Effect of cisplatin, lipopolysaccharide, muramyl dipeptide and recombinant interferon-gamma on murine macrophages in vitro. I. Macrophage-mediated tumor cell lysis

Murine macrophage monolayers treated with cisplatin, lipopolysaccharide (LPS), muramyl dipeptide (MDP) or recombinant interferon-gamma (rIFN gamma) were observed to have significantly increased tumoricidal activity. rIFN gamma had synergistic effects with cisplatin, LPS or MDP in activating macrophages. However, MDP showed much more pronounced synergism with cisplatin and LPS than with rIFN gamma. Supernatants collected from these activated macrophage monolayers also showed increased tumoricidal activity. Tumor cell lysis mediated by cisplatin-treated macrophages did not require priming with rIFN gamma though it may be necessary as a first signal for the increased macrophage activation with LPS and MDP.     

Inhibition of prostaglandin E2-stimulated cAMP accumulation by lipopolysaccharide in murine peritoneal macrophages.

Treatment of murine peritoneal macrophages with 100 nM prostaglandin E2 (PGE2) produced a rapid biphasic increase in intracellular cAMP that was maximal at 1 min and sustained through 20 min. Pretreatment of macrophages with 100 ng/ml of lipopolysaccharide (LPS) for 60 min prior to PGE2 decreased the magnitude of cAMP elevation by 50%, accelerated the decrease of cAMP to basal levels, and abolished the sustained phase of cAMP elevation. The effect of LPS was concentration-dependent, with maximal effect at 10 ng/ml in cells incubated in the presence of 5% fetal calf serum and at 1 microgram/ml in the absence of fetal calf serum. LPS also inhibited cAMP accumulation in cells treated with 100 microM forskolin, but the decrease was about half that seen in cells treated with PGE2. LPS concentrations that inhibited cAMP accumulation produced a 30% increase in soluble low Km cAMP phosphodiesterase activity while having no effect on particulate phosphodiesterase activity. The nonspecific phosphodiesterase inhibitor, 3-isobutyl-1-methylxanthine, as well as the more specific inhibitors rolipram and Ro-20-1724 were effective in inhibiting soluble phosphodiesterase activity in vitro, producing synergistic elevation of cAMP in PGE2-treated cells, and blocking the ability of LPS to inhibit accumulation of cAMP. Separation of the phosphodiesterase isoforms in the soluble fraction by DEAE chromatography indicated that LPS activated a low Km cAMP phosphodiesterase. The enzyme(s) present in this peak could be activated 6-fold by cGMP and were potently inhibited by low micromolar concentrations of Ro-20-1724 and rolipram. Using both membranes from LPS-treated cells and membranes incubated with LPS, no decrease in adenylylcyclase activity could be attributed to LPS. Although effects of LPS on the rate of synthesis of cAMP cannot be excluded, the present evidence is most consistent with a role for phosphodiesterase activation in the inhibitory effects of LPS on cAMP accumulation in murine peritoneal macrophages.     

Lysosomal trafficking functions of mucolipin-1 in murine macrophages

Background  

Mucolipidosis Type IV is currently characterized as a lysosomal storage disorder with defects that include corneal clouding, achlorhydria and psychomotor retardation. MCOLN1, the gene responsible for this disease, encodes the protein mucolipin-1 that belongs to the "Transient Receptor Potential" family of proteins and has been shown to function as a non-selective cation channel whose activity is modulated by pH. Two cell biological defects that have been described in MLIV fibroblasts are a hyperacidification of lysosomes and a delay in the exit of lipids from lysosomes.     

The fate of E. coli lipopolysaccharide after the uptake of E. coli by murine macrophages in vitro

The fate of bacterial lipopolysaccharide (LPS) after the uptake of Escherichia coli by macrophages in vitro was studied. The LPS of the galactose epimerase-deficient E. coli J5 mutant was specifically radiolabeled with [3H]galactose by growing the organism in a basic salts medium containing galactose. Control bacteria were uniformly radiolabeled by growth in [14C]glucose and unlabeled galactose-containing medium. Surface constituents of E. coli were also labeled with 125I. After in vitro phagocytosis of labeled E. coli by murine peritoneal exudate macrophages, the rate of exocytosis of LPS, as assessed by release of 3H over a 72-hr period, was considerably reduced in comparison with other bacterial constituents (14C and 125I release). The [3H]galactose-labeled material exocytosed from macrophages and that remaining intracellularly (obtained from macrophage lysates) were isolated by cesium chloride (CsCl) density gradients and were shown to have altered density profiles as compared with purified E. coli LPS. The macrophage-"processed" [3H] galactose-containing fractions from CsCl density gradients of culture supernatants or macrophage lysates were capable of clotting Limulus amebocyte lysate. The [3H]galactose material obtained from 48-hr macrophage lysates and culture supernatants could also induce a lethal response in actinomycin D-treated mice. These data suggest that bacterial LPS may be selectively retained by the macrophage and that the post-phagocytic events that result in bacterial degradation are not accompanied by the degradation of LPS. Furthermore, although the LPS may be modified by the macrophage, it retains its biologic activity.     

Lysosomal acid phosphatase is transported to lysosomes via the cell surface.

Lysosomal acid phosphatase (LAP) is transported as a transmembrane protein to dense lysosomes. The pathway of LAP to lysosomes includes the passage through the plasma membrane. LAP is transported from the trans-Golgi to the cell surface with a half-time of less than 10 min. Cell surface LAP is rapidly internalized. Most of the internalized LAP is transported back to the cell surface. On average, each LAP molecule cycles greater than 15 times between the cell surface and the endosomes before it is transferred to dense lysosomes. At equilibrium approximately 4 times more LAP precursor is present in endosomes than at the cell surface. Exposing cells to reduced temperature or weak bases such as NH4Cl, chloroquine and primaquine decreases the steady-state concentration of LAP at the cell surface. The recycling pathway is operative at greater than or equal to 20 degrees C and does not include passage of the Golgi/trans-Golgi network. LAP is transferred with a half-time of 5-6 h from the plasma membrane/endosome pool to dense lysosomes, from where it does not recycle to the endosome/plasma membrane pool at a measurable rate.     

beta-Glucosidase inhibition in murine peritoneal macrophages by conduritol-B-epoxide: an in vitro model of the Gaucher cell

Murine peritoneal macrophages were cultured in the presence of conduritol-B-epoxide, a specific covalent inhibitor of beta-glucosidase. The inhibition was found to be dose and time dependent. Upon removal of the inhibitor from the culture medium, beta-glucosidase activity recovered to half maximum by 2.2 days. Treatment of macrophages with this inhibitor for 15 days did not affect cell viability, lysosomal enzyme release to the medium, or levels of intracellular lysosomal enzymes, other than beta-glucosidase activity. This inhibition results in the accumulation of glucocerebroside. In vitro studies on the pathobiology of such macrophages whose beta-glucosidase activity has been reduced may be useful toward understanding the pathogenesis of Gaucher disease.     

Imbalanced accumulation of ribosomal RNA in macrophages activated in vivo or in vitro to a cytolytic stage

Previous studies have shown that peritoneal murine macrophages activated in vivo and in vitro to a tumoricidal stage have a depressed rate of RNA synthesis. In attempting to clarify the differences in RNA metabolism between noncytotoxic and tumoricidal macrophages, we have studied the relative accumulation of various species of RNA in macrophages activated in vivo and in vitro with the use of agarose gel electrophoresis. Macrophages activated in vitro to a cytotoxic stage with supernatants containing lymphokines (LK) and traces of lipopolysaccharide (LPS) have an imbalanced accumulation of mature ribosomal RNA (rRNA), with a decreased accumulation of 28S rRNA compared to 18S rRNA. In contrast, macrophages primed in vitro with LK free of detectable endotoxins that exhibit suppressive rather than tumoricidal activity do not manifest a decreased 28S:18S rRNA ratio. The conclusion that the decreased 28S:18S rRNA ratio was associated with the activation of macrophages to a cytolytic stage was supported by the finding that cytotoxic macrophages activated in vivo by i.p. injection of Propionibacterium acnes (formerly designated C. parvum) also demonstrated a decreased accumulation of 28S comparable with that observed in in vitro-activated macrophages. Moreover, activated macrophages that lost their cytolytic activity upon prolonged in vitro culture had an augmented accumulation of 28S rRNA. These results provide the first direct evidence that the expression of cytolytic activity is associated with modulation of a specific class of RNA. The unbalanced accumulation of rRNA appears to be a late molecular event in the activation process occurring during the transition from primed to cytotoxic macrophages, because inflammatory and primed macrophages had normal rRNA accumulation. A model of macrophage activation accounting for these results is proposed.     

A physiological model to study iron recycling in macrophages

Following erythrophagocytosis (EP) of senescent red blood cells (RBCs), heme iron is recycled to the plasma by tissue macrophages. This process is critical for mammalian iron homeostasis but remains elusive. We characterized a cellular model using artificially-aged murine RBCs and murine bone marrow-derived macrophages (BMDMs) and study mRNA and protein expression of HO-1, ferroportin and ferritin after EP. In vitro ageing of RBCs was obtained by raising intracellular calcium concentration. These RBCs exhibit several features of erythrocyte senescence including externalization of phosphatidyl-serine, specific binding and phagocytosis by BMDMs. During the first hours of EP, we observed a rapid increase of HO-1 and ferroportin mRNAs and proteins, whereas ferritin protein expression was progressively induced with no major changes in RNA levels. At later stages after EP, a different pattern of expression was observed with a net decrease of ferroportin, a sustained high level of HO-1, and a strong increase in ferritins. Taken together, these results suggest that after EP, iron is rapidly extracted from heme and exported by ferroportin. Surprisingly, the gene expression profile at late stages after EP, which is indicative of iron storage, is reminiscent of what is observed in inflammation. However, phagocytosis of artificially-aged red blood cells seems to repress the proinflammatory response of macrophages.     

Brucella abortus lipopolysaccharide in murine peritoneal macrophages acts as a down-regulator of T cell activation

Macrophages play a central role in host immune responses against pathogens by acting as both professional phagocytic cells and as fully competent APCs. We report here that the LPS from the facultative intracellular Gram-negative bacteria Brucella abortus interferes with the MHC class II Ag presentation pathway. LPS inhibits the capacity of macrophages to present hen egg lysozyme (HEL) antigenic peptides to specific CD4(+) T cells but not those of OVA to specific CD8(+) T cells. This defect was neither related to a decrease of MHC class II surface expression nor to a deficient uptake or processing of HEL. In addition, B. abortus LPS did not prevent the formation of SDS-resistant MHC class II complexes induced by HEL peptides. At the cell surface of macrophages, we observed the presence of LPS macrodomains highly enriched in MHC class II molecules, which may be responsible for the significant down-regulation of CD4(+) T cell activation. This phenomenon may account for the avoidance of the immune system by certain bacterial pathogens and may explain the immunosuppression observed in individuals with chronic brucellosis.     

Altered cell-averaged microviscosity of murine peritoneal macrophages undergoing activation in vivo or in vitro

The cell-averaged microviscosity of intact murine peritoneal mononuclear phagocytes in various stages of activation was assessed by quantifying fluorescent depolarization of 1,6-diphenyl-1,3,5-hexatriene. Macrophages activated in vivo with Mycobacterium bovis, strain BCG, were significantly more fluid than resident peritoneal macrophages, responsive macrophages elicited with thioglycollate broth, proteose peptone broth, or fetal bovine serum, or primed macrophages elicited with pyran copolymer, MVE-2. Specifically, the cell-averaged microviscosity decreased from a mean of 3.47 +/- .07 eta 25 degrees C (poise) (range of 3.32 to 3.67 p) to 2.62 eta 25 degrees C. Exposure of responsive macrophages in vitro to bacterial endotoxin plus hybridoma supernatants containing macrophage-activating factor or purified recombinant interferon gamma resulted in decreased microviscosity; the largest effect was seen after 24 hr. Macrophages primed in vivo with MVE-2 and treated in vitro with endotoxin also developed decreased microviscosity. Similar changes in microviscosity were observed in a plasma membrane-enriched fraction isolated from macrophages activated in vitro with interferon gamma and endotoxin, thus suggesting that the cell-averaged measurements reflected changes in membrane viscosity. The optimum concentration of MAF-inducing decreased overall microviscosity was identical to that for inducing tumoricidal capacity. Taken together, the data indicate activation of lytic capacity in murine macrophages is closely associated with decreased cell-averaged microviscosity and that this change reflects, at least in part, decreased microviscosity of the plasma membrane of these cells.     

Internalization and degradation of receptor-bound interferon-gamma by murine macrophages. Demonstration of receptor recycling

Although the interferon-gamma (IFN-gamma) receptor on murine and human mononuclear phagocytes has been defined and partially characterized, very little data exists which describes the ultimate fate of receptor-bound ligand. The current studies were specifically designed to define the metabolic processes which act on murine recombinant IFN-gamma following its interaction with murine macrophages at physiologic temperatures. Ligand internalization was demonstrated by comparing binding of [125I]IFN-gamma to macrophages at 4 degrees C and 37 degrees C. When binding was carried out at 4 degrees C, 96% of the cell-associated [125I]IFN-gamma remained accessible at the plasma membrane and could be stripped from the cell by exposure to pronase. In contrast, at 37 degrees C, only 35% of the cell-associated radioactivity was pronase strippable. Macrophages degraded [125I]IFN-gamma into trichloroacetic acid-soluble material at 37 degrees C at a constant rate of 7000 molecules/cell/hr over a 12-hr time period. The amount of IFN-gamma degraded correlated with the amount of IFN-gamma bound to the cell surface. The receptor was neither up- nor down-regulated by ligand or by other agents known to regulate macrophage functional activity such as IFN-alpha, IFN-beta, lipopolysaccharide, or phorbol myristate acetate. The constant uptake of IFN-gamma by macrophages was due to the presence of an intracellular receptor pool (62% of the total receptor number) and to a mechanism of receptor recycling. Evidence for the latter was obtained using lysosomotropic agents which blocked degradation but not binding and internalization of ligand and caused the intracellular accumulation of receptor. By comparing the relationship between receptor occupancy and biologic response induction, two activation mechanisms became apparent. Induction of certain functions, such as H2O2 secretion, appeared to require only a single round of receptor occupancy. However, induction of more complex functions such as nonspecific tumoricidal activity appeared to require three to four rounds of receptor occupancy. These results thus support the concept that IFN-gamma internalization and receptor recycling are essential in the induction of nonspecific tumoricidal activity by macrophages.     

CD40 ligation activates murine macrophages via an IFN-gamma-dependent mechanism resulting in tumor cell destruction in vitro

We have shown previously that agonistic anti-CD40 mAb induced T cell-independent antitumor effects in vivo. In this study, we investigated mechanisms of macrophage activation with anti-CD40 mAb treatment, assessed by the antitumor action of macrophages in vitro. Intraperitoneal injection of anti-CD40 mAb into C57BL/6 mice resulted in activation of peritoneal macrophages capable of suppressing B16 melanoma cell proliferation in vitro, an effect that was greatly enhanced by LPS and observed against several murine and human tumor cell lines. Anti-CD40 mAb also primed macrophages in vitro to mediate cytostatic effects in the presence of LPS. The tumoristatic effect of CD40 ligation-activated macrophages was associated with apoptosis and killing of tumor cells. Activation of macrophages by anti-CD40 mAb required endogenous IFN-gamma because priming of macrophages by anti-CD40 mAb was abrogated in the presence of anti-IFN-gamma mAb, as well as in IFN-gamma-knockout mice. Macrophages obtained either from C57BL/6 mice depleted of T and NK cells by Ab treatment, or from scid/beige mice, were still activated by anti-CD40 mAb to mediate cytostatic activity. These results argued against the role of NK and T cells as the sole source of exogenous IFN-gamma for macrophage activation and suggested that anti-CD40 mAb-activated macrophages could produce IFN-gamma. We confirmed this hypothesis by detecting intracytoplasmic IFN-gamma in macrophages activated with anti-CD40 mAb in vivo or in vitro. IFN-gamma production by macrophages was dependent on IL-12. Taken together, the results show that murine macrophages are activated directly by anti-CD40 mAb to secrete IFN-gamma and mediate tumor cell destruction.     

Deficient in vitro and in vivo phagocytosis of apoptotic T cells by resident murine alveolar macrophages

Apoptotic lymphocytes are readily identified in murine lungs, both during the response to particulate Ag and in normal mice. Because apoptotic lymphocytes are seldom detected in other organs, we hypothesized that alveolar macrophages (AMphi) clear apoptotic lymphocytes poorly. To test this hypothesis, we compared in vitro phagocytosis of apoptotic thymocytes by resident AMphi and peritoneal macrophages (PMphi) from normal C57BL/6 mice. AMphi were deficient relative to PMphi both in percentage containing apoptotic thymocytes (19.1 +/- 1% vs 96 +/- 2.6% positive) and in phagocytic index (0.23 +/- 0.02 vs 4.2 +/- 0.67). This deficiency was not due to kinetic differences, was seen with six other inbred mouse strains, and was not observed using carboxylate-modified polystyrene microbeads. Annexin V blockade indicated that both Mphi types cleared apoptotic T cells by a mechanism involving phosphatidylserine expression. By contrast, neither mAb blockade of a variety of receptors (CD11b, CD29, CD51, and CD61) known to be involved in clearance of apoptotic cells, nor the tetrapeptide RGDS (arginine-glycine-aspartic acid-serine) blocked ingestion by either type of macrophage. To confirm these studies, apoptotic thymocytes were given intratracheally or i.p. to normal mice, and then AMphi or PMphi were recovered 30-240 min later. Ingestion of apoptotic thymocytes by AMphi in vivo was significantly decreased at all times. Defective ingestion of apoptotic lymphocytes may preserve AMphi capacity to produce proinflammatory cytokines in host defense, but could contribute to development of autoimmunity by failing to eliminate nucleosomes.     

A fibronectin-related synthetic peptide, Pro-Ala-Ser-Ser, inhibits fibronectin binding to the cell surface, fibronectin-promoted cell migration in vitro, and cell migration in vivo

The biological activity of the amino acid sequence consisting of the immediate carboxyl terminus side of the Arg-Gly-Asp-Ser (RGDS) amino acid sequence in the cell-binding domain of intact fibronectin (FN) molecules was examined using synthetic peptides [RGDS, Gly-Arg-Gly-Asp-Ser-Pro (GRGDSP), Arg-Gly-Asp-Ser-Pro-Ala-Ser-Ser-Lys-Pro (RGDSPASSKP), Pro-Ala (PA), Pro-Ala-Ser (PAS), Pro-Ala-Ser-Ser (PASS), and Pro-Ala-Ser-Ser-Lys (PASSK)]. These peptides were applied to the primary mesenchyme cells (PMCs) of the sea urchin, Clypeaster japonicus. In vitro immunohistochemistry indicated that the binding of exogenous FN to the PMC surface was inhibited by the peptides RGDSPASSKP and PASS, but not by RGDS, GRGDSP, PA, or PAS. PASS and RGDS introduced into the blastocoel also inhibited PMC migration in vivo. FN-promoted PMC migration in vitro was also inhibited by PASS and RGDS. The present results indicate that the PASS peptide inhibits FN binding to the PMC surface and promotes PMC migration, suggesting that the FN molecule uses the PASS amino acid sequence to bind to the PMC surface and to promote PMC migration in the blastocoel.     

Activity of murine peritoneal macrophages upon weak red and infrared laser irradiation in vitro and in vivo

The phagocytic activity and production of proinflammatory cytokines by murine peritoneal macrophages was examined upon in vitro and in vivo exposure to low doses (up to 150 mJ/cm²) of red (650 nm) and infrared (850 nm) light generated by semiconductor lasers. Infrared irradiation proved stimulatory in a dose-dependent way for phagocytosis and killing of enteropathogenic Escherichia coli as well as for cytokine production. Red light markedly enhanced cytokine production but was detrimental to macrophage activity and at higher doses caused their destruction.     

In vitro and in vivo induction of nitric oxide by murine macrophages stimulated with Bordetella pertussis

Abstract Nitric oxide (NO) exhibits potent antimicrobial activity in vitro. The function of NO in host defenses in vivo, however, is presently unclear. Experiments were undertaken to determine the production of NO in vitro from murine peritoneal and alveolar macrophages, and murine macrophage cell line (J774A.1) stimulated with Bordetella pertussis or pertussis toxin (PT). In addition, we determined circulating levels of NO in the sera and bronchoalveolar lavage (BAL) fluids of mice infected intranasally with B. pertussis . The results of this study showed that in vitro murine peritoneal macrophages induce production of NO in response to B. pertussis and PT. In addition, murine macrophage cell line, J774A.1 also induces NO production after stimulation with B. pertussis . NO production was also detected in alveolar macrophages from mice infected intranasally with B. pertussis . Finally, a significant increment of circulating levels of NO was noted, in the sera but not in the BAL fluids, of mice infected intranasally with B. pertussis .     

Lysosomal glycogen accumulation in rat liver and its in vivo kinetics after a single intraperitoneal injection of acarbose, an alpha-glucosidase inhibitor

A single intraperitoneal injection of acarbose (400 mg/kg) into rats caused lysosomal accumulation of glycogen in the liver, mimicking the cytological characteristics of human glycogen storage disease type II (Pompe's disease). The animal model is therefore useful for studying the pathogenesis of the disease. In the present study, we applied this model to examine the lysosomal hydrolytic pathway of glycogen in vivo. To quantify the lysosomal glycogen, the lysosome-rich fraction was rapidly prepared from liver homogenate by agglutination in the presence of Ca2+. Then the fraction was treated with alpha-amylase in isotonic medium to remove cytosolic glycogen, followed by transfer to hypotonic conditions in the presence of Triton X-100 to destroy total glycogen. The amount of lysosomal glycogen was calculated from the difference between the glycogen levels measured before and after the treatment under hypotonic conditions, and then it was corrected based on measurements of the intactness (%) of lysosomes and the recovery (%) of the lysosomal marker enzyme (beta NAGase). We observed no measurable lysosomal glycogen in normal liver by this method, and this was confirmed by electron microscopy. After administration of acarbose, the lysosomal glycogen level increased to 2.5 mg/g liver within 2 days, and then decreased gradually at a rate of 0.4 mg/day/g. The accumulation of glycogen in the lysosomes at an initial velocity of 1.5 mg/day/g liver may be considered as the amount of glycogen that would normally be degraded by acid alpha-glucosidase. Therefore, assuming that the liver breaks down about 40 mg glycogen/day/g, we estimated that about 3% of the glycogen would be hydrolyzed by the lysosomal pathway.     

The synthesis of murine ferrochelatase in vitro and in vivo.

Ferrochelatase (protohaem ferro-lyase, EC 4.99.1.1), the terminal enzyme of the haem-biosynthetic pathway, is an integral membrane protein of the mitochondrial inner membrane. When murine erythroleukaemia cells are labelled in vivo with [35S]methionine, lysed, and the extract is immunoprecipitated with rabbit anti-(mouse ferrochelatase) antibody, a protein of Mr 40,000 is isolated. However, when isolated mouse RNA is translated in a cell-free reticulocyte extract, a protein of Mr 43,000 is isolated. Incubation of this Mr 43,000 protein with isolated mitochondria resulted in processing of the Mr 43,000 precursor to the Mr 40,000 mature-sized protein. Addition of carbonyl cyanide m-chlorophenylhydrazone and/or phenanthroline inhibits this processing. These data indicate that ferrochelatase, like most mitochondrial proteins, is synthesized in the cytoplasm as a larger precursor and is then translocated and processed to a mature-sized protein in an energy-required step.     

Regulation of macrophage chemokine expression by lipopolysaccharide in vitro and in vivo.

The host response to Gram-negative LPS is characterized by an influx of inflammatory cells into host tissues, which is mediated, in part, by localized production of chemokines. The expression and function of chemokines in vivo appears to be highly selective, though the molecular mechanisms responsible are not well understood. All CXC (IFN-gamma-inducible protein (IP-10), macrophage inflammatory protein (MIP)-2, and KC) and CC (JE/monocyte chemoattractant protein (MCP)-1, MCP-5, MIP-1alpha, MIP-1beta, and RANTES) chemokine genes evaluated were sensitive to stimulation by LPS in vitro and in vivo. While IL-10 suppressed the expression of all LPS-induced chemokine genes evaluated in vitro, treatment with IFN-gamma selectively induced IP-10 and MCP-5 mRNAs, but inhibited LPS-induced MIP-2, KC, JE/MCP-1, MIP-1alpha, and MIP-1beta mRNA and/or protein. Like the response to IFN-gamma, LPS-mediated induction of IP-10 and MCP-5 was Stat1 dependent. Interestingly, only the IFN-gamma-mediated suppression of LPS-induced KC gene expression was IFN regulatory factor-2 dependent. Treatment of mice with LPS in vivo also induced high levels of chemokine mRNA in the liver and lung, with a concomitant increase in circulating protein. Hepatic expression of MIP-1alpha, MIP-1beta, RANTES, and MCP-5 mRNAs were dramatically reduced in Kupffer cell-depleted mice, while IP-10, KC, MIP-2, and MCP-1 were unaffected or enhanced. These findings indicate that selective regulation of chemokine expression in vivo may result from differential response of macrophages to pro- and antiinflammatory stimuli and to cell type-specific patterns of stimulus sensitivity. Moreover, the data suggest that individual chemokine genes are differentially regulated in response to LPS, suggesting unique roles during the sepsis cascade.