Factors influencing the efficiency of cationic liposome-mediated intravenous gene delivery.

We have characterized the relationships between the design of cationic liposomes as a gene transfer vehicle, their resulting biodistribution and processing in animals, and the level and sites of gene expression they produce. By redesigning conventional cationic liposomes, incorporating cholesterol (chol) as the neutral lipid and preparing them as multilamellar vesicles (MLV), we increased the efficiency of cationic liposome:DNA complex (CLDC)-mediated gene delivery. Expression of the luciferase gene increased up to 1,740-fold and of the human granulocyte-colony stimulating factor (hG-CSF) gene up to 569-fold due to prolonged circulation time of injected CLDC, and increased uptake and retention in tissues. The level of gene expression per microgram of DNA taken up per tissue was 1,000-fold higher in lung than in liver, indicating that in addition to issues of delivery and retention of injected DNA, tissue-specific host factors also play a central role in determining the efficiency of expression. Vascular endothelial cells, monocytes, and macrophages are the cell types most commonly transfected by intravenous injection of CLDC.

Fetal gene transfer by transuterine injection of cationic liposome-DNA complexes.

In utero injection of cationic liposome-DNA complexes (CLDCs) containing chloramphenicol acetyltransferase, beta-galactosidase (beta-gal), or human granulocyte colony-stimulating factor (hG-CSF) expression plasmids produced high-level gene expression in fetal rats. Tissues adjacent to the injection site exhibited the highest levels of gene expression. Chloramphenicol acetyltransferase expression persisted for at least 14 days and was reexpressed following postnatal reinjection of CLDCs. Intraperitoneal administration of the hG-CSF gene produced high serum hG-CSF levels. X-gal staining demonstrated widespread beta-gal expression in multiple fetal tissues and cell types. No toxic or inflammatory responses were observed, nor was there evidence of fetal-maternal or maternal-fetal gene transfer, suggesting that CLDCs may provide a useful alternative to viral vectors for in utero gene transfer.

Liposome-based therapy of human ovarian cancer: parameters determining potency of negatively charged and antibody-targeted liposomes.

Liposomes containing cytotoxic agents may be highly efficacious for intracavitary therapy of malignancies such as ovarian carcinoma, which resides principally in the peritoneal cavity. We have examined in vitro the cytotoxicity of a variety of liposome-drug formulations against OVCAR-3, a human ovarian cancer cell line. Two drugs tested, methotrexate-gamma-aspartate and 5-fluoroorotate, show increased cytotoxicity on various cultured cell lines following encapsulation in liposomes and can be considered liposome-dependent agents. With the optimal lipid composition used in this study, the maximal increase in potency on OVCAR-3 is 2.6-fold for methotrexate-gamma-aspartate and 5.2-fold for 5-fluoroorotate. Studies on liposome-cell association suggest a low capacity of OVCAR-3 to bind and internalize phospholipid vesicles, which limits the in vitro potency of liposomes for these cells. OC-125, a monoclonal antibody recognizing an antigen common to a number of human ovarian cancers (CA-125), has been coupled covalently to the liposome surface. Liposomes bearing OC-125 and containing methotrexate-gamma-aspartate show an 8-fold increase in potency against OVCAR-3 cells in a 96-h growth inhibition assay. Briefer exposure of tumor cells to treatment accentuates the advantage of targeted liposomes. The cytostatic effect of 1 h exposure to OC-125 liposomes is 100-fold greater than the equivalent exposure to free drug and equal to the maximal cytostatic effect achieved with free drug for 96 h. Attachment of OC-125 antibody also confers upon liposomes the capacity to recognize OVCAR-3 cells growing as an ascites tumor in nude mice. After i.p. injection, control liposomes bind tumor cells in relatively low numbers, while fluorescent OC-125 liposomes can be observed bound specifically to tumor cell masses for periods of days.

Immunomodulatory and toxic effects of free and liposome-encapsulated tumor necrosis factor alpha in rats.

Tumor necrosis factor alpha has potent immunomodulatory and antitumor activity, but its therapeutic applications may be limited by its significant host toxicity. We showed that liposome-encapsulated recombinant human tumor necrosis factor alpha (rHuTNF-alpha) retained full anticellular activity in vitro. We then assessed the immunomodulatory and toxic effects of two different doses of i.v. free or liposome-encapsulated rHuTNF-alpha in normal rats. Both free and liposome-encapsulated rHuTNF-alpha significantly enhanced alveolar macrophage- and blood monocyte-mediated interleukin 1 release and tumor cell lysis, as well as natural killer cell cytotoxicity, when compared to buffer-treated controls. However, administration of rHuTNF-alpha in liposomes substantially reduced tumor necrosis factor alpha-mediated toxicity. Animals receiving liposome-encapsulated rHuTNF-alpha showed significantly less tissue injury, gastric retention, and circulating leukocyte shifts than animals receiving free rHuTNF-alpha. In addition, liposome-based delivery significantly increased lung and liver uptake of rHuTNF-alpha. Therefore, liposome-encapsulated rHuTNF-alpha retains immunomodulatory activity, significantly reduces toxic inflammatory effects, and may allow targeting of tumor necrosis factor alpha to selected organs after i.v. administration.

Endocytosis and intracellular processing accompanying transfection mediated by cationic liposomes.

Cationic liposomes mediate efficient transfection of mammalian cells, but the manner in which cells internalize and process cationic liposome-DNA complexes has not been well characterized. We exposed several cell types, including human and murine erythroleukemia cells. African green monkey kidney cells (CV-1), isolated rat alveolar type II cells and alveolar macrophages to DNA-cationic liposome complexes containing N-(1-2,3-dioleyloxypropyl)-N,N,N-triethylammonium (DOTMA) and Dioleylphosphatidylethanolamine (DOPE). The morphology of liposome-cell interactions was assessed by electron microscopy. Liposome preparations were complexed to colloidal gold particles or to both plasmid DNA and gold particles. Cells treated with DOTMA liposome-DNA complexes demonstrated endocytosis of the liposome-DNA complexes in coated pits, which were seen in early endosomes, late endosomes, and lysosomes. In isolated alveolar type II cells, the gold-labelled DOTMA lipid apparently mixed with the contents of lamellar bodies. In most cells, gold particles were dispersed throughout the cytoplasmic matrix. In a small proportion of CV-1 and U937 cells, a membrane system resembling the endoplasmic reticulum developed within the nucleus. This novel structure was also present in nuclei after they were isolated from CV-1 cells and then mixed with DOTMA-containing liposomes. Membranes which form after exposure to DOTMA-containing liposomes were 10 nm in thickness as compared to the approx. 8 nm thickness of endogenous cellular membranes. Based on these morphologic observations, we propose that the main route of entry of cationic liposomes into cells is by endocytosis. In some instances, the endosomal compartment releases its cationic liposome-DNA contents into the cytoplasmic matrix. Occasionally, liposomes may enter the nucleus by fusion with the nuclear envelope, creating vesicular and reticular intranuclear membranes. It is not clear at present which, if any of these morphological observations correlates with transfection mediated by cationic liposomes.

Targeting of anti-Thy 1.1 monoclonal antibody conjugated liposomes in Thy 1.1 mice after intravenous administration.

125I-labeled liposomes, conjugated to an anti-Thy 1.1 monoclonal antibody (MRCOX7), demonstrated up to 7.4-fold greater lymph node uptake than liposomes conjugated to non-specific monoclonal antibody (R-10) after intravenous injection into Thy 1.1 (AKR-J) mice. Uptake of anti-Thy 1.1-conjugated liposomes by the lymph nodes of AKR-J mice was 3-times greater than their uptake by lymph nodes of Thy 1.2 (AKR-Cu) mice. Lymph node localization of anti-Thy 1.1-liposomes was equal to that of control monoclonal antibody-liposomes in Thy 1.2 mice. Conjugation to either monoclonal antibody substantially increased liposome clearance by the liver, while decreasing liposome uptake in a number of organs outside the reticuloendothelial system. Changes in liposome size and phospholipid composition did not significantly alter these results. Administration of a large predose of unconjugated liposomes prior to injection of MRCOX7-conjugated liposomes increased blood levels and reduced liver uptake of the monoclonal antibody-liposome conjugates, but did not further enhance lymph node uptake. This study demonstrates that targeting of liposomes by conjugation to the appropriate monoclonal antibody, can significantly increase their uptake in lymph nodes which contain high levels of cells expressing the target antigen. However, conjugation to monoclonal antibody also increases clearance of liposomes by the liver. To increase the uptake of monoclonal antibody-conjugated liposomes in target tissue, substantial reduction of their clearance by the reticuloendothelial system will be required.

Evaluation of the role of lipoprotein metabolism genes in systemic cationic liposome-mediated gene transfer in vivo.

Germ line gene disruption and gene insertion are often used to study the function of selected genes in vivo. We used selected knockout and transgenic mouse models to attempt to identify lipoprotein-related genes and gene products that regulate the process of intravenous cationic liposome-DNA complex (CLDC)-based gene delivery. Several observations suggested that proteins involved in lipoprotein metabolism might be important in influencing the delivery and/or expression of CLDC. First, in vitro transfection of either K562 or CHO cells by CLDCs was enhanced by the presence of a functional low-density lipoprotein receptor (LDLR). Second, pretreatment of mice with 4-aminopyrazolopyrimidine (4APP), an agent that alters lipoprotein profiles in mice, significantly decreased expression of luciferase (luc) after intravenous injection of CLDC-luc complexes in mice. Therefore, we tested mouse model systems either deficient for, or overexpressing, selected genes involved in lipoprotein metabolism, for their potential to regulate intravenous, CLDC-based gene delivery. Although homozygous knockout mutation in the apoE gene caused a significant decrease in gene expression in many tissues of apoE-deficient mice, mice with homozygous deletion of both the apoE and LDLR genes showed wild-type levels of gene transfer efficiency. Thus, a secondary event, produced by homozygous deletion of apoE, but compensated for by the concomitant deletion of LDLR, and/or effects resulting from strain-related, genetic background differences, appeared to play a significant role in mediating intravenous, CLDC-based gene delivery. Secondary alterations resulting from germ line knockouts, as well as epigenetic effects produced by strain differences, may limit the ability to assign specific, gene transfer-related functions to the deleted gene.

Gene expression along the cerebral-spinal axis after regional gene delivery.

We demonstrate here that intracerebroventricular or spinal cord (intrathecal) injection of either plasmid DNA alone or cationic liposome: DNA complexes (CLDCs) produces significant levels of expression of both reporter genes and biologically relevant genes in nonparenchymal cells lining both the brain and the spinal cord. Gene expression was identified both within the spinal cord and the brain after intracerebroventricular or intrathecal injection of either CLDCs or plasmid DNA alone. Intracerebroventricular or intrathecal injection of CLDCs containing the beta-galactosidase (beta-Gal) gene produced patchy, widely scattered areas of beta-Gal expression. The chloramphenicol acetyltransferase (CAT) reporter gene product reached peak levels between 24 hr and 1 week postinjection, and was still present at significant levels 3 weeks after a single intracerebroventricular or intrathecal injection. Intrathecal injection of the human granulocyte colony-stimulating factor (G-CSF) gene produced high levels of hG-CSF activity in both the spinal cord and the brain. Intracerebroventricular injection of CLDCs containing the murine nerve growth factor (NGF) gene increased mNGF levels in the hippocampus, a target region for cholinergic neurons in the medial septum, and increased cholinergic neurotransmitter synthetic enzyme choline acetyltransferase (ChAT) activity within the brain, a well-characterized effect of both purified and recombinant NGF protein. These findings indicate that intracerebroventricular or intrathecal injection of CLDCs can produce significant levels of expression of biologically and therapeutically relevant genes within the CNS. Efficient gene transfer into the CNS will facilitate the evaluation of gene function and regulation within the brain and spinal cord. We attempted to transfer and express genes within the brain and spinal cord by direct CNS injection of either DNA alone or CLDCs into the intraventricular and subarachnoid compartments. We show that intracerebroventricular or spinal cord (intrathecal) injection of either plasmid DNA alone or CLDCs produces significant levels of expression of both reporter genes and biologically relevant genes in nonparenchymal cells lining both the brain and the spinal cord. Intrathecal injection of the hG-CSF gene produced high levels of hG-CSF activity in both the spinal cord and the brain. Intracerebroventricular injection of CLDCs containing the murine NGF gene increased mNGF levels in the hippocampus, and increased cholinergic neurotransmitter synthetic enzyme ChAT activity within the brain. Locoregional diffusion of gene products expressed by transfected meningeal lining cells into brain and spinal cord parenchyma could potentially target secreted proteins within brain and spinal cord regions relevant to neuropathological states while limiting peripheral side effects.

Topical gene delivery to murine skin.

We topically applied naked plasmid DNA containing the luciferase or chloramphenicol acetyltransferase cDNA directly to mouse skin. Gene expression was detected in skin samples as early as 4 h after DNA application, plateaued from 16 to 72 h post-application, and had decreased significantly by 7 d post-application. Reporter gene activity following topical DNA delivery was comparable with that produced by intradermal injection of DNA. Plasmid DNA at concentrations > or =0.25 microg per microl were required to achieve maximal expression levels. Reporter gene expression following topical administration was largely confined to the superficial layers of the epidermis and to hair follicles. Surprisingly, certain cationic liposomes inhibited the efficiency of cutaneous gene transfer. This technique provides a simple, clinically relevant approach to deliver genes to the skin, with potential application in treating a variety of cutaneous disorders.

Efficient gene expression in skin wound sites following local plasmid injection.

Transfection of the skin by local gene delivery, as well as widespread transfection of systemic tissues following intravenous injection of cationic liposome/DNA complexes have been reported. Here, we show that surgically wounded mouse skin can be transfected either by local injection of DNA alone or by intravenous injection of optimized cationic liposome/DNA complexes; however, direct cutaneous injection produces much higher levels of gene expression in the skin, which is targeted to dermal and subdermal layers. High levels of chloramphenicol acetyltransferase activity were present from 3 h to 2 wk following direct injection of a gene expression plasmid into wounded skin and were maintained at detectable levels up to 8 wk after injection. Expression of transferred chloramphenicol acetyltransferase as well as beta-GAL genes was localized to fibroblasts, macrophages, and adipocytes as determined by histochemistry and immunohistochemistry. Further- more, local injection of a human granulocyte- colony-stimulating factor gene expression plasmid produced high levels of the biologically relevant human granulocyte-colony-stimulating factor protein in wounded mouse skin. This efficient and simple method of site-specific gene transfer into wounds may lead to the development of cutaneous gene therapy directed against disorders of abnormal cutaneous wound healing.

Monocyte-mediated lysis of HIV-infected tumor cells.

Peripheral blood monocytes (PBM) can selectively lyse malignant or virus-infected cells. We investigated the effects of target cell infection with HIV-1 on PBM cytolytic function. Cytokine-activated PBM lysed uninfected, HSV-1-infected or vaccinia virus-infected tumor cells, but did not lyse the same cell lines when infected with the human immunodeficiency virus type 1 (HIV-1). HIV did not impair PBM viability, and actinomycin D (Act D) pretreatment of HIV-infected target cells restored their susceptibility to PBM-mediated lysis. Either antibody to CD4 (Leu3a) or a recombinant vaccinia virus that induces expression of the HIV envelope protein, also inhibited target cell lysis by PBM. These studies indicate that CD4 can function as a mediator of PBM cytolytic function, and that target cell expression of the HIV-1 envelope protein may inhibit monocyte-mediated antitumor responses.

Aerosolized pentamidine as second line therapy in patients with AIDS and Pneumocystis carinii pneumonia.

The use of aerosolized pentamidine was investigated in ten patients with the acquired immunodeficiency syndrome (AIDS) and Pneumocystis carinii pneumonia (PCP) who had previous or concurrent severe adverse reactions or contraindications to trimethoprim-sulfamethoxazole or parenteral pentamidine. A dose of 600 mg pentamidine in 6 ml sterile water, aerosolized in a small-particle producing jet nebulizer was administered for 25 minutes once daily for an average of 10.5 days to these ten patients. All patients improved their arterial O2 saturation and showed clinical and roentgenographic improvement within six to 21 days of aerosol pentamidine therapy. No adverse systemic reactions occurred. The results of this small open trial indicate that aerosolized pentamidine is effective and can be given safely to AIDS patients with PCP who have had adverse reactions to trimethoprim-sulfamethoxazole or parenteral pentamidine.

Occupational exposure to aerosolized pentamidine.

Occupational exposure to aerosolized pentamidine has raised questions regarding transmission of tuberculosis and the effect of the drug itself. To estimate the exposure of a health care worker, we measured the ambient concentration of aerosolized pentamidine in field conditions in 36 m3 unventilated treatment room. The amount of pentamidine averaged in three different environmental air samples over a four-hour period was 4.5 +/- 3.6 x 10(-5) mg/m3. This amount is very small compared to the doses received by the patients in whom long-term adverse effects are few. The greater risk to health care workers is probably transmission of tuberculosis from undiagnosed cases, especially in populations with an increased incidence of tuberculosis. Tuberculosis control measures such as improved ventilation and masks should also decrease exposure to ambient air pentamidine until toxicity studies determine long-term adverse effects, if any, of aerosolized pentamidine.

The folate antagonist, methotrexate, is a potent inhibitor of murine and human cytomegalovirus in vitro.

Cytomegalovirus (CMV) is a major source of morbidity for immunocompromised patients, such as AIDS patients. The folic acid antagonists have not been explored as potential antiviral agents against CMV. We examined the effects of methotrexate, compared to acyclovir and ganciclovir, on both murine CMV (MCMV) and human CMV (HCMV) in vitro. Using a plaque assay in mouse embryo cells or human foreskin fibroblasts for MCMV and HCMV respectively, we found that methotrexate, in micromolar concentrations, was a potent inhibitor of both viruses. This effect was due to folic acid antagonism since folinic acid abrogated the antiviral effect of methotrexate, but not ganciclovir. Cellular toxicity due to methotrexate appeared insufficient to account for the antiviral effects. The ability of methotrexate to inhibit CMV in vivo merits exploration.

Interaction of antimycobacterial and anti-pneumocystis drugs with phospholipid membranes.

Liposomes can be used as carriers of drugs in the treatment of viral, bacterial and protozoal infections. The potential for liposome-mediated therapy of Mycobacterium avium-intracellulare complex infections, one of the most common opportunistic infections in AIDS, is currently under study. Here, we have investigated the effect of the lipid-soluble antimycobacterial drugs ansamycin, clofazimine and CGP7040 on the thermotropic behavior of liposomes composed of dipalmitoylphosphatidylcholine (DPPC) or dipalmitoylphosphatidylglycerol (DPPG) using differential scanning calorimetry (DSC). In the presence of ansamycin (rifabutine), the peak gel-liquid crystalline phase transition temperature (Tm) of DPPG was reduced, as was the sub-transition temperature (Ts), whereas the Tm of DPPC was reduced only slightly. The temperature of the pre-transition (Tp) of DPPC was lowered, while the pre-transition of DPPG was abolished. Ansamycin also caused the broadening of the transition endotherm of both lipids. Equilibration of the drug/lipid complex for 1 or 5 days produced different thermotropic behavior. In the presence of clofazimine, the cooperativity of the phase transition of DPPG decreased. Above 10 mol% clofazimine formed two complexes with DPPG, as indicated by two distinguishable peaks in DSC thermograms. The Tm of both peaks were lowered as the mole fraction increased. Clofazimine had minimal interaction with DPPC. In contrast, CGP7040 interacted more effectively with DPPC than with DPPG, causing a reduction of the size of the cooperative unit of DPPC even at 2 mol%. The main transition of DPPC split into 3 peaks at 5 mol% drug. The pre-transition was abolished at all drug concentrations and the sub-transition disappeared at 10 mol% CGP7040. These studies suggest that maximal encapsulation of clofazimine in liposomes would require a highly negatively charged membrane, while that of CGP7040 would necessitate a zwitterionic membrane. We have also investigated the interaction of the water-soluble antibiotic pentamidine, which has been used against Pneumocystis carinii, the most lethal of AIDS-related opportunistic pathogens. Aerosol administration of this drug leads to long-term sequestration of the drug in the lungs. The DPPG/pentamidine complex exhibited a pre-transition at 3.5 degrees C, an endothermic peak at 42 degrees C, and an exothermic peak at 44.5 degrees C, followed by another endothermic peak at 55 degrees C. The exotherm depended on the history of the sample, requiring pre-incubation for several minutes below the 42 degrees C transition. These observations suggest that upon melting of the DPPG chains at 42 degrees C, the DPPG crystallizes as a DPPG/pentamidine complex that melts at 55 degrees C.

Systemic co-administration of depsipeptide selectively targets transfection enhancement to specific tissues and cell types.

Depsipeptide, a histone deacetylase (HDAC) inhibitor, kills tumor cells much more effectively than normal cells, and can produce significant antitumor activity in human cancer patients. Depsipeptide also increases the expression of lipoplex-delivered genes in cultured tumor cells, as well as following direct intra-tumoral injection. We now show that co-intravenous (i.v.) injection of depsipeptide with polyethylenimine (PEI):DNA complexes significantly increases the expression of PEI-delivered genes in normal, as well as in tumor-bearing mice. At the tissue level, depsipeptide-mediated enhancement of gene expression was selectively targeted to the lung, liver and spleen. At the cellular level, depsipeptide significantly increased the expression of the i.v., PEI co-delivered wild-type human p53 gene in metastatic breast cancer cells, but not in adjacent normal cells. Thus, the ability of depsipeptide to enhance the expression of systemically delivered genes is selectively targeted at both the tissue and cellular levels, without requiring the use of ligand- or promoter-based approaches. Analyzing HDAC-based targeting of gene expression may identify host genes that control the expression of systemically delivered genes.

Incorporation of recombinant gamma interferon into liposomes enhances its ability to induce peritoneal macrophage antitoxoplasma activity.

In this study we compared the ability of free- and liposome-incorporated murine recombinant gamma interferon (rIFN-gamma) to enhance peritoneal macrophage H2O2 release and antitoxoplasma activity in vitro. rIFN-gamma was efficiently (37 to 47%) incorporated into multilamellar vesicles composed of phosphatidylglycerol/cholesterol in a 2:1 molar ratio. The amount of rIFN-gamma incorporated into multilamellar vesicles and added to macrophages (0.1 to 1,000 U/ml) was quantitated with [3H]rIFN-gamma. The concentration of liposomal rIFN-gamma required to enhance macrophage H2O2 release (1 U/ml) and maximally inhibit Toxoplasma gondii growth (10 U/ml) was one-tenth the concentration required for free rIFN-gamma (10 and 100 U/ml, respectively). This increase in potency was observed in both thioglycolate-elicited and resident peritoneal macrophages. Control liposomes containing encapsulated buffer had no effect on the potency of free rIFN-gamma. The duration of macrophage activation induced by 24 h of liposomal rIFN-gamma treatment was also considerably longer than that induced by free rIFN-gamma (2 days versus less than 1 day). These data indicate that liposomal rIFN-gamma is more active than free rIFN-gamma as an inducer of macrophage microbicidal properties in vitro. This enhanced activity, combined with the potential for selective delivery of liposomal rIFN-gamma to phagocytic cells in vivo, may improve the therapeutic efficacy of rIFN-gamma in infections characterized by parasitization of phagocytes.

Enhanced human monocyte cytotoxicity by platelet-activating factor.

The capacity of platelet-activating factor (PAF) to enhance human monocyte cytotoxicity for WEHI 164 cells was examined. Spontaneous monocyte cytotoxicity was 24 +/- 2% (mean +/- SEM, n = 9). Preincubation of monocytes with 1 pM-1 nM PAF for 18 hr significantly enhanced cytotoxicity in a dose-related manner, whereas less enhancement was observed at PAF concentrations above 1 nM. Maximal PAF-induced cytotoxicity was 68 +/- 6%, which was similar to that induced by optimal concentrations of tumour necrosis factor (TNF) and interferon-gamma. The specific PAF antagonist kadsurenone inhibited PAF-induced cytotoxicity but not TNF-induced cytotoxicity. The inactive PAF analogues lysoPAF and enantioPAF did not increase monocyte cytotoxicity. Two observations suggest that TNF mediates PAF-induced cytotoxicity: specific anti-TNF antibodies inhibited PAF-induced cytotoxicity toward WEHI 164 cells, and PAF did not enhance cytotoxicity to TNF-resistant cells. PAF represents a distinct class of phospholipid monocyte activators that increase monocyte cytotoxicity by TNF-dependent mechanisms.

Aerosol administration of antiviral agents to treat lung infection due to murine cytomegalovirus.

Cytomegalovirus (CMV) pneumonia causes significant morbidity and mortality in bone marrow transplant recipients and in patients with AIDS. 9-(1,3-Dihydroxy-2-propoxymethyl) guanine (ganciclovir) and phosphonoformic acid (PFA) demonstrate activity against CMV in human infections, although recurrent CMV and systemic drug toxicity frequently develop. We examined the efficacy of aerosol administration of antiviral agents against murine CMV (MCMV) infection. Animals were inoculated with MCMV intranasally and were treated with oral ganciclovir; with aerosolized ganciclovir, PFA, or ribavirin; or with buffer. MCMV in lung and salivary gland homogenates was quantified by plaque assay. Oral ganciclovir (200 mg/kg per day) reduced titers of MCMV in both tissues by greater than 95%. Aerosolized ganciclovir, 100 and 200 mg/kg per day, reduced lung titers of MCMV by 93% and 97%, respectively. Aerosolized PFA, 20 and 200 mg/kg per day, reduced lung titers of MCMV by 60% and 68%, respectively. Aerosolized ganciclovir and PFA inhibited replication of MCMV in salivary glands substantially less than did oral administration of either agent. Our results suggest that aerosol administration of antiviral agents can potently and selectively inhibit replication of MCMV in the lung.

Cationic lipid-mediated transfection of liver cells in primary culture.

We describe transfection of DNA into parenchymal and individual non-parenchymal cell populations from adult rat liver in early primary culture, using cationic lipid as the carrier. All cell populations were transfectable, although lipid requirements varied by cell type and, for hepatocytes, with the age of the culture. For hepatocytes in early primary culture (2-10 hours after plating), pure DOTMA (N-[1-(2,3-dioleyloxy)propyl]-N,N,N-trimethylammonium chloride) was strikingly more effective than commercial formulations (Lipofectin or TransfectACE) containing components in addition to, or other than DOTMA. For hepatocytes fully adapted to culture (approximately 48 hours after plating), pure DOTMA and Lipofectin were similarly effective. Under optimal conditions, about 10% of hepatocytes expressed the transfected reporter gene. CAT expression in hepatocytes doubled from 48 hours to 7 days after transfection. The effect of culture substratum on transfection efficiency also was examined. The presence of basement membrane-like matrix (EHS gel) reduced uptake of the DNA-lipid complex. However, cells in early culture that were transfected on collagen and then replated on EHS gel, displayed significantly greater reporter gene activity than did cells maintained throughout on collagen. In contrast to hepatocytes, non-parenchymal cells (lipocytes, Kupffer cells and endothelial cells, respectively) were transfected most efficiently by Lipofectin; DOTMA alone was inactive. The methods described will facilitate studies of gene regulation in individual liver cell populations.