In September 1964, I went from Marburg to Paris for the third year of my medical studies. Up to that point, I had no idea about immunology. My fascination for this discipline came during a two-week intensive course on Immunology at the Ancien Faculté de Médicine in Paris (Fig. 1).
The teachers, Felix-Pierre Merklin and Paul Berthaux, put together a useful script  and gave good lectures. However, the really exciting thing about this course, which took place in a small, rather somber room, with limestone walls and light from above, was a commemorative plaque on which was written: “In this room, Charles Richet and Paul Porter discovered in 1902 the anaphylactic shock in dogs”, referring to a deadly breakdown of circulation, resulting from repeated injections of a foreign, non-infectious protein extracted from jellyfish. This discovery of a paradox immune reaction, which did not protect but rather induced immunopathology, won Charles Richet the Nobel Prize in Physiology or Medicine in 1913 . This discovery stood in stark contrast to the previously discovered immunological protection obtained with Jenner’s cowpox vaccination, Pasteur’s attenuated Anthrax vaccine and Behring’s induction of immunity against diphtheria and tetanus toxins . It was this first immunological paradox that stirred my mind and sparked my long-lasting interest in immunopathology.
Richet’s observation is now known as the immediate, or type 1, hypersensitivity reaction. Coombs and Gell  later added three additional immunopathological reactions: type 2 hypersensitivity resulting from cytotoxicity mediated by antibody and complement, type 3 hypersensitivity, resulting from immune complex-mediated vasculitis and glomerulonephritis, and type 4 hypersensitivity, a delayed type of hypersensitivity mediated by sensitized T cells, which control self-tolerance and allograft rejection and are in manifold ways implicated in autoimmunity, immunodeficiency and tumor growth.
In autumn 1967, I passed the state examination in medicine and defended my medical thesis at the German Cancer Center in Heidelberg. In January 1968, I returned to Paris for a first post-doctoral fellowship at the Centre de Recherches Scientifiques sur le Cancer in Villejuif. The reason I went there was my interest in the immunological enhancement of tumor growth, another paradoxical phenomenon that had just been described by Guy-André Voisin . The term derives from the observation that transplanted tumors grow faster in animals with high titers of cytotoxic antibodies, and injection of a hyperimmune serum specific for antigens expressed on the tumor cell surface enhances tumor growth. The phenomenon was linked to the IgG fraction of the hyperimmune serum, possibly to certain IgG subclasses. However, at that time, nobody understood the pathophysiology. I joined the laboratory of Iwan Chouroulinkov in the Immunology Department of Pierre Grabar, and we could confirm the effect of immunological enhancement in a murine tumor transplant model. In this lymph sarcoma tumor model, females and castrated or estrogen-treated males succumb faster to the tumor than normal males. Sera from mice with enhanced tumor growth showed paradoxically high cytotoxic antibody titers , and injection of tumor cells together with this immune serum enhanced tumor growth.
In May 1968, I was physically and emotionally involved in the Paris student uprisings, which, although it obviously did not solve the paradox of immunologic enhancement, it did provide suspense and room for some creative thought. At the end of 1968, I returned to Heidelberg. I finished my medical internship, married and looked for a laboratory where I could continue my studies on immunological enhancement. At that time, the laboratory of Joseph D. Feldman at Scripps Clinic and Research Foundation, in La Jolla Ca., investigated the effects of enhancing alloantisera in a skin allograft model in Lewis rats . The DFG granted me a two-year research fellowship to join Dr. Feldman’s laboratory from 1970–72. We started to produce hyperimmune alloantisera in Lewis rats by repeated immunizations with BN strain lymphocytes, skin allografts and spleen cells. This hyperimmune Lewis-anti-BN alloantiserum delayed the rejection of BN skin grafts in young Lewis rats. We isolated the IgG fraction of this hyperimmune serum, labeled it with 125I and a control IgG fraction with 131I. The mixture of both labelled IgG fractions was injected into Lewis rats that carried BN skin allografts, which were treated either with enhancing alloantiserum or with control serum. The results were clear-cut: The allospecific 125IIgG specifically enriched at the graft site . Pre-treatment with enhancing alloantiserum prevented binding of the labeled allospecific 125IIgG to the graft and delayed its rejection by 4–6 days. The mechanism of graft prolongation obviously involved an afferent blockade of alloantigen recognition by host effector cells. At that time, the type of effector cells involved was not clear. We suspected cytotoxic T cells. To prove this assumption, we had to raise a xenoantiserum in rabbits with specificity for rat thymocytes . Next, we established a 51Cr release assay with BN embryonic fibroblast as targets for allospecific cytotoxic lymphocytes from Lewis graft recipients. The results clearly showed that allospecific cytotoxic effector lymphocytes expressed thymic antigens and appeared first in the draining lymph nodes around day 4 post-transplant and on day 8 in the spleen, when graft rejection was completed. Parallel to the prolonged skin allograft survival in animals treated with enhancing alloantiserum, the appearance of cytotoxic T cells in draining lymph nodes and the spleen was delayed. The results were published in 1972 in a detailed J Exp Med paper . While having proven evidence for an afferent blocking effect of enhancing antibodies at the allograft site, we could not exclude some transient central tolerance induction at the level of the effector cells, since kidney transplants benefited much longer from enhancing alloantisera. Ultimately, the reason why hyperimmune-enhancing sera did not cause an acute graft rejection but instead prolonged allograft survival remained a mystery.
As my stay in La Jolla came to its end, I made an intriguing observation, which would occupy my scientific interest for the next 10 years. During the testing for allospecific cytotoxic T lymphocytes on 51Cr labeled BN embryonic fibroblasts, I noticed background cytotoxicity, which was equally present in allotransplanted and in naïve animals. To calculate allospecific 51Cr-release, this background cytotoxicity had to be subtracted. This spontaneous or natural cell-mediated cytotoxicity varied from animal to animal and could reach up to 20 % over spontaneous 51Cr release (Fig. 2).
Interestingly, spontaneous cytotoxicity was mediated by lymphocytes that were not eliminated by our anti-thymocyte serum . I was convinced that this spontaneous killer cell activity must have something to do with the natural tumor surveillance.
In December 1972, I returned to Germany, or, more precisely, I joined the Department of Clinical Immunology of Helmuth Deicher at Hannover Medical School. I started my training in internal medicine and in parallel established a cytotoxic tumor model with human melanoma cells as targets. Jean-Pierre Cesarini from Paris advised me how to grow melanoma cell lines; Jochen Kalden was supportive in setting up the 51Cr- release assay in our joint laboratory, and Prof. H. Deicher arranged for me to be able to follow metastatic melanoma patients in the Department of Clinical Immunology. With the help of several doctoral fellows, it rapidly became clear that peripheral blood lymphocytes from healthy individuals exhibited cell-mediated cytotoxicity against various melanoma cell lines . At the same time, Herberman’s group in Bethesda, Wigzell’s, Hellström’s and Perlmann’s groups in Sweden, De Vries in Holland, Riethmüller in Munich and Hersh and Gutterman in Texas made similar observations. They all convened at an international symposium on immunological reactions to melanoma antigens in April 1974 in Hannover . At that time, cellular cytotoxicity measured against tumor cells in a short-term (4–8 hrs.) 51Cr-release assay was not yet attributable to a certain lymphocyte subset: it was called either cell-mediated cytotoxicity (CMC) or spontaneous cell-mediated cytotoxicity (SCMC). Only from 1976 onwards, when the effector cell was better characterized the term natural killing (NK) was introduced by Wigzell’s and Herberman’s group and prevailed over CMC and SCMC. Lysis of target cells coated with IgG antibodies was usually 5–10 x more efficient than SCMC lysis and was referred to by Peter Perlman as antibody dependent cell-mediated cytotoxicity (ADCC). The responsible effector lymphocytes (“K” cells), turned out to be identical to NK cells [13, 14]. The term micro-cytotoxicity (MC) assay was reserved for long-term (24–72 hrs.) tumor lysis assays involving mononuclear phagocytes and an optical read-out.
As shown in Fig. 3, CMC activity was similar in primary melanoma and healthy controls; it decreased in patients with disseminated tumor growth and was very low in chronic lymphatic leukemia .
When I first presented these data in September 1973 at an EORTC melanoma meeting in Brussels, Janine Pavie from Francois Kourilsky’s lab in Paris presented similar results using an MC assay. After the session, Kourilsky invited me to come to Paris to characterize the spontaneous CMC effector cell type in a joint effort. We agreed on a 4-month sabbatical, and, in spring 1974, I moved to Kourilsky’s lab at Hôpital St. Louis, Paris. I established the Cr-release assay for non-adherent cytotoxic effector cells while Janine Pavie continued the MC test. Wolf-Hervé Friedman helped us with rosetting techniques to detect T cells, Fc-receptor and C’-receptor positive lymphocytes. Jean-Pierre Cesarini and Regine Roubin took electro micrographs of the isolated “K/NK” cells and traced them in the peritumoral infiltrate of primary melanoma .
Still today, our NK cell isolation protocol is integrated in an automated NK-separation machine (Milteny Biotec) and consists of removal of erythrocytes and granulocytes by Ficoll gradient, removal of adherent and phagocytic cells by plastic adherence and iron bead phagocytosis, removal of T cells by antibodies (no longer by sedimentation of E-rosettes) and removal B cells by passage through anti-Ig columns. We were able to characterize the NK cells as a lymphocyte subset, which did not share the characteristics of adherent monocytes nor did it, express the E-rosette receptor of T cells or surface Ig of B-cells. These NK cells were slightly larger than T cells, had some cytoplasmic granules, expressed Fc-receptors and accounted for 5–8 % of peripheral blood lymphocytes (Fig. 4).
Our first results on the characterization of spontaneous killer lymphocytes appeared in autumn 1974 in the Behring proceedings of a symposium on immunological reactions to melanoma antigens . A detailed report followed in 1975 in the Journal of Immunology .
We called the enriched cell fraction tentatively “null cell fraction” ; in addition to NK cells, it contained some pluripotent stem cells , and as shown later, some plasmacytoid DCs, which upon stimulation produce large amounts of type-I Interferon . At the same time, Kiessling and Wigzell published similar findings on NK cells in mice  (without referring to our findings in humans). In 1980, Stuart Schlossman’s laboratory produced the first monoclonals specific for T and B cells and, in 1985, the first NK-cell specific monoclonal . When in 1983 Klaas Kaerre identified killer-inhibitory receptors on NK cells, which prevent the killing of MHC-I expressing autologous and allogeneic target cells, NK cells entered the mainstream of immunological research .
In autumn 1976, I defended my Habilitation in Clinical Immunology and Internal Medicine at Hannover Medical School with a cumulative study on “Forms of cellular cytotoxicity during transplant rejection and tumor growth – effector cells and mechanisms”. From then on, I worked preferentially in human disease models. In collaboration with Wolfgang Leibold from the Hannover Veterinary School, we studied NK cell activity in different human lymphatic tissues and in several animal species .
My interest shifted from tumor models to the physiology of the immune system and immunopathological variants of autoimmunity and immunodeficiency. The rationale behind was the hope to eventually find patients who totally lack NK cell activity. However, to our disappointment, NK cell activity was detectable in all of the patients with rheumatic diseases as well as in immunodeficiencies such as common variable immunodeficiency (CVID), DiGeorge syndrome, Wiskott-Aldrich syndrome and others. Finally, in 1982, we observed for the first time a total absence of NK cell activity in the blood of 4 out of 8 infants with SCID, an observation that later on was used to phenotypically classify SCID patients in T+/-, B+/- or NK+/- .
With these data, published in 1983, I applied in September 1983 for the vacant position of head of a small Division of Immunopathology at the Freiburg University Hospital. My application was successful. I was appointed in summer 1984 and got the chance to reorganize the Division, which under my predecessor Helmuth Schubothe had been focused on Immunohematology. From the very beginning, I changed the focus of the Division from Immunohematology to Rheumatology and Clinical Immunology. The patient ward was renamed after “Paul Ehrlich” and the building after “Paul Langerhans”, two immunologists who had made significant contributions to immunology during their time in Freiburg. In 1985, we started two outpatient clinics: one for inflammatory rheumatic diseases, collagenoses and vasculitis and one for immunodeficiencies (ID), notably CVID, X-linked agammaglobulinemia (XLA), combined ID, DiGeorge’s syndrome, and secondary IDs. For the ID outpatient clinic, it was paramount that Michael Schlesier and Ruth Dräger came with me from Hannover to set up an innovative cellular immunology program for phenotypic and functional testing of human lymphocytes.
At the same time, the HIV epidemic reached Freiburg, and, to cope with the clinical challenges, we were assigned an additional infectious ward, and Peter Vaith became the responsible clinical consultant.
In the non-HIV immunodeficiency outpatient clinic, a cohort of CVID patients increased steadily. CVID patients often present with simultaneous signs of antibody deficiency and autoimmunity – and here was the third immunological paradox that excited my scientific interest for years to come.
Since 1986, we have introduced many new monoclonal reagents and innovative techniques to phenotypically and functionally characterize lymphocytes from CVID patients. One of the new monoclonals was directed against the inducible costimulator ICOS. We obtained the ICOS reagents (antibody and recombinant ligand) in late 1999 from Richard Kroczek, RKI, Berlin. Ruth Dräger tested them on T cells from our CVID patient cohort and discovered in 4 patients ICOS-deficiency as the first monogenetic immunodeficiency in CVID . Ruth Dräger, Michael Schlesier and myself trained many talented medical students and young physicians in our immunodeficiency project, among them Stefan Feske (New York) who unraveled the Ca++ chanels on lymphocytes.
While the ID clinic made significant progress in diagnosing and classifying PIDs, the rheumatology section received increasing attention due to many new therapeutic options for inflammatory rheumatic diseases. To develop rheumatology research, we were lucky to obtain support from Claus Eichmann of the MPI. Together, we applied for a Clinical Research Unit of Rheumatology (KFR), which was granted by the DFG in 1988. As chairpersons, we hired Inga Melchers as an immune geneticist and Ulrich Krawinkel as a molecular biologist. Ulrich Krawinkel’s group with Gerd Pluschke, Peter Hemmerich and Anna von Mikecz studied the TCR Valpha and Vbeta repertoire of synovial and peripheral T cells from RA patients . In collaboration with Michael Schlesier’s group, they also analyzed TCR V-region sequences of autoreactive human T cell clones with specificity for U1-RNP and two novel autoantigens: ribosomal L7 and MMP19 .
Inga Melchers’s group with Rita Rzepka, Ulrike Rudolphi, Bernd Lang, Harald Illges and others studied intensely MHC associations of different rheumatic diseases as well as the autoreactive B cell repertoire . In 1992, Ulrich Krawinkel accepted an offer for a chair of Immunology at the University of Konstanz. He continued to collaborate with us closely until his much too early and tragic death in 1999. In the same year, the last joint doctoral fellow Johannes Donauer published his thesis .
In 1993, Hermann Eibel, an experienced B cell immunologist who trained in the laboratories of Georges Köhler and Michael Reth, became the successor of Ulrich Krawinkel. He started with his post-doc Willi Aicher, and several doctoral fellows among them Christian Haas and Bodo Grimbacher, to work on transcription factor EgR1-expression in synovial fibroblasts. Then he became increasingly interested in the B cell aspects of our CVID cohort. When Klaus Warnatz returned from a post-doctorate in Denis Carson’s lab in San Diego, Hermann offered him a postdoctoral position in the KFR. Then, in 2000/01, the discovery of ICOS deficiency sparked great enthusiasm among many of us. Bodo Grimbacher, who had just returned from a post-doctorate in Jennifer Puck’s lab, Klaus Warnatz, Michael Schlesier, Hermann Eibel and myself decided to engage more deeply with the genetic and immunological research of primary antibody deficiencies. The first financial support for an immunological analysis of our CVID cohort came in 1997/8 from the Landesforschungs-Programm Baden-Württemberg. Then, Hermann Eibel, with the help of Anne-Marie Perner, initiated an EU-funded research consortium (IMPAD), which was supported from 2000–2004 and established a beautiful primary antibody deficiency (PAD) network throughout Europe.
IMPAD eventually paved the way in 2002 for our successful application of the SFB 620 “Immunodeficiency: Clinic and Animal Models”, which for the following 12 years interconnected immunodeficiency research in Freiburg between the University Hospital, the Medical Faculty Institutes and the MPI.
Finally, the SFB 620 was instrumental for our success in the national competition for the establishment of a Center for Chronic Immunodeficiencies (CCI) under Stephan Ehl’s brilliant leadership. All of you in the audience give daily testimony to how rewarding and necessary translational research in clinical immunology can be.
Keynote lecture given on occasion of the 14th Freiburg Immunology Meeting 2019.