Antigen in tissues: V. Effect of endotoxin on the fate of, and on the immune response to, serum albumin and to albumin—antibody complexes*

Bacterial endotoxin was injected into rat hind footpads together with bacterial flagellin and ¹²⁵I-labelled human serum albumin (HSA); the latter was used unmodified or heat denatured (H.HSA) or as an HSA—antibody complex. Endotoxin did not affect the trapping, retention nor localization of the labelled HSA in the popliteal and aortic lymph nodes, whether the antigen had been injected as HSA, H.HSA or as an HSA—antibody complex.

If endotoxin was injected at the same time as: (1) flagellin, there was an increased production of anti-flagellin antibody; and (2) H.HSA or HSA—antibody complex, detectable amounts of anti-HSA antibody were produced. When H.HSA and endotoxin were injected, the primary response was long lived yet the period of induction of antibody formation and of antigen persistence in the lymphoid tissues was short. If, during the primary antibody response to H.HSA, the animals were challenged with HSA, equally strong secondary antibody responses occurred with an HSA—antibody complex or with HSA alone.

The results were interpreted in terms of the tissue localization pattern of H.HSA (medullary macrophage) and HSA—antibody complex (medulla and lymphoid follicles). It was suggested that: (1) induction of antibody formation and priming of cells for a secondary antibody response might occur following localization of the antigen in the medulla and that antigen localization in the lymphoid follicles might not be a strict requirement for this; and (2) the follicular localization of antigen might be the preferential mechanism for the firing of a secondary antibody response.

     

Induction of thyroiditis in guinea-pigs by intravenous injection of rabbit anti-guinea-pig thyroglobulin serum: II. Studies with fluorescent antibody technique

The localization of rabbit γ-globulin was traced in the thyroids of guinea-pigs receiving rabbit anti-guinea-pig thyroglobulin serum intravenously. The fluorescent antibody technique was used. Rabbit γ-globulin was found in the interstitium and pericapsular region ¼ hour after the injection, and could be seen there for 24 hours. Five, 10 and 20 days after injection the interstitium was free of specific green-fluorescence. The first sign of specific fluorescence, associated with the epithelial cells, was seen as rounded droplets of fluorescent material after 1 hour. Thereafter, these droplets could be found, increasing in size, at all time intervals later studied. Twenty-four hours after injection green-fluorescent material was seen in some of the follicular lumina. However, most of these did not display any specific fluorescence. Granulocytes showing specific fluorescence were first observed 1 hour after injection, but this phenomenon was most evident at 12 and 24 hours. Both eosinophil and neutrophil granulocytes were found in the infiltrates, but it could not be decided whether both of these cell types were involved in the uptake of rabbit γ-globulin. The findings were discussed in relation to the previous reported observation of strongly PAS-positive material in the granulocytes. Together the observations suggested the uptake of both antigen and antibody, most probably as an antigen—antibody complex. The relevance of the observations in relation to thyroid physiology, as well as to experimental thyroiditis after active immunization and to chronic nonspecific human thyroiditis is discussed.     

The passive transfer of delayed hypersensitivity in the guinea-pig: II. The ability of passively transferred antibody to cause local inflammation and retention of antigen and the role of these phenomena in the passive transfer of delayed hypersensitivity

It is known that serum increases the passive transfer of delayed hypersensitivity to certain antigens, such as bovine γ-globulin, in the guinea-pig. This synergic effect of serum in the passive transfer of delayed hypersensitivity to bovine γ-globulin was partly, but not completely, produced by serum to haemocyanin when a mixture of bovine γ-globulin and haemocyanin was used for skin testing. It was concluded that part of the synergic action of serum was due to a local inflammatory reaction.

The ability of serum to cause local retention of antigen in the skin was studied using antigens labelled with radioactive iodine. Immune serum favoured the local retention of antigen. The passive transfer of antiserum to bovine γ-globulin, egg albumin and haemocyanin specifically increased the retention of antigen two- to twelve-fold. This ability of serum to cause the local retention of antigen at the site of intradermal injection was present in serum taken 3 weeks after immunization with bovine γ-globulin in Freund's complete adjuvant but absent in serum taken at 1 week. Antiserum also altered the distribution of antigen at the skin site. Autoradiography showed that it increased the area over which an appreciable concentration of antigen occurred.

Active immunization with bovine γ-globulin had a slight effect on the total amount of antigen retained in the skin after intradermal injection. It had a greater effect on the distribution of antigen. In control guinea-pigs 87 per cent of the total amount of bovine γ-globulin retained at 20 hours was found within a radius of 6.5 mm of the centre of injection. In contrast in guinea-pigs immunized with bovine γ-globulin in Freund's complete adjuvant 43 per cent was found beyond this radius. A similar change in the distribution of human serum albumin was seen in guinea-pigs immunized with bovine γ-globulin when the albumin was mixed with bovine γ-globulin. This indicated that factors other than the formation of immune precipitates were sometimes responsible for the local retention of antigen. The total amount of purified protein derivative of tuberculin (PPD) retained and its distribution in the skin was uninfluenced by immunization.

It was concluded that the synergic effect of immune serum on the passive transfer of delayed hypersensitivity was due in part to some aspect of the inflammation caused by antibody antigen reaction and in part to the local retention of antigen caused by antibody.

     

Antibody isotypes mediating antigen retention in passively immunized mice.

Antibody isotypes vary in their capacity to mediate retention of a readily catabolized protein antigen, human serum albumin (HSA) in spleen, popliteal lymph node (PLN) and hind foot. Hyperimmune anti-HSA mouse sera were separated into fractions highly enriched for IgM, IgG1 and IgG2 via differential elution from protein A-Sepharose. These fractions were used to immunize normal mice passively. Twenty-four hours later the mice were injected with radio-iodinated HSA into the hind footpad. When the amount of HSA retained in the spleen 6 days later was determined, the potency of various antibody fractions to mediate retention could be ranked IgG2=IgG1 > IgM. The amount of HSA retention mediated by various fractions correlated well with autoradiographic evidence demonstrating localization of HSA in splenic follicles. The localization pattern in PLN was similar to the spleen except that the IgM-containing fraction mediated follicular localization of HSA to a considerable degree. In tendons of the hind foot, IgG1 mediated HSA retention five times better than IgG2 or IgM fractions. The amount of radioactivity found in the liver varied inversely with HSA retention in other locations. The results demonstrate differences in antibody isotype requirements for antigen localization in spleen, regional lymph node and collagenous sites of the hind foot.     

The localization of non-microbial antigens in the draining lymph nodes of tolerant, normal and primed rabbits

Soluble haemocyanin (HCy) or human serum albumin (HSA) labelled with ¹²⁵I at 8 and 0·7 μc/μg, respectively, were injected into the footpads of rabbits in doses just sufficient to elicit a primary response in normal animals, and the distribution of radioactivity in the popliteal lymph nodes between 6 hours and 21 days later was studied by autoradiography. The recipient rabbits had either been primed by a single prior injection of unlabelled antigen, or made putatively tolerant by repeated neonatal administration of antigen, or were previously untreated. Localization of antigen in germinal centres, in a typical dendritic pattern, was marked in the primed animals throughout the period of observation; in those tolerant animals which had no detectable serum antibody initially and made no detectable antibody response such localization was not seen at any time; in the animals that had no previous contact with antigen there was no localization in germinal centres until about the time when antibody became detectable in the serum. Localization of radioactivity in medullary sinus macrophages did not differ significantly between the three groups.

It is concluded that localization of these soluble antigens on the dendritic web in lymphoid follicles occurs as a consequence of the presence of circulating antibody. Uptake of the antigens by medullary macrophages, however, can occur in the absence of antibody. Although the degree of labelling of medullary macrophages was not evidently affected by the presence of antibody in these experiments, it is emphasized that the antibody levels, even in the primed animals, were low, and that this finding is unlikely to apply when the amount of antibody present is relatively much greater than the amount of antigen injected.

     

Influence of splenectomy in rats on the formation of 19S and 7S antibodies

Adult rats were immunized with human γ-globulin, in heat aggregated form intravenously, and with complete adjuvant intravenously or in the footpads. In all animals an intravenous booster injection was given on day 25. Splenectomy or sham splenectomy was performed 7 days prior to or 3 and 21 days after immunization, and 19S and 7S antibody activities were determined separately in the sera of these animals in the primary and the secondary response. Splenectomized young rats immunized with human γ-globulin in complete adjuvant intravenously or in the footpads were similarly studied.

The formation of 19S and 7S antibody was affected separately depending upon the route of antigen administration and time of splenectomy. Furthermore, the inhibitory effect of splenectomy on antibody formation was more pronounced in young than in adult rats.

     

The mechanism of anaphylaxis: I. Production and biological properties of `mast cell sensitizing' antibody

In addition to the usual serum precipitating antibody, immunized rats produce a thermolabile antibody that is able to sensitize mast cells and cause damage to these cells on reacting with the specific antigen. Due to this property this antibody is called `mast cell sensitizing' antibody (MCSAb). Like human reagins, it is destroyed by heating at 56° and when injected into homologous skin will remain at the site of injection for as long as 1 month. In contrast, the antibody disappears very quickly from serum when injected into recipient rats. Although its production can be induced in various ways, the administration of Bordetella pertussis organisms along with the antigen is particularly effective. MCSAb seems to be the product of a primary response, since its production is prevented or very much reduced by previous or additional contacts with antigen.

The mechanism of passive sensitization of mast cells with MCSAb was studied by incubating pieces of rat mesentery with antigen in vitro, using antigen-induced mast cell damage or histamine release as a measure of sensitization. Sensitization increases with time of contact between cells and antibody and can attain a maximum within the first few minutes depending upon antibody concentration. Sensitizations at 37° and 4° proceed at about the same speed. Nonspecific γ-globulin inhibits sensitization with MCSAb when used in large amounts. Rabbit, rat, human and bovine γ-globulin slowed down sensitization, whereas horse γ-globulin was without effect. Sensitization with MCSAb can be completely reversed by washing the sensitized mesentery in Tyrode's solution. The time required to remove MCSAb by washing was inversely proportional to the time of contact between cells and antiserum during sensitization. Anaphylactic phenomena induced with MCSAb seem to be primarily due to histamine and 5-hydroxytryptamine release. The possibility of a relationship between the methods of immunization leading to the production of mast cell sensitizing antibody and the establishment of a delayed hypersensitivity is considered.

     

Antigen localization in lymphopenic states: I. Localization pattern following chronic thoracic duct drainage

Adult rats, depleted of thoracic duct lymph for 5–7 days, were tested for their ability to localize ¹²⁵I-labelled polymerized flagellin from Salmonella adelaide. Labelled antigen was injected into both hind footpads 6–12 hours after completion of drainage, and the regional nodes were excised 24 hours later. Grain counts on identically exposed autoradiographic sections from regional nodes were used to assess differences in antigen distribution between depleted and nondepleted rats. The uptake of antigen by medullary macrophages was no different in the two groups. However, the uptake of antigen by primary lymphoid follicles was reduced by thoracic duct drainage to levels one-fourth that observed in normal rats.

Two procedures were found capable of improving follicular antigen uptake in the chronically depleted rat: (1) regional inoculations of 0.01 ml of specific antiflagellar immune serum at a titre of 1:400 1 hour prior to antigen injection, and (2) daily return by intravenous infusion of washed autogenous thoracic duct lymphocytes collected during drainage. Regional injections of both viable and non-viable lymphocytes were ineffective in improving follicular antigen uptake in the depleted animal.

The results show that depleted rats lack a serum factor, presumably an opsonin, important in determining antigen distribution patterns. It seems likely that this factor is normally manufactured by small lymphocytes.

     

Antigen in tissues: II. State of antigen in lymph node of rats given isotopically-labelled flagellin, haemocyanin or serum albumin

1. Three antigens, flagellin from Salmonella adelaide, haemocyanin from Maia squinado blood, and human serum albumin (HSA), each labelled with radioactive iodide, have been injected singly into rats to give a primary or a secondary response. After hind footpad injection, the popliteal and aortic lymph nodes were removed and fractionated in sucrose medium into a nuclear debris fraction, a large granule fraction and a cell supernatant fraction. The large granule fraction was divided into an extract and residue after nine cycles of freezing and thawing.

2. At intervals after injection of antigen, determinations were made of the radioactivity in the node which was: (a) bound to sub-cellular components: (b) macromolecular but not bound to sub-cellular components, or (c) associated with low molecular weight components. Except as mentioned below, the patterns of behaviour of antigen injected into normal or primed rats were essentially similar. Notable features were: flagellin to a large extent became bound to sub-cellular components; after HSA injection, most radioactivity was in a macromolecular but unbound form; whereas nodes from haemocyanin injected rats had the highest proportion of label which was attached to low molecular weight components. Injection of the same dose of HSA with rat anti-HSA serum had a profound effect; more than 100 times more label was recovered in the large granule fraction and the overall distribution of isotope resembled that exemplified by flagellin.

3. With each antigen the time at which the highest proportion of total radioactivity in the node was found associated with low molecular weight component was 2–3 days after injection. In each case, the large granule extract showed the highest proportion of radioactivity in this state.

4. The extent to which the isotope remained associated with specific antigen in the tissues was studied by testing the ability of any fraction to react with specific immune serum to the antigen. With flagellin a large proportion (up to 80 per cent) of radioactive substances associated with the large granule residue reacted with antiserum. In the case of haemocyanin, a very small proportion only of the radioactivity in this fraction reacted with antiserum but this proportion increased as the serum antibody titre rose. The greatest contrast in the behaviour of the fraction occurred with HSA. After injection of HSA alone there was very little specific reaction (about 12 per cent) shown by this fraction. After injection of HSA with antiserum, the proportion reacting was five-fold higher. This increased ability of the radioactive fraction in the case of HSA to react specifically with antiserum could be correlated with the increased localization of the antigen in the lymphoid follicles of the node, as revealed by radioautography.

     

Prolonged antigen half-life in the lymphoid follicles of specifically immunized mice

The kinetics of clearance of ¹²⁵I from the popliteal lymph nodes and feet of human serum albumin (HSA)-immunized mice was studied following the injection of [¹²⁵I]-HSA into the hind footpads. Antigen was cleared from both locations rapidly for the first few days. The antigen half-life (T½) during this period was only a matter of hours. By the end of the first week, however, the rate of clearance in both sites had changed markedly. The antigen T½ in the node between the first and sixth week was 8.1 weeks (95% confidence interval between 5.1 and 20 weeks) and the antigen T½ in the foot was 6.1 weeks (95% confidence interval between 3.7 and 16.6 weeks). There was, however, about twenty times more radioactivity in the feet than in the popliteal nodes. Autoradiography of popliteal lymph nodes revealed that initially antigen was trapped in the medulla, subcapsular sinus, superficial cortex and around lymphoid follicles. During the first few days antigen was cleared from all sites except the follicles. The radioactivity initially trapped in the medulla, subcapsular sinus, and superficial cortex appeared to have been associated with macrophages. Studies with peritoneal macrophages indicated an antigen T½ in these cells of 2 h (95% confidence interval between 1.5 and 3 h). The initial rapid clearance of antigen trapped and catabolized by macrophages and the long-term retention of antigen in the follicles is probably attributable to trapping and retention by follicular dendritic cells. The large pool of antigen trapped in the foot did not appear to serve as a depot to replace antigen degraded in the node, since amputation of the foot did not alter the level of antigen retained in the node. The long antigen T½ in the lymph node follicles indicates that antigen is available in the lymph node to play a role in the maintenance and regulation of immune responses for many months or even years.     

The localization of immunoglobulin and immune complexes in lymphoid tissue

Aggregated and monomeric forms of human γ-globulin (HGG) were prepared by heating at 63°, ultracentrifugation and subsequent separation according to solubility in 0.62 M sodium sulphate. These two forms were injected intradermally into guinea-pigs'' ears and their distribution in the draining auricular nodes determined at different times following injection by staining cryostat sections with fluorescein labelled anti-HGG. Monomeric HGG showed no precise localization; aggregated HGG localized rapidly in the phagocytic macrophages of the sinuses and medulla and after a few hours'' delay in the germinal centres in a dendritic pattern, the latter persisting for up to 4 weeks. With doses of less than 10 μg, aggregated HGG was not seen in the medulla but germinal centre staining was readily visible, possibly due to a concentrating effect. Prior injection of a large dose of monomer either locally or systemically did not alter the pattern of staining produced by subsequent injection of aggregated HGG.Aggregated human serum albumin, colloidal carbon and streptococcal cell walls did not localize in germinal centres in the same way.Monomeric rabbit IgG anti-HSA injected alone did not localize, but when combined with HSA in antigen excess to form soluble immune complexes it localized in germinal centres.It is concluded that germinal centres contain receptors, probably at cell surfaces, for IgG aggregated by mild heat or by complexing with antigen but not for unaltered native IgG, and it is suggested that this may be a means of disposal of aggregated or complexed IgG.     

Morphological and kinetic studies on antibody-producing cells in rat lymph nodes

A study of the secondary response of regional lymph nodes to immunization with diphtheria toxoid adsorbed to aluminium phosphate has been made using a combined immunofluorescent and autoradiographic technique. No difference was detected in the kinetic pattern of the specific antibody-containing cells compared to the general population of those containing γ-globulin. The morphology of the replicating antibody-containing cells and their progeny are described. A population of large basophilic, DNA-synthesizing cells which did not contain γ-globulin was found to be present in the nodes. Rates of cell division were examined in rats immunized with HSA. The blast cells were found to have a rapid turnover and their progeny, the plasma cells, a short life span.     

Antigen localization in lymphopenic states: II. Further studies on whole body X-irradiation

The gross and microscopic distribution of ¹²⁵I polymerized flagellin from Salmonella adelaide was studied in adult rats at various times following 800 r whole body X-irradiation. Injections of radioactive antigen were made in both hind footpads, and the popliteal lymph nodes were excised for autoradiographic study 1 day later. This dose of irradiation caused a progressive decline in the ability of lymphoid follicles of popliteal nodes to capture and retain antigen. Irradiation had no detectable effect upon antigen uptake by whole lymph nodes or upon the number of grains overlying the phagocytic cells of the medullary sinuses of popliteal nodes.

Various substances capable of restoring follicular antigen uptake in the irradiated rat were studied by means of injecting the test substance into one hind footpad 1 hour prior to the injection of antigen into both feet. The distribution of antigen in each popliteal node was compared, each animal thus acting as its own control. It was found that 0.01 ml of specific anti-flagellar immune serum, or 0.25 ml of normal adult rat serum significantly improved follicular antigen uptake when tested ten days after irradiation. Foetal calf serum, homologous lymphocytes, and the media from pooled concentrated lymphocyte cultures were without demonstrable effects when given by regional injection. Shielding of the popliteal nodes at the time of irradiation improved follicular antigen uptake, whereas shielding of the femoral bone marrow and appendix was ineffective. No agent found capable of improving follicular antigen capture in the irradiated rat significantly altered footpad retention of antigen, whole organ counts of the popliteal nodes, or antigen localization in the phagocytic cells of the lymph node medulla.

The results favour the interpretation that the follicular antigen trapping mechanism is extremely sensitive to changes in levels of opsonins; that substances present in normal adult rat serum act as `follicular opsonins'; that these substances decline exponentially following irradiation; and that these substances are secreted by small lymphocytes or their progeny.

     

Preparation of specific anti-chicken immunoglobulin sera

Specific anti-chicken immunoglobulin sera were prepared in guinea-pigs by intra-lymph node injection of antigen–antibody complexes. The antigen used was chicken γ-globulin antibody directed against a soluble antigen, bovine serum albumin.     

Localization by immunofluorescence of gamma-globulin allotypes in lymph node cells of homozygous and heterozygous rabbits

The cellular production of two rabbit γ-globulin allotypic specificities, A4 and A5, determined by allelic genes was investigated by the fluorescent antibody method. The 7S γ-globulin fractions of precipitating antisera were conjugated to fluorescein isothiocyanate and to lissamine rhodamine B sulphonyl chloride. Frozen sections of lymph nodes from eighteen rabbits, A4—A4 and A5—A5 homozygotes and A4—A5 heterozygotes, were studied after exposure to the fluorescent antibody conjugates. The conjugates, each specific for antigenic determinants of 7S γ-globulin, reacted specifically with the cytoplasm of plasma cells and intrinsic cells of the germinal centres. The rabbit anti-A4 conjugate reacted only with lymph node cells of A4—A4 and A4—A5 rabbits; the rabbit anti-A5 conjugates reacted only with cells of A5—A5 and A4—A5 rabbits; the horse anti-rabbit γ-globulin conjugates reacted with cells of all three genotypes. By a variety of techniques, identical cellular localization of the two allotypes, A4 and A5, was found in the A4—A5 heterozygotes. Less than 1 per cent of the cells in any heterozygous lymph node section contained one allotype without the other.     

The Localization of Antigen in Lymph Node Follicles of Congenitally Athymic Nude Rats

We have examined the postulated dependence on T cells of follicular retention of antigen by studying antigen retention in the draining lymph nodes of congenitally athymic, nude rats after local injections of horseradish peroxidase (HRP). The lymphoid tissues of these rats contained germinal centres and follicular dendritic cells (FDC) that were ultrastructurally identical to those seen in euthymic rats and expressed the differentiation antigen MRC OX2. Nude rat FDC captured and retained locally injected antigen on their surfaces, but as with euthymic rats, only in the presence of previously injected anti-HRP antibody. This demonstrates that the FDC mature both morphologically and functionally in the absence of a thymus or T cells. However, in contrast to euthymic rats, there was no detectable antigen retention in nude rats that had been actively immunized by repeated intraperitoneal injections with HRP for 3 months. The lower number of germinal centres observed in athymic animals compared with their euthymic littermates could thus be explained by deficient production of specific antibody of the isotype necessary for follicular localization of environmental antigens.     

Antigens in immunity: VIII. Localization of 125I-labelled antigens in the secondary response

Two flagellar antigens, intact flagella and monomeric flagellin from Salmonella adelaide were labelled with ¹²⁵I by the Chloramine-T method. They were injected into the hind foot-pads of rats, and the localization in the draining lymph nodes was studied by scintillation counting and autoradiography. Injected rats belonged to one of four groups: normal, unimmunized adult rats (NI); rats that had been given an unrelated antigen 6 weeks previously (HI); rats that had been given a priming dose of the same antigen (though unlabelled) 6 weeks previously (AI); and rats that had been given passive antibody by intraperitoneal injection (PI).

Follicular localization was more rapid in AI and PI groups than in NI or HI rats. With flagellin, but not with flagella, the final follicular concentration reached was also greatly increased. No differences were observed between NI and HI rats, or between AI and PI rats.

In primary lymphoid follicles, the antigen was distributed throughout the follicle in a diffuse network, presumably of macrophage fibrils. In secondary follicles, the antigen localized in a crescentic cap occupying the superficial aspect of the follicle.

The study stressed the importance of antibody acting as an opsonin in determining details of antigen localization.

     

The development of specific antibody-containing cells and the localization of extracellular antibody in the follicles of the spleen of rabbits after administration of free or liposome-associated albumin antigen

In order to study the localization pattern of specific antibody-containing cells and extracellular antibody in the spleen during a primary immune response, the antigen human serum albumin (HSA) was injected intravenously in rabbits, either free in solution, or associated with the surfaces of liposomes as a model of membrane-associated antigens. Demonstration of specific antibody-containing cells was performed by incubation of sections of spleen with HSA-HRP conjugates, followed by peroxidase cytochemistry. Specific anti-HSA antibody-containing cells were detected already at 4 days after injection of the antigen. The bulk of these cells localized initially in the outer parts of the periarteriolar lymphocyte sheaths (PALS) and around the terminal arteriolar branches. Both extracellular antibody and specific antibody-containing cells were also found in the follicles of the spleen. Arguments are given that extracellular antibody precedes the development of specific antibody-containing cells in the follicle. This extracellular antibody probably represents antigen-antibody complexes trapped in the follicles as soon as the antigen in the circulation is complexed by the first antibodies produced during the immune response. The localization pattern of specific antibody-containing cells and extracellular antibody did not differ markedly when rabbits injected with free or liposome-associated antigen were compared. Results are discussed particularly with respect to the role of follicles in the immune response.     

Complement-mediated follicular localization of T-independent type-2 antigens: the role of marginal zone macrophages revisited.

In this study we demonstrate a hitherto undescribed phenomenon, namely that thymus-independent type-2 antigens (TI-2 Ag) localize in splenic follicles within 1 h after administration. The follicular localization of 2,4,6-trinitrophenyl (TNP)-Ficoll was not antibody mediated. In addition in case of high-dose administration we observed a relatively large amount of TI-2 Ag in marginal zone macrophages. However, after low-dose administration we observed a preferential localization of TNP-Ficoll in the splenic follicles. Detection of TNP-haptenated Ag in cryostat sections of murine spleens was performed with a high-affinity TNP-specific monoclonal antibody conjugated to beta-galactosidase. Within minutes after injection the TI-2 Ag localized in the marginal zone, attached to marginal zone macrophages and B cells. Twenty minutes after injection the Ag was also detected in the follicles and gradually accumulated there until 7 h after injection. Thereafter, the amount of follicular Ag gradually decreased but was still detectable up to 14 days after immunization. The follicular localization of TNP-Ficoll was complement dependent in contrast to the binding to and uptake by marginal zone macrophages. Double staining revealed that Ag was bound by macrophages, B cells and follicular dendritic cells. Haptenated thymus-dependent (TD) Ag localized exclusively in the red pulp macrophages. In vivo macrophage elimination drastically increased the amount of TNP-Ficoll in the follicles, and enhanced the humoral immune response at low doses of Ag. Moreover, complement deprivation of mice abrogated the localization of TI-2 Ag in the follicles, and led to a decreased humoral TI-2 immune response. In conclusion, we demonstrate for the first time that TI-2 Ag localize in follicles. Moreover, the presented results provide further evidence that B cells and follicular localized Ag play an important role in the induction of humoral TI-2 immune responses.     

Haemorrhagic reactions elicited at sites of passive cutaneous anaphylaxis by the intravenous injection of aggregated γ-globulin

In guinea-pigs injected intradermally with a small amount of antibody and challenged 2 hours later, by the intravenous route, with a mixture of homologous antigen and aggregated γ-globulin, haemorrhagic reactions of the Arthus type develop at the sites of intradermal sensitization. This effect was obtained with γ-globulins of different species (human, rabbit and horse) by using different techniques for aggregation (heat, mercaptoethanol—urea and bis-diazobenzidine) and was always correlated with the ability of the aggregated globulin to fix complement.

Fluorescein labelled aggregates of γ-globulin were detectable in the wall of vessels at sensitized sites.

In experiments performed with guinea-pig antibodies, the localizing effect was observed only with γ₁, whereas the γ₂, Arthus-producing fraction proved completely ineffective.

Histamine and histamine liberators are not sufficient for eliciting the effect obtained with sensitizing antibody plus homologous antigen. It is postulated, therefore, that other effects occurring at the site of specific sensitization may also be responsible for the phenomenon.