Clinical infection control in gene therapy: a multidisciplinary conference.

Gene therapy is being studied for the treatment of a variety of acquired and inherited disorders. Retroviruses, adenoviruses, poxviruses, adeno-associated viruses, herpesviruses, and others are being engineered to transfer genes into humans. Treatment protocols using recombinant viruses are being introduced into clinical settings. Infection control professionals will be involved in reviewing the safety of these agents in their clinics and hospitals. To date, only a limited number of articles have been written on infection control in gene therapy, and no widely available recommendations exist from federal or private organizations to guide infection control professionals. The goals of the conference were to provide a forum where gene therapy experts could share their perspectives and experience with infection control in gene therapy and to provide an opportunity for newcomers to the field to learn about issues specific to infection control in gene therapy. Recommendations for infection control in gene therapy were proposed.

Preclinical development strategies for novel gene therapeutic products.

With over 220 investigational new drug applications currently active, gene therapy represents one of the fastest growing areas in biotherapeutic research. Initially conceived for replacing defective genes in diseases such as cystic fibrosis or inborn errors of metabolism with genes encoding the normal, or wild-type, gene product, gene therapy has expanded into other novel applications such as treatment of cancer or cardiovascular disease, where the risk:benefit profiles may be more acceptable in relation to the severity of the disease. Different types of vectors, including modified retroviruses, adenoviruses, adenovirus-associated viruses, and herpesviruses and plasmid DNA, are used to transfer foreign genetic material into patients' cells or tissues. Developing a toxicology program to determine the safety of these agents, therefore, requires a modified approach that encompasses the pharmacology and toxicity of both the gene product itself and the vector system used for delivery in the context of the application for the clinical trial. In general, the issues involved in designing and developing appropriate preclinical testing to determine the safety of these products are similar to those encountered for other recombinant molecules, including protein biotherapeutics. Limitations to some of the typical toxicology studies conducted for a traditional drug development program may exist for these agents, and nontraditional approaches may be required to demonstrate their safety. Many factors may affect the safety and clinical activity of these agents, including the route, frequency, and duration of exposure and the type of vector employed. Other safety considerations include quantitation of the duration and degree of expression of the vector in target and other tissues, the effects of gene expression on organ pathology and/or histology, evaluation of trafficking of gene-transduced cells or vector after injection, and interactions of the host immune system with the transduced cell population. Because of the unique concerns regarding each of these therapies, the Center for Biologics Evaluation and Research encourages sponsors to obtain toxicity data whenever possible while evaluating the pharmacologic activity of the vector in a species or animal model relevant to their clinical indication. Sponsors are encouraged to discuss preclinical study design and results with the Center during product development to facilitate early identification of safety concerns prior to entry of these novel agents into the clinical setting and to ensure an uninterrupted course of development while addressing issues required for licensure.

Safety assessment of biotechnology-derived pharmaceuticals: ICH and beyond.

Many scientific discussions, especially in the past 8 yr, have focused on definition of criteria for the optimal assessment of the preclinical toxicity of pharmaceuticals. With the current overlap of responsibility among centers within the Food and Drug Administration (FDA), uniformity of testing standards, when appropriate, would be desirable. These discussions have extended beyond the boundaries of the FDA and have culminated in the acceptance of formalized, internationally recognized guidances. The work of the International Committee on Harmonisation (ICH) and the initiatives developed by the FDA are important because they (a) represent a consensus scientific opinion, (b) promote consistency, (c) improve the quality of the studies performed, (d) assist the public sector in determining what may be generally acceptable to prepare product development plans, and (e) provide guidance for the sponsors in the design of preclinical toxicity studies. Disadvantages associated with such initiatives include (a) the establishment of a historical database that is difficult to relinquish, (b) the promotion of a check-the-box approach, i.e., a tendancy to perform only the minimum evaluation required by the guidelines, (c) the creation of a disincentive for industry to develop and validate new models, and (d) the creation of state-of-the-art guidances that may not allow for appropriate evaluation of novel therapies. The introduction of biotechnology-derived pharmaceuticals for clinical use has often required the application of unique approaches to assessing their safety in preclinical studies. There is much diversity among these products, which include the gene and cellular therapies, monoclonal antibodies, human-derived recombinant regulatory proteins, blood products, and vaccines. For many of the biological therapies, there will be unique product issues that may require specific modifications to protocol design and may raise additional safety concerns (e.g., immunogenicity). Guidances concerning the design of preclinical studies for such therapies are generally based on the clinical indication. Risk versus benefit decisions are made with an understanding of the nature of the patient population, the severity of disease, and the availability of alternative therapies. Key components of protocol design for preclinical studies addressing the risks of these agents include (a) a safe starting dose in humans, (b) identification of potential target organs, (c) identification of clinical parameters that should be monitored in humans, and (d) identification of at-risk populations. One of the distinct aspects of the safety evaluation of biotechnology-derived pharmaceuticals is the use of relevant and often nontraditional species and the use of animal models of disease in preclinical safety evaluation. Extensive contributions were made by the Center for Biologics Evaluation and Research to the ICH document on the safety of biotherapeutics, which is intended to provide worldwide guidance for a framework approach to the design and review of preclinical programs. Rational, scientifically sound study design and early identification of the potential safety concerns that may be anticipated in the clinical trial can result in preclinical data that facilitate use of these novel therapies for use in humans without duplication of effort or the unnecessary use of animals.

Antitumor effects of human recombinant interleukin-1 alpha and etoposide against human tumor cells: mechanism for synergism in vitro and activity in vivo.

Recombinant human interleukin 1 alpha (rh IL-1 alpha) and etoposide (VP-16) synergize for direct growth inhibition of several human tumor cell lines in vitro. Our previous studies demonstrated that VP-16 increased the number of membrane-associated IL-1 receptors (IL-1Rs) and also enhanced the internalization of receptor-bound rh IL-1 alpha. The purposes of this study were to test our hypotheses that these events were critical to the synergy between rhIL-1 alpha and VP-16, to determine whether rhIL- 1 alpha and VP-16 synergize to increase superoxide (SO) anion radical production in vitro since SO anion has been implicated in the toxic effects of IL-1, and to investigate the antitumor efficacy of the combination against tumors in vivo. A375/C6 melanoma cells and OVCAR-3 ovarian carcinoma cells were tested with IL-1 receptor antagonist (IL-1 ra) before exposure to rhIL-1 alpha, VP-16 and rhIL-1 alpha plus VP-16. The synergistic or antagonistic effects were assessed by MTT assay. SO production was measured by reduction of cytochrome C. Athymic female mice bearing the A375/C6 melanoma were treated by rhIL-1 alpha, VP-16, and rhIL- 1 alpha+VP-16. The antitumor effects were evaluated by quantitating tumor growth and survival time. Pretreatment with the IL-1ra abrogated the synergistic effects of rhIL-1 alpha and VP-16. The production of SO radical by A375/C6 cells was increased 2.5 fold by the combination of rhIL-1 alpha and VP-16, and the addition of exogenous SOD blocked the synergy between rhIL-1 alpha and VP-16. However, when A375/SOD15 cells which over-expressed manganese superoxide dismutase (MnSOD) after MnSOD cDNA transfection were exposed to rhIL-1 alpha and VP-16, in vitro antagonism was observed. In vivo studies demonstrated that the combination of rhIL-1 alpha and VP-16 delayed tumor growth better than either agent alone, although long-term survival was not improved because of substantial toxicity. Our results suggest that the synergistic antitumor effects of IL-1 alpha and VP-16 may be due to IL-1R modulation and increased internalization of IL-1-IL-1R complex by VP-16 treatment, as well as to a subsequent increase in SO anion radical production from the tumor cells exposed to both drugs. Thus, the combination of IL-1 alpha and VP-16 might prove useful for the treatment of malignant disease in vivo, if the increased toxicity can be reduced or managed.

TNF-alpha is a principal cytokine involved in the recruitment of NK cells to liver parenchyma.

Isolated murine splenic NK cells and the cultured murine endothelioma cell line, eEND2, were used to study the effects of cytokines on NK cell/endothelial cell adhesion. Treatment of eEND2 cells with TNF-alpha induced a marked increase (four- to sevenfold) in adherence of NK cells, as compared with control cultures of endothelioma cells or eEND2 cells treated with IL-1 alpha or IL-6. TNF-alpha induction of NK cell adherence to eEND2 was dose dependent with rapid kinetics, reaching a maximum at concentrations between 10 and 1000 U/ml after a 2-h incubation. TNF-alpha treatment of L929 fibroblasts or CL-2 hepatoma cells did not result in increased NK cell adhesion. The concentration range of TNF-alpha that was found to maximally augment NK cell adhesion to eEND2 also induced NK cell chemokinetic activity. The relevance of these in vitro results was subsequently analyzed in vivo. Initial studies confirmed that a single dose of polyinosinic-polycytidylic acid and poly-L-lysine stabilized in carboxymethyl cellulose (poly-ICLC), augmented hepatic NK activity and resulted in a 2.2-fold increase in the number of liver-associated NK cells. Concomitant treatment of mice with a TNF-alpha neutralizing antisera eliminated both the hepatic influx of NK cells and the increase in poly-ICLC-induced liver NK activity. These results suggest that TNF-alpha is a principal cytokine involved in the in vivo recruitment and localization of parenchymal NK cells after treatment with a biological response modifier, and that this regulation seems to occur via alterations in NK cell/endothelial cell interactions.

A method for obtaining and culturing large numbers of purified organ-derived murine endothelial cells.

Fibroblast growth factor 1 (FGF-1)-coated collagen-gelatin sponges were affixed to various tissues to generate vascular beds, in which the vessels originated in the tissue to which the sponges were affixed. Organ-derived endothelium was obtained from vascularized sponges implanted in or on the skin, peritoneal wall, abdominal mesentery, epimysium, spleen, and liver. Collagenase digestion yielded single-cell suspensions that were analyzed by flow cytometry. Approximately 25% of the cells were positive for the endothelial cell (EC) markers MECA-32 and Sca-1 and for uptake of diIAcLDL. Similar results were obtained when sponges were implanted in several different mouse strains, although there was some evidence of heterogeneity in the degree of vascularization and EC recovery. Long-term cultures of high purity were obtained when the ECs were grown on mitomycin C-treated L929 feeder layers, in medium supplemented with cis-hydroxyproline and FGF-1. These cells have been utilized in preliminary studies of T cell-EC binding. Thus we have developed a generalized method for the recovery and culture of organ-derived murine endothelial cells. This technique should greatly improve the feasibility of studies of the interactions between murine endothelial and immune effector cells.

Quantitative analysis of rod-cored vesicles and dense granules of large granular lymphocytes in the liver, spleen, and peripheral blood of rats.

Large granular lymphocytes (LGL) comprise a natural defense system in the liver and exert an inhibitory effect on tumor cell metastasis. In order to demonstrate the maturation of LGL in the liver from the morphological aspect, we evaluated electron-microscopically the frequency of 0.2 micron vesicles (rod-cored and "empty" vesicles) and dense granules in LGL from the liver, spleen, and peripheral blood of the rat. Both of these cell organelles are characteristic to LGL and may relate to natural killer-mediated cytolysis. On the average, there were 12.7 of the 0.2 micron vesicles and 4.3 rod-cored vesicles (RCV) per cell section in the liver, 6.6 0.2 micron vesicles and 1.6 RCV in the spleen, and 8.6 0.2 micron vesicles and 0.9 RCV in the peripheral blood. The number of 0.2 micron vesicles per cell section ranged from 0 to 19 with the exception of a few higher instances. Therefore, LGL were divided into vesicle-rich (> 9 0.2 micron vesicles per cell section) and vesicle-poor (< 8 per cell section) populations. Hepatic LGL consisted mainly of a vesicle-rich population while splenic LGL consisted mainly of a vesicle-poor population, and peripheral blood contained equal proportions of both populations. In addition to diversity with regard to the number of 0.2 micron vesicles, LGL obtained from various organs also displayed heterogeneity in the number and size of dense granules.(ABSTRACT TRUNCATED AT 250 WORDS)

Purification and cloning of a novel serine protease, RNK-Tryp-2, from the granules of a rat NK cell leukemia.

We have biochemically purified a 27-kDa serine protease (designated RNK-Tryp-2) from the granules of the rat large granular lymphocyte leukemia cell line (RNK-16) which has tryptase activity. Utilizing molecular sieve chromatography and reverse-phase HPLC, we purified RNK-Tryp-2 to homogeneity and sequenced 33 NH2-terminal amino acids. Oligonucleotide primers were used in the PCR to generate a 528-bp cDNA clone encoding a novel serine protease from RNK-16 mRNA. This cDNA clone was used to isolate an 884-bp RNK-Tryp-2 cDNA from an RNK-16 lambda-gt11 library. The open reading frame predicts a mature protein of 233 amino acids which does not have potential sites for N-linked glycosylation. The cDNA encodes a leader peptide of at least 25 amino acids. The characteristic Ile-Ile-Gly-Gly amino acids of the N-terminus, and the His, Asp, and Ser amino acids that form the catalytic triad of serine proteases, are conserved. The amino acid sequence has less than 45% identity with any other member of the serine protease family, indicating that RNK-Tryp-2 is distinct protease. Southern blot analysis suggests the existence of one or more related genes. A single 1.3-kb mRNA transcript was detected by Northern blot analysis of total cellular RNA from the in vivo passaged RNK-16, rat splenocytes, lung and liver nonparenchymal cells, as well as in highly purified rat LGL and T cells. RNK-Tryp-2 is a novel serine protease that is expressed in the granules of large granular lymphocytes.

Substrate-specific proteases (BLT-esterase) are localized predominantly in the natural killer cells of unprimed mice.

In leukocytes isolated from unprimed mice, the levels of extractable N alpha-Cbz-Lys-thiobenzylesteresterase (BLT-esterase) closely correlated with the number of natural killer (NK) cells. The spleens of mice that exhibit severe combined immunodeficiency (SCID) contained much higher levels of this enzyme than other mouse strains. Treatments that resulted in a local accumulation of NK cells (as assessed by lytic activity) produced a concomitant increase in BLT-esterase activity. However, short-term in vitro treatment of spleen cells with interferon (IFN)-alpha/beta indicated that BLT-esterase levels correlated more closely with absolute numbers of NK cells than with their lytic capacity. There was a very good correlation between the numbers of cells bearing the NK phenotype (NK-1.1+) and BLT-esterase levels. Cells positively sorted using the NK-specific antibodies NK-1.1 and LGL-1 had high enzymatic activity. The BLT-esterase levels were high in both the NK-1.1+/LGL-1- and NK-1.1+/LGL-1+ subsets. Highly purified CD4+ and CD8+ T cells and sIg+ B cells demonstrated negligible enzyme, as did populations of cells highly enriched for macrophages or neutrophils. However, it should be stressed that the inbred mice used on this study have been maintained in a pathogen-free facility. It would be anticipated that mice maintained under less stringent conditions could exhibit appreciable levels of BLT-esterase activity in their T cells. Nonetheless, BLT-esterase is present at high levels in NK cells and cannot be regarded as a T cell-specific enzyme.

c-Ets-1 protein is hyperphosphorylated during mitosis.

The ets-1 and ets-2 proto-oncogene products can serve as transcription factors and become phosphorylated in response to Ca(2+)-mediated signals. We have examined expression of Ets proteins during the cell cycle in cells synchronized by centrifugal elutriation or nocodazole-induced mitotic block. Both methods revealed the presence of a hyperphosphorylated isoform of Ets-1 during the mitotic phase. This isoform showed a characteristic mobility shift and was observed during mitosis in each of four cell lines (three human T-cell lines and a human astrocytoma) that express ets-1. In elutriated cells, only a small portion of the Ets-1 in cells from the G2/M fractions was hyperphosphorylated, while in nocodazole-arrested cells more of the Ets-1 was shifted. When cells were released from nocodazole arrest, this isoform disappeared within 1-2 h as cells completed mitosis and entered G1. This suggests that hyperphosphorylated Ets-1 is present transiently during early mitosis, before or around the time of the metaphase-anaphase transition. Exposure of unsynchronized cells to okadaic acid resulted in a dramatic hyperphosphorylation of virtually all Ets-1, suggesting that changes in cellular phosphatase activity are important for cell cycle regulation of Ets-1. Hyperphosphorylated Ets-1 appears to arise from multiple phosphorylations on serine in the exon 7-encoded domain of the protein and did not appear to alter sequence-specific DNA-binding activity. Although enhanced phosphorylation of Ets-2 was detected in nocodazole-arrested cells, no Ets-2 hyperphosphorylation was seen.

In vivo distribution and cytokine gene expression by enriched mouse LAK effector cells.

Lymphokine activated killer (LAK) cells administered in combination with interleukin 2 (IL2) can mediate antitumor activity in tumor-bearing mice and advanced cancer patients. Relatively little is known about the mechanism by which adoptively transferred LAK cells plus IL2 mediate these antitumor effects in vivo, and it remains unclear to what extent the actual LAK effector cells can accumulate in tumors. In the present study, enriched cytolytic LAK effector cells were obtained by fractionation of bulk LAK cell cultures on Percoll density gradients. About 95% of the total lytic activity was recovered from the 55% of cells isolated in fraction 2 (Fr2). The cells recovered in Fr2 are mostly large, proliferating lymphoblasts that express either the NK-associated surface markers NK1.1 (38%) or LGL-1 (31%), or the cytotoxic T cell phenotype, Lyt2 (39%). The cytolytic lymphoblasts obtained from Fr2 were radiolabelled with either 111Indium-Oxine (111InOx) which labels all cells in the population, or with 125Iododeoxyuridine (125IUdR) which labels only proliferating cells, and injected iv into mice bearing murine renal cancer (Renca). 111InOx-labeled Fr2 cells migrated mostly to spleen (28%) and liver (35%), with approximately 5% of the injected label detectable in the Renca-bearing kidney by 24 hrs. In contrast, Fr2 cells labeled with 125IUdR, which labels only the proliferating blasts thought to include the actual effector cells, exhibited a very different localization pattern. 125IUdR-Fr2 cells were retained in the lungs at higher levels than were 111InOx-Fr2 cells and very little label was detectable in liver (6%), spleen (3%), or tumor bearing kidney (2%) at 24 hrs. These results suggest that most of the large, proliferating lymphoblasts are cleared from the body by 24 hrs and very few localize into even large tumors. Subsequently, Northern blot analyses performed on bulk LAK cells revealed a potent induction of mRNA for TNF alpha by 6 hrs and for IFN gamma by 48 hrs. The intensity of gene expression for both cytokines was increased in Fr2 as compared to the unfractionated bulk LAK cells or to non-cytolytic cells obtained from Fr3. Overall, these results suggest that at least some of the antitumor effects mediated by LAK cells occur by the release of cytokines that synergize with exogenous IL2 for the activation of host effector cells.

An improved in vitro assay to quantitate chemotaxis of rat peripheral blood large granular lymphocytes (LGL).

We have developed an improved method to study the directed migration, or chemotaxis, of rat peripheral blood large granular lymphocytes (LGL) in vitro. A modified Boyden chamber technique was used to measure chemotaxis of LGL through polycarbonate filters that had been coated with different basement membrane components. LGL were found to adhere to collagen types I and IV, laminin and fibronectin. However, only collagen type IV was not in itself chemotactic for LGL. Migrated cells could be identified both morphologically and phenotypically as LGL on collagen type IV-coated filters after incubation with a chemotactic stimulus. LGL were found to display chemotaxis to a number of different stimuli, including the classical chemoattractant agents N-formyl-methionyl-leucyl-phenylalanine, leukotriene B4, and complement fragments present in activated sera. However, the degree of response to these stimuli was much less than that of isolated peripheral blood neutrophils or monocytes. In contrast, all three cell types showed increased chemotaxis to the diacyl glycerol analog 1-oleoyl 2-acetyl glycerol (OAG), which induced a 4-14 fold stimulation of migration. Induction of chemotaxis of LGL by OAG was time and dose-dependent, as confirmed using checkerboard assays. In summary, we have developed a rapid, quantitative method to measure chemotaxis of LGL in vitro. This technique may now be utilized to identify naturally occurring chemoattractants for LGL and to study the intracellular and regulatory events associated with LGL migration.

Augmentation of mouse liver-associated natural killer activity by biologic response modifiers occurs largely via rapid recruitment of large granular lymphocytes from the bone marrow.

A variety of biologic response modifiers (BRM) can potently augment NK activity in nonlymphoid organs. By using the liver as a model organ, we have shown that this augmentation of organ-associated NK activity is coincident with a 10- to 15-fold increase in the number of large granular lymphocytes (LGL) which can be isolated. The present study was designed to investigate the mechanism by which BRM induce this increase in liver-associated LGL and the coincident increase in hepatic NK activity. Initial studies confirm that a single dose of the pyran copolymer, maleic anhydride divinyl ether (MVE-2), augmented hepatic NK activity and increased the number of liver-associated LGL from 3 x 10(4)/liver to 5 x 10(5)/liver (a 17-fold increase). Multiple injections of MVE-2 further augmented total liver-associated NK activity and LGL number (to 13 x 10(5)/liver). As expected, both the NK activity and detectable LGL were eliminated by treatment of the mice with antiasialo GM1 (asGM1) serum. Three possible mechanisms for the BRM-induced increase in liver-associated LGL have been investigated, including 1) the rapid proliferation of resident hepatic LGL, 2) the redistribution of mature LGL from peripheral sites such as the spleen, or by 3) a rapid output and subsequent hepatic localization of LGL or their precursors recently derived from the bone marrow (BM). Our results demonstrated that the contribution of in situ proliferation to the BRM-induced increase in liver-LGL was relatively small, since the number of cells expressing NK-associated markers (i.e., asGM1, Thy-1.2, and NK1.1) and in G2/M phase (as assessed by propidium iodide uptake) was only 4 to 8%. Further experiments demonstrated that splenectomy before the administration of MVE-2 did not inhibit the augmentation of liver-associated NK activity. This result argued against a recruitment of mature LGL from the spleen. In contrast, selective depletion of the BM following administration of 89Sr decreased the ability of MVE-2 to augment liver-associated NK activity by greater than 80%. This procedure also significantly decreased the ability of Propionibacterium acnes (85%) and multiple doses of IL-2 (49%) to augment liver-associated NK activity. These results demonstrate that the rapid augmentation of liver-associated NK activity by BRM is largely due to localization and accumulation in the liver of LGL recently derived from the BM.

Effect of cytokines on polymorphonuclear neutrophil infiltration in the mouse. Prostaglandin- and leukotriene-independent induction of infiltration by IL-1 and tumor necrosis factor.

The i.p. injection of mice with highly purified recombinant human rIL-1 alpha or beta resulted in the rapid influx of a large number of polymorphonuclear neutrophils (PMN) into the peritoneal cavity. Significant increases in the number of PMN were induced by doses of IL-1 which ranged from 0.005 to 5 ng/injection. Interestingly the dose response for PMN influx was bell-shaped because 50 ng of IL-1 did not result in a significant increase in peritoneal PMN. IL-1 induced PMN infiltration was detectable by 1 h with peak levels of PMN obtained by about 2 h, followed by a subsequent decline by 24 h. Other cytokines, IL-2, IFN-gamma, IFN alpha beta, granulocyte-CSF, granulocyte-macrophage-CSF, IL-3, TNF-alpha, and TNF-beta were compared to IL-1 for their ability to induce a PMN influx into the peritoneum. Only TNF-alpha or TNF-beta (lymphotoxin) were able to induce a significant influx of PMN within 2 h. However, based on total protein administered, about 100 times more TNF than IL-1 was required to produce a comparable PMN infiltration. Intraperitoneal injection of inhibitors of the cyclooxygenase or lipoxygenase pathways did not inhibit the IL-1-induced influx of PMN. Also, neither IL-1 nor TNF triggered an increase in PG or leukotriene release from peritoneal cells in vitro. Furthermore, direct peritoneal injection of leukotriene B4, a potent PMN chemoattractant in vitro, did not induce any significant increase in PMN in the peritoneal cavity indicating that chemotactic activity alone is insufficient for inducing peritoneal infiltration. These results suggest that the local production of very low levels of IL-1 in vivo would be sufficient to initiate a sequence of events that results in a rapid accumulation of PMN. Because IL-1 was not chemotactic for PMN in vitro, our data suggest that IL-1 induces production of factors that are chemotactic for PMN. Alternatively, IL-1 may act on other stages of the complex sequence of events that regulates the emigration of PMN into tissue sites in vivo. The synergy apparent in PMN influx when suboptimal concentrations of IL-1 and TNF were injected suggests that the local production of very low concentrations of these cytokines in situ could play a critical role in the emigration of PMN during infection.

Activation of liver macrophages following phenobarbital treatment of rats.

Phenobarbital is a potent inducer of hepatic cytochrome P-450 and is a tumor promoter in the two-stage model of liver carcinogenesis. In the present studies, we show that phenobarbital also induces an accumulation of activated macrophages in the livers of treated rats. These macrophages are larger and more stellate than resident Kupffer cells and are highly vacuolated. In addition, macrophages isolated from livers of phenobarbital-treated rats display increased phagocytosis of sheep red blood cells, chemotaxis toward the complement fragment C5a and enhanced production of hydrogen peroxide. Biologically active mediators released by activated macrophages have been implicated in tumor promotion as well as in the regulation of cytochrome P-450-mediated drug biotransformation. We propose that the activated macrophages that accumulate in the liver following phenobarbital treatment may contribute, at least in part, to the biological responses to this drug.

Functional and biochemical properties of rat Kupffer cells and peritoneal macrophages.

Functional and biochemical techniques were used to further characterize heterogeneity between rat Kupffer cells and peritoneal macrophages. Both macrophage cell types were found to phagocytize antibody coated sheep red blood cells in a time-dependent manner. However, Kupffer cells were two to three times more phagocytic than were peritoneal macrophages. In contrast, the peritoneal cells released significantly more superoxide anion in response to the complement cleavage product, C5a and the phorbol ester tumor promoter, 12-0-tetradecanoyl-phorbol-13-acetate, and produced more hydrogen peroxide than did the liver macrophages. Both cell types responded chemotactically to C5a. These results suggest that macrophages may develop specialized functions depending on the needs of their local environment. Using one and two dimensional SDS-polyacrylamide gel electrophoresis, we also compared the production of newly synthesized proteins by Kupffer cells and peritoneal macrophages. In general, the macrophages were found to produce similar types and numbers of proteins with some exceptions. These included proteins that were unique to peritoneal macrophages and other proteins observed only in Kupffer cells. The production of these proteins in liver macrophages did not appear to correlate with levels of functional activation, but may be more related to the tissue origin of the cells.

Potential role of activated macrophages in acetaminophen hepatotoxicity. I. Isolation and characterization of activated macrophages from rat liver.

Twenty-four hours following treatment of rats with the analgesic acetaminophen (1.2 g/kg), we observed an infiltration of mononuclear cells into centrilobular regions of the liver in the absence of necrosis. To determine whether acetaminophen induces the accumulation and activation of mononuclear phagocytes, we compared the morphologic and functional characteristics of macrophages obtained from livers of acetaminophen-treated rats with those of resident macrophages (Kupffer cells) from untreated control animals. Macrophages were isolated from rat livers by combined collagenase/pronase perfusion, selective digestion, and differential centrifugation on a metrizamide gradient. Acetaminophen treatment resulted in a twofold increase in macrophage yields from the liver compared with controls. Macrophages isolated from treated animals were generally larger than resident Kupffer cells, were highly vacuolated, and adhered to culture dishes more rapidly. Liver macrophages from both treated and untreated rats phagocytized sheep red blood cells (sRBC) in a time-dependent manner, reaching a maximum after 60-75 min incubation with sRBCs. However, macrophages from livers of acetaminophen-treated rats phagocytized two to three times more sRBC than did resident Kupffer cells. Using the Boyden chamber technique, both macrophage populations were found to be chemotactic to a number of stimuli including the complement fragment, C5a, and synthetic collagenous peptides related to tissue breakdown products. Levels of migration of macrophages from livers of acetaminophen-treated rats were four to seven times greater than those of resident Kupffer cells. In addition, compared with resident Kupffer cells, macrophages from acetaminophen-treated rats released 30% more superoxide anion in response to the stimulus, 12-O-tetradecanoyl-phorbol-13-acetate. Taken together, these results suggest that acetaminophen treatment of rats leads to the recruitment and activation of macrophages in the liver.

Potential role of activated macrophages in acetaminophen hepatotoxicity. II. Mechanism of macrophage accumulation and activation.

Treatment of rats with acetaminophen (1.2 g/kg) results in the accumulation of activated macrophages in the centrilobular regions of the liver. To study the mechanism by which these cells accumulate and become activated, we examined the release of chemotactic and activating factors from cultured hepatocytes treated with acetaminophen (10-100 microM). We found a dose- and time-related generation of Kupffer cell and monocyte chemotactic activity by acetaminophen-treated hepatocytes. The maximum response was observed with a 25% dilution of medium collected 24 hr following treatment of hepatocytes with acetaminophen. Using a checkerboard assay, the factor in conditioned medium was determined to induce chemotaxis as well as chemokinesis in both Kupffer cells and monocytes. The hepatocyte-derived chemotactic factor was also found to be stable to freeze-thawing but to lose activity following heat or trypsin treatment. These results, together with our findings that chemotactic activity was eluted in the void volume following Sephadex G-25 size exclusion chromatography, suggested that the chemotactic factor released by hepatocytes is a large molecular weight protein. The release of Kupffer cell activating factors by acetaminophen-treated hepatocytes was also examined. Hepatocyte-conditioned medium was found to stimulate Kupffer cell phagocytosis and superoxide anion release, two characteristics of activated macrophages. These effects were maximal with conditioned medium collected from hepatocytes 24 hr following treatment with 50-100 microM acetaminophen. Acetaminophen alone had no effect on chemotaxis, phagocytosis, or superoxide anion production by Kupffer cells or monocytes. Taken together, these results suggest that macrophage accumulation and activation in the liver following acetaminophen treatment may be mediated, at least in part, by factors released from hepatocytes.

Accumulation of activated mononuclear phagocytes in the liver following lipopolysaccharide treatment of rats.

Lipopolysaccharide (LPS) is a toxic bacterial cell wall component that is rapidly cleared from the portal circulation by Kupffer cells. To determine if interaction with LPS causes the accumulation and activation of mononuclear phagocytes (MNP) in the liver, we compared the morphological and functional characteristics of MNP obtained from livers of rats treated with LPS (5 mg/kg, intravenously [IV]) with normal resident Kupffer cells. MNP were isolated from rat livers by combined collagenase/pronase perfusion, selective digestion, and differential centrifugation on a metrizamide gradient. MNP obtained from livers of LPS-treated rats were found to display morphologic and functional characteristics of activated macrophages. These cells were generally larger than resident cells, were highly vacuolated, and adhered to culture dishes more rapidly. Both cell types phagocytized sheep red blood cells (sRBC) in a time-dependent manner, reaching a maximum after 60-75 min incubation with sRBC. However, MNP from livers of LPS-treated rats phagocytized 10-15 times more sRBC than resident Kupffer cells from untreated animals. Employing the Boyden chamber technique, both cell types were also found to be chemotactic to a number of stimuli including the complement fragment, C5a, the tumor promoter, 12-0-tetradecanoyl-phorbol-13-acetate (TPA), and collagenous peptides related to tissue breakdown products. MNP from livers of LPS-treated rats were generally 10-15 times more responsive to the chemoattractants than resident Kupffer cells. In addition, both resident Kupffer cells and MNP from LPS-treated rats were found to release superoxide anion in response to stimulation by C5a and TPA. Taken together these results suggest that LPS treatment of rats leads to the recruitment and activation of MNP in the liver.