[Piet de Somer, the University of Leuven and the Belgium poliovaccine in 1956-57]. / Piet De Somer, het Leuvense Rega instituut en het Belgisch poliovaccin in 1956-57.

In the years following WW II, all 'Western' countries were struck by recurrent epidemics of infantile paralysis (poliomyelitis). In the early 1950s, a vaccine developed by Jonas Salk in Pittsburgh, became available in the U.S. and Canada. In 1953-54 central virology laboratories in Sweden, Denmark and France were already well advanced in setting up local production lines of the vaccine. At that point in time, the Catholic University of Leuven, on the initiative of the young microbiology professor, Piet De Somer, and in collaboration with the pharmaceutical concern R.I.T. (Recherches et Industries Thérapeutiques, Genval, Belgium), erected a new, multidisciplinary medical research institute, the Rega Institute. One of the research units to be headed by De Somer was destined to introduce the relatively new discipline of virology. As a test case, De Somer decided to venture on developing a production line of the Salk vaccine. In less than one year's time, the project was successful, such that Belgium became one of the first European countries to be self-supporting for its vaccine supply and to be able to initiate a large-scale vaccination campaign. The planning, preparation and execution of the project was accompanied by an extensive correspondence of De Somer with experts and other concerned parties in Belgium and abroad. This correspondence has been preserved and allows for a detailed reconstruction of the remarkable achievement.

Interferon gamma, a mediator of lethal lipopolysaccharide-induced Shwartzman-like shock reactions in mice.

The involvement of cytokines in the pathogenesis of a generalized, Shwartzman-like lethal inflammatory response to bacterial lipopolysaccharides (LPS) was studied by testing the ability of cytokines or neutralizing anticytokine antibodies to modify the course of the syndrome. The reaction was elicitable in non-SPF NMRI mice by two consecutive injections of S. marcescens LPS: a first injection in the footpad, followed after 24 h by an intravenous dose; the size and route of the preparatory LPS dose were found to be critical. Treatment with mAbs against IFN-gamma was found to completely prevent the reaction. Treatment with IFN-gamma on the other hand, rendered the mice more sensitive to elicitation of the reaction. In contrast, systemic administration of IFN-alpha/beta exerted a desensitizing effect. The role of endogenous cytokines in the pathogenesis of this generalized Shwartzman reaction was also documented by a study of the cytokine levels in the serum of the mice. In comparisons between mice given lethal and nonlethal induction schedules, a good correlation was found between mortality rates and height of IFN or TNF levels, but no correlation was seen with IL-6 levels. Also, in mice that were protected by anti-IFN-gamma antibody, serum IFN and TNF were undetectable, whereas IL-6 levels were as high as in unprotected mice. These data provide evidence that among the cytokines that govern the inflammatory response to LPS, endogenous IFN-gamma occupies a key position. These findings therefore also open perspectives for clinical application of IFN-gamma antagonists.

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.

Interleukin 6 production in experimental cerebral malaria: modulation by anticytokine antibodies and possible role in hypergammaglobulinemia.

The production of Interleukin 6 (IL-6) was studied during experimental cerebral malaria (ECM) induced by Plasmodium berghei ANKA (PbA) infection. IL-6 is present in the serum of mice with ECM, the highest concentrations being observed in mice with full-blown neurological syndrome. High IL-6 levels were also observed, however, in the absence of pathology in nonlethal malaria infection. These data suggest that IL-6 is produced in large amounts during malaria infection, but does not play a major role in the pathogenesis of ECM. A modulation of IL-6 production in ECM was achieved by in vivo treatment with other anticytokine antibodies: antibodies to interferon (IFN-gamma) or to tumor necrosis factor (TNF) abolished the rise of IL-6, while anti-IL-3 and anti-granulocyte/macrophage colony-stimulating factor antibodies only partially prevented this rise, suggesting that the two cytokines IFN-gamma and TNF are important intermediates in IL-6 production. Passive immunization against IL-6 did not prevent ECM, but significantly reduced serum IgG levels in malaria-infected mice. Thus, by its effects on B cells, IL-6 may be involved in hypergammaglobulinemia and immune-complex diseases, e.g., glomerulonephritis observed during malaria infection.

Identification of the human 26-kD protein, interferon beta 2 (IFN-beta 2), as a B cell hybridoma/plasmacytoma growth factor induced by interleukin 1 and tumor necrosis factor.

A factor that promotes the growth of certain B cell hybridomas and of plasmacytomas is shown to be produced by normal human fibroblasts and by a line of human osteosarcoma cells (MG-63) after treatment with IL-1 or TNF. The hybridoma-plasmacytoma growth factor (HPGF) is identified with a 26 kD protein whose mRNA was previously shown to be induced in the same cells by the same inducers. First, poly(A)-rich RNA extracted from IL-1-treated cells could be enriched in HPGF-mRNA content by hybridization to 26 kD cDNA. Second, MG-63-derived HPGF purified to electrophoretic homogeneity was subjected to amino acid sequence analysis, whereby the NH2-terminal sequence was found to match the nucleotide sequence of a 26 kD cDNA clone.

[Penicillin in Belgium 1945-1952]. / Penicilline in België 1945-1952.

Penicillin, discovered now eighty years ago (1929) by Alexander Fleming in London, was developed during the world war II into a revolutionizing drug by Howard Florey and Michael Chain in Oxford. At first, industrial production of penicillin was exclusively in the hands of a consortium of large U.S. pharmaceutical companies. However, the war being ended, European entrepreneurs likewise ventured to set up penicillin production units. Amongst them, in Belgium, was Jacques Lannoye, director and co-owner of 'Papeteries de Genval' and of a modest pharmaceutical company, called 'Soprolac'. Through his connections with several medical faculty professors of the Catholic University of Leuven, Lannoye came in touch with Piet De Somer, then a young researcher at the Leuven 'Institute of Bacteriology', with an interest in production of penicillin. A years-long collaboration followed, from which emerged a booming antibiotic and vaccine factory, 'RIT' (Recherche et Industrie Thérapeutiques) in Genval, as well an industry-supported research laboratory, the later Rega Institute, at the University of Leuven. From 1947 to 1952, while coping with the practical problems of setting up large-scale production of penicillin, De Somer maintained a lively correspondence with some other players in the field, sharing with them the ups and downs of the enterprise. Fortunately these letters have been preserved in the archives of the Rega Institute, such that they allow for a reconstruction of this interesting episode in the medical history of Belgium.

A brief history of interferon's trajectory to clinical application, and personal reminiscences of a large-scale human interferon production initiative.

In a first part of this historical article, the author overviews how, over the half century since the first in vitro demonstration of interferon, systems for large-scale production of human interferon have been elaborated and how these efforts have allowed for evaluation of the clinical potential of the different molecular types of interferon. In a second part the author reflects in more detail on research activities in Belgium aimed at production of human interferon.

Xenotransplantation: where are we in 2008?

Xenotransplantation holds promise to solve the ever increasing shortage of donor organs for allotransplantation. In the last 2 decades, major progress has been made in understanding the immunobiology of pig-into-(non)human primate transplantation and today we are on the threshold of the first clinical trials. Hyperacute rejection, which is mediated by pre-existing anti-alpha Gal xenoreactive antibodies, can in non-human primates be overcome by complement- and/or antibody-modifying interventions. A major step forward was the development of genetically engineered pigs, either transgenic for human complement regulatory proteins or deficient in the alpha1,3-galactosyltranferase enzyme. However, several other immunologic and nonimmunologic hurdles remain. Acute vascular xenograft rejection is mediated by humoral and cellular mechanisms. Elicited xenoreactive antibodies play a key role. In addition to providing B cell help, xenoreactive T cells may directly contribute to xenograft rejection. Long-term survival of porcine kidney- and heart xenografts in non-human primates has been obtained but required severe T and B cell immunosuppression. Induction of xenotolerance, e.g. through mixed hematopoietic chimerism, may represent the preferred approach, but although proof of principle has been delivered in rodents, induction of pig-to-non-human primate chimerism remains problematic. Finally, it is now clear that innate immune cells, in particular macrophages and natural killer cells, can mediate xenograft destruction, the determinants of which are being elucidated. Chronic xenograft rejection is not well understood, but recent studies indicate that non-immunological problems, such as incompatibilities between human procoagulant and pig anticoagulant components may play an important role. Here, we give a comprehensive overview of the currently known obstacles to xenografting: immune and non-immune problems are discussed, as well as the possible strategies that are under development to overcome these hurdles.

CTLA-4 blockade in murine bone marrow chimeras induces a host-derived antileukemic effect without graft-versus-host disease.

We studied the effect of CTLA-4 blockade on graft-versus-leukemia and graft-versus-host responses in a mouse model of minor histocompatibility-mismatched bone marrow transplantation. Early CTLA-4 blockade induced acute graft-versus-host disease. Delayed CTLA-4 blockade resulted in a lethal condition with lymphosplenomegaly, but with stable mixed T-cell chimerism, unchanged alloreactive T-cell frequencies and absent anti-host reactivity in vitro. In contrast, multiorgan lymphoproliferative disease with autoimmune hepatitis and circulating anti-DNA auto-antibodies were documented. Splenic lymphocytes exhibited ex vivo spontaneous proliferation and a marked proliferative response against host-type dendritic cells pulsed with syngeneic (host-type) tissue-peptides. Both phenomena were exclusively mediated by host and not donor T cells, supporting an autoimmune pathogenesis. Selectively host-derived T-cell immune reactivity was equally documented against leukemia-peptide-pulsed dendritic cells, and this was paralleled by a strong in vivo antileukemic effect in anti-CTLA-4-treated and subsequently leukemia-challenged chimeras. In conclusion, delayed CTLA-4 blockade induced a host-derived antileukemic effect, occurring in the context of an autoimmune syndrome and strictly separated from graft-versus-host disease. Both antileukemic and autoimmune responses depended on the allogeneic component, as neither effect was seen after syngeneic bone marrow transplantation. Our findings reveal the potential of using CTLA-4 blockade to establish antileukemic effects after allogeneic hematopoietic stem cell transplantation, provided autoimmunity can be controlled.

Interferon-gamma in autoimmunity.

Autoimmune disorders are characterized by abrogation of self-tolerance, resulting in emergence of activated self-reactive lymphocyte clones that trigger or maintain inflammatory reactions in specific organs. Interferon-gamma (IFN-gamma), as well as other cytokines, plays an important role as a regulator of the activation of self-reactive lymphocytes and of bystander and accessory cells that are involved in the autoimmune inflammatory response. In experimental models of autoimmunity, endogenous IFN-gamma has invariably been found to profoundly affect the disease course. However, it acts in one way in some diseases and in the opposite way in others.

Human hepatoma cells produce an 85 kDa gelatinase regulated by phorbol 12-myristate 13-acetate.

Several human cell lines were studied for the production of gelatinases. Diploid fibroblasts, the melanoma cell line Bowes, the MG-63 osteosarcoma cell line and the human hepatoma cell line Malavu all constitutively produced a 67 kDa gelatinase. Gelatinolytic enzymes were quantified by a sensitive zymographic substrate conversion assay. Upon induction with phorbol 12-myristate 13-acetate (PMA), the human hepatoma cell line secreted considerable amounts of an 85 kDa gelatinase activity. The induction process was time- and dose-dependent. It represented a true increase in production per individual cell and was associated by a marked change of the cell morphology. The effect of various proteinase inhibitors and the maximal activity of the enzyme near neutral pH demonstrate that it is a neutral metalloproteinase. Characterization studies showed the 85 kDa gelatinase to be transformed to lower molecular weight, active forms by treatment with p-aminophenylmercuric acetate (APMA) or trypsin.

Modes of action of Freund's adjuvants in experimental models of autoimmune diseases.

Freund's adjuvants are irreplaceable components of induction protocols of many experimental animal models of autoimmune disease. Apart from the early studies done in the 1950s and 1960s, no further direct investigation on the mode of action of these adjuvants has been undertaken. It is generally assumed that incomplete (IFA) and complete Freund's adjuvant (CFA) act by prolonging the lifetime of injected autoantigen, by stimulating its effective delivery to the immune system and by providing a complex set of signals to the innate compartment of the immune system, resulting in altered leukocyte proliferation and differentiation. Here, we review evidence collected from various types of studies that provide more insight in the specific alterations of the immune response caused by IFA and CFA. Early events include rapid uptake of adjuvant components by dendritic cells, enhanced phagocytosis, secretion of cytokines by mononuclear phagocytes, and transient activation and proliferation of CD4+ lymphocytes. The mycobacterial components within CFA signal T lymphocytes to assume a Th1 profile so that strong delayed-type hypersensitivity against autoantigens develops. In the absence of mycobacteria, T-lymphocyte differentiation tends to assume a Th2 profile with strong antibody production only. The mycobacterial component also accounts for a morphologic and functional remodeling of the haemopoietic system that develops over a period of several weeks and that is characterized by a drastic expansion of Mac-1+ immature myeloid cells. These cells have been found to be associated with enhanced disease in some models but with reduced disease in others. Thus, in experimental autoimmune diseases, CFA-mediated activation of the innate immune compartment is important not only by regulating the early induction phase but also by providing a surplus of effector and regulator cells in the late phase.

The protective effect of IFN-gamma in experimental autoimmune diseases: a central role of mycobacterial adjuvant-induced myelopoiesis.

The study of animal models for organ-specific autoimmune disease contributes to our understanding of human diseases such as multiple sclerosis and rheumatoid arthritis. Although experimental autoimmune diseases develop spontaneously in certain strains of mice, others need to be induced by administration of organ-specific autoantigen, often together with complete Freund's adjuvant (CFA), containing heat-killed mycobacteria. In the two types of models, the role of endogenous interferon-gamma (IFN-gamma) has extensively been investigated by using neutralizing anti-IFN-gamma antibodies and by employing mice genetically deficient in IFN-gamma or its receptor. In these studies disease-promoting as well as disease-protective roles of endogenous IFN-gamma have been described. Remarkably, in most models that rely on the use of CFA, there is abundant evidence for a protective role. Here, we review evidence that this role derives from an inhibitory effect of IFN-gamma on myelopoiesis elicited by the killed mycobacteria. These findings explain the bimodal role of IFN-gamma in different models of autoimmune disease and raise questions regarding the clinical relevance of these models.

The effects of NO synthase inhibitors on murine collagen-induced arthritis do not support a role of NO in the protective effect of IFN-gamma.

DBA/1 mice deficient in expressing the interferon-gamma (IFN-gamma) membrane receptor (IFN-gammaR KO mice) are more susceptible to collagen-induced arthritis (CIA) than wild-type mice, indicating that endogenous IFN-gamma plays a protective role in the pathogenesis of CIA. In IFN-gammaR KO mice, nitric oxide (NO) production during CIA is impaired. Because NO is known to exert immunosuppressive and anti-inflammatory effects in certain model systems, the protective effect of IFN-gamma might be mediated by NO. Here, we tested in wild-type mice whether inhibition of NO production by metabolic inhibitors, aminoguanidine (AG) and L-N-(1-iminoethyl)lysine (L-NIL), could mimic the ablation of the IFN-gamma receptor. A high-dose regimen of AG supplied in the drinking water inhibited NO production, disease development, and anticollagen antibody production but was also associated with transient body weight loss. At a dose and time regimen that still inhibited NO production but did not cause body weight loss, AG failed to affect disease scores. Treatment with L-NIL, which more specifically than AG affects inducible NO production, caused a slight increase in anticollagen antibody production although not significantly affecting disease occurrence. These data indicate that the diminished capacity of the IFN-gammaR KO mice to produce NO following immunization with collagen is unlikely to account for their higher susceptibility to CIA.

Bimodal role of endogenous interleukin-6 in concanavalin A-induced hepatitis in mice.

Acute concanavalin A (Con A)-induced hepatitis in mice is an animal model for hepatic injury induced by activated T cells. The evolution of hepatic involvement can be followed from hour to hour by measuring serum transaminase levels. We investigated the possible role of endogenous interleukin-6 (IL-6) in this model. We found serum IL-6 levels and splenic IL-6 mRNA during Con A-induced hepatitis to be significantly lower in interferon-gamma (IFN-gamma)-deficient mice, which are resistant against the Con A-induced syndrome, than in wild-type ones, suggesting that systemic IL-6 production favors development of hepatic injury. However, IL-6-deficient mice proved to be more susceptible to the disease than wild-type mice, indicating that endogenous IL-6 plays a predominantly hepatoprotective role. Experiments in which wild-type mice were treated with anti-IL-6 antibodies, before or after Con A challenge, allowed us to reconcile these contrasting observations. The antibody injections resulted in a biphasic alteration of serum IL-6 levels, initial neutralization being followed by rebound increased levels due to accumulation of IL-6 in the form of antigen-antibody complexes. The effect of antibody on disease severity differed depending on the time of injection. Antibody injection at 2.5 h post Con A resulted in delayed disease manifestation, whereas treatment initiated before Con A resulted in accelerated disease. We conclude that endogenous IL-6 plays a bimodal role. IL-6 present before Con A challenge as well as that induced in the very early phase after Con A injection triggers hepatoprotective pathways. Continuation of IL-6 production beyond this early phase, by some other pathway, seems to be harmful to hepatocytes.

Transcriptional control of the human MCP-2 gene promoter by IFN-gamma and IL-1beta in connective tissue cells.

Human monocyte chemotactic protein-2 (MCP-2) is a member of the CC chemokine family. It is produced by mononuclear leukocytes, diploid fibroblasts, and tumor cells after induction with IL-1beta or IFN-gamma. To understand the transcriptional regulation of the gene, we have analyzed the structure and function of the promoter region. The sequence of the 5'-flanking region was determined and the transcription start site was found to be located at 68 nucleotides upstream of the ATG translation start codon. 5'-Deletion mutants were generated and transfected into E6SM diploid fibroblasts and MG-63 osteosarcoma cells. Expression was measured by luciferase assay in transfected unstimulated cells and after stimulation with IL-1beta, IFN-gamma, or a combination. The region between nucleotides -143 and -73 (relative to the transcription initiation site), containing putative cis-elements for GATA-1, H-APF1, AP-1, and GAS, is important for basal transcription levels in both cell lines. Stimulation for 18 h with IL-1beta alone failed to affect expression of any of the constructs both in diploid fibroblasts and in osteosarcoma cells. In both cell lines IFN-gamma increased the activity of all mutants that possessed the region between -340 and -301. In MG-63 cells, stimulation with the combination of IL-1beta and IFN-gamma caused an additional increase in expression of the constructs from -340 onward. Finally, the presence of transcription factors in nuclear extracts of MG-63 cells and their specificity to bind to various oligonucleotide probes in this [-340; -301] region were evidenced by electromobility shift assays. These results show that IFN-gamma, produced by lymphocytes and NK cells, induces the transcription of the MCP-2 gene in fibroblasts and thereby can indirectly contribute to recruitment of various leukocyte cell types to inflammatory sites.

Granulocyte chemotactic protein-2 and related CXC chemokines: from gene regulation to receptor usage.

Chemokines contribute to the inflammatory response by selective attraction of various leukocytic cell types. Human GCP-2 was originally identified by amino acid sequence analysis as a CXC chemokine co-produced with IL-8 by osteosarcoma cells. Furthermore, the complete coding domain of human GCP-2 was disclosed by means of RT-PCR. Similarly, mouse GCP-2 was isolated from fibroblastoid and epithelial cells and completely identified by sequence analysis. Human and mouse GCP-2 share 61% identical amino acids. Both chemokines occur as multiple NH2-terminally truncated forms. The shorter forms of mouse, but not those of human, GCP-2 showed a higher neutrophil chemotactic potency and gelatinase B releasing capacity. Mouse GCP-2 was a more potent neutrophil activator than human GCP-2, natural mouse KC, and MIP-2. Human GCP-2 was not chemotactic for monocytes, lymphocytes, or eosinophils. Quantitative studies of mRNA expression in diploid fibroblasts revealed GCP-2 induction by IL-1beta. Human GCP-2 induced [Ca2+]i increase in neutrophils, which was reciprocally desensitized by IL-8, GROalpha, and ENA-78. Human GCP-2 induced [Ca2+]i increases and chemotactic responses in both CXCR1- and CXCR2-transfected cells. Finally, GCP-2 provoked neutrophil accumulation and plasma extravasation in rabbit skin. In humans, GCP-2 complements the activity of IL-8 as neutrophil chemoattractant and activator but it constitutes a major neutrophil chemokine in the mouse. GCP-2 induces neutrophil chemotaxis and activation but it might also contribute to detrimental tissue damage in sepsis, acute respiratory distress syndrome, acute hypersensitivity reactions, and autoimmune diseases. It might also influence the invasive capacity of GCP-2-secreting tumor cells.

Bacterial expression of a single-chain antibody fragment (SCFV) that neutralizes the biological activity of human interferon-gamma.

Studies with animal models have indicated that neutralizing antibodies against human interferon-gamma (HuIFN-gamma) may be used to treat a number of diseases in man. A major handicap for the implementation of this form of therapy is the immunogenicity of antibodies of non-human origin. Antibody fragments that do not contain parts of the most immunogenic regions may help to circumvent this problem. Therefore, we have constructed several antibody fragments [VH (variable fragment of the heavy chain), Fv (variable fragment) and scFv (single-chain Fv)] derived from a murine hybridoma (D9D10) which produces a neutralizing antibody against HuIFN-gamma. cDNAs encoding the variable domains of the L and H chains of D9D10 were cloned by PCR-based techniques in a suitable E. coli expression system. Bacterial clones are described that produce either VH, Fv or scFv. The Ig fragments were secreted into the periplasm and leaked into the culture supernatant. By SDS-PAGE and immunoblot analysis, the fragments were shown to be of the expected size (15, 14 and 30 kDa for VH, Fv and scFv, respectively). Functionality of the recombinant Ig fragments was tested by an ELISA for HuIFN-gamma binding and by a neutralization assay for the antiviral activity of HuIFN-gamma. Fv as well as scFv, but not VH, were found to bind to HuIFN-gamma and to neutralize its antiviral activity. Since it is found that scFv proteins are more stable at physiological temperatures than the Fv, it may have potential usefulness for the treatment of diseases in which overproduction of IFN-gamma plays a crucial role.

Effect of VH and VL consensus sequence-specific primers on the binding and neutralizing potential of a single-chain FV directed towards HuIFN-gamma.

We have previously reported on the cloning and bacterial expression of a biologically active scFv antibody fragment (scFv-D9D10) derived from the mouse anti-human interferon-gamma (HuIFN-gamma) antibody, D9D10. Since the variable (V) regions were isolated by means of VH and VL consensus sequence-specific PCR primers and cloned in an expression vector relying on primer-incorporated restriction sites, some amino acids (aa) at the N- and C-terminal ends of the cloned V domains were expected to differ from the corresponding ones in the natural D9D10 antibody. Therefore, the naturally occurring sequences of both V domains were isolated by means of traditional cDNA synthesis procedures. In comparison with scFv-D9D10, the "natural" V sequences differed in three aa in VH and three in VL. The V domains of scFv-D9D10 were adapted to their natural sequence by means of PCR-directed mutagenesis to yield scFv-D9D10N. Comparison of the binding and neutralizing potentials of both antibody fragments did not reveal differences in either of both activities. In addition, their affinities for HuIFN-gamma were found to be equal. These results show that murine VH and VL consensus-specific primers can yield antibody fragments having functional properties equivalent to those of the natural scFv. Information on the impact of the use of V-specific primers on kinetics of interaction between the recombinant antibody and the corresponding antigen is important for the development of most engineered antibodies or their fragments.

Evidence that folliculo-stellate cells mediate the inhibitory effect of interferon-gamma on hormone secretion in rat anterior pituitary cell cultures.

Interferon-gamma (IFN-gamma) is known to inhibit the release of ACTH, PRL, and GH by rat anterior pituitary (AP) cells, stimulated by appropriate hypothalamic releasing factors. In the present study we examined the mechanisms underlying this inhibition. Dose-response studies, revealing a maximal inhibitory effect with an IFN-gamma dose as small as 10 U/ml, suggested the existence of a limiting intermediate step. In addition, in perifusion experiments with aggregates of established hormone-secreting tumor cell lines (AtT-20 and GH3), IFN-gamma had no inhibitory effect, suggesting that an accessory cell type was involved. Studies on differentially enriched cell populations of normal rat AP, obtained by velocity and buoyant density sedimentation, indicated that the inhibitory effect of IFN-gamma on stimulated ACTH and GH release was absent in those populations that contained only few folliculo-stellate (FS) cells. The presence of a minimal proportion of FS cells was found to be necessary for the inhibition to be manifest. This was seen in monolayers, but also in cultures of AP cell aggregates, which are well known to closely mimic the behavior of the AP gland in vivo. Definitive evidence for the role of FS cells was obtained by reconstitutive coculture experiments; FS cell-poor populations, which by themselves resisted the inhibitory effect of IFN-gamma, became sensitive when cocultured with an FS cell-rich population. Basal ACTH and GH release were not influenced by preincubation with IFN-gamma in either original or fractionated AP cell populations. In contrast, basal PRL release was inhibited in both systems. In cultures of AP cell populations, separated by velocity sedimentation, this inhibition showed a pattern similar to that observed for stimulated release of ACTH and GH, i.e. more inhibition in fractions with a higher proportion of FS cells. However, in FS cell-poor cultures, inhibition of basal PRL release did occur, although to a lesser degree than in FS cell-rich cultures. Our results indicate that IFN-gamma affects AP hormone secretion through the FS cell. In addition, they suggest that IFN-gamma and the FS cell constitute a system through which the pituitary gland perceives changes in the activation state of the immune system.