Macrophage component gp160, a major trypsin-sensitive surface glycoprotein.

Macrophages secrete a large number of proteases, implying in vivo exposure of the cell surface to proteolytic conditions. Mild trypsin treatment of 125I-labeled guinea pig peritoneal macrophages preferentially cleaves one surface component of apparent 160,000 mol wt. Similar trypsin treatment of macrophages with 3H-labeled carbohydrate surface moieties also cleaves a single 3H-labeled 160,000 mol wt glycoprotein, referred to as gp160. Nonreducing sodium dodecyl sulfate (SDS)-electrophoresis established that gp160 of trypsinized cells remains assembled in the membrane as a multichain disulfide-bonded molecule. gp160 was purified by detergent extraction, L. culinaris lectin affinity chromatography and DEAE-cellulose chromatography. The corresponding molecule from trypsinized cells was purified by the same procedure. Reducing SDS-electrophoresis of purified trypsinized 125I-labeled gp160 revealed two proteolytic fragments with apparent molecular weights of 85,000 and 71,000. Thus, mild trypsin treatment of macrophages preferentially cleaves a single surface protein, possibly at a single site. Because the two fragments of gp160 are accessible to lactoperoxidase and trypsin, both must be exposed on the membrane surface. The reactive carbohydrate site was found on the 85,000 mol wt fragment, which alone contains the 3H-label introduced into intact cells by neuraminidase, galactose, oxidase, and [3H]KBH4.

Modulation of CD4 antigen on macrophages and microglia in rat brain.

Mononuclear phagocytes which express the HIV entry receptor CD4 have been implicated as possible sites of virus replication in brain, but there is still considerable uncertainty as to which cells in the CNS express CD4 Ag. Although it is not susceptible to HIV infection the rat provides a model to define expression of the CD4 Ag on MO in brain. We report that the CD4 epitopes W3/25 and OX35 are found only on monocytes, MO, microglia, and occasional lymphocytes and not on neurons, other glia, or endothelium. CD4 Ag levels are modulated during microglial differentiation, after reactivation after local inflammation, and within the intact blood brain barrier. MO and microglia also express other potential plasma membrane binding and entry sites for HIV viz Fc and complement receptors that are regulated independently of CD4.

Novel macrophage receptor for glucose-modified proteins is distinct from previously described scavenger receptors.

A high-affinity macrophage receptor has been identified that recognizes proteins modified by a common in vivo process, long-term nonenzymatic reaction of glucose with proteins (AGE proteins). This receptor for glucose-modified proteins is now shown to be distinct from previously described scavenger receptors, using competition and crosscompetition experiments between AGE-modified protein and a variety of in vitro-modified scavenger receptor ligands, including unmodified BSA, unmodified low-density lipoproteins (LDL), acetyl-LDL, maleyl-BSA, and formaldehyde-treated BSA. Furthermore, the specific pattern of AGE-protein receptor inhibition by the polyanionic compounds polyinosinic acid, polyadenylic acid, polyglutamic acid, polycytidylic acid, fucoidin, and heparin was distinctly different from that of acetyl-LDL. By thus selectively recognizing a time-dependent in vivo protein modification, macrophages may preferentially degrade senescent macromolecules, thereby having an important role in the regulation of extracellular protein turnover.

The subcellular distribution of antigen in macrophages.

The intracellular fate of phagocytosed antigens in cells from peritoneal exudate in CBA mice has been studied by using (126)I and (131)I labeled antigens. After uptake of labeled antigen, cells were homogenized and the subcellular fractions were analyzed by isopycnic centrifugation in a sucrose gradient. The uptake of heat-denatured BSA (c BSA) by these cells in vivo is 3.5 microg/mg c BSA injected/10(8) cells. The uptake by cells in animals which were exposed 2 days earlier to 900 r whole body irradiation is slightly lower but does not differ significantly. 90% of the phagocytosed material is degraded within 2-3 hr, the residual 10% is retained at least over an 8 hr periods. Using a pulse and chase technique, with (125)I and (131)I c BSA in vitro and in vivo it was shown that newly phagocytosed antigen is found mainly in a lysosomal turnover compartment of a density 1.19 g cm(-3). Antigen which has been in the cells for longer was found in a denser fraction (1.26 g cm(-3)). In a comparison of nhrmal and X-irradiated cells it can be shown that after irradiation with 900 r less c BSA is found in this storage compartment. Binding of the antigen to the subcellular fractions, and its behavior towards several detergents has been studied. Subcellular fractions do not have the increased immunogenic capacity of antigen enclosed in living macrophages. Two synthetic polypeptide antigens, poly(D-Tyr, D-Glu, D-Ala) and poly-(L-Tyr, L-Glu) have a different subcellular distribution from c BSA, BSA, or bovine gamma globulin. Apart from also being found in the 1.26 storage compartment the polypeptide antigens are mainly located in a 1.15 compartment and only to a small extent in the 1.19 compartment. The half-life of these antigens in the cells is much longer than the half-life of the protein antigens studied. The finding of several subcellular compartments is discussed in connection with the functions possibly performed by macrophages.

Localization of immunological complexes fixing beta1C (C3) in germinal centers of lymph nodes.

28 human and 60 experimentally stimulated rabbit lymph nodes were studied by means of light microscopy and immunofluorescence. 21 of the 28 human lymph nodes showed well-developed germinal centers. IgM, IgG, and the beta(1C) component of complement were found in the same distribution within germinal centers when examined in serial cryostat sections. 36 rabbits were stimulated with Brucella antigen, and 24 rabbits with BSA. A strikingly consistent correlation between distribution and appearance of specific staining for rabbit beta(1C), IgM, and IgG was observed; when lymph nodes were stimulated with BSA, antigen and specific antibody were present. Treatment of unfixed sections with citrate-buffered saline at low pH resulted in complete elution of immunoglobulins, beta(1C), and BSA from rabbit germinal centers, and in marked diminution of IgG and IgM in human germinal centers, while at the same time plasma cells remained strongly fluorescent. Specific selective fixation of heterologous (human) complement in rabbit germinal centers positive for beta(1C), IgG, IgM, and BSA was also obtained. These data present strong evidence for the existence within germinal centers of antigen-antibody complexes which fix at least the beta(1C) component of complement in vivo. The possibility of complete elution of immunoglobulins from rabbit germinal centers can be taken as evidence that, at least for 20 days after primary and secondary stimulation, a major component of the immunoglobulins present in germinal centers is not produced locally but accumulates at the surface of cells.

Dietary fish oil modulates macrophage fatty acids and decreases arthritis susceptibility in mice.

B10.RIII and B10.G mice were transferred from a diet of laboratory rodent chow to a standard diet in which all the fat (5% by weight) was supplied as either fish oil (17% eicosapentaenoic acid [EPA], 12% docosahexaenoic acid [DHA], 0% arachidonic acid [AA], and 2% linoleic acid) or corn oil (0% EPA, 0% DHA, 0% AA, and 65% linoleic acid). The fatty acid composition of the macrophage phospholipids from mice on the chow diet was similar to that of mice on a corn oil diet. Mice fed the fish oil diet for only 1 wk showed substantial increases in macrophage phospholipid levels of the omega-3 fatty acids (of total fatty acid 4% was EPA, 10% docosapentaenoic acid [DPA], and 10% DHA), and decreases in omega-6 fatty acids (12% was AA, 2% docosatetraenoic acid [DTA], and 4% linoleic acid) compared to corn oil-fed mice (0% EPA, 0% DPA, 6% DHA, 20% AA, 9% DTA, and 8% linoleic acid). After 5 wk this difference between the fish oil-fed and corn oil-fed mice was even more pronounced. Further small changes occurred at 5-9 wk. We studied the prostaglandin (PG) and thromboxane (TX) profile of macrophages prepared from mice fed the two diets just before being immunized with collagen. Irrespective of diet, macrophages prepared from female mice and incubated for 24 h had significantly more PG and TX in the medium than similarly prepared macrophages from male mice. The increased percentage of EPA and decreased percentage of AA in the phospholipids of the macrophages prepared from the fish oil-fed mice was reflected in a reduction in the amount of PGE2 and PGI2 in the medium relative to identically incubated macrophages prepared from corn oil-fed mice. When this same fish oil diet was fed to B10.RIII mice for 26 d before immunization with type II collagen, the time of onset of arthritis was increased, and the incidence and severity of arthritis was reduced compared to arthritis induced in corn oil-fed mice. The females, especially those on the fish oil diet, tended to have less arthritis than the males. These alterations in the fatty acid pool available for PG and leukotriene synthesis suggest a pivotal role for the macrophage and PG in the immune and/or inflammatory response to type II collagen.

Macrophages as origin of factor increasing monocytopoiesis.

Earlier investigations had indicated that the factor increasing monocytopoiesis (FIM), present in the serum of mice and rabbits during the onset of an inflammatory response, is released by cells of the inflammatory exudate. The present study was performed to determine which cells produce and secrete this factor and to establish the kinetics of its production and secretion. FIM was assayed in vivo by intravenous injection of samples into untreated mice and monitoring the course of the number of blood monocytes in the recipients. FIM was assayed in vitro by adding samples to cultures of the macrophage cell line PU5 and determining the rate of proliferation of the cells. The results show that only macrophages contain and synthesize FIM. This factor is secreted upon exposure to a phagocytic stimulus, and after the release of preformed FIM, macrophages secrete newly synthesized FIM. Granulocytes and lymphocytes neither contain nor secrete FIM. The characteristics of FIM derived from macrophages are in all aspects similar to those of FIM in serum. Macrophage-derived FIM is a protein with a molecular weight between 10 and 25 X 10(3), its activity is cell-lineage specific and dose dependent, and it stimulates monocyte production in the bone marrow. Macrophage-derived FIM is not identical to either CSF-1 or IL-1, and has no chemotactic activity. Taken together, the present results show that FIM occurring in serum during an inflammatory response originates from macrophages at the site of the inflammation. In this way the macrophages themselves regulate the supply of circulating blood monocytes that can migrate to the site of injury when needed.

Activated macrophages release a factor which lyses malignant cells but not normal cells.

When rat macrophages are activated by exposure to bacterial endotoxin they acquire the capacity to lyse sarcoma cells in tissue culture. Although showing no immunological specificity this lytic effect was only detectable on malignant cells; normal cells were unaffected. This tumor-selective lytic effect is mediated by a labile supernatant product.

Dual origin of mouse spleen macrophages.

The present study concerns the isolation, characterization, origin, and kinetics of spleen macrophages. The spleen was first perfused in situ to remove monocytes from the vascular bed and then dissected and treated with collagenase. The macrophages in the cell suspension thus obtained were characterized morphologically and cytochemically and then quantitated. The spleen cell suspension was incubated for 24 h in Leighton tubes to obtain an enriched glass-adherent population of macrophages for characterization and [3H]thymidine-labeling studies. Almost all of the adhering macrophages were esterase positive, had Fc and C3b receptors, and ingested EIgG and opsonized bacteria. In vitro labeling with [3H]thymidine showed that approximately 5% of the mononuclear phagocytes in the spleen synthesize DNA and must be considered to be dividing cells. The course of the number of labeled monocytes and macrophages after a single injection of [3H]thymidine indicates migration of monocytes into the spleen, where they become macrophages. Calculation of the influx of monocytes into the spleen and of the local production of macrophages by DNA-synthesizing mononuclear phagocytes showed that under steady-state conditions, 55% of the population of spleen macrophages is supplied by monocyte influx and 45% by local production. This means that there is a dual origin of spleen macrophages. The mean turnover time calculated with the value for the efflux of spleen macrophages is 6.0 d.

Absence of platelet membrane glycoproteins IIb/IIIa from monocytes.

Two-dimensional gel electrophoresis, immunoprecipitation, and crossed immunoelectrophoresis were used in the investigation of glycoproteins IIb/IIIa in platelets, monocytes, and monocyte-derived macrophages from human blood. All techniques detected the glycoproteins in platelets but not in the mononuclear phagocytes. Similar results were obtained by immunochemistry using a monoclonal antibody against the platelet glycoproteins IIb/IIIa (revealed by a gold-labeled second antibody) which bound heavily to the platelet but not to the monocyte surface. The biochemical techniques used for the analysis of mononuclear phagocytes would have reliably detected the level of glycoproteins IIb/IIIa contributed by a 5% contamination with platelets, calculated on a per cell basis. We conclude that human monocytes and monocyte-derived macrophages lack glycoproteins IIb/IIIa. Our results further indicate that centrifugal elutriation yields monocyte preparations with minimal contamination by platelets. It seems likely that the positive results obtained by other authors were due to the presence of platelets or fragments on the monocytes.

Immunoinhibition of intracellular protein digestion in macrophages.

Specific anti-(rabbit cathepsin D) serum, previously shown to inhibit cathepsin D, arrested the intracellular digestion of sheep IgG and radiochemically labeled hemoglobin and proteoglycan in rabbit alveolar macrophages. In the presence of antiserum, cells remained viable, but became very vacuolated. Both sheep IgG and hemoglobin were demonstrated immunocytochemically in vacuoles most of which could also be shown to contain cathepsin D. When the antiserum was removed, cells regained their normal morphology, and digestion of endocytosed proteins returned to normal. These results indicate that cathepsin D can be inhibited within lysosomes of viable cells, in which it plays a major role in the intracellular digestion of certain proteins.

Downregulation of tumor necrosis factor receptors on macrophages and endothelial cells by microtubule depolymerizing agents.

Exposure of murine and human macrophages and human umbilical vein endothelial cells to micromolar concentrations of five microtubule (MT)-depolymerizing agents (colchicine, nocodazole, podophyllotoxin, vincristine, and vinblastine) resulted in a loss of binding sites for iodinated TNF-alpha. The reduction amounted to 40-60% by 1 h and approximately 75% by 2-4 h. In 1 h, specific binding was reduced 50% by 0.1-5 microM of these drugs at 37 degrees C, but not at 4 degrees C. Inactive isomers of colchicine were ineffective, as were microfilament-destabilizing cytochalasins. The active agents did not compete with TNF-alpha R for binding. Antiserum against TNF-alpha did not neutralize the effect of colchicine and nocodazole. PGE1 and dibutyryl-cAMP could not mimic, and cyclooxygenase inhibitors could not prevent the drug effects. All the binding sites were regenerated within 3 h after removal of nocodazole, which binds tubulin reversibly, whereas little recovery was found even 18 h after the removal of colchicine, which binds tubulin irreversibly. These findings suggested that MT disassembly was responsible for the observed downregulation of TNF-alpha R. The protein synthesis inhibitor cycloheximide inhibited binding of TNF-alpha to a similar extent and with a similar time course as colchicine in the absence of added ligand. Neither drug affected binding of IFN-gamma to macrophages, nor binding of TNF-alpha to human polymorphonuclear leukocytes. Thus, an intact MT network appears to be important in maintenance of the steady state of TNF-alpha R on those cells in which TNF-alpha R turns over rapidly in the absence of ligand. The antiinflammatory actions of MT-depolymerizing agents may result in part from their interference with the ability of such cells to respond to TNF-alpha.

A novel, NH2-terminal sequence-characterized human monokine possessing neutrophil chemotactic, skin-reactive, and granulocytosis-promoting activity.

A factor able to induce an early local inflammation in rabbit skin was detected in the supernatant of mitogen-stimulated human blood leukocytes. The factor was different from IL-1 which, although present in the supernatants, was chemically separable from the factor and induced a late rather than an early skin response. Other biological effects of the principal factor were its in vitro chemotactic effects on granulocytes and its ability to induce rapid granulocytosis upon intravenous injection in rabbits. When tested under the same conditions, IL-1 beta did not act chemotactically and induced granulocytosis at a later time. The factor was purified to homogeneity and identified by electrophoretic mobility as a protein of Mr 6,500. Amino acid sequence analysis revealed the presence of an uncontaminated NH2-terminal sequence identical to a segment of the sequence previously predicted from the cDNA clone (3-10C) copied from an mRNA isolated from human leukocytes and coding for a protein of unknown function. The NH2-terminal sequence of the factor also showed extensive homology to that of the platelet factors beta-thromboglobulin (beta TG) and platelet factor 4 (PF-4). Studies done to identify the cell source of the factor revealed that it was produced by adherent mononuclear cells but not by platelets, while the opposite was true for beta TG.

Identification of macrophage external membrane proteins and their possible role in cell adhesion.

Starch-activated mouse peritoneal macrophages (STpMAC) plated on plastic demonstrate the adhesive properties typical for activated pMAC: attaching as round cells and, within 15 min, spreading out with marginal membrane ruffles. These attached STpMAC were labeled by lactoperoxidase-catalysed 125I surface iodination, sodium dodecyl-sulfate-lysed, and the lysates electrophoresed on polyacrylamide gels which were examined by autoradiography. The STpMAC morphological phenotype correlates with the labeling of a particular protein (195,000, estimated mol wt). Normal pMAC (NpMAC), from unstimulated mice, do not spread and do not display the 195,000 band. Both pMAC band patterns, including the 195,000 band, are relatively resistant to trypsin digestion, as is pMAC adhesion itself trypsin-resistant. Neither class of pMAC exhibits fibronectin (Cell Adhesion Factor, LETS protein) which is a component in the adhesive matrix of cells forming trypsin-sensitive monolayers. When pMAC are tested against antifibronectin antibody, these cells do not give immunofluorescent staining. In summary, two functions in pMAC adhesion, enzyme resistance and the ability to spread, appear related to molecular properties distinctive for pMAC surface protein.

Identification of phagocytosis-associated surface proteins of macrophages by two-dimensional gel electrophoresis.

Two-dimensional PAGE (P. Z. O'Farrell, H. M. Goodman, and P. H. O'Farrell. 1977. Cell. 12:1133-1142) has been employed to assess the effects of antibody-dependent phagocytosis on the cell surface protein composition of RAW264 macrophages. Unilamellar phospholipid vesicles containing 1% dinitrophenyl-aminocaproyl-phosphatidylethanolamine (DNP-cap-PE) were used as the target particle. Macrophages were exposed to anti-DNP antibody alone, vesicles alone, or vesicles in the presence of antibody for 1 h at 37 degrees C. Cell surface proteins were then labeled by lactoperoxidase-catalyzed radioiodination at 4 degrees C. After detergent solubilization, membrane proteins were analyzed by two-dimensional gel electrophoresis. The resulting pattern of spots was compared to that of standard proteins. We have identified several surface proteins, not apparently associated with the phagocytic process, which are present either in a multichain structure or in several discretely charged forms. After phagocytosis, we have observed the appearance of two proteins of 45 and 50 kdaltons in nonreducing gels. In addition, we have noted the disappearance of a 140-kdalton protein in gels run under reducing conditions. These alterations would not be detected in the conventional one-dimensional gel electrophoresis. This evidence shows that phagocytosis leads to a modification of cell surface protein composition. Our results support the concept of specific enrichment and depletion of membrane components during antibody-dependent phagocytosis.

Ferritin particles in macrophages and in associated mast cells.

In a variety of tissues (lymph node and glandular stroma), mast cells have been found in close and often intimate association with macrophages containing numerous ferritin-like particles in their cytoplasm and within cytoplasmic vacuoles (siderosomes). Phagocytic vacuoles in a given macrophage differed markedly. Some contained abundant Prussian blue-reactive material and others contained periodic acid-Schiff reactive substance at the light microscope level, and ultrastructurally some were filled with ferritin particles and others were not. Ferritin-like particles have also been observed occasionally in the mast cells associated with macrophages and even within the matrix of some of the granules in these mast cells.

The architecture of actin filaments and the ultrastructural location of actin-binding protein in the periphery of lung macrophages.

A highly branched filament network is the principal structure in the periphery of detergent-extracted cytoskeletons of macrophages that have been spread on a surface and either freeze or critical point dried, and then rotary shadowed with platinum-carbon. This array of filaments completely fills lamellae extended from the cell and bifurcates to form 0.2-0.5 micron thick layers on the top and bottom of the cell body. Reaction of the macrophage cytoskeletons with anti-actin IgG and with anti-IgG bound to colloidal gold produces dense staining of these filaments, and incubation with myosin subfragment 1 uniformly decorates these filaments, identifying them as actin. 45% of the total cellular actin and approximately 70% of actin-binding protein remains in the detergent-insoluble cell residue. The soluble actin is not filamentous as determined by sedimentation analysis, the DNAase I inhibition assay, and electron microscopy, indicating that the cytoskeleton is not fragmented by detergent extraction. The spacing between the ramifications of the actin network is 94 +/- 47 nm and 118 +/- 72 nm in cytoskeletons prepared for electron microscopy by freeze drying and critical point drying, respectively. Free filament ends are rare, except for a few which project upward from the body of the network or which extend down to the substrate. Filaments of the network intersect predominantly at right angles to form either T-shaped and X-shaped overlaps having striking perpendicularity or else Y-shaped intersections composed of filaments intersecting at 120-130 degrees angles. The actin filament concentration in the lamellae is high, with an average value of 12.5 mg/ml. The concentration was much more uniform in freeze-dried preparations than in critical point-dried specimens, indicating that there is less collapse associated with the freezing technique. The orthogonal actin network of the macrophage cortical cytoplasm resembles actin gels made with actin-binding protein. Reaction of cell cytoskeletons and of an actin gel made with actin-binding protein with anti-actin-binding protein IgG and anti-IgG-coated gold beads resulted in the deposition of clusters of gold at points where filaments intersect and at the ends of filaments that may have been in contact with the membrane before its removal with detergent. In the actin gel made with actin-binding protein, 75% of actin-fiber intersections labeled, and the filament spacing between intersections is consistent with that predicted on theoretical grounds if each added actin-binding protein molecule cross-links two filaments to form an intersection in the gel.(ABSTRACT TRUNCATED AT 400 WORDS)

The membrane proteins of the vacuolar system I. Analysis of a novel method of intralysosomal iodination.

A method has been developed to deliver an idoinating system into the confines of the phagolysosome, allowing us to study the nature of the phagolysosomal membrane. Lactoperoxidase (LPO) is covalently coupled to carboxylated latex spheres (LPO-latex) in a stable, enzymatically active form. The addition of LPO-latex to cultured macrophages leads to their rapid attachment, ingestion, and enclosure in a plasma membrane-derived phagocytic vacuole. These organelles rapidly fuse with preexisting lysosomes and are converted to phagolysosomes (PL) that demonstrates both acid phosphatase and lactoperoxidase activities. The exposure of LPO-latex containing cells to 125I- and an extracellular peroxide-generating system, glucose oxidase-glucose, at 4 degrees C leads to incorporation of label into TCA-precipitable material. The incorporated cel-associated label was present as monoiodotyrosine, and negligible amounts were found in lipids. Cell viability remained > 99%. Autoradiography at both the light and EM level revealed that > 97% of the cells were labeled, and quantitative analysis demonstrated the localization of grains to LPO-latex containing PL. PL were separated on sucrose gradients, and their radiolabel was confined almost exclusively to the membrane rather than soluble contents. SDS-polyacrylamide gel electrophoretic analysis of the peptides iodinated from within PL demonstrated at least 24 species with molecular weights ranging from 12,000 to 250,000. A very similar group of proteins was identified on the plasma membrane (PM) after surface iodination, and on latex phagosomes derived from iodinated PM. No novel proteins were detected in PL, either immediately after phagosome-lysosome fusion or after 1 h of intracytoplasmic residence. We conclude that the membrane proteins accessible to LPO-catalyzed iodination on the luminal surface of the PL and on the external face of the PM are similar, if not identical.

Distribution of actin-binding protein and myosin in macrophages during spreading and phagocytosis.

Actin-binding protein (ABP) and myosin are proteins that influence the rigidity and movement, respectively, of actin filaments in vitro. We examined the distribution of ABP and myosin molecules in acetone-fixed rabbit lung macrophages by means of immunofluorescence. The staining for both of these proteins in unspread cells was quite uniform, but was reduced in the nucleus and concentrated slightly in the periphery. The peripheral accumulation of staining attenuated in uniformly spread cells, although filopodia and hyaline veils definitely stained. In cells fixed during ingestion of yeast particles, the brightest staining correlated with the disposition of organelle-excluding pseudopodia initially surrounding the yeast. After phagocytosis was complete and the yeasts resided in intracellular vacuoles, no concentration of staining around the ingested yeasts was detectable. We conclude that ABP and myosin molecules are components of the structural unit of the cell responsible for spreading and phagocytosis, the hyaline cortex, a region known to be rich in actin filaments. The findings are consistent with the theory that these molecules control the rigidity and movement of filaments in the periphery of the living macrophage.

Appearance and distribution of ferritin in mouse peritoneal macrophages in vitro after uptake of heterologous erythrocytes.

Mouse peritoneal macrophages have been studied in vitro after ingestion of treated rat, rabbit, or sheep erythrocytes. Under light microscopy, phagocytic vacuoles persist up to 24 h. Macrophages lose benzidine reactivity about 5 h after red cell ingestion, and they become prussian blue positive at 2 days. Ultrastructural studies show little or no ferritin in control macrophages not fed erythrocytes. In contrast, after red cell ingestion, ferritin is widely distributed in the cytoplasmic matrix and in some cytoplasmic granules by 48 h. The Golgi complex, pinocytic vacuoles, endoplasmic reticulum, nuclei, and mitochondria do not contain ferritin. Between 2 and 4 days, ferritin in cytoplasmic granules increases, concomitant with decrease in the ferritin in the cytoplasmic matrix. Evidence is presented suggesting that ferritin in the cytoplasmic matrix is translocated into cytoplasmic granules by autophagy. Polyacrylamide gel studies on macrophages after uptake of red blood cells labeled with radioiron confirm that macrophages produce radiolabeled ferritin by 4 days.