-irradiated and were sacrificed at intervals of 1 day to 40 weeks after irradiation. Immunohistochemical analysis of kidney cryosections was performed using a polyclonal vWF antibody or monoclonal CD45 antibody (leukocyte common antigen). The amount of glomerular vWF staining and CD45 staining in the renal cortex (percentage surface coverage) was quantified using a computerized image analyzer. The mean glomerular vWF staining in the nonirradiated kidneys was 34.4 +/- 6.2% (mean +/- SEM, 10 weeks after sham treatment). After irradiation, the expression of glomerular vWF increased gradually from 10 weeks to 53.4 +/- 3.6% at 40 weeks. The total number of leukocytes in the renal cortex of nonirradiated mice at 10 weeks after sham treatment was low, with a mean area of 1.0 +/- 0.09%, whereas in the irradiated kidneys the relative tissue area covered by leukocytes increased to 7.6 +/- 2.1% at 40 weeks. These alterations preceded impairment of renal function. The extent to which these changes are causally related to impairment of function will be the subject of future study using specific antithrombotic and anti-inflammatory agents.

Inhibition of lymphoma invasion and liver metastasis formation by pertussis toxin.

We have examined whether pertussis toxin, an agent known to inhibit entry of normal lymphocytes into tissues, affects invasion and metastasis formation by malignant lymphoma and T-cell hybridoma cells. The toxin reduced invasion in vitro in hepatocyte cultures to 20% of control values. Inhibition was maximal after pretreatment for 2 h with approximately 100 ng/ml. The effect of pretreatment with 1 to 5 micrograms toxin/ml for 4 h persisted for at least 5 days, despite a more than 100-fold increase in cell number. The proliferation rate was not affected. Liver metastasis formation after tail vein injection of TAM2D2 T-cell hybridoma cells in syngeneic AKR mice, measured as liver weight, was reduced to 10 to 25% of controls after pretreatment of the cells for 4 h with 1 microgram pertussis toxin/ml. Metastasis to kidneys, ovaries, and lymph nodes was not, or less evidently, affected. With MB6A lymphosarcoma cells no effect was seen after treatment with 1 microgram/ml, but a significant reduction of the liver tumor burden to approximately 50% of controls was achieved by treatment with at least 5 micrograms toxin/ml. Spleen metastasis by MB6A cells was not affected. These results provide evidence for a similarity in invasion mechanisms of normal and malignant lymphoid cells, and they suggest that invasiveness is an important factor in the formation of lymphoma metastases, particularly in the liver.

Adhesion of tumor cells to hepatocytes: different mechanisms for mammary carcinoma compared with lymphosarcoma cells.

Mechanisms of adhesion between tumor cells and hepatocytes, which are likely to play a role in liver metastasis formation, were studied in vitro. TA3 mammary carcinoma and MB6A lymphosarcoma cells were added to rat hepatocytes that had been cultured for 24 hours. Adhesion was quantified by counting adherent cells seen in sections of pelleted, Epon-embedded culture fragments. Adhesion of TA3, but not of MB6A cells, was inhibited by antibodies prepared from an antiserum raised against sinusoidal face-enriched liver plasma membranes. Detergent-solubilized liver components, affinity purified on immobilized inhibitory antibodies, neutralized inhibition, whereas a subfraction separated from this material with the use of immobilized noninhibitory antiliver antibodies had no neutralizing activity. Adhesion of MB6A but not of TA3 cells was inhibited by the calcium ionophore A23187 and the local anesthetic procaine. The calmodulin inhibitor trifluoperazine inhibited adhesion of MB6A cells more strongly than that of TA3 cells. Finally, adhesion of TA3 cells was dependent on external calcium, whereas in the case of MB6A cells calcium could be replaced by magnesium. These observations suggested that adhesion of the two tumor cell types to hepatocytes involved distinct hepatocyte surface molecules and required distinct biochemical machinery.

Interactions between lymphoid tumor cells and isolated liver endothelial cells.

Interactions were studied between highly metastatic murine MB6A lymphosarcoma cells and rat liver endothelial cells that had been isolated by collagenase perfusion and purified by unit gravity sedimentation. Experiments were performed on the day of isolation. MB6A cells were observed to adhere to the endothelial cells. Addition of rat serum had a striking effect: The endothelial cells spread over the MB6A cell surface, engulfing the tumor cells. The factor involved was nondialyzable and also was present in rat plasma. Similar interactions were seen with highly metastatic ESb and MDAY-D2 lymphoma cells and with nonmetastatic Eb cells. Low-metastatic GRSL 34 leukemia and TA3/Ha ascites mammary carcinoma cells did not adhere to the endothelial cells. With this in vitro model the molecular mechanisms of adhesion to liver endothelium were studied. As a first step, univalent antibodies against MB6A cells were found to inhibit adhesion, indicating involvement of specific cell surface molecules.

Antigen-activated T lymphocytes infiltrate hepatocyte cultures in a manner comparable to liver-colonizing lymphosarcoma cells.

Previously we have described the infiltration of lymphosarcoma cells into hepatocyte cultures. The interaction between tumor cells and hepatocytes was comparable to that occurring during the formation of liver metastases. Presently we report that antigen-activated T lymphocytes, but not unstimulated T cells, infiltrate hepatocyte cultures in a manner comparable to the lymphosarcoma cells. Thus, liver-colonizing lymphoid tumor cells and activated non-transformed T cells apparently have common characteristics, that enable them to infiltrate between liver cells. A comparative study of activated and non-activated T cells may aid in elucidating these characteristics. In the intact liver activated lymphocytes did not infiltrate, probably because they were not arrested.

Infiltration of lymphosarcoma cells into hepatocyte cultures: inhibition by univalent antibodies against liver plasma membranes and lymphosarcoma cells.

The number of murine MB6A lymphosarcoma cells that infiltrated rat hepatocyte cultures was found to be diminished after treatment of the lymphosarcoma cells with univalent antibodies raised against these tumour cells (anti-MB6A Fab), and also after treatment of the hepatocyte cultures with univalent antibodies directed against rat liver plasma membranes (anti-LPM Fab). The inhibition of infiltration by anti-MB6A Fab and an anti-LPM Fab raised against sinusoidal face-enriched membranes could be entirely attributed to their interference with adhesion of MB6A cells to the exposed surface of the hepatocytes, because infiltration of the adherent cells was not inhibited. Anti-LPM Fab raised against contiguous face-containing LPM, on the other hand, inhibited the adhesion to the exposed surface and the subsequent infiltration of adherent cells. These observations suggest that specific membrane constituents of both MB6A cells and hepatocytes take part in liver infiltration, and that there may be two different hepatocyte components involved, one mediating adhesion to the exposed surface and the other taking part in the infiltration process proper.

Effect of tubulin-binding agents on the infiltration of tumour cells into primary hepatocyte cultures.

The influence of tubulin-binding agents on the infiltration of murine MB6A lymphosarcoma and TA3 mammary carcinoma cells into primary rat hepatocyte cultures was studied. Colchicine, nocodazole and vinblastine reduced the number of infiltrating lymphosarcoma cells, probably by interfering with the adhesion of these cells to the exposed hepatocyte surface. However, the subsequent infiltration of cells that did adhere was not affected or even slightly stimulated. The reduced adhesion appears to be due to an effect on both the MB6A cells and the hepatocytes. In contrasts, adhesion of TA3 cells was not reduced and infiltration was markedly enhanced by these agents, due to an effect on the TA3 cells but probably not on the hepatocytes. These observations support previously described morphological evidence for differences between the infiltration mechanisms of the two tumour cell types. It is concluded that the system within the hepatocytes involved in adhesiveness of the exposed surface to MB6A cells is distinct from that mediating other types of adhesion. The tendency of TA3 cells to invaginate hepatocytes may be due to disturbances in tubulin-dependent processes.

Infiltration of tumour cells into cultures of isolated hepatocytes.

MB 6A lymphosarcoma and TA3 mammary carcinoma cells have previously been shown to infiltrate from liver blood vessels, where they had been arrested, into the liver parenchyma. The same tumour cells were previously added to cultures of isolated hepatocytes, 24 h after their isolation. Both tumour cell types adhered to both dorsal and lateral surfaces of hepatocytes. The lymphosarcoma cells rapidly infiltrated between the hepatocytes. They first extended pointed pseudopods between the liver cells, and when the tumour cell body intruded, they deeply invaginated the liver cells at an interhepatocyte boundary. The MB A6 cells accumulated between and under the hepatocytes, and after 24 h virtually all cells were contained within the cultures. TA3 cells also invaginated hepatocytes, not only at interhepatocyte boundaries, but all over the exposed surface. They did not extent pseudopods. The process was much slower than with MB 6A cells: After 24 h a few TA3 cells were completely encircled by hepatocytes. These observations indicate that the mechanism of infiltration is different from the 2 tumour cell types. Part of the TA3 cell did not invaginate the hepatocytes. Several of these cells spread on th hepatocyte surface, attaining a flattened shape. TA3 cells formed extensive tight junctions with the hepatocytes, sometimes sealing an intercellular lumen that resembled both a tumour acinus and a bile canaliculus. Also desmosomes were occasionally formed. The hepatocyte cultures appear to be a suitable model for studying the mechanism of liver infiltration, not only of tumour cells, but also leucocytes.

Mammary-carcinoma cells in mouse liver: infiltration of liver tissue and interaction with Kupffer cells.

Interactions between TA3 mammary-carcinoma cells and liver cells were studied with the electron microscope in mouse livers that had been perfused with a defined medium containing the tumour cells. Infiltration of liver tissue by the TA3 cells proceeded in the following steps. First, numerous small protrusions were extended through endothelial cells and into hepatocytes. Next, some cells had larger processes deeply indenting hepatocytes. Finally a few tumour cells became located outside the blood vessels. Two variant cell lines, TA3/Ha and TA3/St, differing in cell coat and surface charge, did not differ in the extent of infiltration. TA3/Ha cells were often encircled by thin processes of liver macrophages (Kupffer cells). Encircled cells were initially intact, but later some of them degenerated. These observations suggest that TA3/Ha cells were phagocytized by the Kupffer cells. Encirclement appeared to be inhibited after only 30 min, when many cells were still partly surrounded. Encirclement of TA3/St was much less frequent. After injection of tumour cells intra-portally in vivo, similar results were obtained, which demonstrated the validity of the perfused liver model. TA3/Ha cells formed much fewer tumour nodules in the liver than TA3/St cells.

Invasion of lymphosarcoma cells into the perfused mouse liver.

Liver of (C57BL X DBA)F1 mice were perfused in situ with a synthetic hemoglobin-free medium, to which murine lymphosarcoma cells were added. All cells were arrested. They were only found in the sinusoids, predominantly in periportal areas. Many lymphosarcoma cells penetrated the walls of the sinusoids with protrusions that extended into and through the endothelial cells. The protrusions often also invaded hepatocytes, and some cells migrated out of the sinusoids. The percentage of the cells that penetrated endothelium was constant and reproducible (68 +/- 4%) in experiments lasting longer than 90 minutes, but the percentage of these cells that also invaded hepatocytes varied greatly. Parellel experiments in vivo yielded similar results, except that the number of cells that invaded hepatocytes was generally much lower. The advantages of the perfused liver as a model for experimental study of the invasion mechanism were evaluated.