Prologue

Innate Alloimmunity: Evolution and Revolution in Organ Transplantation

In this work, that is, Part 2 of the book Innate Alloimmunity, experimental and clinical data, as well as innovative conceptual notions and novel therapeutic strategies, reflecting an ongoing evolution and revolution in the field of organ transplantation, are collected, quoted, and discussed. In this regard, transplantology follows that has been and still is regarded as the leading basic science discipline in the field of organ transplantation.

Charles Janeway’s Pattern Recognition Hypothesis

As outlined in Part 1 (section 2.3.5), the revolution in immunology began with the exceptional article of the late Charles Janeway in 1989, entitled “Approaching the Asymptote? Evolution and Revolution in Immunology,” in which he proposed the famous “pattern recognition theory” [1]. This theory provided a conceptual framework for our current understanding of innate immune recognition and its role in the activation of adaptive . Before Janeway, there was the hypothesis, referring to adaptive immunity only, and stating that T and B equipped with many -specific receptors are able to distinguish just between self and nonself. Then, Janeway, on purely theoretical grounds, proposed that an innate must exist, in terms of an evolutionarily highly conserved rapid first line of host defense against pathogens, preceding adaptive immunity. Specifically, he proposed that the innate immune system determined the origin of recognized by T and B cells and instructed the latter to initiate the response if the antigen was of microbial origin. He further postulated that antigen-presenting cells (APCs) – later identified as dendritic cells (DCs) – were equipped with pattern recognition receptors (PRRs), that recognized unique features of microbial molecules derived from pathogens, the pathogen-associated molecular patterns (PAMPs). When PAMPs were present, for example, derived from an infection, then DCs were stimulated to activate T cells. Specifically, Janeway proposed that the costimulatory signal required for activation was inducible on APCs by conserved microbial products, thus placing the activation of adaptive immunity under the control of pathogen-sensing mechanisms. In a subsequent 1992 article, Janeway stressed again that the innate immune system evolved strictly to discriminate infectious nonself from noninfectious self [2]. Remarkably, only a few years later, following Janeway’s original suggestion, pattern recognition receptors, the Toll-like receptors (TLRs), were discovered! As described in full detail in Part 1, Subchapter 2.3, the pioneering work of these discoveries has to be attributed in particular to 3 researchers and their groups: Jules Hoffmann [3], Bruce Beutler [4], and Shizuo Akira [5].

XXVIII Today, Janeway’s original model is generally recognized to be largely correct. However, the concept turned out to be too simplistic, and importantly, it could not explain all immune responses. In particular, it could not explain the robust T-cell-mediated alloimmune response leading to allograft rejection, a process in the apparent absence of microbial infection. If, for example, the innate immune system is able to discriminate noninfectious self from infectious nonself, in order to respond to self with tolerance and to mount an immune response against nonself, then how is it possible that the same system is able to discriminate pathogenic microbes from commensal and other nonpathogenic microbes? As is well known, the adult human intestine contains trillions of bacteria that are not immunologically attacked and eliminated. So, there was a problem with Janeway’s model. The major questions raised included: (1) how does the innate immune system interpret the microbial environment, allowing the discrimination of harmful pathogenic from harmless nonpathogenic microbes; and (2) why does the immune system reject a noninfectious nonself organ, as in the way an allograft is rejected? Notably, only a few years later, in 1994, the answer to these questions came from 2 hypotheses, nearly simultaneously published, that showed a way out of this dilemma [6,7].

The Danger Hypothesis, the Injury Hypothesis, and Their Unification with Janeway’s Hypothesis

Polly Matzinger tackled Janeway’s model by proposing her famous “danger hypothesis” [6]. Her model (described in more detail in subsection 1.2.3.2), proposed on purely theoretical grounds, suggested that the primary driving force of the immune system is the need to detect and protect against danger. However, in this model, danger equals tissue destruction or tissue injury. Our “injury hypothesis” (described in more detail in section 1.2.2), proposed on the basis of statistically significant data from a clinical trial in kidney transplant patients [7], postulated that it is the primary injury to an allograft that, in addition to its foreignness, initiates and induces an adaptive alloimmune response. Thus, the 2 hypotheses postulated the same underlying scenario. Tissue injury, that is, the injurious inflammatory tissue environment, alerts the immune system and is a mandatory prerequisite to mounting an efficient immune response against foreign antigens.

The unification of the 2 models, the pattern recognition hypothesis on one hand and the danger/injury hypothesis on the other hand, seemed heavily opposed, but a growing number of recent reports now show activation of innate immune events by injury-induced (normally hidden) host endogenous molecules [8-16]. Both models may synergize the quality and extent of the innate immune response. The unification of Janeway’s self-nonself model with the injury/danger model has reconciled many immunologists and led to new thinking in immunology. In fact, the injury/danger model can now explain why the innate immune system is able to mount an efficient immune response against harmful pathogenic microorganisms but not against harmless nonpathogenic microorganisms. The presentation of microbial antigens in

XXIX the context of pathogen-induced tissue injury triggers an efficient immune response, not simply the foreignness of microbial antigens. Likewise, the injury/danger model can also explain why the innate immune system is sometimes able to mount an efficient immune response against other nonself, for example, foreign tissue such as transplanted alloantigens, but sometimes not, for example, in the case of fetal semi-alloantigens [17]. The system distinguishes between an injured transplant (rejection) and a noninjured fetus (tolerance). Again, it is the presentation of alloantigens in the context of tissue injury that triggers an efficient alloimmune response and not simply the foreignness of allogeneic tissue, as reflected, for example, by an HLA-mismatch.

Accordingly, modern notions in immunology now hold that the immune system is directed against dangerous microbes, which cause inflammatory tissue injury. Apparently, the innate immune pathways not only scan the cellular environment for signs of invading pathogens, but also recognize the damage caused by them, by using those special pattern recognition receptors such as TLRs that are able to recognize PAMPs. Likewise, as comprehensively described in Part 2, increasing evidence suggests that the innate immune defense system is not only directed against pathogen-mediated tissue injury but against any tissue injury, including allograft injury. In fact, growing evidence supports the notion that allograft injury induces innate immune events that precede adaptive alloimmunity, where donor-derived and recipient-derived DCs translate innate immunity to adaptive alloimmunity. In this scenario, PRRs recognize injury-induced host endogenous molecules, the damage-associated molecular patterns (DAMPs). Notably, the term “DAMPs” was coined by our group [18] and subsequently Matzinger’s group [19] (compare also, Introduction 3.1).

Current Notions in Modern Immunology and DAMPs

Current notions in modern immunology hold that DAMPs play a central and crucial role in the initiation of adaptive immune responses. Via different cellular and molecular pathways, DAMPs lead to the generation of fully immunostimulatory DCs that promote adaptive T-cell-dependent immunity, a scenario that is dependent on DAMP-induced/ mediated (1) efficient antigen processing and /cross-presentation; (2) up-regulation of costimulatory molecule expression; and (3) creation of an inflammatory milieu (for recent reviews, see 14-16).

As outlined in Part 1 of the book, our understanding has grown tremendously of the cellular and molecular mechanisms by which the innate immune system molecules sense specific molecular patterns from components of invading organisms of both bacterial and viral origin. Various PRRs are now known to be part of this important sensing system that allows the host to mount an effective immune response to eliminate microbes and establish an effective adaptive immune response for long-lasting immunity. What is fascinating today is the realization that certain PRRs, which were postulated more than 10 years ago, also detect various injury-induced endogenous DAMPs in the absence of microbial infections, including DAMPs generated in injured allografts. There seems to

XXX be a complex orchestration of and collaboration between various membrane-bound and cytosolic PRRs on one side and their cognate DAMPs on the other side, which result in vigorous injury-induced innate immune responses. Most likely, this complex orchestration and collaboration of intragraft PRRs and DAMPs finely regulate the maturation process of donor-derived and recipient-derived DCs, a process that is associated with the acquisition of immunostimulatory capacities. There is no doubt anymore. During recent years, we have continued “approaching the asymptote” and are currently experiencing an ongoing evolution and revolution in organ transplantation, a topic covered by this book.

XXXI References to the Prologue

1) Janeway CA Jr. Approaching the asymptote? Evolution and revolution in immunology. Cold Spring Harb Symp Quant Biol 1989; 54: 1-13. 2) Janeway CA Jr. The immune system evolved to discriminate infectious nonself from noninfectious self. Immunol Today 1992; 13: 11-16. 3) Lemaitre B, Nicolas E, Michaut L, et al. The dorsoventral regulatory gene cassette spatzle/Toll/cactus controls the potent antifungal response in Drosophila adults. Cell 1996; 86: 973 983. 4) Poltorak A, He X, Smirnova I, et al. Defective LPS signaling in C3H/ HeJ and C57BL/10ScCr mice: Mutations in Tlr4 gene. Science 1998; 282: 2085–2088. 5) Takeda K, Kaisho T, Akira S. Toll-like receptors. Annu Rev Immunol 2003; 21: 335 376. 6) Matzinger P. Tolerance, danger, and the extended family. Annu Rev Immunol 1994; 12: 991-1045. 7) Land W, Schneeberger H, Schleibner S, et al. The beneficial effect of human recombinant superoxide dismutase on acute and chronic rejection events in recipients of cadaveric renal transplants. Transplantation 1994; 57: 211-217. 8) Oppenheim JJ, Tewary P, de la Rosa G, Yang D. Alarmins initiate host defense. Adv Exp Med Biol 2007; 601:185-194. 9) Bianchi ME. DAMPs, PAMPs and alarmins: all we need to know about danger. J Leukoc Biol 2007; 81: 1-5. 10) Beutler B. Neo-ligands for innate immune receptors and the etiology of sterile inflammatory disease. Immunol Rev 2007; 220:113-128. 11) Lotze MT, Zeh HJ, Rubartelli A, et al. The grateful dead: damage-associated molecular pattern molecules and reduction/oxidation regulate immunity. Immunol Rev 2007; 220: 60-81. 12) Klune JR, Dhupar R, Cardinal J, et al. HMGB1: endogenous danger signaling. Mol Med 2008; 14: 476- 484. 13) Zhang X, Mosser DM. activation by endogenous danger signals. J Pathol 2008; 214: 161- 178. 14) Kono H, Rock KL. How dying cells alert the immune system to danger. Nat Rev Immunol 2008; 8: 279- 289. 15) Manfredi AA, Capobianco A, Bianchi ME, Rovere-Querini P. Regulation of dendritic- and T-cell fate by injury-associated endogenous signals. Crit Rev Immunol 2009; 29: 69-86. 16) Martinon F, Mayor A, Tschopp J. The inflammasomes: guardians of the body. Annu Rev Immunol 2009; 27: 229-265. 17) Kanellopoulos-Langevin C, Caucheteux SM, Verbeke P, Ojcius DM. Tolerance of the fetus by the maternal immune system: role of inflammatory mediators at the feto-maternal interface. Reproductive Biology and Endocrinology 2003; 1:121. 18) Land W. Allograft injury mediated by reactive oxygen species: from conserved proteins of Drosophila to acute and chronic rejection of human transplants. Part III: Interaction of (oxidative) stress-induced heat shock proteins with Toll-like receptor-bearing cells of innate immunity and its consequences for the development of acute and chronic allograft rejection. Transplantation Rev 2003; 17: 67- 86. 19) Seong SY, Matzinger P. Hydrophobicity: an ancient damage-associated molecular pattern that initiates . innate immune responses. Nat Rev Immunol 2004; 4: 469-478

XXXII 1 Early Appreciation in the 1990s: The Injured Allograft as an Acutely Inflamed Organ and the First Clues to the Existence of Innate Alloimmunity

1.1 Introduction

Solid organ transplants are, by definition, injured organs, but this was not taken as a condition that would play any role in allograft rejection. It took until 1994, when we published a clinical trial in which we discussed, for the first time, the possibility that an initial allograft injury, the postischemic reperfusion injury (ie, ischemic reperfusion injury, abbreviated IRI, an abbreviation that will predominantly be used throughout this book), led to the development of acute allograft rejection and contributed to the development of chronic allograft rejection.

Today, in 2009, it has become clear that, like all multicellular animals, human beings have evolved a biological warning and defending system that detects any tissue injury, including any injury to allografts: the innate immune system. In vertebrates, this system subsequently activates the adaptive immune response. In the transplant setting, 2 innate immune systems, that of the donor and that of the recipient, activate the adaptive alloimmune response in the recipient, a scenario, for which, in 2002, I coined the term Innate Alloimmunity.

Using clinical data from the Munich superoxide dismutase (SOD) trial published in 1994 [1,2] and with respect to the Injury Hypothesis, we wrote several review articles in the 1990s (1996 [3,4] and 1999 [5-7]), in which we discussed in depth the first clues of the existence of innate alloimmunity (interestingly, without actually mentioning the name “innate immunity” expressis verbis). These reviews included experimental and clinical findings and observations published shortly before and during this time in support of our notion that any injury to the allograft, in particular oxidative injury, led to the initiation of acute allograft rejection and contributed to the development of chronic allograft rejection.

3 I consider it interesting enough to refer to these review articles in this chapter, because they reflect the early appreciation of allograft injury as a dominant predictor of allograft outcome nicely. Moreover, they appear to impressively mirror the overall knowledge in those days about the first scientific clues of the existence of innate alloimmunity, although almost all those supporting findings and observations cited in these papers were obtained in other research fields not directly related to innate alloimmunity. As a matter of fact, our articles in the 1990s attempted to collect and review all those publications deriving from different research fields and to order them in terms of a reasonably intrinsic interrelationship, so that they presented plausible arguments in favor of the notion that it was primarily the (oxidative) allograft injury and not the allograft foreignness that initiated those pathways that determined acute and chronic outcomes of allografting.

Accordingly, in the following, at first, I would like to remind the reader of some historical aspects about the Injury Hypothesis posed in 1994 and a few similar concepts subsequently published, and then continue to describe the impact of IRI on early events after organ transplantation as we saw it during this time [3,5,6].

Let me add here that the second major impact of this canonical injury on allograft outcome, that is, on nonspecific, late, chronic events after organ transplantation as we understood it in the 1990s [4,7], will be mentioned in Subchapter 6.2 in more detail.

1.2 Historical Remarks

1.2.1 The Munich superoxide dismutase trial story

1.2.1.1 General remarks

Indeed, the first clue of the existence of innate alloimmunity came from clinical observations made during the Munich SOD trial [1]. This prospective, randomized, double-blind, placebo-controlled trial was performed at the Division of Transplant Surgery of the Klinikum Großhadern between January 1987 and October 1988. One hundred seventy-seven patients who received a renal transplant from a deceased donor were entered into this trial to receive either recombinant human superoxide dismutase (rh-SOD) or placebo at the time of operation.

The study was a retrospective analysis of clinical data collected during the originally performed, prospective, randomized, double-blind trial. The endpoints of this trial, therefore, were chosen prospectively as far as assessment of early graft function was concerned and not chosen prospectively but chosen before this analysis (“retrospectively prospective”) as far as observations of all early and late clinical events after transplantation were concerned.

4 Figure 1.2.1. Data from the Munich SOD trial: A prospective, randomized, double-blind, placebo- controlled trial in kidney-transplant patients under CsA-based immunosuppressive treatment.

In this trial, kidney-transplant, CsA-treated patients who received 200 mg rh-SOD during surgery before allograft reperfusion (blue columns) showed the following data, in comparison with placebo- treated patients (yellow columns): (1) statistically significant reduction in acute rejection episodes to 18.5%; (2) statistically significant reduction in acute irreversible allograft rejection to 3.7%; (3) statistically significant improvement of the 5-year allograft survival time. Interestingly, the beneficial long-term effect of just 1 intraoperative injection of rh-SOD was still demonstrable 8 years after transplant. In contrast, as shown on the right side of the figure, the original aim of the study, improvement of early allograft function, was not reached as indicated by the incidence of ATN observed and the number of hemodialysis sessions required postoperatively in both patient cohorts; no difference whatsoever was noted. Abbreviations: ATN, acute tubular necrosis; CsA, Cyclosporine A; No. hemodialysis pop., number of hemodialysis sessions postoperatively; rh-SOD, recombinant human superoxide dismutase; rej, rejection. Source: Figure traced and slightly modified from a figure published in: Land W, Messmer K. The impact of ischemia/reperfusion injury on specific and non-specific, early and late chronic events after organ transplantation. Late chronic events. Transplant Rev 1996; 10: 236-253.

The substance used, rh-SOD, was identical to human SOD, obtained from cultivated yeast cells using a recombinant DNA technique manufactured by Grünenthal, Aachen, Germany, licensed by Chiron, Los Gatos, CA, USA.A single dose of 200 mg rh-SOD or 200 mg sucrose as a placebo was given intravenously during surgery as a rapid infusion, 2 to 10 minutes before reperfusion of the renal transplant.

5 Immunosuppressive induction treatment in non-high risk patients consisted of a low-dose cyclosporine (CsA)-based triple drug regimen (CsA, azathioprine, and steroids); in patients with increased immunologic risk, polyclonal antilymphocyte preparations were added to this triple protocol (quadruple drug induction treatment). Immunosuppressive maintenance treatment consisted either of CsA monotherapy (mean CsA dose: 4 mg/kg/body weight/day) or CsA/steroids double-drug therapy. The results obtained showed that rh-SOD exerted a beneficial effect on acute rejection events, as indicated by a statistically significant reduction in the incidence of (1) first acute rejection episodes from 33.3% in controls to 18.5%, as well as (2) early irreversible acute rejection from 12.5% in controls to 3.7%. In addition, with regard to long-term outcome, there was a statistically significant improvement of the 4-year graft survival in rh-SOD–treated patients to 74% (with a projected half-life of 15 years) compared with 52% in controls (with an extrapolated half-life of 5 years). Remarkably, this beneficial long-term effect could be demonstrated even 8 years after the administration of rh-SOD (Figure 1.2.1).

1.2.1.2 A few anecdotes

There are several anecdotal but remarkable stories around the Munich SOD trial. Of interest, for example, is the launching phase of the clinical trial that started in 1986. The first step came from my colleague and friend, Konrad Messmer (Figure 1.2.2), who, at that time, was the Chairman of the Institute of Experimental Surgery at the University of Heidelberg, Germany. In the late 1960s and early 1970s, we worked together at the Institute of Surgical Research in Munich under Professor Walter Brendel and knew each other very well. In Heidelberg, pathophysiological studies on a model of IRI in hamsters represented a main topic of research in his institute.

Together with Karl Arfors, International Pharmacia Experimental Medicine, of Uppsala, Sweden; Leopold Flohe and Hans Barth from Grünenthal, Aachen, Germany; and Gianfranco Rutili, Pharmacia, Uppsala, Konrad Messmer visited me in my office at the Division of Transplantation Surgery, Klinikum Großhadern, University of Munich. At that time, the Division of Transplantation Surgery belonged to the 10 leading transplant centers worldwide, with a rate of more than 200 kidney transplants per year.

The purpose of this visit was to discuss a clinical trial in kidney-transplant patients by testing a new compound, the free radical scavenger, rh-SOD. The purpose of the planned trial was to treat renal allograft reperfusion injury with rh-SOD, intra-arterially administered to the donor organ during surgery shortly before the start of reperfusion (later on, we switched to intravenous application via central venous catheter). The aim of the study (the primary endpoint) was improvement

6 of early renal allograft function as a consequence of impairment/minimization of IRI to the allograft.

At a round table with experts in my office, after a relatively brief discussion, there was soon agreement on the initiation and performance of such a single-center, prospective, randomized, double-blind and placebo-controlled clinical trial in kidney- transplant patients receiving CsA-based immunosuppressive therapy. It was further decided to ask Gregory Bulkley from the Johns Hopkins University in Baltimore, MD, USA, to participate in the trial as an advising and counseling expert. He already had experience with successful use of rh-SOD in a pig model of IRI in kidney allografts.

The background of the planned study, however, was different. The reperfusion of the renal allograft should be used as a model for successful resuscitation after cardiac arrest. A potentially significant improvement by rh-SOD of early renal function via successful treatment of IRI was planned as a surrogate marker for improved function of vital organs, such as the heart and lung, after successful resuscitation with an rh-SOD–containing resuscitation solution. In fact, the use of rh-SOD as a compound to be added to such solutions was the real marketing interest, and this would have negative consequences for the future development of the drug, as explained in the following.

After all the necessary preparatory steps were carefully done, the clinical trial started at my transplant division (ward H5) in 1987. A young fellow, Helmut Schneeberger, who was not involved in the patients’ care at that time, was selected to monitor the ongoing trial as the only unblinded person. He also was responsible

Figure 1.2.2. Konrad Messmer, MD, PhD.

Professor Dr. med. Dr. h.c. mult. Konrad Messmer is Emeritus Professor of Experimental Surgery at the Ludwig-Maximilians- University in Munich (LMU) and a renowned scientist in the field of surgical research. Between 1963 and 1981, he was educated at the Institute of Surgical Research at the LMU in Munich; in 1981, he became the Chairman of the Department of Experimental Surgery at the University of Heidelberg, Germany; and in 1990, he became the Chairman of the Institute of Surgical Research at the Medical School, LMU in Munich. He has been the recipient of numerous awards and honors and served as the President as well as Honorary Member of many international and national scientific societies. Currently, he is the Dean of the Section of Medicine at the European Academy of Science and Arts. He has published more than 700 scientific articles in main surgical and biomedical sciences journals, proceedings, and textbooks. In addition, he served as the Editor-in-Chief of European Surgical Research (Karger), Progress in Applied Microcirculation (Karger), and presently for TATM Transfusion Alternatives in (LMS group, Paris). (Wiley Blackwell)

7 for correct practical performance of the study and for collecting all the data of the study patients. For this purpose, he developed a computerized database only for this clinical trial. Other colleagues working in the Division of Transplant Surgery at that time actively participated in the study by being mainly responsible for the care of the study patients, including Stefan Schleibner, an excellent and profoundly experienced transplant physician, as well as Wolf-Dieter Illner and Dietmar Abendroth, 2 immensely experienced, excellent, and skillful transplant surgeons.

Unfortunately, the results of the clinical trial were not what we had hoped for. The primary endpoint of the study, early renal allograft function, proved not to be statistically significantly improved at the end of the trial. In fact, in view of this disappointing study outcome, there was a deep depression in all of the colleagues involved in the trial. The bitter consequence was that the participating pharmaceutical industry declined to continue such trials, for example, with the use of other dosages of rh-SOD.

Fortunately, however, this was not the end of the SOD trial, and its course finally converted to a happy end. This is the story.

At the 1990 Congress of The Transplantation Society held in San Francisco, I was invited to deliver a lecture on long-term results in kidney transplantation. I asked Helmut Schneeberger to look at all the long-term data so far collected in our database, including the data obtained from those patients who were previously enrolled in the SOD trial. Interestingly, he found that the SOD-treated patients experienced fewer acute rejection episodes and better long-term survival. Over the next years, a careful and detailed retrospective analysis of the trial data confirmed these preliminary observations, which led to the publications mentioned above.

Of course, at that point, we wanted to restart similar clinical trials with rh-SOD in our patients, and we planned a multicenter trial. However, to our disappointment, the pharmaceutical industry that participated in the Munich SOD trial lost any marketing interest. This kind of reluctance to produce and develop an antioxidative agent for clinical use in transplant patients has not changed today, although, in the meantime, I approached the big pharmaceutical companies involved in the production of immunosuppressive agents and tried to convince them to start development of an antioxidative agent. Obviously, the time-limited use of an antioxidant—in comparison with the long-term use of immunosuppressive drugs currently applied to transplant patients—did not evoke any marketing interests (see also below, Subchapter 7.6).

8 1.2.2 The Injury Hypothesis in relation to acute and chronic allograft rejection

1.2.2.1 General remarks

Clearly, nobody in our group could explain our clinical observations based on experimental data available at that time and published earlier. What was clear to all of us was the fact that the attenuation of IRI to allografts with the use of a free radical scavenger resulted in the reduction of acute rejection episodes and improvement of long-term allograft outcome. Certainly, the inversion of an argument was easy to make: the nonspecific injury to an allograft contributes or even leads to acute rejection and contributes to the development of chronic rejection. But what were the underlying mechanisms? In this situation and in view of clinical data that were produced by an interdisciplinary team consisting of transplant surgeons, researchers in the field of reperfusion injury, and representatives of the pharmaceutical industry, I dared to pose a working hypothesis, today known as the Injury Hypothesis, which was modified and extended several times later and which is referred to in the following.

1.2.2.2 Original version in 1994

The Injury Hypothesis as posed in Transplantation in 1994 held that the reactive oxygen species (ROS)-mediated reperfusion injury to an allograft, in addition to its degree of foreignness, initiated and induced the adaptive alloimmune response resulting in acute rejection – predominantly via activation of antigen-presenting cells (APCs). Furthermore, in this Transplantation article, I discussed the possibility that the ROS-induced injury contributed to the development of alloatherosclerosis of donor organ vessels (chronic rejection) via endothelial injury-induced proliferation of smooth muscle cells [1]. The original figure and legend of the working hypothesis as published in Transplantation is reshown here (Figure 1.2.3).

Interestingly, as indicated by a frame within the figure, we already described an immunologic system as an entity in its own right, activated by allograft reperfusion injury, and subsequently leading to an (adaptive) immune response. In the center of the system, we postulated up-regulated activity of APCs. However, we missed an opportunity to call this system “innate immunity.”

1.2.2.3 Modification and extension in 1996-1999

In 1996, together with Konrad Messmer, we discussed for the first time a role for reperfusion injury-induced secretion of as mediators of innate immunity (in this article, in fact, we used the term natural immunity instead of innate immunity)

9 Figure 1.2.3. The working hypothesis (original text) – as posed in 1994 in view of the beneficial effect of rh-SOD on both early and late rejection events following kidney transplantation.

“One could speculate that free radical-mediated reperfusion injury of the graft has the potential to up-regulate HLA-DR expression, adhesion molecule expression, and the phagocytic activity of APCs, leading to an increased graft immunogenicity. By ameliorating the reperfusion injury, rh-SOD interferes with “up-regulated” immunogenicity, and thus reduces the incidences of immunological graft loss and acute rejection episodes. With regard to the development of chronic obliterative rejection arteriosclerosis, free radical-mediated reperfusion injury causes acute endothelial damage contributing (together with chronic -mediated endothelial damage) to smooth muscle cell proliferation in the intima of graft vessels. By ameliorating this acute reperfusion-induced endothelial injury, rh-SOD reduces stimuli (possibly growth hormones) for smooth muscle cell proliferation, and thus leads to retardation of the process of chronic obliterative transplant vasculopathy.” Abbreviations: Ag, antigen; APC, antigen-presenting cell; rh-SOD, recombinant human superoxide dismutase. Source: A redrawn, slightly modified figure from: Land W, Schneeberger H, Schleibner S, et al. The beneficial effect of human recombinant superoxide dismutase on acute and chronic rejection events in recipients of cadaveric renal transplants. Transplantation 1994; 57: 211-217. and their contribution to the establishment of an inflammatory milieu in allografts [3]. We wrote: “Cytokines appear to play a critical role in the response to injury... In allogeneic organ transplantation, the dominant role of cytokines has been seen in their capability to mediate and regulate specific immune reactions, especially at the efferent arc of the response. However, today, an additional role must be discussed: their contribution as proinflammatory cytokines to the establishment of

10 an acute inflammatory state in postischemically reperfused organ transplants. In particular, the group of mediators of natural immunity are addressed here... Thus, the initial events leading to an allograft rejection does not appear to be primary specific antigen-recognition by T cells [3].” Inversely, it can be indirectly concluded from this statement that the initial events leading to allograft rejection appear to be a primary nonspecific recognition of other non-antigen molecules by other non-T cells. However, we did not mention this, although, retrospectively considered, we were very close to the early mechanisms of innate alloimmunity.

In addition, a role for injury up-regulated costimulatory molecules on APCs such as dendritic cells (DCs), as suggested by the coincident expression of adhesion molecules, was also noted in this article [3]. In fact, in the early 1990s, it was believed that the engagement of adhesion molecules such as intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1), and leukocyte/lymphocyte function associated antigen-3 (LFA-3) during T-cell receptor (TCR)-peptide/ major complex (MHC) interactions was of utmost importance in providing a costimulatory signal to T cells that led to appropriate and efficient antigen-driven activation of T cells.

Moreover, in 1996, we continued to discuss the Injury Hypothesis with regard to chronic allograft dysfunction/chronic rejection [4]. In this review, we described the impact of the primary nonspecific IRI on late chronic events leading to chronic transplant failure (chronic rejection), exemplified herein by—as it was called at that time—chronic obliterative transplant arteriosclerosis. In this respect, we coined the terms alloatherogenesis and alloatherosclerosis by proposing that all risk factors contributing to native atherosclerosis, that is, autoatherosclerosis, may also contribute to atherosclerosis of the allograft vessels, that is, alloatherosclerosis, and that, in this situation, additional ongoing subclinical, chronic cell/antibody- mediated alloimmune events are responsible for the accelerated development of alloatherosclerosis in comparison with the more slowly developing autoatherosclerosis. Also in this paper and others, we formulated the “Two Endothelial Cell Hit” hypothesis in the development of alloatherogenesis. This hypothesis held that 2 pivotal injurious events to an allograft, the primary nonspecific reperfusion injury and the subsequent specific injury mediated by an acute rejection episode, were of major importance for the accelerated development of alloatherosclerosis, the underlying mechanism in both situations being the induction of endothelial dysfunction or activation, leading to subsequent / adhesion molecule cascade-induced vessel wall remodeling.

Moreover, in this article, we further discussed the possibility that chronic allograft dysfunction, as indicated by the development of alloatherogenesis, represented a result of multifactorial antigen-dependent and antigen-independent events that can injure the endothelium, the latter events including alloatherogenic factors like

11 high donor age, cytomegalovirus (CMV) infection, hypertension, hyperlipidemia, diabetes mellitus, and long-term use of steroids and/or CsA.

This “Multiple Endothelium Hit” hypothesis, which served as an explanation for the well-known rapid development of alloatherosclerosis in certain categories of transplant patients, gained much credibility in light of innate immunity (eg, multiple activation of Toll-like receptor [TLR]4-bearing endothelial cells (ECs) after recognition of damage-associated molecular patterns (DAMPs) generated in the course of simultaneously or subsequently occurring, multiple but different, chronic- repetitive endothelial injuries). In fact, the injuries were believed to be responsible for the development of chronic allograft dysfunction as a result of multiple activation of donor-derived and recipient-derived innate immune cells, associated with the development of alloatherosclerosis and allofibrosis.

In 1999, I contributed 3 additional articles to this issue [5-7]. In particular, I discussed the possibility that the main mechanism of IRI-induced acute allograft rejection was in the initial T-cell alloactivation via up-regulation of (DC)- mediated stimulation, costimulation, and adhesion. In fact, I sketched the scenario that the nonspecific injury contributed to T-cell alloactivation via the activation of DCs, which, shortly before that time, had been recognized as the main regulators and modulators of immunity. The IRI up-regulated expression of MHC products, costimulatory molecules, and adhesion molecules on DCs may be the underlying mechanisms. The IRI, by leading to the disintegration of the donor organ’s vessel walls, may facilitate communication among DCs and T cells and the travel of these cells into and out of the graft. DCs and T cells may meet either in the allograft itself or in the secondary lymphoid tissue of the recipient.

In regard to chronic allograft dysfunction, I extended our concept by proposing 3 mechanisms by which this nonspecific injury contributed to the initiation and progression of chronic rejection: (1) an ROS-mediated disturbance of the microcirculation in renal allografts, which may also contribute to the existence of anaerobic metabolism, and to delayed graft function as a risk factor for chronic dysfunction; (2) an ROS-mediated acute endothelial injury at the allograft vessel walls, which led, in analogy to the response-to-injury hypothesis, to alloatherogenesis as a characteristic sign of chronic allograft dysfunction; and (3) an ROS-mediated global injury to the allograft, which represented, in analogy to the danger hypothesis, the primary key event that initiated alloimmune responsiveness via T-cell activation by facilitating costimulatory processes. These clinically manifesting or subclinically occurring immune events, which were initiated by IRI as the driving force, contributed specifically to the development of chronic allograft dysfunction.

12 1.2.3 Similar hypotheses posed in the 1990s (1994-1996)

1.2.3.1 General remarks

In the time between 1994 and 1996, a phenomenon occurred that has repeatedly been observed in many other scientific disciplines over hundreds of years: the simultaneous or nearly simultaneous publication of similar papers on the same topic by independently working scientists. Notably, in 1994, only 3 months after our publication in Transplantation, Polly Matzinger published her famous “Danger Hypothesis,” which was very similar to our Injury Hypothesis. In addition, Nick Tilney in 1999 and Phil Halloran in 1991 and 1996 published similar hypotheses with regard to a relation between ischemia and alloimmunity. In the following, these articles are briefly referred to [8-11].

1.2.3.2 Polly Matzinger

In Part 1 (Section 2.3.5),I outlined that Charles Janeway proposed that the old immunologic paradigm of “self”-“nonself” discrimination had reached the limits of its usefulness. Janeway argued that the innate immune system was the real gatekeeper of whether the immune system responded or not. He also argued that the innate immune system used ancient pattern recognition receptors (PRRs) to make these decisions, recognizing a pathogen by its unchanging characteristics.

In her 1994 article titled “Tolerance, Danger, and the Extended Family” [8] (Figure 1.2.4), Matzinger went several steps further by laying out the idea that APCs respond to danger signals - most notably from cells undergoing injury, stress, or “bad cell death” (as opposed to apoptosis, controlled cell death). The alarm signals released by these cells let the immune system know that there was a problem requiring an immune response. She argued that T cells and the immune response they orchestrate occurred not because of a neonatal definition of “self,” as in the previous model, nor because of ancient definitions of pathogens, as in Janeway’s argument, but because of a dynamic and constantly updated response to danger as defined by cellular damage. Indeed, the Danger Model is quite broad, covering topics as diverse as transplantation, maternal/fetal immunity, , cancer treatments, and vaccines, but Matzinger pointed out that although it offered an explanation of how an immune response was triggered and how it ended, it did not (yet) offer an explanation of why the immune system responded in different ways to different situations. In her review article, she hypothesized that tissues sent signals to the immune system that determined the immune response appropriate for that tissue, and her lab is still working on experiments to test that hypothesis. The Danger Model has meanwhile gained full universal acceptance, although in earlier times, some immunologists, following Janeway’s ideas more directly, believed that the immune response was

13 mainly fueled by pathogen-associated molecular patterns (PAMPs) expressed/ released by pathogens and recognized by evolutionarily-conserved PRRs expressed on innate immune cells.

Figure 1.2.4. Polly Matzinger, PhD

Professor Polly Matzinger is now a Section Head at the National Institute of and Infectious Diseases, National Institutes of Health, Bethesda, MD, USA. Matzinger and her coworkers refer to their laboratory lab’s as the Ghost Lab when listing their affiliation in papers. The nickname was given to the lab by her colleagues when Matzinger first arrived at the NIH, because she spent the first 9 months studying a new field, chaos theory, that she thought might apply to the immune system, and the lab sat empty. The official name of her laboratory is the T-Cell Tolerance and Memory Section of the Laboratory of Cellular and Molecular Immunology. The laboratory typically hosts 6 to 8 post-doctoral and 1 to 2 post-baccalaureate researchers at any one time. From this knowledge pool up to 3 research articles are published each year, including original research articles and theoretical pieces. A majority of the publications are her single-author published works, theoretical in nature and published in highly cited journals.

1.2.3.3 Tilney’s group

Experimental data suggesting a relation between IRI and alloresponsiveness were also published in the 1990s by the group of Nicholas L. Tilney at the Harvard Medical School, Department of Surgery, Brigham and Women’s Hospital in Boston, MA, USA (Figure 1.2.5).

In 1999, the group published a paper on “Exploitation of the continuum between early ischemia/reperfusion injury and host alloresponsiveness: indefinite kidney

14 allograft survival by treatment with a soluble P-selectin ligand and low-dose cyclosporine in combination” [9]. The group had already previously shown that sPSGL, a soluble glycoprotein ligand for P-selectin and E-selectin, reduces the events associated with ischemia/reperfusion injury of the kidney. In this study, the researchers attempted to differentially modulate early inflammatory influences and later host alloresponsiveness in an Lewis Brown Norway F1 (LBN-F1) rat renal graft model by treatment with sPSGL in combination with a marginally effective dose of CsA. Four experimental groups were studied: group 1 = control animals receiving vehicle only; group 2 = sPSGL alone; group 3 = low-dose CsA; and group 4 = sPSGL plus low-dose CsA. Grafts were removed at 1, 3, 5, and 7 days (n = 3/time point) and assessed by histology, immunohistology, and reverse transcriptase-polymerase chain reaction (RT-PCR). Long-surviving grafts in recipients of groups 3 and 4 were followed functionally for more than 28 weeks. In these experiments, the following results were obtained. Graft function was prolonged indefinitely in recipients in group 4, all of which survived for more than 200 days. In contrast, survival of animals in groups 1 and 2 was not increased substantially, whereas only 4 of 17 animals in group 3 (23.5%) survived more than 24 days (P < .01). Five days after engraftment, necrosis was relatively minimal in group 4 organs but pronounced in the other groups. By immunohistology, numbers of infiltrating CD4+ and CD8+ T cells and ED1+ were significantly diminished in group 4 allografts compared with those of the other groups. Serial assessment of chemokine and cytokine mRNA expression confirmed these findings. The long-term effects of CsA treatment alone were compared with those of sPSGL in combination with CsA. Proteinuria remained virtually absent in group 4 recipients. Morphologically, the few long-surviving grafts in group 3 showed signs of chronic rejection; those in group 4 remained relatively normal. Tilney’s group concluded that, although treatment with sPSGL alone

Figure 1.2.5. Nicholas L. Tilney, MD.

Professor Nicholas L. Tilney is the Francis D. Moore Professor of Surgery at Harvard Medical School and a Senior Surgeon at Brigham and Women’s Hospital. He was educated at Harvard College and Cornell University Medical School. He trained in surgery at The Peter Bent Brigham Hospital and in cellular immunology at Oxford University. He received additional training in surgery and transplantation biology at the University of Glasgow. From 1976 to 1992, he was Director of the Renal Transplant Division at the Brigham. He also served as Director of the Surgical Research Laboratory at Harvard Medical School, combining clinical practice with research. His research in transplantation biology was continuously funded by the National Institutes of Health between 1974 and 2002. He has been the recipient of numerous awards and honors and was President of The Transplantation Society, 2006-2008. He is the author of Transplant: From Myth to Reality (Yale Press) and A Perfectly Striking Departure: Surgeons and Surgery at the Peter Bent Brigham Hospital 1912-1980 (Science History Publications).

15 showed no apparent influence on the acutely rejecting transplants, at least by the parameters examined in this study, it produced indefinite survival of kidney grafts when used in combination with low-dose CsA. Clearly, the data published already in 1999 supported impressively the influence of early nonspecific injury on later immunologic rejection.

1.2.3.4 Halloran’s group

Philip Halloran published similar theories in the 1990s. In 1990, Shoskes, Parfrey, and Halloran wrote an article in Transplantation on “Increased major histocompatibility complex antigen expression in unilateral ischemic acute tubular necrosis in the mouse” [10] (Figure 1.2.6). This paper was dedicated to the clinical problem of acute tubular necrosis. In fact, at that time, it was already well known from many studies of renal transplant recipients that ATN predisposed to a higher rate of graft loss, apparently due to rejection, but the mechanism of this effect was unknown. The authors suggested in this article that one possibility was increased immunogenicity of the graft. To study this possibility, they examined the expression of MHC antigens in kidneys damaged by ischemia, using a mouse model of ischemic ATN, produced in the left kidney of male CBA mice by temporary clamping of the vascular pedicle for up to 60 minutes. Class I and class II MHC expression was quantified by the extent of binding of monoclonals in radioimmunoassay after 1 to 35 days in both kidneys. Major histocompatibility complex induction was localized by indirect immunoperoxidase staining. Specific steady-state mRNA for beta 2 microglobulin and class II was quantified by Northern blotting using [32] P-labeled probes. Changes in MHC expression were assessed by comparing the ischemically injured left kidney to the control right kidney. The investigators found that, by day 1, ATN was evident

Figure 1.2.6. Philip Halloran, MD, PhD, OC

Professor Philip Halloran received his MD from the University of Toronto in 1968. He trained in internal medicine and nephrology in Toronto and completed his PhD in immunology and immunogenetics in London, England at the London Hospital/ University of London in 1976. From 1975 to 1987, he was a clinician and investigator at the Toronto Hospital. From 1987 to 2003, he was the Director of the Division of Nephrology & Immunology at the University of Alberta. He is currently the Director of the Alberta Transplant Applied Genomics Centre, a Professor in the Departments of Medicine and Medical Microbiology & Immunology at the University of Alberta, and the recipient of a Canada Research Chair in Transplant Immunology. In addition, he has been the recipient of numerous awards and honors. Dr. Halloran was the founding Editor-in-Chief of the American Journal of Transplantation, the official journal of the American Society of Transplantation and the American Society of Transplant Surgeons. The journal launched its first issue in May, 2001 and is now first in its field with an impact factor of 6.002.

16 by histology, but there was no change in MHC expression. By day 3, class I was increased in the left kidney by 3-fold to 6-fold over the right. In tissue sections, the class I increase was localized to tubular epithelial cells. Starting on day 7 and persisting to day 35, class II was increased by 1.5 to 3 times for the ischemic kidney over the control, primarily in interstitial cells but also in tubular cells. This increase in class II was associated with the appearance of Thy 1.2-positive cells in the interstitial areas. Increased antigen expression was preceded by increased steady-state mRNA. Halloran’s group concluded that unilateral ischemic ATN causes increased MHC expression in tubular cells and the accumulation of an inflammatory infiltrate, both of which may contribute to the increased rate of rejection and graft loss in ischemically injured kidneys.

In 1996, Shoskes and Halloran published an article on “Delayed graft function in renal transplantation: etiology, management and long-term significance” [11]. This contribution was devoted to the clinical problem of delayed allograft function after transplantation of a kidney from a deceased donor. At that time, this clinical complication was not uncommon and often associated with poor graft outcome. The investigators reviewed the biology of reperfusion injury and delayed graft function in renal transplantation, as well as its prevention, management, and long-term effects as published in the medical literature at that time. They found that delayed graft function was clearly associated with poor allograft survival and was caused by an interaction of ischemic and immunologic factors. Technical and pharmacological maneuvers were able to improve early function rates. The response to ischemic injury was complex, and probably able to increase graft immunogenicity and promote the chronic proliferative changes seen in chronic allograft nephropathy. The authors concluded that improvement in early renal function should be a primary goal in renal transplantation to enhance early and long-term results. They proposed that basic research into the injury response may yield insights into renal pathophysiology.

1.2.3.5 Concluding remarks

Of note, these early publications on a possible role of the initial injury in the development of an adaptive immune response (here, of the influence of an allograft injury on the development of an adaptive alloimmune response) were not necessarily accepted, and in particular not by the leading “conservative” immunologists and researchers in the field of (HLA) matching. In fact, the older and the more renowned they were, the more heavily they attacked this innovative concept. One fact that aggravated the isolation of the groups was that they did not intensely communicate with each other. For example, during the years 1994 to 1997, we did not know of Polly Matzinger’s danger hypothesis, and she was not aware of our work. Early recognition of the contributions of each of us would have made the

17 defending nature of our discussions with other colleagues much easier. However, this omission was corrected in the late 1990s.

Likewise, the “Multiple Endothelium Hit” hypothesis, which served as an explanation for the well-known rapid development of alloatherosclerosis in certain categories of transplant patients, was not broadly accepted as a valid foundation for the scenario of developing chronic allograft dysfunction.

Nevertheless, this theory has gained much credibility in light of innate immunity. One may now discuss that multiple activation of intragraft PRR-bearing ECs and/or myofibroblasts after recognition of DAMPs, generated in the course of simultaneously or subsequently occurring, multiple, chronic or repetitive endothelial and/or interstitial injuries, may lead to rapid development of chronic allograft dysfunction. In this sense, allograft failure may be interpreted as an aggravated and ongoing innate immune response to multiple allograft injuries associated with the development of alloatherosclerosis and allofibrosis. More details of this concept will be addressed in Chapter 6.

1.2.4 Résumé

Retrospectively, the time between 1994 and 1999 can be regarded as the foundation of the concept of innate alloimmunity. Although this term was not coined at that time, certain experimental data and clinical observations were already interpreted in terms of a possibility by which the initial injury to an allograft may be implicated in the development of acute and chronic allograft rejection. Interestingly, the year of the discovery of the Toll protein and the Toll pathway operating within the innate immune defense system of Drosophila melanogaster (see Part 1, Subsection 3.4.4.2) was the same year in which we published our first review article on the impact of ischemia/ reperfusion injury on allograft rejection by referring to natural immunity in relation to mechanisms of acute rejection, namely, the year 1996. While we discussed possible relations between allograft injury, function of activated APCs (later appreciated to be DCs), and acute allograft rejection, we missed thinking about the initial activation of these cells by injurious events via recognition of DAMPs by PRRs.

I remember one of the famous “ski-immunology meetings” in Austria in the late 1990s, originally launched by Prof. Walter Brendel and attended by leading transplantologists and immunologists, where I presented the data linking allograft injury to allograft rejection in terms of a puzzle where only the last piece of the puzzle was lacking. Today, we know that this last piece was the recognition of DAMPs by PRR-bearing DCs followed by subsequent activation of these cells.

18 1.3 The Impact of Postischemic Reperfusion Injury on Acute Allograft Rejection–As Seen in the 1990s

1.3.1 Introductory remarks

As a consequence of our observations from the SOD trial, we presaged that the oxidative ROS-mediated injury to allografts must have a great impact on both acute and chronic rejection events. In fact, in view of these data, we realized that, in the past, knowledge in transplant immunology was based on a large variety of theoretical considerations, experimental data, and clinical observations that had persistently omitted reperfusion-induced biological or pathophysiological events. The point was reached where it became clear to us that we had to reconsider the mechanisms of acute and chronic rejection, this time in light of the recognition that every transplant is a stressed, injured organ. Of course, we knew at that time that we were dealing with stressed organs to be transplanted. They are exposed to a sequence of ischemic injuries, starting already in the potential donor, for example, during shock conditions in the course of brain injury, continuing during organ removal and preservation, and culminating in the process of reperfusion during implantation after successful revascularization in the recipient.

Today, we know that any injury to an allograft matters, but I think that the ROS- mediated oxidative injury plays a dominant role and may be interpreted as the canonical injury to allografts that predominantly initiates and induces innate alloimmunity via its stringent impact on T-cell alloactivation as we understood it in the 1990s and as we understand it today as the leading event in the activation of the donor’s and recipient’s innate immune system. Accordingly, the chapter presented here refers to our review articles published in the 1990s.

1.3.2 The nature of reactive oxygen species-mediated postischemic reperfusion injury

Organ transplantation necessarily involves discontinuation of organ blood supply; the graft becomes temporarily ischemic. Overall ischemic damage of a vascularized allograft was earlier recognized as being the sum of (1) initial periods of warm ischemia (eg, occurring under shock conditions in the potential brain-injured donor or during organ removal from living donors), (2) cold ischemia during organ perfusion with hypothermic preservation solutions (“cold shock”), (3) anoxic preservation time per se, that is, the cold ischemia time during cold storage of the donor organ, (4) the “warm/cold” anastomosis time during implantation of the donor organ until completion of revascularization, and finally—as the culmination of all these stressful events— (5) reperfusion of the allograft in the recipient associated with the generation of reperfusion-induced, cell-damaging and tissue-damaging ROS.

19 The sum of these 5 events is referred to in this Part 2 of this book as the postischemic reperfusion injury (IRI).

Notably, with regard to skin allografts, instead of an acutely reintroduced oxygen supply during reperfusion associated with ROS generation, a prolonged local hypoxic environment occurs that is probably also associated with generation of (almost undetectable) ROS and lasts until revascularization.

The tissue changes caused by ischemia were already well known in the early 1990s. Upon depletion of energy-rich phosphates (stores of adenosine-5’-triphosphate [ATP]), the active ion transmembrane transport systems are inhibited, resulting in intracellular accumulation of ions accompanied by influx of water and swelling of cells, in particular, ECs. Translocation of water from the intravascular space results in local hemoconcentration and increased viscosity. The flow properties of blood are further impaired by stiffening of blood cells, particularly of leukocytes, caused by impaired viscoelastic properties of the cells from acidosis. Swelling of ECs, the reduced intravascular lumen, and impaired flexibility of cells leads to obstruction of capillary flow. Of note, the “no-reflow” phenomenon was intensely investigated by the Munich group and described by us in more detail [3]. This phenomenon is initiated by ischemia, and its degree of severity depends on the ischemia time [12]. The functional importance of the “no-reflow” phenomenon cannot be separately estimated, because the obstruction to microvascular flow only becomes apparent at the time of reperfusion, which in turn, affects the microcirculation, and therefore, accentuates the initial ischemic damage. In 1990-1991, it had clearly been shown by intravital microscopic studies that the damage caused by tissue ischemia depends on the length of ischemia and the changes inflicted upon reperfusion and reoxygenation. Although reperfusion of the ischemic tissue is essential for its survival, it increases the damage caused by ischemia [13]. Hence, the ischemic injury is superimposed on reflow-associated injury (“reflow paradox”) [14]. Whereas “no reflow” involves the capillaries in particular, the “reflow paradox” manifests itself through activation of leukocytes and their adherence to the endothelial surface of postcapillary venules, a process that impairs endothelial integrity and allows extravasation of plasmatic macromolecules into the extravascular space (ie, leukocyte-endothelium interaction). As will be discussed below, the establishment of a cytokine/chemokine/ adhesion molecule cascade represents the underlying mechanism of this scenario of leukocyte-endothelium interaction (Figure 1.3.1).

Of interest is our early description of the mechanisms leading to microvascular and tissue reperfusion injury as illustrated in the original figure in which hypoxia was chosen as the starting condition (Figure 1.3.2) [3]. (As we will see in Subchapter 2.4, recent research has revealed a dominant role for hypoxia in the induction of a life-saving antihypoxia response program via activation of the transcription factor, hypoxia-inducible factor-1alpha [HIF-1alpha]. Of note, mammalian cells respond to

20 Figure 1.3.1. Oversimplified diagram: Mechanism of the phenomenon of leukocyte-endothelial interaction as discussed by us in the 1990s.

In the 1990s, the phenomenon of IRI-induced leukocyte-endothelial interaction was already well- known, in particular, the involvement of adhesion molecules such as selectins and integrins. Later, further investigations revealed the additional participation of chemokines. At sites of IRI-induced tissue , rolled along the endothelium of postcapillary venules, collected inflammatory signals, arrested, and transmigrated. This figure roughly illustrates the phenomenon in an oversimplified diagram that is not inclusive, and other molecules not shown here mediate each of these events for distinct leukocyte types under different inflammatory conditions. Thus, sequential steps in leukocyte emigration are controlled by specific adhesion molecules on neutrophilic leukocytes and ECs. First, circulating leukocytes are tethered to the activated endothelium. Subsequently, leukocytes are slowed by rolling over the activated endothelium, which is facilitated by interaction of selectins and selectin ligands, that is, selectins and their counterreceptors. Efficient conversion from rolling to firm adhesion is dependent on the time a leukocyte spends in close contact with the endothelium. Upon arrest, integrins bound to their ligands, can signal into the , and stabilize the adhesion. This results in a firmer adhesion of leukocytes to the activated endothelium, which is again mediated by integrins and integrin ligands. Finally, leukocytes transmigrate across the endothelial monolayer. The basal lamina is depicted as separate from the remainder of the extracellular matrix, because migration across the subendothelial basal lamina is believed to be a separate step controlled by distinct molecules. Notably, in all steps in leukocyte trafficking over activated endothelium, heparan sulfate is involved; in addition, fibronectin and fibrinogen are involved in migration steps within the extracellular matrix. Abbreviations: EC, endothelial cells; ECM, extracellular matrix; IRI, postischemic reperfusion injury. Sources: adapted from: Land W, Messmer K. The impact of ischemia/reperfusion injury on specific and non-specific, early and late chronic events after organ transplantation.Early events. Transplant Rev 1996; 10: 108-127; and Land W. Immunsuppressive Therapie. Georg Thieme Verlag, Stuttgart, New York, 2006.

21 hypoxia by activating this transcription factor, which, as a major regulator of energy homeostasis and cellular adaptation to low oxygen stress, allows a cell to survive in the hypoxic environment. Excitingly, with regard to the topic of this book, ROS efficiently contribute to the stabilization of HIF-1α).

Upon reperfusion injury and reintroduction of molecular oxygen, oxygen radicals are formed within 3 to 5 minutes after reperfusion, most probably by ECs. The endothelium is activated, a phenomenon called endothelial dysfunction by many researchers. The activated endothelium in turn activates leukocytes, polymorphonuclear in particular. Twenty to 30 minutes after ischemia/ reoxygenation, leukocyte adherence is significantly enhanced and remains at high levels for about 2 to 3 hours [12,14,15]. Upon activation/adherence, leukocytes release further oxidants and mediators such as the activity factor (PAF), PAF-like

Figure 1.3.2. Pathophysiological mechanisms proposed by us in the early 1990s for the development of microvascular postischemic reperfusion injury.

Notice the vicious circle between postischemic reperfusion injury-induced oxygen radicals (produced in the course of the reaction from hypoxanthine → uric acid), which activate leukocytes, which, in turn, produce oxygen radicals after activation (compare also text in section 1.3.2). Source: The figure is traced and redrawn from a figure published by: Land W, Messmer K. The impact of ischemia/reperfusion injury on specific and non-specific, early and late chronic events after organ transplantation. Early events. Transplant Rev 1996; 10: 108-127.

22 substances, and leukotrienes, which impair endothelial integrity and favor leakage of macromolecules from postcapillary venules and development of interstitial edema. Intravital microscopic studies on striated skeletal muscle show that postischemic tissue damage (as indicated by cell nuclei stained with propidium iodide) directly depends on the number of granulocytes sticking to the postcapillary venules [16]. It should be noted from Figure 1.3.2 that oxygen free radicals, leukocytes, and mediators are linked in a vicious circle that is activated by IRI.

To elucidate the importance of free oxygen radicals, leukocytes, and mediators, the group of Konrad Messmer during the early 1990s conducted a series of intravital microscopic studies using strategies including antioxidants, mediator inhibition, and inhibition of leukocytes [12,14-23]. At that time, a causative role for leukocytes in the development of IRI was discussed, which was already known from earlier findings, showing that prevention of leukocyte accumulation in postischemic tissue by depletion with antineutrophil serum or leukocyte filters results in marked attenuation of microvascular injury. Of course, of major interest was the underlying mechanism of leukocyte adhesion, which was revealed by further targeted studies. Adhesion of leukocytes to the endothelial surface was achieved by engagement of adhesion proteins on both neutrophils and ECs, a process that could be inhibited with monoclonal (mAbs) directed against the surface-adhesive receptor MAC-1 (CD11b/CD18) and leukocyte function associated antigen-1 (LFA-1) as well as by antibodies directed against the intercellular adhesion proteins, ICAM-1 or E-selectin.

A series of intravital microscopic studies performed by the Munich group evaluated different strategies to interfere with leukocyte-mediated IRI in striated muscle, liver, and small intestine. All strategies tested proved successful in inhibiting or preventing postischemic leukocyte adherence and macromolecular leakage. Antioxidant therapy with sodium dismutase was particularly efficient in reducing macromolecular leakage [24], whereas inhibition of leukotriene B (LTB) synthesis and PAF antagonism resulted in equally efficient inhibition of both leukocyte adherence and macromolecular leakage [17,23]. Monoclonal antibodies directed against the heterodimeric leukocyte adhesion molecules or their subunits also proved efficient in preventing both postischemic leukocyte sticking and macromolecular leakage [20- 25]. The same was true when the ligand to integrin molecules was specifically blocked with an anti-ICAM-1-antibody [25]. In the dorsal skin fold model of Balb/C mice, the intermittent leukocyte/endothelial interaction (rolling), which precedes firm leukocyte adherence, was mediated by P-selectin, but not by L-selectin [26].

As a matter of fact, in the 1990s, the IRI-induced phenomenon of leukocyte/ endothelial interaction gained tremendous attention by physiologists, experimental surgeons, and transplant researchers around the world [27-29]. Investigators in many laboratories were fascinated by this interplay involving the initial slowing

23 or “rolling” of neutrophils along the endothelium during the early moments of reperfusion, followed by firm attachment and amplification of the neutrophil response, and culminating in the diapedesis of neutrophils into the parenchyma where neutrophil-parenchymal cell interactions contributed to the necrotic process. Interestingly, adhesion molecules played a key role in the early phases of neutrophil adherence and activation (Figure 1.3.1). In fact, findings and mechanisms associated with the phenomenon of leukocyte/endothelial interaction, such as disturbance of the microcirculation, were defined as primary pathophysiologic processes initiated by the action of adhesion molecules. Today, when looking back to this time, many of those processes may be interpreted as secondary events following the primary activation of the innate immune system, including adhesion molecule expression that was up-regulated in activated cells of this system. In fact, notions and knowledge in science and research may change very rapidly.

1.3.3 Sources of reactive oxygen species production and mechanisms of their generation under ischemic conditions

1.3.3.1 General remarks

In the 1990s, ROS and reactive nitrogen species (RNS) such as hydrogen peroxide, free superoxide anions, free hydroxyl radicals, the free radical nitric oxide, and the very toxic peroxynitrite, were well known to be continuously produced within cells during the lifespan, either as a result of the mitochondrial electron transfer processes or as a by-product of the catalytic activity of several enzymes [30]. In particular, the interest in superoxide radicals became apparent after the discovery of the enzyme, SOD, by Fridovich [31], an enzyme that removes superoxide anions (compare Part 1, Subsection 1.3.2.5). A new recognition in the 1990s was that ROS did not always act as toxic molecules. It became clear that under normal circumstances, namely in low concentrations, ROS operated as physiologically important mediators—so-called second messenger molecules—and by this, regulated biological signaling processes in cell functions [32,33]. In this way, and under the control of intracellular antioxidative defense systems, they were responsible for cell homeostasis, normal growth, and metabolism [34]. Produced by mitochondria during normal metabolism and by ECs, macrophages, and leukocytes during responses to injury and inflammation, these molecules, only in higher concentrations, play the role of “devils in biology” and are used as efficient weapons in host defense. In healthy aerobic organisms, however, generation of ROS and RNS is essentially balanced by an antioxidative defense system that has already been alluded to in Part 1 (Subsection 1.3.2.5) [30]. (Of note, modern notions about these extraordinary properties of ROS in its dual role as a distinct regulator of cell functions and toxic cell-destroying molecules have gained considerable attention in the new millennium and will be addressed in more detail in Subchapter 2.5).

24