Insoluble Immune Complexes Research Articles

35] Chronic arthritis in rabbits

Publisher Summary Acquisition of experimental data pertinent to the human disease is made more difficult by the rather sketchy and incomplete knowledge available of the pathophysiology of this disease. Long-term retention of antigen in avascular collagenous structures depends on its trapping in the form of insoluble immune complexes maintained in antibody excess by the high antibody concentration found in the joint cavity. Such an environment is provided in part by the active local synthesis of antibody. This feature may be of paramount importance for the retention of antigen, because both antigen and antibody are free to diffuse out of the tissue when strong chaotropic solutions or antigen in excess are used to solubilize the sequestered immune complexes, suggesting that they are retained within the collagen fiber meshwork mainly on the basis of their size and poor solubility. Quantitation of the cellular exudate within the joint cavity provides a simple and objective method of assessing the magnitude of the inflammatory response. Soon after the initial acute Arthus reaction abates, the synovial cavity may contain very small amounts of synovial fluid, precluding direct-cell quantitation.

Monoclonal antibody characterization of Plasmodium falciparum antigens in immune complexes formed when schizonts rupture in the presence of immune serum.

When Plasmodium falciparum parasites are cultured with some immune sera, merozoites are agglutinated by antibodies to form immune clusters of merozoites and prevent their invasion into erythrocytes. Within these immune clusters of merozoites, several antigens that are normally found in the soluble fraction after detergent extraction accumulate in relatively insoluble immune complexes. From mice immunized with these immune complexes, we obtained hybridomas secreting monoclonal antibodies (mAb) that react with various immune clusters of merozoites antigens, including mAb 3D5, which recognizes a 101-kDa antigen (p101) and mAb, 5E3, which recognizes a 113-kDa antigen (p113). Both mAb reacted with antigens at the surface of schizonts, in the vacuolar space, and at the surface of merozoites before their release from schizont-infected cells. Both p101 and p113 were synthesized by mature trophozoites and young schizonts. In pulse-chase experiments, p113 was processed to 100-, 70-, 55-, and 50-kDa products. Both p101 and p113 appeared in the culture medium when schizont rupture occurred in normal culture medium but were found in immune complexes when schizont rupture occurred in the presence of immune serum. Antibodies in immune complexes, when dissociated with acid and used to probe immunoblots, reacted with affinity-purified p101 and p113. Antigens such as these, which are accessible at the parasite surface and react with antibodies present in immune serum that inhibits parasite invasion, are logical candidates to study in the search for a vaccine against the erythrocytic stages of malaria.

Immune complexes bind to cultured rat glomerular mesangial cells to stimulate superoxide release. Evidence for an Fc receptor.

Contractile mesangial cells (MC) possess a number of macrophage-like characteristics, including oxygen radical generation. We suggest that under certain conditions MC may serve as immune effector cells in glomerulonephritis. Immune complex (IC) deposits are a hallmark of glomerulonephritis. Because IC elicit oxygen radicals from other cell types and because oxygen radicals can induce glomerular injury, we measured release of O2- by cultured rat MC in response to IC and, in separate experiments, the binding of IC to MC. Soluble and insoluble IC markedly stimulated dose- and time-dependent, saturable O2- release. Specific antibody (Ab) alone or mixtures of nonimmune Ab and antigen had no significant effect. IC-induced O2- release was not affected by cytochalasin B, an inhibitor of phagocytosis. Binding studies with radioiodinated IC demonstrated specific binding with an affinity of 1.56 X 10(6) M-1 and 1.02 X 10(5) receptors per cell. Both binding and O2- release required the Fc region of Ab. IC formed with F(ab')2 fragments did not bind specifically to or stimulate O2- release by MC. Cultured cells from rats depleted of bone marrow-derived phagocytes by irradiation produced amounts of O2- similar to cells from normal rats. These results provide evidence that IC affect the biology of the contractile glomerular MC in a manner that is dependent on the Fc region of Ab and suggest that MC structure and function may be altered at sites of injury.

Respiratory burst in murine peritoneal macrophages: Differential sensitivity to phorbol diesters by macrophages in different states of functional activation

Activation of macrophages either in vivo or in vitro can modulate the capacity to generate and secrete reactive oxygen intermediates including H 2O 2 and O 2 −. Thus, the cellular and biochemical components requisite for execution of the respiratory burst must be regulated during the activation process. In the present report, we have examined murine peritoneal macrophages in different stages of activation for their sensitivity to stimulants of respiratory burst known to activate protein kinase c (i.e., phorbol dibutyrate or diacylglycerol). The results demonstrated that more highly activated macrophages showed, in addition to greater magnitude of H 2O 2 or O 2 − production, a two- to fourfold greater sensitivity to these stimuli. While more active macrophages also exhibited a higher rate of H 2O 2 secretion, the time at which secretion was measured did not account for or modulate the heightened sensitivity. The increased sensitivity to stimulation was dependent upon the stage of activation and not on the agent used to elicit the macrophages. Increased sensitivity of the more active macrophage populations was also seen when physiologic stimuli (i.e., insoluble immune complexes or unopsonized zymosan) were used. These findings indicate that macrophage activation for H 2O 2 secretion modulates the sensitivity to stimulation such that more H 2O 2 is produced in a shorter time and at a lower concentration of stimulus, thereby heightening the inflammatory response in several independent ways. Because all the stimuli employed in the present study have in common the ability to activate protein kinase c (either directly or indirectly), the data also suggest that this form of macrophage activation may involve, at least in part, modulation of the stimulus-response coupling mechanisms which utilize this enzyme.

Relationship between Murine Macrophage Fc Receptor-Mediated Phagocytic Function and Competency for Activation for Non-Specific Tumor Cytotoxicity

The relationship between Fc receptor (FcR) function and activation of murine macrophage populations for non-specific tumor cytotoxicity was studied. Oil-elicited inflammatory peritoneal macrophages (PMФ) from C3HeB/FeJ mice had higher FcR function upon harvest than resident PMФ from the same strain or elicited PMФ from genetically deficient C3H/HeJ mice. C3HeB/FeJ inflammatory PMФ were uniformly responsive to activation by MAF and the complement activators: LPS, Poly I:C, cobra venom factor (CVF) and zymosan for tumoricidal activity. Resident cells from the same strain and C3H/HeJ-elicited PMФ were uniformly unresponsive to the same activators. In vitro culture of C3HeB/FeJ resident PMФ with fetal bovine serum for 24–48 h produced unregulation of FcR function which coincided with a conversion from an unresponsive to a responsive state for tumoricidal activity. Reconstitution of the FcR function of C3H/HeJ-elicited PM In vitro during 24–48 h culture with lymphokine or Poly I:C also coincided with the restoration of responsiveness to activation by LPS, CVF, and zymosan for tumor cytotoxicity. Thus, the consistent temporal relationship between upregulated FcR function and the capacity of macrophages to respond to activation for non-specific tumoricidal activity may be more than coincidental. Preincubation of responsive C3HeB/FeJ-elicited PMФ with insoluble immune complex or heat-aggregated IgG was shown to blockade FcR-mediated phagocytosis and to abrogate LPS-mediated tumoricidal activity. Interestingly, FcR blockade by IgG-opsonized sheep erythrocyte conjugates selectively inhibited activation by MAF, LPS, and Poly I:C, but had no inhibitory effect on activation by CVF or zymosan. Similar blockade of C3b receptors produced an identical pattern of selective inhibition of activation. This selective inhibition of non-specific tumoricidal activity by FcR/C3bR blockade suggests the existence of two pathways for antibodyindependent activation of macrophages.

Formation of soluble immune complexes by complement in sera of patients with various hypocomplementemic states. Difference between inhibition of immune precipitation and solubilization.

To examine whether the ability of complement to form soluble immune complexes plays a role in preventing immune complex-mediated diseases, we analyzed the capacity of complement to inhibit immune precipitation (IIP) and to solubilize preformed immune aggregates (SOL) in 23 sera of patients with various hypocomplementemic states, and we correlated the results of these studies with the clinical syndromes found in the various patients. In sera with deficiency or depletion of early classical pathway components, IIP was profoundly altered, whereas SOL was delayed but in the normal range. In contrast, in sera with C3 depletion but intact classical pathway and in properdin-deficient serum, IIP was initially preserved, whereas SOL was abolished. Since the incidence of immune complex diseases in various hypocomplementemic states correlates with the severity of IIP defects, but not with reduced SOL, it is suggested that IIP is an essential biological function of complement that prevents the rapid formation of insoluble immune complexes in vivo.

Mutual inhibition of the binding of Clq and protein A to rabbit IgG immune complexes

A complex of rabbit IgG antibody with horseradish peroxidase covalently linked to Sepharose 4B was used as an insoluble immune complex for studying the binding of complement factor Clq, protein A from Staphylococcus aureus, and its IgG-binding fragments AB and B, to rabbit IgG. It was shown that protein A (mol. wt approx. 42,000) and fragments AB and B (mol. wts approx. 14,000 and 7000, respectively) inhibited the binding of Clq to insoluble immune complex at 4°C. However, at 37°C fragment B did not inhibit this binding. On the other hand, Clq, when bound to an insoluble immune complex, almost completely blocked the binding of protein A and fragment B at both temps. The higher affinity of Clq for its C H2-binding site than of fragment B for its C H2-binding site may explain the displacement of the latter from the C H2 domain. The mutual inhibition of the binding of Clq and protein A (and its smaller fragments) indicates that the binding sites for Clq and protein A are closely located in the C H2 domain.

Gamma-Interferon enhances the secretion of arachidonic acid metabolites from murine peritoneal macrophages stimulated with phorbol diesters.

Macrophage activation in vivo has been associated with qualitative and quantitative alterations in the release and metabolism of arachidonic acid. In the present study, we examined the effect of in vitro macrophage activation with recombinant gamma-interferon (IFN-gamma) on arachidonic acid secretion induced by exposure to a variety of stimulating agents. Secretion stimulated by challenge with unopsonized zymosan, insoluble immune complexes, the calcium ionophore A23187, or combinations thereof was unaltered in IFN-gamma-treated macrophages. However, when phorbol diesters active as tumor promoters were employed as challenge agents, arachidonate secretion was enhanced as much as 10-fold over that seen in nonactivated controls. The enhanced secretory response to PMA was detectable as early as 1 hr after exposure to IFN-gamma, reached a maximum within 3 to 6 hr, and subsequently declined to control levels even in the continued presence of the agent. Treatment with IFN-gamma did not alter the pattern of individual metabolites produced by macrophages challenged with either zymosan or PMA. Finally, the sensitivity to phorbol diesters was also increased by treatment with IFN-gamma (ED50 reduced from 35 ng/ml to 4 ng/ml). Thus, IFN-gamma could prime macrophages for a substantially amplified response to phorbol esters. Because the cellular mediator of PMA action has been identified as a Ca++, phospholipid-dependent protein kinase, a role for this enzyme in macrophage functional development is indicated.

Erythrocyte-bound immune complexes trigger the release of interleukin-1 from human monocytes

We have demonstrated herein that immune complexes (IC) bound to O + human erythrocytes trigger the release of interleukin-1 (IL-1) from human monocytes in vitro. The efficiency of monocytes stimulation by erythrocyte-bound IC is in sharp contrast to the lack of stimulation of monocytes by soluble IC. The IL-1 activity triggered by erythrocyte-bound IC was lower than that induced by insoluble IC or opsonized fluorescent latex beads, but was greater than that induced by anti-D-treated O + human erythrocytes. The magnitude of IL-1 release induced by insoluble IC and erythrocyte-bound IC were similar up to 18 hr of incubation. The IL-1 activity decreased in the supernatant of monocytes exposed to erythrocyte-bound IC for 40 hr, while the activity of IL-1 induced by insoluble IC remained at a high level. The physiological and physiopathological significance of triggering IL-1 release by erythrocyte-bound IC are discussed.

Alteration of Fc-receptor phenotype and proliferative capacity of Fc-IgG-receptor positive lymphocytes through interaction with soluble immune complexes of patients with SLE or RA.

Soluble circulating immune complexes (CIC) are a common finding in patients with systemic lupus erythematosus (SLE), rheumatoid arthritis (RA) or other autoimmune diseases. The predominant immunoglobulin class of most CIC is IgG, which enables these CIC to bind to Fc-IgG-receptor expressing cells. In this study the interaction between soluble CIC from patients with SLE or RA and Fc-IgG-receptor positive lymphoid cells from healthy individuals was investigated. Similar to the effect observed with insoluble immune complexes, soluble CIC interact with Fc-IgG-receptor positive lymphoid cells and can induce a modulation of Fc-receptor expression. Fc-IgG-receptors are lost and Fc-IgM-receptors are expressed on the same cells after IC interaction and culturing the cells for 24 h. Simultaneously with this change of Fc-receptor phenotype the originally Fc-IgG-receptor positive cells demonstrate a decreased ability to proliferate upon mitogen stimulation. This change of phenotype and proliferative capacity correlates with the content of CIC in the sera of patients with SLE or RA.

Fc-receptor mediated protein phosphorylation in murine peritoneal macrophages

The effect of Fc receptor engagement on protein phosphorylation in murine peritoneal macrophages has been investigated. Treatment of macrophage cultures with insoluble immune complexes resulted in enhanced phosphorylation of six proteins at 73, 66, 53, 37, 31 and 25 kD. Comparison of the protein phosphorylation patterns induced by immune complexes with those induced by agents which mimic the actions of well known intracellular second messengers (i.e., A23187, dibutyryl cAMP, or phorbol myristate acetate) revealed substantial similarity between Fc receptor induced events and those induced in response to phorbol diesters. There were, however, two phosphorylated proteins which were only seen following stimulation with immune complexes. Thus, more than one kind of protein kinase activity appears to be involved in Fc receptor mediated stimulation of macrophage function.

The effects of cleavage of the inter-chain disulphide bonds of rabbit IgG on its ability to bind C1q

Immune complexes prepared from rabbit anti-ovalbumin IgG in which the interchain disulphide bonds had been reduced and then blocked with N-(iodoacetylaminoethyl)-8-naphthylamine-1-sulphonic acid retained the ability to bind 125I-labelled C1q. This ability was lost when a small alkylating agent (iodoacetamide) was used to block the cleaved disulphide bonds. The ability of the IgG to form insoluble immune complexes was partially compromised when iodoacetamide was used to block the disulphide bonds, but was unimpaired when N-(iodoacetylaminoethyl)-8-naphthylamine-1-sulphonic acid was used. These data are consistent with the suggestion that access to the C1q binding site in IgG in immune complexes is modulated by movement of the Fab arms, which may block access to the site.

Structural studies of Fc receptors. III. Isolation and molecular weight analysis of the component chain of Fc gamma receptor of macrophage.

Surface receptors of guinea pig peritoneal macrophages specific for the Fc region of IgG (Fc gamma receptor) were isolated and identified as a surface-radioiodinated component with a molecular weight of 44,000 that bound in an Fc-specific manner to IgG2 of guinea pig immunoglobulin immobilized in any of the following three different ways: IgG2 antibody in insoluble immune complex, IgG2 antibody bound to antigen-coupled Sepharose, and IgG2 covalently coupled to Sepharose. In order to obtain the Fc gamma receptor retaining the binding activity, the Fc-binding component was isolated by IgG2 affinity chromatography in which mild acidic buffer (pH 5.0-4.0) was chosen to elute the component bound to the affinity column. Forty-five to sixty-two percent of the eluted radioactivity was shown to rebind to the IgG2-affinity column. The bound fraction showed a single radioactive peak of 44,000 daltons in sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The Fc-binding component isolated by the affinity chromatography behaved similarly in gel filtration in the presence of a detergent, as did the detergent-solubilized Fc gamma receptor before isolation by affinity chromatography. These results suggested that the Fc gamma receptor was isolated in a native form. Furthermore, it was confirmed that the isolated Fc gamma receptor is distinct from actin or the actin-like protein (DNase I-binding protein) which had been reported to bind to IgG-affinity column.

Inhibition of macrophage tumoricidal activity by immune complexes and altered erythrocytes.

Engagement of the macrophage membrane by biologic particles including insoluble immune complexes inhibited the development of lymphokine-mediated nonspecific tumoricidal activity by murine macrophages. The degree of inhibition was dependent on the dose of particles and the lymphokine concentration. Inhibition was not due to macrophage cell death or to diminution of cell adherence after ingestion of the immune complexes. Soluble immune complexes were not inhibitory, although approximately 10% of the complexes became cell-associated. Monomeric or heat-aggregated IgG was also not inhibitory. IgG-opsonized erythrocytes (EA) were inhibitory and inhibition was dependent on the degree of opsonization. In contrast, nonopsonized erythrocytes (E), which did not bind to macrophages, were not inhibitory. Phagocytosis of glutaraldehyde-treated E or E carrying IgM antibody and complement (EAC) also led to a reduction of tumorilytic activity. Insoluble immune complexes were inhibitory when added either before or after lymphokine. Phagocytosis was neither sufficient nor necessary to cause inhibition because 1) ingestion of polystyrene latex beads did not diminish tumoricidal activity, and 2) macrophages plated on IgG-coated surfaces were inhibited with respect to the tumoricidal function. Inhibition was not affected when indomethacin (10(-6) M) was included in the assay, which indicated that prostaglandins were not involved in the process. Thus, macrophage tumoricidal responsiveness may be compromised by interaction of biologic substances with macrophage plasma membranes. This process may thereby inactivate an important host defense mechanism against neoplastic cells.

Immune-complexes solubilization: Effect of antigen-antibody ratio and relative role of alternative and classical complement pathways

Insoluble immune-complexes (IC) are solubilized by complement. A functional alternative pathway (AP) is required for solubilization. Classical pathway (CP) has an enhancing effect, but the effect of antigen-antibody (Ag-Ab) ratio and the relative role of AP and CP at these different Ag-Ab ratios is controversial. We reinvestigated these factors in a human model (IC prepared with tetanus toxoid and affinity purified human Ab solubilized by normal serum—NHS). We present evidence that (1) IC prepared at Ag excess are only partially solubilized. (2) IC prepared at large Ag excess are not solubilized by AP. (3) AP prepared at Ab excess are solubilized exclusively by AP. (4) IC prepared at equivalence are solubilized by AP and CP acting in synergy.

Requirements for the binding of human plasma fibronectin to the C1q subunit of the first component of complement.

We have shown previously that [125I]fibronectin (Fn) binds to solid-phase C1q in a dose-dependent manner. When C1r and C1s were added, the binding of Fn to C1q was abolished; removal of C1r and C1s restored Fn binding to C1q. In this report, we have systematically examined the optimal conditions that favor the Fn-C1q interaction. Our studies show that purified native 125I-Fn binds to C1q in a specific, saturable manner. Maximal 125I-Fn binding to C1q and gelatin occurs at low ionic strength (mu = 0.05) and drops sharply as the ionic strength is increased. At mu = 0.20, the binding to C1q is inhibited by 95%, whereas the binding to gelatin is inhibited by 50%. Optimal binding of Fn to C1q and gelatin occurs between pH 5.5 and 7.5, is decreased by 45% at 4 degrees C, and increases with incubation time; saturation of binding occurs in 60 min at 37 degrees C. Scatchard analysis of binding at mu = 0.05 indicates a single class of high affinity binding sites (Kd = 3.7 X 10(-8) M +/- 0.36 SD). The Kd of the reaction when C1q is bound to either plastic or immune complexes is essentially the same, and, in both cases, increases as the ionic strength of the medium is increased. Finally, significant binding of Fn to C1q can be demonstrated at physiologic ionic strength employing either insoluble immune complexes containing C1q or chemically cross-linked C1q.

Relationship between epitope density and immunoprecipitation of multivalent antigens by bivalent antibody: Immunoprecipitation of adenylylated glutamine synthetase by anti-AMP antibodies

Escherichia coli glutamine synthetase (GS) preparations composed of 12 adenylylated subunits (GS 12 −) are almost completely precipitated by sheep Anti-AMP immunoglobulin G (IgG), whereas glutamine synthetase preparations containing 6 adenylylated subunits (GS 6 −) are only partially precipitated by the antibodies ( R.J. Hohman, S.G. Rhee, and E.R. Stadtman, 1980, Proc. Nat. Acad. Sci. USA 77, 7410–7414). By means of 125I-labeled anti-AMP antibodies and double immunoprecipitation techniques, in which rabbit antiserum to sheep IgG or anti-GS antibodies were used to precipitate soluble immune complexes, it was demonstrated that under optimal conditions, both the soluble and insoluble immune complexes obtained with either GS 6 − or GS 12 − contain 0.5 mol antibody/mol adenylylated subunit. In agreement with the lattice theory of immuno-precipitation, soluble immune complexes are formed in antibody excess. Scatchard plots of binding data indicate that under conditions of antibody excess, one antibody molecule is bound to each AMP moiety of GS 12 −, whereas GS 6 − binds a maximum of only 0.68 antibody molecule/adenylylated subunit. We propose that with some species of GS 6 −, the distribution of adenylylated subunits favors monogamous interactions of the bivalent antibody with two subunits within the same GS molecule and thereby leads to the formation of small, soluble, immune complexes. Other explanations are considered. Only 30% of the antibody population that recognizes unconjugated 5′-AMP binds to the AMP moiety of adenylylated GS. Anti-AMP antiserum can be fractionated on a GS 12 −-Sepharose matrix into two subpopulations of antibody with strikingly different immunoprecipitation characteristics. Conversely, species of GS with various states of adenylylation ranging from 0 to 8 were separated from a GS 6 − preparation by means of affinity chromatography on an anti-AMP antibody-Sepharose matrix. Under optimal conditions, antibodies purified by affinity chromatography precipitated a smaller fraction of a GS 6 − preparation than did unfractionated antiserum. Competence of the purified antibody was nearly restored to that of the unfractionated serum by the addition of an enhancement factor present in the IgG fraction of nonimmune serum. The enhancement factor was not required for complete precipitation of GS − 12 by purified antibodies. Contrary to most antibody-antigen reactions, immunoprecipitation of GS 6 − with anti-AMP antibodies is greater at 30 °C than at 4 °C.

Complement-mediated fragmentation of soluble and insoluble immune complexes containing porcine anti-DNP antibodies

Complement-mediated release of soluble immune complexes and immune precipitates containing DNP-PSA and precipitating or non-precipitating porcine anti-DNP antibodies was studied. A decrease in the average size of soluble immune complexes indicating their fragmentation was observed during incubation in excess human serum, the extent of the complex release was found to be in direct proportion to the time of incubation. The effect was complement-dependent. In the second part of the study, complement-dependent solubilization of the immune precipitates of the precipitating antibody preparation was compared to the solubilization of the precipitates of the non-precipitating antibody formed in the presence of PEG. Although, both types of precipitates activated complement in about the same extent, complexes of non-precipitating antibody were solubilized much slower than those of the precipitating one. As avidity of both antibody preparations to the antigen was high, the observed differences in the rate of the complex solubilization probably reflected differences in the structure of the two types of complexes.

The Mechanisms which Produce Vacuolated and Degranulated Eosinophils

Summary. As blood eosinophils in patients with an eosinophilia may contain refractile‘vacuoles’(vacuolated eosinophils), and some cells contain only a few granules (degranulated eosinophils) which show ultrastructural changes, experiments were done to see whether similar morphological alterations could be induced in blood eosinophils from normal people. Overnight culture of normal eosinophils with serum treated with cobra venom factor induced refractile‘vacuoles’and ultrastructural changes in granule matrix material. On the other hand, concanavalin A, 10 μ/ml for 2 h, led to the formation of larger perinuclear vacuoles containing granule contents, and ultrastructural changes were more marked in granule crystalloid material. Cytochalasin B inhibited these effects of concanavalin A. Neither of these stimuli caused a marked reduction in the number of cytoplasmic granules, but lysolecithin and insoluble (but not soluble) immune complexes produced large endocytic vacuoles into which partially solubilized and whole eosinophil granules were discharged.These results suggest that eosinophil granules can be affected in two distinct ways. In one, the granule matrix is altered leading to the light microscopy appearances of small refractile ‘vacuoles’. In the second, the crystalloid material is dispersed into the matrix. These apparently distinct steps precede the release of the granule contents into large cytoplasmic spaces or to the outside of the cell. Also, as patients' eosinophils may show one or both of these alterations it is suggested that they are distinct events. As concanavalin A induced solubilization of granules did not result in release to the outside, it is also proposed that an additional stimulus may be required to produce exocytosis of eosinophil granule contents. Further studies may show how these processes are linked, and their relative importance in eosinophil effector functions.

Fibroblasts and immune complexes. 1. Phagocytosis of in vitro-prepared insoluble immune complexes by cultured human fibroblasts.

ABSTRACTCultured human fibroblasts were incubated with in vitro‐prepared insoluble immune complexes consisting of horseradish‐peroxidase (HRP)‐anti HRP for 4, 24 or 48 hours. These incubated fibroblasts were then processed for electron microscopy. Peroxidase‐positive immune complexes were observed both on the surfaces and within the cytoplasms of the fibroblasts. However, the frequency of interiorized immune complexes or phagosomes was not high. It was speculated that, in connective tissue, fibroblasts can also adhere to and trap immune complexes in vivo, thus eliminating them from tissues under certain conditions.