Q: Dr. Tsimikas, warmest congratulations on your study on the effect of oxidation-specific antibodies on foam cell formation and atherosclerosis in a murine model, published in the Journal of the American College of Cardiology in October 2011. Would you mind highlighting the key elements and findings of your study?
A: Dr. Herrmann I would like to thank you on behalf of my coauthors for the opportunity to describe our study. This was a collaborative study of several laboratories and investigators at UCSD, Chicago and Germany, including Drs Witztum, Friedmann, Li, Hammond, Getz, Reardon, Torzewski and all of their talented investigators. In particular, my friend and colleague Joe Witztum deserves the lion’s share of credit for developing many of the concepts that have allowed us to reach this near-clinical level of translational application of oxidation-specific antibodies. In a nutshell, what we have shown is that human oxidation specific antibodies, and specifically human antibody IK-17, provides atheroprotection in two models of experimental atherosclerosis ranging from 29-46% reduction in the progression of atherosclerosis. Importantly, this was carried out with both IK-17 Fab fragment and IK17 single chain fragments, both of which lack immunological properties of intact antibodies other than the ability to inhibit uptake of oxidized LDL by macrophages, and therefore isolates the therapeutic effect specifically to the antigen binding site of the antibody. Furthermore, we were able to show that intraperitoneal infusion of these antibodies or over-expression with an adenovirus resulted in reduction of macrophage foam cell formation, which appears to be the mechanism of reduced progression of atherosclerosis in this study. Thus, the therapeutic effect would be very early in the inflammatory process of atherosclerosis and should be impactful clinically as it would minimize all downstream effects of inflamed macrophages in the vessel wall. These findings suggest that using fully human oxidation-specific antibodies may be relevant to patients with ST elevation myocardial infarction, acute coronary syndromes, percutaneous coronary intervention and perhaps long-term sustained therapy to minimize the complications of atherosclerosis in patients with recurrent events on maximally tolerated therapies.
Q: The title of your study starts with human oxidation-specific antibodies. Please tell us more about the journey you went through to generate these.
A: Our laboratory has been interested in the oxidation hypothesis for many years. Dr. Witztum has led the conceptual formulation and proof of the role of the innate immune system in mediating atherogenesis. Specifically, it has been demonstrated that oxidation- specific epitopes, such as oxidized phospholipids and malondialdehyde epitopes, are proatherogenic and also activate the immune system in producing “autoantibodies” to such epitopes. In that regard, Dr. Witztum’s group first identified that glycated LDL was immunogenic and documented hapten-specific autoantibodies to glycated LDL, and other glycated proteins, in patients. His group followed this up by demonstrating that both mice and humans with atherosclerosis develop autoantibodies to these oxidations-specific epitopes, as occur on OxLDL. Most recently, it was also shown that mice have substantial numbers of “natural” antibodies that recognize such epitopes, up to 30% of all IgM antibodies, that are evolutionarily conserved and may have homeostatic functions, for example in removing apoptotic cells, which similar to OxLDL also express such oxidation-specific epitopes. This work initially led to the generation of hybridomas that secreted murine monoclonal antibodies to oxidation-specific epitopes, such as antibody MDA2, which we have used for molecular imaging of atherosclerosis, as well as E06, which we have used to generate a human ELISA to measure oxidized phospholipids on apoB containing lipoproteins (OxPL/apoB). In fact, the OxPL/apoB assay, which predicts future death, MI and stroke in unselected populations has been licensed to QUEST by UCSD, and it will be made available to the research and clinical communities in the future.
Continuing on this work with murine monoclonal antibodies, we hypothesized that these antibodies could also be cloned from patients since circulating autoantibodies have been demonstrated in a large number of clinical studies. Therefore, in the late 1990’s and early 2000 we screened a phage display Fab library prepared from a human subject and were able to isolate 120 clones that were secreting human oxidation-specific antibodies. IK17 was the 17th clone from that screening that had all of the characteristics and immunological properties of a fully human antibody that bound to both MDA-LDL and copper oxidized LDL, which were the two key models of oxidation-specific epitopes that had been well characterized previously. Therefore, we have been using IK17 for a variety of applications, including molecular imaging of atherosclerosis and now therapeutic applications.
Q: In the current work, you went through two studies with the first being a continuous infusion of IK17-FAB in LDLR-/- mice. What were the dynamics of this infusion protocol – were you able to sustain the IK17-FAB infusion?
A: We initially decided to use an intraperitoneal infusion in of IK17-Fab in LDLR-/- mice that were cholesterol fed, a model which most closely reflects human disease. The idea here is that this would be easily transferred to a clinical application by infusing such antibodies in patients intravenously or subcutaneously. Using this approach, we were able to sustain IK17-Fab plasma levels at approximately 40 micrograms per milliliter, which is what is the in the range of what is seen therapeutically with other antibodies for non-cardiovascular applications. However, because this a fully human antibody infused in the mouse, we discovered that the mice developed, and very quickly, anti-IK17 antibodies and immune complexes. Sustained plasma levels of IK17 were present for 4-8 weeks but then at 14 weeks the levels were markedly diminished due to the generation of anti-IK17 antibodies and immune complexes. Despite this immunological effect to the infused antibody, a 29% reduction in the progression of atherogenesis in the en face measurement across the entire aorta was noted. Interestingly, the data points were all fairly well clustered together suggesting a consistent effect across most of the mice.
Q: Despite these shortcomings you observed a reduction in the extent of atherosclerosis with IK17-FAB infusion, at least by en face analysis of Sudan IV-stained aortas. How extensive was this reduction and was there any correlation with the anti-IK17 immune response?
A: Quite Interestingly, despite this significant reduction in atherosclerosis there was a significant correlation of murine plasma IgG anti-IK17 antibodies with the extent of en face lesion formation in the IK17-Fab treated mice, with an R-value of 0.87. This suggests that the mice with the most potent immune response to the infused IK17 had the greatest extent of atherosclerosis, i.e. the least therapeutic benefit from the infused antibody. This is another piece of evidence showing the therapeutic effect of IK17-Fab.
Q: Next, you took the elegant approach of adenoviral expression of IK17-scFv in LDLR -/-Rag1 -/-. Please tell us more details about this. How does it work? How well does it work to develop functional IK17-scFv?
A: Because of the murine immunological response to human IK17, we needed to develop a novel and clean experiment that would address the hypothesis without this effect. Obviously, we could not go straight to patients at this point, therefore we decided to use LDLR-/-RAG1-/- mice, which lack B and T cells, and therefore cannot mount an immunological response by developing antibodies to IK17. This model also develops atherosclerotic lesions when cholesterol fed, although less than LDLR-/- mice with intact RAG 1. We were able to successfully generate an adenovirus overexpressing the IK17 as a single chain fragment, documented that it secreted sufficient plasma levels to block uptake of OxLDL by macrophages and then, that these levels could be sustained when the adenovirus was infused every two weeks. Interestingly, the effect of the adenovirus was transient despite the fact there was no immunological response to the antibody, suggesting that the promoter used to express the adenoviral construct was deactivated, probably by methylation. However, we were able to reinject the mice with the adenoviral construct repeatedly, which allowed us to generate sustained and adequate plasma levels of fully functional IK17 single chain over the entire time course of the intervention period.
Q: How did you verify the properties adenoviral expression of IK17-scFv – in vitro and in vivo?
A: We were able to verify the properties of the adenoviral expressed IK17-scFv from cells in culture, as well as the IK17 expressed from the liver into plasma of mice infected with the adenoviral construct by use of ELISA assays to confirm the titers of the IK17-scFv, as well as IK17-scFv’s functional properties, by evaluating the capacity of the IK17-scFv to inhibit the binding of OxLDL or MDA-LDL to macrophages.
Q: You used Rag -/- mice that lack both B and T cell to overcome the immune response. How did this turn out? Would this have been feasible for the first study as well?
A: Using the LDLR-/-/RAG1-/- mice as a model we were able to document an even better therapeutic response from the IK17, now having a 46% reduction in atherosclerosis compared to the control adenovirus overexpressing green fluorescent protein. This suggests that compared to ip injection of IK17 Fab, where the anti-IK17 response significantly neutralized the therapeutic efficacy of IK17, lack of immunological response to the scFv IK17 led to an even more pronounced therapeutic effect. These two studies together document nicely the therapeutic effect of the IK17 lacking the Fc portion and really suggest the effect was mediated through its antigen binding effects. With additional resources, we could have injected the adenovirus also in LDLR-/- mice but since the data looked fairly consistent across the two studies we decided we had enough evidence to prove our hypothesis.
Q: By immunostaining you demonstrated quite nicely, that plasma from mice with adenoviral expression of IK17-scFv stains human atherosclerotic plaques in exactly those areas where you would expect it. If so, how did this translate into effects on atherosclerosis in mice?
A: We were able to show that plasma from mice with secreted IK17 single chain could stain atherosclerotic lesions. This documents that not only is the antibody functional but it can bind to and detect these epitopes in the vessel wall. We believe the mechanism of benefit of such antibodies is to bind to oxidized LDL, either the small amount that may be in the circulation or the much greater amount present in the vessel wall, and prevent its uptake into macrophages. Therefore, the atherosclerotic plaque has diminished available oxidation-specific epitopes and the downstream pro-inflammatory and proatherogenic effects are mitigated.
Q: Intriguingly, you noted a reproducible study effect by Sudan IV staining but not by cross-sectional analysis at the aortic origin. How do you explain and interpret this?
A: It should be noted that the Sudan staining was a study of lesions in the aorta, while the cross sections were at the aortic root—different anatomical sites. Why differences were noted is not entirely clear, but we do know from many studies that there are site specific effects of therapeutic modalities, some of which have effects on the aortic root more than the enface aorta or some of which have effects on the innominate artery versus others. In fact, it’s unusual to have completely concordant results in multiple territories. A second explanation is that, particularly in the LDLR-/-/RAG1-/- mice, they had very small amounts of atherosclerosis in the aortic root and therefore this may have been insensitive to detect significant changes with the antibody. Interestingly, in human studies that we have not published yet, the IK17 epitopes are very prevalent in advanced necrotic core lesions and it’s possible that the IK17 epitopes may not have been fully expressed in the aortic root compared to the en face aorta.
Q: Finally, you used peritoneal macrophages from the study mice as representatives to study the effect of adenovirally expressed of IK17-scFv on the foam cell process. What did you find?
A: The study with peritoneal macrophages was quite illustrative in pointing towards a potential mechanism of benefit. One effect of these antibodies that don’t have the Fc portion is that they bind to their antigen extracellularly and block its ability to be recognized by scavenger receptors. The peritoneal macrophage experiment tested exactly this and showed that when mice were cholesterol fed and the peritoneal macrophages were removed after treatment with single chain IK17 expressed with the adenovirus, that lipid accumulation in foam cells was markedly diminished as assessed by Oil Red O staining. This suggests that IK17 acts to block the ability of scavenger receptors to take up oxidized lipids therefore diminishing foam cell formation. What is not clear, however, is what happens to the epitopes that are bound up by the antibody. Are they cleared by additional reticuloendothelial system pathways such as in the liver and spleen thereby not accumulating in the vessel wall? This will be an area of future study to more fully understand this mechanism.
Q: At the end then, was the hypothesis you set out with confirmed? What did we learn from this study?
A: We believe our hypothesis was confirmed and builds on approximately thirty years of work on trying to understand the role of oxidation and how it’s intimately integrated in with responses by the innate immune system. We believe that the effect of the immune system, particularly in the generation of autoantibodies, can be used as a tool to develop therapeutic modalities that will minimize the effects of oxidation-specific epitopes in the vessel wall.
Q: How can we apply the current observations? What is the next step? What is your next step?
A: At this point, using therapeutic oxidation-specific antibodies would be a new paradigm in our therapeutic arsenal for treating cardiovascular disease. We like to be able to obtain funding, either NIH or pharmaceutical based, to be able to now move closer to phase I human studies. That would include generating sufficient quantities of antibodies and performing toxicology studies and ultimately testing this in patients. I believe that a potential clinical application would be in ST elevation myocardial infarction and acute coronary syndromes, and/or percutaneous coronary intervention where patients need very rapid plaque stabilization to minimize recurrent events. For example, we use high dose statins now but they may take several weeks to have an effect on the vessel wall, whereas an infusion of antibody that directly treats the plaque disease would act very quickly and should be able to be tested fairly easily in a randomized trial. We think this approach would be quite useful as a hypothesis in patients and look forward to one day being able to expand our translational work to the clinical arena.
Q: Dr. Tsimikas, thank you so much again for this interview! We deeply appreciate your time and the personal tour of your study. Again, congratulation on this important study!
A: Thank you very much for the opportunity to provide details of our study. I look forward to one day seeing the successful use of such antibodies in humans in minimizing effects of active cardiovascular disease.
European Society of Cardiology
European Heart HouseLes Templiers2035 Route des CollesCS 80179 Biot
06903Sophia Antipolis, FR
Our mission: To reduce the burden of cardiovascular disease
© 2017 European Society of Cardiology. All rights reserved