The purpose of this project was to utilise recent advances in cellular and molecular immunology as well as in drug metabolism to develop new in vitro assays for the prediction of adverse immune reactions induced by pharmaceutical agents including new biotechnology-derived products (monoclonal antibodies, recombinant cytokines and growth factors) and by environmental chemicals. We were considering immunotoxic reactions induced by drugs not primarily designed to act on the immune system as well as those driven by immunosuppressive or immunomodulatory agents. To facilitate the cooperative links, the programme was divided in 6 sections: cytokines and growth factors, monoclonal antibodies, autoimunity, allergy, immunosuppression, and drug metabolism.
The major objectives of the project were as follows:
(1) To develop in vitro tests to evaluate the ability of biotechnology-derived substances such as monoclonal antibodies and recombinant cytokines to trigger specific or non-specific immune responses or to induce immunosuppression.
(2) To take advantage of human T cell clones to define optimal conditions for measuring T cell responsiveness in the different types of drug-induced hypersensitivity reactions.
(3) To develop in vitro tests predictive of the immunogenicity of drugs or drug-metabolites for T lymphocytes.
(4) To investigate the use of ex vivo cells derived from susceptible strains of rats and mice in the development of in vitro assays for autoimmune potential of xenobiotics. To validate these tests against patients with drug-induced immunopathological manifestations.
(5) To investigate the potentially suppressive effects of medicinal substances and chemicals on in vitro models of immune responsiveness.
(6) To reproduce and utilize in vitro the metabolic pathways that may be generated in vivo by certain classes of drugs known to act on the immune system.
1. Cytokines as mediators of immunotoxic reactions
One major problem in the toxicological evaluation of recombinant proteins produced by genetic engineering in bacteria is their contamination with bacterial toxins. Indeed, bacterial toxins are potent inducers of cytokines which mediate inflammatory reactions. The Limulus amoebocyte lysate test commonly used is not satisfactory because of false negative results. Several investigators (Pietro Ghezzi, Mario Negri Institute, Milano, Italy; Georges Grau, University of Geneva, Switzerland; Eric Tartour/ Wolf Herman Fridman, Institut Curie, Paris, France; Barry van Tits/Wim Buurman, University of Limburg, Maastricht, The Netherlands) obtained data suggesting that the ability of a compound to elicit cytokine production can be reliably assessed in vitro using as indicator cells either peripheral blood mononuclear cells from normal volunteers or cell lines derived from macrophages or endothelial cells. The sensitivity of the assays depends both on the nature of the indicator cells and the system of measurement for cytokine levels. The use of THP1 cells and quantification of tumor necrosis factor (TNF)-alpha mRNA using reverse competitive polymerase chain reaction (PCR) permits detection of lipopolysaccharide (LPS) levels as low as 100 pg/ml (Eric Tartour/Wolf Herman Fridman). Moreover, it was found that treatment of THP1 cells with vitamin D3 dramatically increase their expression of the CD14 LPS receptor and thereby their sensitivity to LPS (Eric Tartour/Wolf Herman Fridman). Beside TNF-alpha synthesis, IL-8 release was found to also represent a sensitive indicator of LPS-induced THP1 cell activation
Other tests under development aim to predict toxic effects of cytokines themselves. Indeed, cytokines like TNF might be of therapeutic value as anti-cancer agents but are highly toxic when administered in vivo. One major breakthrough in this area was the discovery of two different TNF receptors (termed p55 and p75) and the demonstration that one of them (p75) is more important for toxicity than for the antitumor activity. In vitro assays that can distinguish between binding to each of the two receptors are being developed to screen the toxicity of new TNF muteins with improved therapeutic index (Peter Vandenabeele/Walter Fiers, University of Ghent, Belgium). The TNF receptors might be produced in soluble form and it is obviously important to define the factors involved in the release of these soluble TNF receptors, especially in response to bacterial toxins. Jonathan Cohen (Hammersmith Hospital, London, UK) obtained evidence that at least two pathways are involved, one being independent on the TNF response itself as C3H/HeJ mice which do not produce TNF after LPS challenge display shedding of TNF receptors in response to LPS. In the context of this programme, TNF muteins which specifically bind either the p55 or the p75 TNF receptor (Peter Vandenabeele/Walter Fiers) were elaborated and tested in models of inflammation (Pietro Ghezzi, Georges Grau, Wim Buurman).
beta cells of the pancreatic islets represent potential targets of the toxic effects of cytokines and George Papadopoulos (University of Ionannina, Greece) observed that the rat beta-cell line RIN-SAH might be useful to predict the diabetogenic potential of recombinant cytokines. Organs retrieved from cadavers for transplantation might also be damaged by cytokines, in the context of the so-called ischemia-reperfusion phenomenon. Indeed, in vivo and ex vivo data indicate that significant amounts of interleukin (IL)-8, a cytokine chemotactic for neutrophils, and of IL-10, an immunosuppressive cytokine, are released upon reperfusion of human liver (Ursula Rauen/Herbert de Groot, University of Essen, Germany and Arnaud Marchant/Michel Goldman, Free University of Brussels, Belgium). In this context, cultures of hepatocytes and endothelial cells were established to study in vitro properties of organ preservation solutions (Ursula Rauen/Herbert de Groot).
Nitric oxide (NO) emerges as a major mediator of the toxicity of recombinant cytokines. The group of Jonathan Cohen developed several assays using a rat smooth muscle cell line and more recently human cell lines to detect the ability of cytokines either alone or in combination to induce NO production. Along the same line, the group of Georges Grau developed an in vitro assay to detect NO synthesis by brain endothelial cells.
2. The bases of allergic reactions to drugs
T cells of the Th2 type (T lymphocytes secreting IL-4, IL-5 and IL-10) play a central role in the development of drug-induced allergies. Individuals at risks for developing of allergic reactions might be identified on the basis of the profile of cytokines produced by their peripheral blood mononuclear cells upon in vitro exposure to pokeweed mitogen: a high IL-4/interferon-gamma index was indeed observed in allergic patients (Mary-Beth Erskine-Grout/ Clive Meredith, BIBRA International, Carshalton, UK). Some drugs such as penicillins directly activate T cells of susceptible individuals, without the need of prior processing by specialized antigen-presenting cells (APCs). By contrast, drugs such as sulfamethoxazole require metabolization and presentation by specialized cells in order to trigger T cell activation in vitro. This results in the production of high levels of IL-5 which mediates eosinophilia, a hallmark of drug-induced allergy (Werner Pichler, Institute for Immunology and Allergy, Bern, Switzerland). T cell reactivity to drugs can be evaluated by cytokine mRNA expression (Klara Miller, BIBRA International, Carshalton, UK; Vera Stejskal, AB Astra, Sodertalje, Sweden) and also by the expression of CD30 antigen on T cells, which correlates with the secretion of TH2-type cytokines (Enrico Maggi, Istituto di Clinica Medica III, Florence, Italy). It was also shown that drugs inducing allergic reactions generate neoantigens by binding to serum proteins and cell membranes (John Coleman, University of Liverpool, UK; Vera Stejskal).
In vivo correlates of in vitro data were analyzed in several settings. Thus, the group of John Coleman used a `haptenated autologous mixed lymphocyte test' to detect T cell sensitization to penicillamine while the groups of Sergio Romagnani and Michel Goldman demonstrated that recombinant interferon-alpha promotes Th1 responses both in vitro and in vivo in patients with the hypereosinophilic syndrome or with Kaposi's sarcoma. A drug known to inhibit HIV replication (nevirapine) was found to induce modulation towards a Th2 profile, probably explaining why treatment with this drug is associated with adverse IgE-mediated reactions.
3. Mercuric chloride as a model compound for chemically-induced autoimmunity
In susceptible strains of rats, mercuric chloride induce a systemic autoimmune syndrome which mimics autoimmune reactions induced in man by drugs such as gold salts. The joint efforts of the groups of Philippe Druet (Institut National de la Santé et de la Recherche Médicale (INSERM) Unit U28, Toulouse, France), Peter Mathieson (University of Cambridge, UK), David Oliveira (Addenbrooke's Hospital, Cambridge, UK) and Jan Weening (University of Amsterdam, The Netherlands) enabled elucidation of the respective roles of B lymphocytes, T lymphocytes and mast cells in the syndrome. Interactions of B lymphocytes with polyclonally activated T lymphocytes of the Th2 type triggers the secretion of autoantibodies. Mast cells are also activated by mercuric chloride and contribute to the syndrome in two ways: (i) they mediate vasculitis in the intestinal mucosa, (ii) they secrete IL-4 which is crucial for the differentiation of T cells into Th2 cells. The effects of mercuric chloride appear to be related to changes in intracellular redox status both in T lymphocytes and mast cells (David Oliveira and Jan Aten/Jan Weening). By contrast, there is no evidence for the involvement of heat shock proteins (Raymond Pieters/Marie Bloksma, Research Institute of Toxicology, Utrecht, The Netherlands). The T cells responsible for the autoimmune syndrome appear to recognize class II MHC molecules (Philippe Druet) and to over express costimulatory molecules such as LFA-1 which plays an important role in the effector functions of T cells.
This model was also used to delineate the genetic factors governing the susceptibility to chemically-induced autoimmunity. Both MHC and non-MHC genes appear to be involved (Philippe Druet) and there are preliminary indications that polymorphism in cytokine genes (IL-4, IL-12) might be involved (David Oliveira).
4. Drug metabolism
A major problem in designing in vitro tests is to reproduce the metabolic pathways that function in vivo. As cytochromes P450 in the liver play an important role in the metabolism of several drugs, efforts are being made to reproduce this pathway in vitro. Moreover, it was found that assessment of the drug acetylation pathway might allow the identification of individuals at risk of severe reactions to exposure to procainamide and sulfamides. Indeed, all tested patients who suffered from Lyell syndrome as a consequence of sulfamide exposure were found to be slow acetylators, indicating that the production and detoxification of the hydroxyl amine is a key event in the development of the disease (Philippe Beaune, INSERM Unit U75, Paris, France; Patrick Dansette, Centre National de la Recherche Scientifique (CNRS)-Associated Research Unit U400, Paris). The same groups demonstrated that assessment of thiopurinemethyl-transferase activity might lead to the prevention of toxic reactions to azathioprine, an antimetabolite drug used in organ transplant recipients. The role of cyclosporine metabolism in its toxicity is now under further investigation with in vitro systems expressing P450. They have also pursued studies on tienilic acid-induced and dihydralazine-induced hepatitis by identifying the epitope recognised by the anti LKM autoantibodies. An important conclusion of those studies was that the covalent binding on human liver microsomes of drugs inducing immunoallergic type hepatitis is very specific while with drugs inducing directly toxic hepatitis the covalent binding is spread on many proteins (Philippe Beaune, patrick Dansette).
Finally, it was found by the group of Ernst Gleichmann (University of Düsseldorf, Germany) that the metabolism of drugs such as gold salts within macrophage lysosomes might be a crucial step in the generation of immunogenic peptides that give rise to immunotoxicity.
5. Immunosuppressive chemicals
Compounds with well established immunosuppressive properties such as cyclosporin A, bis(tri-n-butyltin)oxide (TBTO) and dibutyltinchloride (DBTC) were used to design in vitro assays on peripheral blood mononuclear cells or cell lines, using several read-out parameters including cell viability, lymphocyte proliferation and cytokine mRNA expression (Rob Vandebriel/ Henk van Loveren, Rijksinstitute voor Volksgezondheid and Milieuhygiene (RIVM), Bilthoven, The Netherlands; Barry Hudspith/Clive Meredith, BIBRA International). The main conclusion of these experiments was that both 3H-thymidine uptake and altered cytokine mRNA expression can be considered as indicators of immunotoxicity whereas flow cytometry changes were not helpful. Additional studies by L. Dencker (Uppsala Universiteit, Uppsala, Sweden) on mouse fetal thymus grown as an organ culture in vitro indicated that this system might be useful to screen for immunosuppressive substances.
6. Monoclonal antibodies as immunosuppressive agents
Monoclonal antibodies specific for receptors expressed at the surface of lymphocytes are used to prevent rejection of organ and bone marrow transplants and are being developed for the treatment of autoimmune diseases. A major pitfall linked to the clinical use of rat or mouse monoclonal antibodies is the sensitization to the foreign protein, which leads to loss of any therapeutic effect. Apart from developing sensitive and standardized assays for the identification of anti-xenogenic and anti-idiotypic antibodies, the groups involved in this project have succeeded in developing a new technique for the detection of anti-immunoglobulin IgE antibodies that might give rise to severe immediate hypersensitivity reactions upon repeated monoclonal antibody administration (Daniel Abramowicz/Michel Goldman, Université Libre de Bruxelles, Brussels; Lucienne Chatenoud/Jean-François Bach, INSERM Unit U25, Paris, France). The new generation of `humanized' monoclonal antibodies that were obtained by genetic engineering (Herman Waldmann, University of Oxford, UK) were found less immunogenic although they might still elicit the production of anti-idiotypic antibodies
The toxic effects of monoclonal antibodies can also be related to their cytokine-inducing potential, and assays on human whole blood are being developed as pre-clinical tests to predict the possible risk of cytokine-associated toxicity upon monoclonal antibody injection (Lucienne Chatenoud/Jean-François Bach, INSERM Unit U25, Paris, France; Jean-Pierre Revillard, INSERM Unit U80, Lyon, France). Efforts are being made, in parallel, to evaluate in vitro and in vivo the immunosuppressive potential of new antibodies directed against CD3 (Fabienne Willems/Oberdan Leo, Université Libre de Bruxelles, Brussels, Belgium), CD2 (Dominique Latinne, Cliniques Universitaires Saint-Luc, Brussels, Belgium) and CD4 (Jean-Pierre Revillard, INSERM U80, Lyon, France) T cells. Particular attention was paid to the ability of anti-lymphocyte antibodies developed by the group of Herman Waldmann to induce T cell anergy, a state of specific unresponsiveness of the effector lymphocytes mediating tissue injury, and T cell apoptosis (Jean-Pierre Revillard, Herman Waldmann).
MAJOR SCIENTIFIC BREAKTHROUGHS:
Several aspects of the programme might provide useful informations to the pharmaceutical industry within the EC, as discussed during several meetings with the In Vitro Testing Industrial Platform (IVTIP):
-Development of sensitive in vitro assays for the detection of compounds inducing the production of cytokines
-Development of in vitro assays to predict the toxicity of recombinant cytokines
-Development of a new method for the detection of anti-mouse immunoglobulin IgE antibodies
-Development of in vitro assays to assess the immunosuppressive potential of anti-lymphocyte monoclonal antibodies and chemicals
-Identification of pathogenetic mechanisms involved in autoimmune disorders induced by chemicals
-Development of in vitro models for drug-induced allergic reactions
-Identification and assessment of metabolic pathways involved in drug immunotoxicity.
Funding SchemeCSC - Cost-sharing contracts
SM5 4DS Carshalton
3720 BA Bilthoven
W12 ONN London
151 85 Södertälje
L69 3SG Liverpool
1012 WX Amsterdam
3584 CL Utrecht
BS2 8HW Bristol
CB2 1QP Cambridge