Strategies for Developing Therapeutics that Directly Target Anthrax and its Toxins; Public Workshop
45 Center Drive
National Institutes of Health, Bethesda, MD
June 10, 2004
Participants
ROY BARNEWALL, D.V.M., Ph.D.
SHUKAL BALA, Ph.D.
R. JOHN COLLIER, Ph.D.
DAVID FRUCHT, M.D.
JESSE L. GOODMAN, M.D., M.P.H.
SUE GORMAN, Pharm.D., DABAT
M. DAVID GREEN, Ph.D.
JUDY HEWITT, Ph.D.
NISHA JAIN, M.D.
BRAD LEISSA, M.D.
STEPHEN H. LEPPLA, Ph.D.
JULIE LOVCHIK, Ph.D.
ANTHONY MACALUSO, Ph.D.
JAMES McCORMACK, Ph.D.
KAREN MIDTHUN, M.D.
MARISSA A. MILLER, D.V.M., M.P.H.
CARL NIELSEN, Ph.D.
M. LOUISE M. PITT, Ph.D.
CONRAD QUINN, Ph.D.
JENNY MELLQUIST-RIEMENSCHNEIDER, Ph.D.
DAVID ROSS, M.D., Ph.D.
LEWIS SCHRAGER, M.D.
DAVID STEPHENS, M.D.
KEITH WEBBER, Ph.D.
KAREN WEISS, M.D.
ALEXANDRA WOROBEC, M.D.
DR. MACALUSO: Good morning, my name is Tony Macaluso. I'm a bio-defense project manager at the National Institute of Allergy and Infectious diseases. On behalf of the National Institutes of Health I welcome you to the NIH campus, and also tothis workshop: Strategies for Developing Therapeutics that Directly Target Anthrax and Its Toxins.
This workshop will address issues related to product characterization, proof of concept and safety and efficacy testing in order to expedite the development of these products which are regulated by the FDA.
The Food and Drug Administration is the lead agency for organizing this workshop which is co-sponsored by the National Institutes of Health, the Centers for Disease Control and Department of Health and Human Services.
As the NIH point of contact it has been my pleasure to work with any people who helped make this workshop a success. In particular, I'd like to take this opportunity to thank the lead organizers at FDA for doing an excellent job of putting this workshop together: Dr. Karen Weiss, Dr. Dale Slavin, and Ms. Melanie Whalen.
Before the workshop gets underway there are a few housekeeping issues that I've been asked to address. The first is that you're probably all aware--but just in case--the agenda for this workshop has been compressed into one day, so we will not--I repeat, not--be meeting tomorrow. This change from the original one-and-a-half agenda was necessary due to the closing of most government agencies as a mark of respect for former President Ronald Reagan.
I apologize for the inconvenience this may have had on your travel plans, but I think you'll all agree that this was an unforeseeable circumstance, and the alternative to postponing the meeting really wasn't a very palatable one.
The second item is the availability of slides and transcripts. The transcripts will be available at this website, probably in about two weeks; the slides, probably within one or two days.
We will be having panel sessions after most of the sessions. I would encourage everyone to participate. There are microphones in both aisles. If you prefer to submit your questions in writing there will be some cards available from some of the staff and they will be picked up at the beginning of each discussion session, and also during the discussions sessions. And the moderators can ask the questions for you.
Regarding food: coffee and snacks will be provided during the breaks. Now, we realize that because of the compressed agenda, this meeting will last much longer than we had originally anticipated. To encourage you to stay through the last sessions--for which I happen to be the moderator--we've arranged for more than the usual coffee and cookies for the afternoon break. So I encourage you to chow down, if necessary, in the afternoon so that you don't feel the urge to leave early for dinner.
Also, we've reduced the time available for lunch. So you really will not have enough time to leave the campus for lunch and get back. So I encourage you to use the cafeteria, which is just one flight up.
The agenda for the last session wasn't given. I'd like to just go over that very briefly so you'll know what to expect. In addition to a presentation by Carl Nielsen of Challenges and Opportunities in Product Development, we'll also have short five- to 10-minute talks before the panel session. Those talks--the topics will be the FDA's Proactive Approach with Medical Countermeasures Development, Emergency Use Authorization, Information about the Strategic National Stockpile Program, DHHS Plans for Implementation of Project Bio-Shield, and Opportunities and Resources for Bio-Defense Countermeasures Research and Development that are available through the National Institute of Allergy and Infectious Diseases.
As a reminder--since we are on a compressed schedule--it will be very important for everyone to try to stick to the schedule. I'll ask the moderators to try and push everything along to keep us on that schedule.
And then the last item is the location of restrooms and telephones. There are restrooms at either end of this hall on this floor, and also additional restrooms just one flight up. There's also telephones directly across from this auditorium, and also on the next floor.
At this point, I'd like to turn the podium over to Dr. Karen Weiss, from the center for Drug Development and Research at the FDA.
Karen?
DR. WEISS: Good morning to everyone. I am very gratified to see so many people here early in the morning. Like Tony, I extend my apologies for having to, at the last minute, rearrange the schedule and condense a day-and-a-half of what was going to be a fairly nice, leisurely workshop into a very whirlwind one-day workshop. But, we just have to kind of roll with the punches, and do the best you can.
One other housekeeping rule to mention is that the conference center does not like people to bring food and drink into the conference. So those of you that have it there--just to let you know, that a during the breaks--whatever--finish everything up outside.
I also, like Tony, wanted to extend some deep appreciation to a couple people. This was a fairly--it was a very collaborative, collegial effort to put this workshop on, between four different government agencies: the FDA, the CDC, the NIH and Office of Emergency Preparedness under HHS. And we all worked for a long number of months, on numerous conference calls, to put this together. And, on behalf of the FDA I want to just extend my appreciation to all the participants in the working group.
But, in addition, I'd like to just highlight a couple of people and organizations, particularly: Dr. Dale Slavin, who's a project manager in my office, and volunteered to take on this job in addition to her other full plate of activities, and did a great job just pulling together all the different participants for the various activities that have to be done. And I just owe her a great deal of appreciation for that; the Office of Communications, Training and Manufacturing Assistance--OCTMA--in the Center for Biologics Evaluation and Research--who worked on all the logistics of this conference; and last, but not least, Dr. Tony Macaluso, and the staff at NIH, for making this all possible by the funding through an interagency agreement, as well as being responsible for the food that's out there.
So, the carbs and the caffeine that are going to be with us for the day--which we're going to definitely need--are all due to his continually calling and nagging to get that accomplished. So I definitely appreciate that. It's very difficult to supply food when you're a government agency and doing something in a government facility. So it is a great effort.
And with that, then, I would like to just go ahead and start the session by introducing our first individual who will provide a welcoming--a number of welcoming remarks. I'm very, very please to ask Jesse Goodman to come to the podium.
Jesse Goodman has been the Director of the Center for Biologics Evaluation and Research for a number of years. He's also an infectious disease doctor, and I appreciate him coming and making some remarks. One quick thing is that Janet Woodcock, who was going to come, because of the compressed schedule can no longer be here this early in the morning, and asked me to convey her regrets. So, thank you.
Jesse?
DR. GOODMAN: Good morning, folks. I'll try to be, like, really quick because of the schedule and, if anything, let you get started ahead of time.
But, you know, if was more than two years ago, after the anthrax attacks that all of us in the Public Health Service and those of us who are also clinicians realized there was a need for better therapeutic outcomes than were getting even from antibiotics, even those were better than expected. And some of the discussions that led to this meeting started then.
So there is a real need, you know, for the half or so patients who, despite aggressive treatment and supportive care don't make it.
And there's also a precedent for immune therapies working in acute severe infectious diseases; for example, in use of anti-serum early treatment of pneumococcal disease. So there is some reason to at least believe that there could be therapeutic gains from treatments directed at toxins.
I'd like to say, though, that the field obviously, as most of you know, is littered with failures, as well. And it's been very difficult, once there's disease, to intervene. Excuse me--I got off a plane late last night here--so-- I think, therefore it's important that we really have a good development program; animal studies that not only can prove success and promise of a therapy, but tell us if something isn't going to work.
In addition, you know, if we had looked at the experience after the anthrax attacks, let's say we had given everybody anthrax sera, and had observed a 50 percent survival rate, we would have said that anti-sera saved patients, because the expected survival rate, historically, was 10 or 20 percent. So we'd all be--probably not having this meeting, and be producing tons and tons of anti-sera, which may or may not work, as we know.
So I think while we're looking for better therapies, it's important we use the best methods to evaluate those therapies.
As part of that, I think it's also worth considering whether there's any possibility at all, if such therapies were ever to be used, of field evaluation of them. Even though it's unlikely, for example, that an intravenous immune globulin product would be harmful, it's remotely possible. And certainly some of the chemical therapeutics--toxin inhibitors, etcetera--could potentially be harmful in humans with anthrax disease. So I think a very difficult thing tocontemplate is the idea of controlled clinical trials in a disease with a high public profile and a high mortality rate, like inhalational anthrax. But it's something that I hope, in the discussions and the panels on the last day, that we'll look at. This is the last day, now.
[Laughter.]
You know, and again, just to emphasize, you know, I really do think if we had had immune globulin, for example, or an experimental therapeutic available for those few cases, and had used it, many, many people would be convinced that it had worked and it would become the standard of care. And, again, FDA is quite familiar with circumstances where this has happened--for example, autologous bone marrow transplantation for breast cancer, which appeared to give such excellent results that when subjected to clinical trials did not hold up. And nobody would argue that that's something you want to put patients through unless they benefit from it.
So I think, again, the importance of really good animal models--which are an incredible challenge, especially if you are considering an animal model in a symptomatic disease state. Animal models are problematic enough, even when they're simple. And when you add things like disease and antibiotic therapy, the variability can become incredible.
So, I really look forward to hearing the results of this. There's a terrific group of people. And I know that the Department of Health and Human Services is very, very committed to improving therapy for anthrax, so this is a very important meeting.
So, with that, I guess we start the first session, so Karen will come back up. Thanks very much.
Part I - Introduction to and Pathogenesis of B. anthracis
DR. WEISS: Thank you very much, Jesse, for your opening remarks, and for setting the stage.
We're going to go ahead and get started right now with Session I: Introduction to the Pathogenesis of B. anthracis.
The objective of this session is to provide a critical background to facilitate and focus the workshop. And I have the pleasure of introducing two speakers who are going to be discussing things at this session.
There's no panel planned after the first two speakers, but we'll see what the time frame is like, and if there potentially is some time available after both speakers have completed their talks we might have a minute or so for some questions.
I'm going to go ahead and introduce them both at the same time so I don't have to keep popping up and down, and try to save a little bit of time.
So, the two speakers are, first, Dr. Stephen Leppla, who is right here at the NIH campus. Dr. Leppla has done research on anthrax for well over 24 years and is eminently qualified to give one of these opening presentations.
And following Dr. Leppla, we'll hear from Dr. David Stephens from Emory University. Dr.Stephens is the head of infectious diseases at Emory, and he has worked in the Meningitis and Special Pathogens Group with CDC, and he was a clinical team leader with CDC for the 2001 anthrax outbreak. And he will directly follow Dr. Leppla's presentation.
Dr. Leppla?
Anthrax Toxin as a Target for Therapeutics-Overview
DR. LEPPLA: Thank you. Good morning, it's a pleasure to open the scientific aspects of this meeting. Since this is a very targeted meeting, and all of you, I think, have worked on anthrax, it's a little presumptuous, perhaps, to provide introduction about the basic properties of the organism. So forgive me if I say things which are too obvious to any of you.
So, the organism we're dealing with is bacillus anthracis. As you know, it's a Gram-positive spore former. It infects livestock. And we're here because it can also infect humans.
And it's virulence is sort of a classic in pathogenic microbiology. We think it's fairly simple, in that its virulence is determined by two virulence factors: the poly-glutamic acid capsule, which is anti-phagocytic, and the three-protein component anthrax toxin. And each of these is encoded by a large plasmid, PX01, and PX02.
I'm going to skip these two slides, in the interest of a shortened talk--except just to indicate, since you have this slide in your folders--that this was to indicate the difference between bacillus anthracis and its genomically very closely related neighbors, bacillus cereus and theragensis--but in spite of their very close genetic similarity, they have very different sets of virulence factors. Bacillus cereus having a set of secreted aggressins that are, in large part, transcriptionally controlled by a regulator called PLCR. In contrast, anthracis has the two plasmids I mentioned--PLCR, which is the gene is there but it's been inactivated. So all of these aggressins are not produced by anthracis and, instead, you have the toxin and the poly-D-glutamic acid capsule.
So this is a pathogen that's evolved in a very particular way.
And the infection process as we understand it is that spores enter the body either through the skin, GI tract, or through the lungs. And those spores are engulfed by phagocytes--usually mentioned as macrophages, but I think other phagocytes are probably also involved. The spores are carried in those phagocytes to lymph notes where they geminate. The bacteria then escape from the phagocytes. Toxins are produced--both the capsule and the protein toxins--and these toxins have various activities which clearly suppress the host responses and allow the bacteria to grow to very high numbers in the blood of these infected animals.
The anthrax toxin accumulates as the bacteria grow to high numbers. But I think an interesting feature is that the toxin is not a hugely, highly cytotoxic and rapidly potent killer of cells or tissues. And that's really to the advantage o the bacteria, because its objective is to not only--parenthetically, to kill the host, but to produce a large crop of spores in the herbivores, which are the natural host of the disease, those spores would then be deposited in the soil, where they could remain for long times until another animal comes along and is exposed to them.
And the indications of the effectiveness of the pathogen are that in the best hosts--the large herbivores--the bacteria can grow to very high titres in the blood at the time of death. And, with any luck, the bacteria convert a lot of these into spores, and so you have a large infectious reservoir in the soil. [Slide.]
Most of my talk, and most of this meeting, is about the toxin, but I think I just wanted to remind people of recent work that--I guess I skipped a slide which is not showing up. That's okay--recent work from Tom Kozel's lab, indicating that the capsule plays a previously unrecognized role in immunity to anthrax. So what they did was that they showed that monoclonal antibodies produced to the poly-D-glutamic capsule were protective, in mice, against anthrax infection. And this had not previously been recognized, that antibodies played a potentially protective role.
So I think this is a newly identified target, and I hope and anticipate that there will be increased attention to this target of therapeutics. [Slide.]
Now focusing more on the toxin, again, there are these three protein components--large proteins. The protective antigen is the central player, and its role is to deliver the two enzymatic moieties into the cytosolive cells. Adeolate cyclase, the edema factor, will raise adeolate cyclase, or cyclic AMP levels. The lethal factor is a metalloprotease, which cleaves all of the MEKs.
Now, over the last decade or so, we've developed a detailed picture of how these toxin proteins interact to get inside cells and damage cells. So we now know that the protective antigen binds to cellular receptors. These were identified in the lab of John Young and Jon Collier. And they were first called "anthrax toxin receptor" It was then recognized that they are variants of a molecule called "tumor endothelial marker 8." And, more recently, another isoform capillary morphogenesis protein-2 has been identified.
So the toxin is bound to these receptors. It is then activated protealytically by cleavage with the cellular protease furin, with removal of a fragment. And removal of that fragment allowed the remaining portion of protective antigen to heptamorize into a very tightly associating heptamer. And on the newly exposed surface of this heptamer there are binding sites for the edema factor and lethal factor components.
And John collier has shown that, actually, the binding site spans two of these monomers, and it follows--rather constraints that there's a maximum of three molecules of edema factor and lethal factor that can be bound onto the heptamer.
This heptamer is then internalized--presumably through lipid wrap and endocytosis to a vesicle which becomes acidified, and it then inserts in the lipid membrane and becomes a protein-conducting channel.
And through Dr. Collier's work in particular, that channel is probably the best understood protein-conducting channel now known, through extensive mutagenesis and biophysical and biochemical measurements.
So we believe that the edema factor and lethal factor proteins transduce--pass through the center of this heptamer channel to reach the cytosol. Again, then, the edema factor is an adenolate cyclase, and makes very high levels of cyclic AMP. LF cleaves all except one of the ME- kinases, or MEKs, and the results of these two events--presumably in combination--cause the tissue damage and lethality which observe as the pathogenic effects of infection. [Slide.]
This is genetic evidence that introduces the fact that the toxin--the two toxins are really the dominant virulence factors in bacillus anthracis. And I'm not going to go through the numbers here. This is work from Michelle Mock. But basically, it shows that if you knock out the capsule or the toxin, or individual toxic components, you greatly reduce the virulence of bacillus anthracis for mice. It's going from an LD 50 of 5 scores, up to, essentially, avirulent organisms.
So, to focus separately on the lethal factor and edema factor components of the toxin, we have for a long time thought that the lethal toxin is the major cause of pathogenesis. And the numbers from the previous slide showed that bacterial strains in which LF is genetically inactivated are attenuated at least a thousand-fold. And, furthermore, there's a very large and growing body of evidence that the lethal toxin injected into animals duplicates the symptoms of bacterial infection. And, of course, an even larger body of evidence that antibodies to either--certainly to PA and increasing evidence that antibodies to LF protect against bacterial challenge. So all of this indicates the important role of the lethal toxin in pathogenesis. [Slide.]
The established effects of the toxin in several system provide the ways to study the toxin, but also have find wide use in bioassays for characterizing the toxin and antibodies, as well. So, one of the early discoveries, by Art Friedlander, was that the macrophages from certain limited number of inbred strains of nice exposed to lethal toxin lysed in about 90 minutes.
But, again, this is restricted to certain strains of mice. And that's widely used as a bioassay, as well as a way to study the mechanisms of action. [Slide.]
And perhaps an even older--certainly an even older bioassay is based on the fact that in the Fischer 344 rat, intravenous injection of lethal toxin will kill them in as little as 38 minutes. Other rat strains are much less sensitive. So there's the unique feature of this rat. And this is widely used as a bioassay for protective agents that target the lethal toxin.
Mice are susceptible to the lethal toxin, but they die much more slowly. Typically, they take two to three days to die after lethal toxin administration. [Slide.]
This is our data showing the sensitivity of two strains of inbred mice, the Balb/c being the more sensitive, and the C57 Black being somewhat more resistant. Mentioned here is the fact that though it has been the view that the susceptibility of mice is a function of the susceptibility of their macrophages, to lytic action of lethal toxin. And so, indeed, the Balb/c mice, their macrophages, in vitro, are sensitive to the toxin, whereas the C57 Black macrophages are resistant in vitro. But we've looked at a large number of inbred mouse strains and, in fact, this correlation between sensitivity of macrophages and sensitivity of the whole animals is not a very good correlation. And it's certainly not true over a larger number of strains. [Slide.]
Now, the neglected partner in this pathogen has been the edema toxin. And it's, I think, in part been neglected because the genetic evidence in the slide that I went over quickly, is that if you knock out edema factor from a bacillus anthracis strain, the virulence decreases tenfold. One should ignore tenfold, but in comparison to lethal factor, it seems like a minor player.
But we actually have now shown that the edema toxin injected into mice is, in fact, highly lethal to these mice, and that the edema toxin does produce some clinical signs that are similar to those seen in bacterial infections. [Slide.]
This is some unpublished evidence on the susceptibility of Balb/c mice to the injection--intravenous- injection-of the edema toxin PA and EF. And we find that the LD 50 is about 20 micrograms each of the combination of PA and EF. And at the higher doses, the animals die very quickly, with a wide--showing a wide variety of pathogenic responses, biochemical and histopathological changes--really a wide spectrum than is seen in the lethal toxin-injected mice.
So I'd like to suggest that the edema toxin has probably been ignored, and probably should receive more attention as a target of therapeutics. [Slide.]
So, what kind of therapeutics are people developing? And you all know this very well. I'd like to, for the purpose of the discussion, divide them into these two groups: those which act extracellularly, and those which work inside cells. And perhaps a unique feature of the anthrax toxin system is that lots of things are going on outside the cell, and therefore there are a number of steps which are accessible to macromolecular inhibitors; that includes, of course, antibodies, but also toxin fragments, receptor decoys and others.
And, in contrast, the agents that would block intracellular activities would be typically targeting the enzymatic activities of the adenolate cyclase and the protease. And this is more amenable to standard pharmaceutical approaches, where one looks for small molecule inhibitors that are cell-permeable. [Slide.]
So, again, returning to the scheme of how the toxin gets into cells, let me highlight a number of targets that people are considering for therapeutic approaches. So these could include things like the receptor decoys. It's been shown that if you express--if you provide the extracellular domain of the receptor, it will interact with the toxin and act as a competitive inhibitor and protect.
Similarly, fragments of protective antigen--perhaps Domain IV, or perhaps peptides--mimicking regions of Domain 4 that interact with the receptor could block that interaction and provide protection.
Furin inhibitors are a potential therapeutic. I'll give an example of one in a minute.
Antibodies to PA, which either bound to the receptor-recognition domain in PA, or, you could imagine, antibodies which bind onto the newly formed surface of the PA heptamer. Of course, antibodies to EF or LF have potential value; LF competitors. These could be fragments of LF peptides, and it's been demonstrated in the Collier lab, also, that these could be made more effective by increasing their valency through multimerizing them, because then you get an evidity enhancement.
Dominant negative protective antigen mutants. I think we'll hear more about those later from Dr. collier. And once we get inside the cell, as mentioned, you have two enzymes which are susceptible to development of small-molecule drug inhibitors, either an adenolate cyclase inhibitor, or a protease inhibitor.
And then there's the large class of molecules that could be imagined as dealing with downstream consequences of the toxin action. And these are more in the class of supportive therapies. And as we know more about the pathogenic--the consequences of these two events, we could perhaps select a better set of supportive therapies. [Slide.]
I'm going to give an example of a monoclonal antibody that's been developed. And here I'm just indicating again the potential targets for antibodies--targeting the extracellular steps. And we're guided in this work by work done in the middle-'80s in USAMRIID, where Steve Little and others developed a set of mouse monoclonal antibodies. And the ones that were neutralizing and are best characterized are 14B7, which binds to Domain 4 and prevents toxin binding to receptor, and 1G3, which binds to the newly exposed surface on the PA heptamer and essentially competes with LF and EF binding. That's at this point. [Slide.]
So, I was involved in a small way, and worked from the Lab of George Georgio, and what that group did was to take the 14B7 mouse monoclonal antibody and clone the genes, and produce a single chain antibody, based on the 14B7 sequences. They then went on, by phage display, to engineer a higher affinity variant of that antibody, and they were able to increase the affinity on a monovalent molecule about 40-fold. And so 1H was their improved single-chain antibody.
Then they resorted to the rat model, which I referred to earlier as a test of antitoxins. And again using the monovalent 14B7 they were able to recuse rats which were dying in the control group at 91 minutes--they rescued, in fact, 0 of them--none of them. It was perhaps a small delay in time to death.
But the affinity-enhanced variant of 14B7 saved three of the five rats and delayed the time to death of the other two. [Slide.]
And this molecule--the enhanced 14B7--is actually being developed by Elusys, and they have humanized it, as you would want to do, and made it into a full-size antibody. In the interest of disclosure, I'd just say I have no financial interest in this, although I had a small hand in the early steps of its development.
So I think the antibodies have obviously attracted lots of interest as therapeutics. And that's, in part, because there's lots of expertise and skill in humanizing antibodies and producing them in large amounts. One could imagine that in a scenario where there's a mass exposure of a population one could administer a single dose and provide a number of weeks of protection against infection. It might not be practical to treat large numbers of symptomatic patients, but these antibody reagents could certainly have a potential role in that aspect.
And there's always the concern about antibiotic-resistant strains. Again, these antibodies would remain resistant effective against those strains.
Less mentioned, I think, but deserving notice, is that an antibody product could provide immunity to infants, children, immunocompromised persons for whom vaccines might not be available or effective. [Slide.]
So, I'm going to skip that--another example of an antibody which has been developed by a different company--Alexion--and it's based on the 1G3 molecule. And it has some unique properties which, I guess, are probably evidence in the slides that I provided. [Slide.]
And here's an article from The Washington Post I thought was well written. This is where I get a lot of my information-- [Laughter.] --about anthrax therapeutics. And this was from March, and it was pointing out the work from Elusys that I referred to and the perhaps even better known work from Human Genome Sciences. They also have a human antibody which seems to be highly protective in several models of infection. [Slide.]
And so I highlight the quotes here: "So many companies have responded to the task that the government now confronts an embarrassment of riches." And I think that's true. I think there are so many promising antibodies coming down the road--I think the public press lists at least five companies that have such products--that--and I think all of them have efficacy. I have every reason to expect that they would all be efficacious.
And so the director of my institute, Dr. Fauci, said we would completely break the bank if we committed to purchasing every one of them. So there is going to be a problem in choosing between these products. [Slide.]
I'm now going to turn to a few examples with small-molecule inhibitors targeted to anthrax toxin. And, again, we have two enzymes here which are potential targets. This is a slide prepared for Dr. Fauci which I swiped from him, where he highlighted, in testifying downtown I believe, two drugs which have shown some promise--at least in very early studies.
And this one I had a small hand in. And this is an edema factor inhibitor. So the furin inhibitor--I pointed out furin is a potential target--the work here is from Iris Lindberg in Louisiana, and she has an inhibitor which hexa-D-arginine, which she had been developing as a furin inhibitor. And in this experiment she showed--again in the rat model--that the control rats were dying very promptly after just a few hours. Co-administration of this furin inhibitor did save half the rats. So this is a demonstration that furin inhibitors have potential. [Slide.]
And the other inhibitor I want to just draw attention to is targeted to the adenylate cyclase. This is a model of the edema factor with calmudulin very tightly bound to it, and without. And Wei-Jen Tang, at the University of Chicago screened a number of compounds, and was able to identify adefovir as an inhibitor with nanomolar inhibitory activity against the enzymatic activity of the edema factor. And this shows, in cell culture models, that the adenylate cyclase production induced by edema factor--in the solid symbols--is blocked as you increase the adefovir concentration. And so he's--Wei-Jen Tang is trying to carry this forward as a first-generation, or at least a candidate lead compound from which other edema factor inhibitors might be developed. [Slide.]
Lethal factor is perhaps has been a more popular target for inhibitors because it is a protease, and the pharmaceutical industry has comfort in dealing with--searching for protease inhibitors. The structure of lethal factor has been solved, and a number of inhibitor candidates have been identified. And this is just a model of the active site of the lethal factor protease, in which three different--or I should say two protease inhibitors have been superimposed, along with the natural peptide substrate. [Slide.]
Also published--already, now, two years ago--from Merck was a paper describing a peptide-based fluorescence assay which--it was clearly developed with the intent of screening this company's large family of protease inhibitors. And while we haven't heard anything more from them in public--to my knowledge--I do note that they're presenting this work at the Gordon Conference next month. So there's indication they're continuing to work on inhibitors to lethal factor. [Slide.]
So, again, I've highlighted a number of targets at which one might expect to block the activity of anthrax toxin and thereby protect animals and humans who are infected with bacillus anthracis. I'm impressed, in the short time that this has been--these targets have been under frontal attack by academic and pharmaceutical companies, I think tremendous progress has been made. And I think we can look forward to development of some effective products int he not too distant future.
So that's all I wanted to say. And now I'm going to pass the baton on to the next speaker. [Applause.]
Clinical Aspects of Disease
DR. STEPHENS: Good morning, and I thank you very much for the invitation to be here.
Anthrax, from a clinical perspective, was of historical interest prior to 2001. In the 20 th century, some 18 cases of inhalational or inhalation anthrax were reported. Most of those were in millworkers, associated with, in this country, goat hair importation. There were some cases of cutaneous anthrax occasionally in the midwest. But, certainly, 2001 was an important point in terms of our appreciation and understanding of some of the clinical issues of bacillus anthracis. [Slide.] now, I think there are several lessons from that outbreak--and certainly lessons from my perspective. Those are the impact of surge; the clinical recognition issues and the differential diagnosis of anthrax. Some of the issues of management--and we're obviously focused today on anti-toxin approaches, but some of the issues of antimicrobial management were equally--and remain equally as important; issues of immune response; and certainly gaps, in terms of our ability to rapidly diagnose anthrax, its different clinical manifestations; issues of use of the vaccine; and issues of prophylaxis, in particularly, in terms of drugs. [Slide.]
Now there were 23 cases of inhalational--or 23 cases total of anthrax in the outbreak; 11 inhalation and 12 cutaneous. I'm including one laboratory-acquired case that occurred some months, due to handling of specimens from the outbreak.
However, there was at least a log higher of cases in which there was real concern about anthrax and the differential diagnosis of these cases remains--is, and has been, critical in terms of trying to rule out anthrax. So this required a lot of effort and a lot of involvement of the public health community, and as well as both the state and national level.
There were a lot of individuals evaluated--at least a log higher individuals evaluated for anthrax. 30,000 individuals were started on prophylaxis because of exposure in the areas Florida, Washington, D.C., New Jersey, New York. A number of people were obviously directly affected, and virtually the entire population was impacted. [Slide.]
Now, we won't go into this. This is the--the point, really, of this slide is to emphasize the three clinical forms of disease: cutaneous, inhalation and gastrointestinal. [Slide.]
Most of you know this history well. The initial case report: a 63-year-old man from south Florida who had been on--he became ill on a vacation. He developed fever, myalgias, cough, headache, nausea and vomiting, and then he developed altered mental status and presented to a local hospital in Florida with a diagnosis of meningitis. [Slide.]
His laboratory exam revealed a leukocytosis; LP was remarkable from a neutrophyllic pleicytosis, with large numbers of polys, but also--which is characteristic of anthrax meningitis--large numbers of red cells found in the spinal fluid. And this actually turns out to be a rye stain of the spinal fluid, showing large numbers of polys, and obviously large numbers of Gram-positive bacilli, which rapidly grew B. anthracis. [Slide.]
The investigation: photographer for a tabloid newspaper. He was on a vacation when he became ill. Computer cultures yielded B. anthracis from the surface of the computer. Nasal swabs of a number of individuals in the building also were positive. Prophylaxis was ultimately given and the outbreak began. [Slide.]
I think you're obviously very familiar with this kind of picture, in terms of inhalation anthrax. Incubation period is--from historical records--some 2 to 14 days, with a range of up to 60 days. And most of this has been covered by Steve in his talk.
Again, the importance, though, of mediastinal disease, edema, hemorrhagic mediastinitis, subsequent hemotogenous spread and meningitis should be emphasized. [Slide.]
This is a diagram from a JAMA article looking at the inhalation anthrax in Sverdlovsk; the outbreak associated with a bioweapons plant, looking at number of days after the accident and the onset of inhalational cases following the accident, up to one case at 43 days after exposure. [Slide.]
There were two patients in the 2001 outbreak: one from New York City, and a second one from Connecticut, in which there was probably a longer length of incubation, although the exact time of exposure for those cases is not known. [Slide.]
This is a summary from the article that was published by us in Emerging Infections in 2001. And the second are the 11 cases published in JAMA in 2002, looking at the clinical features of inhalation anthrax. Median age was 56. Most were males. Incubation period--which I guess is the important feature on this particular slide--was four days, in which the incubation period was known. And the median duration of symptoms prior to presentation was 3.5 days, with a range of one to seven days. [Slide.]
Major features, in terms of symptoms were chills, fever, fatigue and malaise. Night sweats, in particular--or sweats in particular--drenching sweats--were also noted in a number of patients; a non-productive cough, nausea and vomiting, a dyspnea, chest discomfort--which was described in the older series, also occurred in these patients; rhinorhea and sore through--upper respiratory symptoms--were infrequent in these individuals. [Slide.]
Fever, tachycardia was common. Very few of these patients were hypotensive on admission. Some subsequently develop hemodynamic instability. But hypotension as a presentation was uncommon. This is very different from, say, meningecoccal septicemia, for example, where hypotension and DIC are common presentations. [Slide.]
The initial laboratory findings: the white count was high, but not excessively high; neutrophilia, however, was present, with greater than 70 percent neutrophils present on the initial WBC. Transaminases, interestingly, were elevated in 10 of the 11 inhalation cases. And hypoxia, by some measure, was noted in 7 of the 11. [Slide.]
The diagnosis was made by blood cultures in eight of the individuals who had not received antibiotics. Interesting, any antibiotic therapy rapidly sterilized the blood and the diagnosis in three patients was established by other, newer technologies: immunohistochemical staining of transbronchial biopsy specimens or pleural biopsy or pleural fluid, and detection of DNA by PCR in blood or pleural, and by the detection of immune response to PA. [Slide.]
The initial radiographic findings in patients with anthrax: the chest x-ray was abnormal in all 11; mediastinal widening, infiltrates or pleural effusion was noted. Mediastinal widening-- considered to be the classic for inhalation anthrax--as not present in everyone. And, again, the chest x-ray findings, although abnormal on admission, the findings were--could be subtle. And on a couple of occasions, the initial abnormalities were, in fact, missed.
And the chest x-ray abnormalities were noted within 48 hours of onset of presentation. [Slide.]
This is Case 1--our index case from Florida. And I think you can appreciate that this individual did have mediastinal widening in this setting. [Slide.]
However Case 2 from Florida was somewhat different in terms of its presentation: presented with infiltrates and a pleural effusion that persisted--and really never did have significant mediastinal adenopathy. [Slide.]
And this is a CT scan showing the large pleural effusions, which were characteristics of these patients, and really complicated their clinical course. [Slide.] This is the case from the Washington area. There--and I think the findings were more subtle in this case. There is some mediastinal widening in this particular patient; maybe an early development of a pleural effusion on the left side, but findings can be, in fact subtle. [Slide.]
CT was more sensitive, showing mediastinal adenopathy in this particular setting. [Slide.]
So, in summary: profound seating and GI symptoms were notable; chest x-rays were uniformly abnormal, although with a variety of findings. And, again, the initial films could be--were subtle, and findings could be missed.
Blood cultures were positive early in the course, before antibiotics.
Pleural effusions were an important feature of the illness. Frequently it required drainage. And, certainly, some of the issues regarding improved survival have to do with aggressive attention to these pleural effusions, and the alleviation of the respiratory compromise that was characteristic of these patients.
Pleural infiltrates were found in over 60 percent of patients. Survival was higher than the 15 percent previously reported. And, again, the emphasis on the newer diagnostic tests--IHC, PCR and serology--were very helpful in understanding the spectrum of this disease. [Slide.]
This is a brief review. This is Jeanette Guarner and Sheriff Zaki at the CDC published this review of the pathology of inhalation anthrax on five fatal and three non-fatal cases--again, emphasizing these serosanguinous pleural effusions, the hemorrhagic mediastinitis, and the presence--in multiple organs, especially in the patients who died--of bacilli--of cell-wall or capsulary antigens in multiple organs. And, again, in this particular study, IHC was an indispensable tool. [Slide.]
And just a couple of quick slides from that paper: hemorrhagic mediastinal lymph node; lots of inflammation and hemorrhage in the mediastinum. And these--this is an IHC looking at antigen positivity in these specimens from patients who died. [Slide.]
This is a series of pleural studies on patients during the outbreak of cell block, looking at pleural reaction. And these also include a pleural biopsy--again with lots of reaction at the pleura, with lots of B. anthracis antigen present in these specimens. [Slide.]
Now, the differential diagnosis of inhalation anthrax includes influenza or a viral syndrome. Several of these patients were thought to have a viral syndrom; actually sought medical attention and then were sent home with that diagnosis.
The atypical causes of pneumonia--from mycoplasma through viral pneumonia, Q fever, psittacosis, Legionnella. And those conditions that are known to give you mediastinal are endothoracic lymphadenopathy, histoplasmosis, coccidiodiomycosis, tuberculosis. And one patient was actually admitted and being worked up for a malignancy. [Slide.]
This is--turning to the cutaneous cases, case one, from New York: a 38-year-old woman, assistant anchor, developed an erythematous papule on her chest; three days. She developed a vesicular, ulcerated, edematous lesion; had headaches, malaise, satellite vesicles; was started on Ciprofloxacin, and by 10/9 of '01 had developed a black eschar. IHC and serology was positive. [Slide.]
And this is just the late features and manifestations of anthrax--cutaneous anthrax--in that particular patient. [Slide.]
The only child in the outbreak was a cutaneous case: actually a seven-month-old child who, after visiting a network studio with his mother, his arm became swollen and he was given augmentin but he remained febrile. He was ultimately admitted to the hospital and had edema and a large black eschar. Interestingly, he and one of the inhalational cases, late in their course, had evidence of hemolytic uremic-like syndrome, with hemolysis and thrombocytopenia. And in this particular patient, IHC and PCR was positive on blood for B. anthracis. [Slide.]
And this is the ultimate course of his particular lesion in this child. [Slide.]
And this just summarizes--the letters--some of the New York cases on onset. This was actually after this second--or the New York Post letter was identified, and resulted from handling of that particular letter. [Slide.]
So, for cutaneous anthrax, the incubation period is 1 to 12 days. The papules are painless. Papules progress to vesicle or bullous formation with surrounding, nonpitting edema. The central vesicle becomes ulcerated and necrotic, and surround--and becomes often surrounded by satellite vesicles, subsequently forming this black eschar, which is characteristically depressed and painless.
Fatigue, chills, fever, regional adenopathy may occur in these individuals. [Slide.]
This is not from the outbreak, but just gives you a better sense of some of the progression of lesions from vesicle and papule formation, to eschar formation over a period of 7 to 10 days. [Slide.]
Again, not from the outbreak, put showing you some of the differences--some of the clinical presentations of--and the earlier ulcerations of cutaneous anthrax. [Slide.]
This actually is from the outbreak, showing an eschar--actually a debrided eschar--on a finger in one of the cutaneous lesions. [Slide.]
And this is actually an early lesion associated with the outbreak, showing you the initial vesicle formation. [Slide.]
And this also is from the outbreak. And as you will--again, showing--this individual also had secondary staphylococcal bacterial infection complicating his anthrax, which were seen in a couple of individuals. This individual also had positive blood cultures for B. anthracis. [Slide.]
So, the differential diagnosis is important. These kind of cases continue to occur. Although the outbreak is obviously over, the issues of differential diagnosis, the issues of unusual rashes and the concern about future cases continues. And there's a lot of--there's importance, obviously, in appreciating what are the most common differential diagnoses--what is the most common of the differential diagnoses of cutaneous anthrax.
Interestingly: spider bites. In a review of Sheriff Zaki, isolated lesions of varicella Zoster were actually quite common in the differential diagnoses--from the New York area in particular, Rickettsial pox; herpes simplex type 1 also is in the differential diagnosis, as are the more traditional lesions associated with cutaneous anthrax: erythema gangrenosum or pyoderma gangrenosum, which also can present like cutaneous anthrax; tularemia, plague--and the importance, again, of the common, but sometimes presenting in an eschar kind of way, in particular staphylococcal infections. [Slide.]
This is actually a case. Cases like this often occur on Saturday night. And this one occurred on Saturday night, and was a child of a laboratory worker, who was real concerned about what this was. It turned out to be a Brown Recluse spider bite in this particular child. But it gives you a sense of the differential diagnosis that is important to consider in these kind of individuals. [Slide.]
Now, this is an article--Conrad Quinn's going to be talking in a minute. There was an opportunity, obviously to examine the immune response in patients who were a part of the outbreak. And this is--the data I'll show you is an article that's in press in Journal of Infectious Diseases. Twenty-two patients comprised this group--this study. Serial serum samples were obtained; humoral response to PA and also LF is being looked at in these individuals, and also toxin neutralization. [Slide.]
And I won't go into this assay. This is from Conrad's paper in Emerging Infectious Diseases, looking at the validated anti-PA IgG ELISA--originally developed to look at vaccine questions, but rapidly adapted in the outbreak to be very useful clinically with a very good sensitivity and a specificity which can be enhanced by a competitive ELISA. [Slide.]
And I won't go into this. [Slide.]
These are data from the outbreak looking at anti-PA IgG antibody in patients with inhalational anthrax, noting that in some individuals the response was quite high, approaching 1,500 mcg per ml. Most of the individuals--the lowest was around 150, as I recall--in terms of peak levels, the peaks generally occurred approximately 30 to 60 days after onset of symptomatology. And in following these patients out to a year, they continue to have levels of anti-PA antibody present. [Slide.]
Cutaneous patients were quite different, however. They--only a couple of them mounted response greater than 100 mcg. Some of them were high early and came down quickly. In a couple of individuals the response was actually quite low, and there was a question of why this was the case. Was this antibiotic suppression? Many of these patients were started on antibiotics fairly quickly.
But the data would suggest a very different picture between cutaneous disease. The one patient shown here, who mounted one of the higher antibody responses, was also a patient who was bacteremic. [Slide.]
There is a very good correlation--in work looking at the levels of anti-PA IgG with toxin neutralization, a good correlation between toxin neutralization and anti-PA antibody. In work done with Shane Croddy and Al Humed at Emory, we've been able to look at specific IgG memory cells in patients--the patients with inhalational disease versus cutaneous disease, and all of the six patients who were available, who survived, had evidence at six months and longer, of memory B cells that were present in individuals with inhalation disease, versus only one--and this was the bacteremic patient--who had evidence of memory B cells with cutaneous disease. [Slide.]
And I won't go into that. [Slide.]
So IgG and anti-PA antibodies in patients with inhalational anthrax were detectible 11 days after symptom onset. Anti-PA was a predictor of toxin neutralization and the development of specific PA memory B cells. And in the cutaneous anthrax patients, the magnitude of anti-PA-specific IgG and toxin neutralization and memory B cell response was less. And there really were two groups-those with a rapid rise and fall, and those with a very low and delayed response. And the reasons for that are not--at least in my mind--clear. [Slide.]
I won't go in--because of the focus of this particular meeting--in terms of the antibiotic issues. I will point out that number of these patients did get protein inhibitors known to have anti-toxin effects; in particular, Clindamycin. And whether that was a component of the increased success, certainly, aggressive antimicrobial therapy, aggressive drainage of pleural effusions, aggressive supportive care were, in my view, key instruments in terms of the success. [Slide.]
There are lots of issues that I think are still out there regarding treatment--the best antimicrobial regimen; the treatment of meningitis, what are you using in the setting of meningitis, a pretty much uniformly fatal disease?
What about steroids? Steroids were used in several of the individuals--especially the individuals with extensive edema and cutaneous disease. Again, we're talking anecdote, in the sense of a limited number of patients, in terms of the data that we have in humans clinically.
Length of therapy, persistence of spores, and how long do you continue prophylaxis. There was, and still is some controversy about how long you should continue prophylaxis. [Slide.]
And I won't go into this. This has to do with issues of long-term use of antibiotics. [Slide.]
After the outbreak, there was a meeting to discuss research priorities. And, obviously, this meeting is a continuation of issues of antitoxin immunotherapy and how to best approach that; also issues of antibiotic therapy, and the importance of animal models and establishing animal models that are reliable and predictors of human disease. As most of you know, there are lots of issues in that particular area.
I won't go into the anthrax vaccine, but in the interest of time I want to acknowledge the role of the CDC, the Meningitis and Special Pathogens Branch, the National Center for Infectious Diseases, the Clinical Team and State Teams, who were very instrumental in collecting a lot of the clinical data; obviously, Conrad Quinn and his laboratory at CDC; Sheriff Zaki and his laboratory; Patter Dull and Carolyn Greene, two former EAS officers who played a significant role in the evaluation of and obtaining specimens on the patients--surviving patients; John Hernigan at CDC and Emory; the local health departments; and individual physicians who were caring for these very sick patients.
So, I appreciate your time.
[Applause.]
DR. FRUCHT: Thank you, Dr. Stephens.
PART II - In Vitro Characterization
DR. FRUCHT: It's my pleasure to introduce you to the In Vitro Characterization session of this workshop.
I'm David Frucht. I'm from the Division of Monoclonal Antibodies, and I'm happy to say that we've assembled an excellent group of speakers and panelists today.
I should say that our speakers will only be covering a subset of the bioassays that are used to characterize the large variety of potential anthrax therapeutics. However, with the group of experts that we'll have on the panel, I'm sure that we'll be able to answer any other questions, or discuss points that aren't covered in the talks.
Our first speaker today is Dr. Conrad Quinn. He's the Chief of the Microbial Pathogenesis and Immune Response Laboratory. Dr. Quinn received his Ph.D. in microbiology in 1989 from the University of Wales, following which he did his post-doctoral training at the NIH. The current focus of his laboratory is the development of validated immunoassays for the diagnosis of anthrax in humans, and for quantitative evaluation of humoral immune responses to anthrax vaccines. In addition, his laboratory performs immunoassays for clinical trials in the CDC anthrax vaccine research program.
Dr. Quinn?
In Vitro Assays to Characterize Anti-toxin Based Therapies
DR. QUINN: Good morning, ladies and gentlemen.
Can I get the first slide? [Slide.]
This morning I'd like to touch on some of the in vitro assays that we have been developing at the CDC for evaluation of toxin therapies. We will focus on one assay in particular--the toxin neutralization assay--because of its broad spectrum application. [Slide.]
Well, I'll start by firstly recapping on some of the things we've heard already this morning from Dr. Leppla and Dr. Stephens.
The causative organism of this disease is bacillus anthracis, which is a Gram-positive spore former. It's a large bacilli within the bacillus cereus group. It can be distinguished from its close cousins by its clear characteristics of the absence of motility; usually penicillin-sensitive, usually gamma-phage sensitive; and it is non hemolytic--which distinguishes it from bacillus cereus.
It also produces a tripartite protein toxin and a gamma-linked poly-D-glutamic acid capsule--which we heard about, again, this morning. And you can see it hear, stained with a McFadden stain on the outside of the organism, growing in serum or blood.
I'd like to focus on these two components here: the tripartite protein toxin, and the acid capsule, because these are its major virulence determinants. [Slide.]
Of course, the proteins can now be produced and purified to high levels of purity for analysis and antigen development and therapeutic molecule development. They are, interestingly, serologically distinct, which is illustrated nicely here by the rather old-fashioned but still very effective Ouchterlony double immunodiffusion technique. And, as we heard from Steve Leppla this morning, these toxins--these three proteins--act in binary combinations of PA and LF to generate lethal toxin: protective antigen and edema factor to generate the edema toxin. The effects of the lethal toxin are now known to be due to its anti-protease activity, which affects cleavage of kinases. It has also been shown to lyse certain macrophage cell lines in vitro; shown first by Dr. Friedlander at USAMRIID in the mid-'80s, and subsequently developed as a neutralization assay by Steve Little at USAMRIID.
We also know that it has an effect on cytokine modulation and perhaps immunosuppression in the early stages of infection. And some of the characteristics of its fatal effect in animals are hypoxic insult. [Slide.]
The edema toxin is known to be an adenylate cyclase, converting ATP to cyclic AMP intracellularly. This has also been demonstrated to have some level of cytokine modulation, and the gross characteristic effects are the edema of the infection, characteristic in its diagnosis. [Slide.]
Again, we saw this morning from Steve the mode of action--or the accepted mode of action from the toxin: as the protective antigen binds the cell receptor, gets cleaved by surface proteases such as furin; the 20-kilo-doltan fragment is lost, leaving this 63-kilo-doltan monomer which then heptamerizes.
This heptamer then complexes with edema factor or lethal factor to form a complex which is internalized through septa-mediated endocytosis. And after acidification of the endosome, is translocated into the cytosol where the two different toxin enzymes exert their different effects. [Slide.]
All three of the toxin proteins have been purified and crystalized. And we know that they have different demand structures--and these become relevant in terms of developing therapies, particularly for protective antigen, which undergrows this conformational shift and change when it forms the heptamer, exposing new sites, and perhaps hiding earlier epitopes.
Lethal factor also--which I'll focus on very briefly--crystalized and became structurally eluded. We see here the ainc atoms buried in the catalytic domain, in Domain 4. And the relevance of these structures, and our understanding of these structures and the conformational changes during intoxication, indicate that there are multiple sites of intervention for developing therapeutics; be they blocking interaction with the receptor, blocking cleavage by the activating proteases such as furin; blocking heptamerization; complex formation; internalization and translocation; as well as the individual enzymatic activities of the proteins themselves. [Slide.]
A very brief review of what's in the literature shows that the small-molecule inhibitors fall into three categories: inhibitors of edema factor themselves, as measured by reduction of adenylate cyclase activity, either intracellular or extracellular. And here we have just two relevant publications from Soleman and Shen.
The second group would be inhibitors of lethal factor, focusing on its endoprotease activity. And we have three representative literature citations here. These fall into aromatic pharmacophores, peptide inhibitors, and also polyphenol catechin.
The third group is innovation of protein interaction. And here I've put the furin inhibitors, such as Hexa-D-arginine--which Steve Leppla referred to this morning; but also complex inhibition, such as the polyvalent protein decoys and dominant negative mutants which have been developed in John Collier's lab. [Slide.] in terms of immune products, these focus on ployclonal or monoclonal antibodies--and these are, again, taken from the literature. In the late '90s Steve Little, et al., developed monoclonals from anti-AVA vaccinated mice, and also PA-specific and LF-specific monoclonals.
There are monoclonal anti-AVA, focusing on protective antigen. And, of course, monoclonals raised against recombinant proteins. And these all are featured prominently in the literature in the last few years. [Slide.]
Steve Leppla also related to Tony Fauci's comment that we can't address or invest in every immune product that's out there. So at the outset of this year, with the CDC, we were mandated by HHS to find out what was out there in terms of what candidates in immune-product development are available; what might be their stage of development and their availability for product development; and to do some sort of initial evaluation, using a uniform technology platform that would allow us to formulate a procurement strategy for later this year or next year. [Slide.]
At the end of last year we put out a request for information, requesting responses by February of this year, in which we proposed to undertake an in vitro analysis of some of these products, using anti-PA analyses, binding assays, but, more importantly, the lethal toxin neutralization assay, which I'll focus on for the rest of this presentation.
This neutralization assay is a functional assay. It's essentially species-independent, and we intend to have this preliminary evaluation of responses to the RFI completed by the end of July this year. [Slide.]
The features of the CDC assays that make it attractive in this context are that we have generated, as part of the anthrax vaccine research programs in the NIH RPA clinical trials, a series of qualified reference standards and reagents. We've also developed standardized technologies for these trials. And if we focus on the neutralization assay, which is lethal toxin-specific, containing both protective antigen lethal factor, we know that this assay is not species or molecule-dependent. We have modeled the response curves using the four parameter logistic log model. And this combination of science and mathematics allows us to generate calculatable endpoints with high precision and with high accuracy.
In some instances, where appropriate, we continue to use the ELISA--for example, for comparing polyclonal or monoclonal antibody products--human antibody products--to anthrax immunoglobulin which is being developed by CDC and HHS as an emergency response measure. [Slide.]
Using this assay, the sort of reportable values that we generate are effective concentrations, giving 50 percent neutralization, and the different ranges around this bioassay curve, such as concentrations giving 90, 95 or 99 percent neutralization. And at the low end of the curve, concentrations giving 1, 5 or 10 percent. And hopefully this will become clearer as I go through the presentation. [Slide.]
Well, let me start first by briefly describing the way the assay works. We have a fixed concentration to protect lethal factor. The lethal factor here is in a stochiametric excess. We have a fixed concentration of cells per well in the bioassay plate, and we present varied dilutions of the test material. We record reporter signal as a surrogate measure of viability against dilution of the product in the plate. [Slide.]
At the upper end of the curve we have our positive neutralization. Control. And, again, a zero neutralization control or 100 percent effective killing.
We then model our standards and our products, using a four-parameter fit sigmoidal curve to transform data. And this four-parameter model allows us to measure or determine the upper asymptotes of this curve, the lower asymptote, and the inflection point, as well as the gradient of this curve.
The inflection point of the four-parameter logistic log model we refer to as the 50 percent neutralization, or ED 50--effective dilution giving 50 percent protection in the cells. [Slide.]
Because we use the four-parameter logistic log model, and developments thereof, we also can pick specifically, and with precision, different points in this curve which give us different measures, which we refer to as the quantitation titer and the threshold titer. [Slide.]
These calculable or reportable value from these slopes are generated from a mathematical interpretation of the four-parameter curve; the first and second derivative. The first derivative measures the slope and the changes in the slope with that original function of the bioassay data. The second derivative measures the rate of change of slope in the original function. [Slide.]
This is shown graphically here, where we have the bioassay curve with this four-parameter logistic log curve fit. And here we have the plot of the first derivative. At these intersections we have the threshold titer. And this is the first point in the curve which, after empirical evaluation, is shown to be statistically significantly above background--or the lower s-asymptote. [Slide.]
Looking at the second derivative, based from the first derivative--so this is totally data-driven--we have the minimum and the maximum points of the second derivative which define a linear or usable portion of this four-parameter logistic curve. And the lower of these we refer to as the quantitation titer. So this has got a higher level of robustness, mathematically, than the threshold titer, but it has lower sensitivity. That's why we opt to use both of them. [Slide.]
So, taken together, then, we have these three reportable values from the neutralization assay: the ED 50, which is the inflection point of the four-p l fit; the threshold titer, which is the lowest point on that curve we report with acceptable precision; and the quantitation titer, which defines the usable portion of that curve. [Slide.]
So if we have the standard curve fitted to each plate, and the test curve, giving a sigmoidal curve, things are fine and dandy. We can report either out our ED 50, and the CT and TT at the lower points of the curve.
More significant for serological responses or vaccine responses for which this assay was developed: the model fit also allows us to evaluate low but reactive sera responses or product responses by back-modeling but constraining to the positive controls. So we can develop theoretical ED 50 should the need arise.These together with the threshold titres are the reportable values for our products. [Slide.]
The benefits of using QT, TT and the four-parameter logistic model fit are that it gives us higher precision of reproducibility than other available methods. The method itself has broad application to other assays; not just the TNA, but ELISA, for example.
And currently it's being developed in SAS. We call it the Taylor Method, after the statistician at CDC who's developing it. The four-PL model, together with the high through-put analysis that this system provides gives us the opportunity to apply consistent and objective QC criteria to anything that we evaluate in this system. [Slide.]
We also apply these assays and comparisons to a rigorous QC criteria. The standard curve provides this QC parameter, together with positive neutralization and negative neutralization controls. But the standard itself must return within an expected range and allowing a bracket of two standard deviations.
There must be a good relative fit of the standard stated to the model. We must have a good distribution of the data points across that model. We must have sufficient depth of curve, with a maximum OD and a minimum OD which are acceptable, showing good viability and good cell density.
And we also have low variability in the standards data, as well.
The negative serum control and the positive neutralization control define the upper and lower limits of the assay, indicating that it has succeeded.
So with these QC criteria, as well as the mathematical interpretation, this is the system which we are evaluating current responses to the RFI of February. [Slide.]
If you look at the way this assay performs in terms of AVR414, which is our human standard reference serum, here we have a small subset of 96 plates. We see that the precision is high, with 7.7 percent CD. Intermediate precision of the assay is also acceptable and is good. We have four here: three test samples and the reference standard itself, all the CDs returning under 30 percent.
The mean goodness of fit of the data to the standards model is high. And the ED 50--the inflection point of this model for the AVA414 standards curve, is robust and reproducible--high precision here, as well.
The assay is sensitive--the assay system is sensitive, with a lower read on the standards curve of 41 nanograms per ml, and a quantification, which is the range between the QT readings on the second derivative of .07 to .3 for this particular standard serum. [Slide.]
We've evaluated this assay in a variety of species, including monoclonal antibodies--Murang developed at CDC. Polyclonal antivaccine antibodies in different species. And what we found is the performance characteristics of macaque, mice and rabbits on monoclonals are very, very similar. [Slide.]
And this is illustrated here, just by giving the basic sigmoid curves from these three standard reagents in the same assay. The data points and error bars are left off for clarity.
But the point here is that the different species generate similar curves--different slopes in some instances, but essentially upper and lower asymptotes and inflection points characteristic of the human response. [Slide.]
In this assay we know also that there is a good correlation between the neutralization efficacy of serum antibody and the ELISA quantification of polyclonal serum antibody in both humans and also in macaques, who have a high correlation coefficient of .84. [Slide.]
So when we put this assay to looking at immunotherapeutic testing, we actually flip the curve around, because we are no longer interested in just dilutions of product or serum, we now want to address concentrations. So we use the same parameters, the same set-up, the same mathematical evolution and development, but now we flip it around so that we convert dilutions to concentrations.
And this is what a typical standard response looks like. We have our sigmoidal curve with the four-parameter logistic log fit. We have our inflection point of the model which gives us 50 percent neutralization OF THE EC50. And we also report EC 1, 5, 95 and 99. And this becomes important where we have curves which are not parallel. [Slide.]
This is data generated--this is actual output from our SAS algorithm--again, based around 414--when the data has been transformed. We see a nice sigmoidal fit. And here we have the reportables at this point on our dilution scale of 1 percent, 5 percent, 50 percent neutralization--95 and 99 percent neutralization.
The untransformed data curve looks like this, and at this point when we map it from dilution back to concentration.
So for this particular polyclonal of human vaccine E serum, we see the EC50s here are 167 nanograms per mil. This is from a small subset of the recent evaluation data. [Slide.]
So of outputs that we generate for each product, including the AVR414, the ED50 dilutional, QT and TT dilutional, and then the concentration values.
And here we can see that for each of these reportable values, we have high precision, as reflected in low CV values here. And, as expected, the EC50--the inflection point of the curve--has the highest precision, manifest in the lower CV, and as we go down to the lower parts of the curve, the precision is lower, but still acceptable. [Slide.]
Other outputs that we capture from this curve are the asymptotes and the slope around the inflection point. And there's a change here from the slide as to what is in the notes. There was a typo on the first copy. [Slide.]
The reason for capturing all of these --as I alluded to--is that some curves will not be parallel. This assay does not address parallelism, but we want to be able to capture--for example, here--the standard; a nice curve which will report an ED50 or EC50 of a particular value.
But another product, or another serum, could have a similar or identical EC50, but very different slope characteristics. So, at this point in the preliminary evaluation, it's important to capture the range from 1 to 99, or 5 to 95, as well as the EC50 and the slope of that curve.
We would expect that the upper and lower asymptotes of the curves are agreeable to similar. [Slide.]
Here we can see shifts to the right for lower potency molecules, and to the left for higher potency molecules. [Slide.]
We do recognize that at this stage in the assays, based on the rationale for which they were developed that there are limitations. The ELISA, for example, which we use extensively to analyze human response to vaccines and infection is restricted to human antibodies with Xc components; and that automatically excludes non-human antibodies, fabs, single-chain fraction variables, mimetics, and other small molecule inhibitors. So there are major limitations to using straightforward binding assays.
The neutralization assay, however, is our broad spectrum application at this time. But it, too, has limitations.
Currently, as designed, it emphasizes the contribution of the PA83 molecule. And it is also possibly limited to PA and lethal factor or receptor-binding therapeutics, but that has to be countered with the fact that it can still be used in those arenas.
And we're currently contemplating modifications to broaden the scope of this assay so that it has more emphasis on anti-LF immuno products. We will be able to differentiate between pre- and post-receptor binding events. We'll be able to develop it for analysis of conformational-dependent events; and also for evaluation of small molecules. This is something we have not yet started in our labs at CDC. [Slide.]
So, to conclude, then, our focus in immunsup product therapy, immune therapeutic product evaluation, is on the neutralization assay, which we have demonstrated to be accurate, precise and robust.
We have a panel of standardized reagents and technologies which allow us a high level of quality control of the assay. And it has a flexible application in that it's species-independent and also quantifiable.
We do recognize the limitations currently, in that it primarily has a PA emphasis, but it can be optimized--and it is optimized--for antibodies other than small molecules.
An important next step in developing this, as well as broadening its scope of interpretation, is to relate what we see in vivo to what this assay tells us in vitro.
So, with that very brief overview of our work at CDC, I'll hand over to the next speaker. [Applause.]
DR. FRUCHT: Our next speaker is Dr. Jennie Riemenschneider. She's a biologist in the Office of Blood Research and Review at CBER, FDA. She received her Ph.D. in molecular virology from Case Western Reserve University. Later, she was an NRC fellow at the U.S. Army Medical Research Institute of Infectious Diseases where she worked on ebola and anthrax.
She joined the FDA in 2002, and has published work on sheep-derived anthrax antitoxins. She's currently studying a therapeutic role for both bovine-derived and human antitoxins.
Dr. Riemenschneider.
Development of Polyclonal Immunoglobulin Products
DR. RIEMENSCHNEIDER: Good morning. Today I have the pleasure of speaking to you about the unique and challenging issues that surround the manufacturing and testing of polyclonal antibody products.
As natural proteins of the immune system, antibodies make ideal drugs. And because of the inherent multi-valency of polyclonals, they have a unique place in the arsenal for the treatment of infectious diseases.
The Office of Blood at CBER has a unique perspective on these products, both from the regulatory and research points of view. As you know, we are responsible for the regulation aspects of such products but, in addition, we also have active research laboratories studying polyclonal antibodies, including those agents of interest to counter-terrorism, such as vaccinia virus and anthrax. [Slide.]
Immune globulins may very well be the first historically used plasma product, with Behring's work on diphtheria antitoxin, which is now over a century old. In 1893 he demonstrated that it was possible to treat diphtheria infection with serum. After additional research, Behring realized that antitoxin characteristic of blood was not found in the blood cells but in the cell-free serum. With his important discovery, Behring laid the foundations of modern immunology. [Slide.]
Immune globulins can be used for a variety of different conditions. First, they can be used in the prevention of a variety of bacterial and viral diseases, and this is especially critical in immune-deficient people.
Immune globulins are also critical treatment to those who have been exposed to certain pathogens, and this is the setting in which polyclonal Immune globulins are most likely to be used for counter-terrorism.
In addition, immune globulins have been used to prevent newborn hemolytic diseases, and also an immune modulation for patients with ITP.
Another critical indication for immune globulins is the role of antitoxins, which is especially important for this discussion. In this setting, polyclonal antibodies have been a critical treatment for diphtheria, as I mentioned, and also botulism and snake and spider invenomation.
Blood plasma contains a mixture of hundreds of different kinds of proteins, only a few of which are of therapeutic interest. To make plasma-derivative products, plasma can be treated in a variety of ways to separate the desirable products--in this case, immune globulin--from others. I'll go into a bit more detail about the manufacturing in a few minutes, but I do want to point out early that the process of obtaining antibodies from plasma is very complex, and a variety of different methods can be used.
The economy of scale of manufacturing immune globulins and the need for a wide spectrum of specificities requires a large donor pool--typically, at least a thousand donors. However, for the manufacture of hyperimmune globulins this number may be less, mainly due to the availability of appropriate donors.
And because they're biological products derived from both humans and animals, there is a safety issue related to the transmission of viruses and other pathogens. And, in addition, adverse reactions can be encountered.
So, as I mentioned, there's different sources for polyclonal immune globulins, human and animals. And they both have their own set of issues that need to be addressed. [Slide.]
For human-derived immune globulins, we must be acutely aware of the potential to transmit diseases. In addition, there's the possibility of unwanted antibodies, such as anti-D or isoagglutinins.
A concern that is extremely relevant for counter-terrorism-related immune globulin is the fact that there may not be a large donor pool, or population for the collection of plasmas manufacture for the desired product. Certain issues need to be considered, such as who and when can be plasma-pheresed to obtain the desired hyper immune plasma? And will new individuals need to be immunized or re-stimulated in order to collect sufficient amounts of plasma?
Animals are the other source of plasma for immune globulin production, however there is a risk that the immune globulins themselves from animals may cause adverse events due to their cross-species immunogenicity. One way to address this is to remove the immunogenic region of the antibody--the Fc, with a method called "despeciation." Residual animal proteins can also be a potential source of undesirable immune reactions, and because of this it is sometimes necessary to test for hypersensitivity and perform desensitization procedures prior to treatment.
And as with the human product, there is potential to transmit infectious agents--in this case, zoonotic agents, such as West Nile virus or rabies--although, to date, neither of these have occurred with plasma derivatives. [Slide.]
So, because polyclonal immune products have been used for several decades, we have extensive experience with these types of products. And this experience even extends to bioterrorism countermeasures, such as vaccinia immune globulin derived from human, and botulism antitoxin derived from horses.
Because these products were licensed many years ago, modern efficacy studies were not performed, and licensure was based on literature and small case studies. And so, as we know, requirements for licensure have changed and will be a major topic of discussion in today's workshop. [Slide.]
So, first, I'd like to focus on the human-derived immune globulins. And as I mentioned already, they're already used for variety of indications. The bulk of these are in a prophylactic setting. This includes the use of IGIV in primary and secondary immune-deficient patients, and also in the prevention of diseases such as hep B, tetanus, CMV and RSV.
Human immune globulins can also be used for treatment, although this is a less common occurrence. Examples of this are in the treatment of infant botulism and tetanus. However, for the purposes of today's discussion, this is an important indication, since this is where hyperimmunes are likely to be used in the treatment of anthrax.
I'd also like to point out that immune globulins can be administered by IM or IV injection, and they're referred to as IG or IGIF, respectively.
And, also, there are already currently a variety of specific or hyperimmune products, several of wich are listed here. [Slide.]
Now, obviously, human plasmas--the starting material for human immune globulins--and there's two classes of plasma recovered in source.
Recovered plasma is made by separating plasma in a donation of whole blood from other components, where source plasma involves the process of removing whole blood and separating red blood cells from plasma. The red cells are then returned to the donor and the plasma is retained for use and further manufacture.
For the manufacture of hyperimmune products, source plasma is often used. Plasma is often collected from those who have antibodies as a result of a prior immunization, such as with tetanus, or those who have antibody from earlier infection, such as with CMV or RSV, and in the case of Rh antibodies from Rh-negative women who have been exposed to Rh-positive pregnancies.
Source material can also be obtained from vaccinees in active immunization programs. And some examples include people vaccinated against rabies for the manufacture of RIG; men vaccinated with the Rh antigen to make anti-Rho-D immune globulin; HPV vaccinated people to make HVIG; and, of special interest is the SIP--or special immunization programs--for laboratory workers who have been a source to general vaccinia immune globulin--although I should point out for this product volunteers and military recruits were also a source of plasma. In fact, military donors are of utmost importance for the generation of polyclonal immune globulins that will be used in a counter-terrorism setting, since they are often immunized prior to active duty in endemic and/or high-risk areas. [Slide.]
Before blood or plasma collection, there are several general issues that need to be addressed with a donor screening questionnaire. And the examples I've shown here were taken from the new Uniform Donor Screening Questionnaire that was recently published on the FDA website. And the link is shown on the bottom of the slide.
And I do want to point out that I'm only showing a few examples of the type of questions that are asked of a donor. And these fall into several general categories, such as general health-related--"How are you feeling today?" on the day of donation; those questions related to viral risk, such as "Have you had a tattoo or piercing within the last year?"; and those more geographically related, such as "Time spent in the UK," which is related to TSE exposure. [Slide.]
Now, for the collection of plasma for counter-terrorism products there are additional concerns that often arise in donors who are in the military or involved in special immunization programs. One of the biggest questions is whether or not the donors have been immunized with live vaccines to stimulate specific immunity, and whether there is potential for viremia at the time of plasma donation for the manufacture of the product.
Also, we need to be aware of whether IND vaccines--either the vaccine given to elicit the desired immune globulin, or other INDs that may be administered around the time of plasma collection, and what is the donor deferral period for that particular vaccine? And then FDA assesses these situations on a case-by-case basis. [Slide.]
In addition, the blood and plasma needs to be tested, and there are a set of specific requirements in the 29 CFR 610.40, which indicates testing must be done for HIV1, 2, HBV, HCV, HTLV1 and 2, and also syphilis. In addition to these agents, the agency may also recommend blood and blood components to be testing for additional agents, depending on the source of the materia; for example, West Nile virus. [Slide.]
So, as I mentioned earlier, the manufacture of immune globulins is a very complex process. And the story of IGIV really starts with the work of Dr. Cohn and his colleagues at Harvard in the early 1940s. The group developed numerous fractionation methods for the large-scale separation of plasma into its components. And the method that was favored involved the use of cold alcohol, sub-zero temperatures as a protein precipitant. John Oncley in Cohn's lab furthered this method to isolate IgG out of the plasma fraction 2+3.
Cohn's method is shown here. And rather than to delve into the specifics of the steps, I just want to highlight the complexity of the manufacturing process. And the arm that's shown in yellow is the part of the process that results in the fractionation of immune globulins, and that side of the process alone is quite complex, and also has many variations. [Slide.]
Now, to emphasize this complexity even more, the Cohn-Oncley method is just one way to isolate and purify immune globulins. The Kistler-Nischmann method is a modified version of Cohn's scheme--shown on the blue side--and was developed in the '60s as a simplified method to Cohn's method, which has fewer steps, but still maintains the basic ethanol precipitation chemistry.
And, also, column chromatography method, including ion exchange, gel filtration and affinity methods--in the absence of alcohol--are also used to isolate immune globulins. [Slide.]
So, one of the most important parts, when you're dealing with human plasma, is to ensure that viral inactivation steps are done. There are several methods in the isolation process itself that help to partition viruses. And these include fractionation, precipitation steps, and column chromatography. There are also intentional viral clearance steps which are employed, which involve solvent detergent treatment, caprylate, heat treatment, nanofiltration, and treatment with low pH.
Now, typically, FDA would recommend two orthogonal steps to clear each type of virus--and by "type of virus" I mean those that are enveloped versus non-enveloped, sensitive or resistant to different types of methods.
Validation studies are done and must show clearance of the actual virus when possible. However, when that's not possible, model viruses can be used. [Slide.]
As I mentioned, adverse events can occur with these products. And this slide shows a list of both common and uncommon--and sometimes rare--adverse reactions that have been associated with immune globulins. And I just want to point out that the most common are mild to moderate headache, fatigue, chills, backache, nausea, low-grade fever.
And one thing that's interesting is some of these adverse reactions, over time, have been linked to certain measurable characteristics of the products. And certain anaphylactoid reactions have happened, and have been associated with the presence of aggregated immune globulin, because this can cause an increase in Complement activation which is measured by a test, ACA, a typical lot-release test.
In addition, hypotensive reactions have occurred, and this has been linked, in some cases, to the presence of pre-kalochrine activator--or PKA-- and kalochrine, which are components of the contact activation system. So, PKA is also a common lot-release test that's performed on these products.
Now, there's CFR-required lot-release testing for human immune globulins, and these can be found in the 610s listed here. And I'm not going to go through them, but do want to point out that under 610.100, Subpart J, there are specific requirements for human immune globulin that are indicated, such as source material, heat stability and so on. [Slide.]
So, in addition to the CFR-required lot-release testing, we often request other parameters to be measured, including tests for molecular distribution of the product into its fragments, monomers, dimers and aggregates; potency, if it's a hyperimmune, and the other tests shown here. [Slide.]
For immune globulin, stability testing is performed to ensure integrity, safety and potency throughout the dating period of the product. An example--examples of typical testing parameters for that stability protocol are shown here, as is a typical testing schedule. [Slide.]
So, as I mentioned at the beginning of my talk, humans are just one source of immune globulin, and animal species are also used to generate these types of products, with the most common today being horses, sheep and goats. [Slide.]
So the considerations with animal-derived immune globulins are somewhat different that those with humans. And, as mentioned, immunogenicity issues can arise because of the immune globulins themselves, or trace impurities that are animal proteins found in the final product.
Because of this, there is sometimes a need to test patients for hypersensitivity, as I mentioned, and perform desensitization procedures.
I'll talk in a minute about despeciation, and also like to point out again that zoonotic infections agents are of concern. [Slide.]
And I put a few slides related to animal husbandry up here just because this is also an important part when you're dealing with animals for the source of immune globulins. And I just wanted to point out a few things. These are all located in the CFR 611 section.
For the laboratory and bleeding rooms for animals, they must be kept free of flies and vermin; and for animal quarters and stables, the same. Food storage area shall be of appropriate construction, fly-proofed, adequately lighted and ventilated, and maintained in a clean, vermin-free and sanitary condition. [Slide.]
The care of the animals need to be addressed. Again, the animal quarters and cages shall be kept sanitary, inspected daily; and competent veterinary care needs to be provided as needed.
There's a quarantine period for animals. Animals shall not be used in processing until they are kept under the appropriate quarantine period of time--here it states at least seven days. And horses and other animals susceptible to tetanus need to be immunized. [Slide.]
Because these animals are going to be immunized against the antigen of choice so that you can develop your product, there's a statement that indicates toxins or other non-viable antigens administered in the immunization of animals shall be sterile, and viable antigens, when used, shall be free of contaminants.
There's issues about blood withdrawals. And the CFR indicates that blood shall not be used if it was drawn within five days of injecting an animal with viable microorganisms. And the blood intended for use as a source for biological product needs to be collected in a clean, sterile vessel. And if it's intended for use as an injectable, it needs to be pyrogen free.
In addition, CBER needs to be notified if there are certain diseases that are suspected or determined to exist, such as foot and mouth disease and Glanders, and the others listed here. [Slide.]
So, as I've mentioned a couple of times now, despeciation is one method that can be used to make the animal-derived product less immunogenic. And so what I'm showing here is digesting with pepsin, where you take an intact immune globulin and the result is an Fc fragment and an Fab prime 2. And so what you'd want to do during manufacturing is isolate this fraction of the molecule and remove these from the final product.
In addition, you could also digest with papain, and the result is 2 Fab molecules.
And when optimizing these procedures, the most important things to consider are time, temperature and the amount of enzyme in the process. [Slide.]
So, with animal-derived immune globulins there's often been the need to test for hypersensitivity using a skin test, and looking for a weal and flare reaction. It's often also--or, not "often" but could also be necessary to do desensitization. And we have required this type of testing for animal-derived products in the past. [Slide.]
So, standard lot-release testing for animal-derived immune globulins is shown here. It's not identical, but similar, to that for humans. [Slide.]
And, again, stability testing is performed to ensure the safety, integrity and potency throughout the dating period. [Slide.]
So, as we know, there's potential hurdles in the licensure, not just polyclonals, but all of these products for counter-terrorism. And efficacy studies cannot be performed in the absence of illness or prior to a bioterrorism event. So, alternative strategies need to be employed, which typically include Phase 4 study commitments.
Clinical safety studies with hyperimmune globulin can be performed, and are typically done in normal volunteers, which allows for common adverse events to be identified, and PK profiles to be investigated. [Slide.]
So, because of the unique nature of CT products, current licensure strategies employ mechanisms such as the Animal Rule--which will be discussed later today--or Accelerated Approval designation, which is found in 21 CFR 601.40 through 46.
In these cases, licensure is based on surrogate markers for efficacy, but also comes with the need for Phase 4 study commitments to validate the surrogate marker. [Slide.]
So, just as a quick summary, I'd like to conclude by saying polyclonal antibodies have the advantage of having multiple specificities against the particular antigen--anthrax, in this case. And a large amount of plasma for the manufacture of immune globulins can be easily obtained, and there are multiple sources, both human and animal.
Plasma fractionation is a well-studied process; has been employed for decades. However, it's important to keep in mind that transmissible agents are of utmost concern, and the manufacturing process must ensure that the viral inactivation steps are effective. [Slide.]
So, I also just want to say thanks to the folks in the lab of Plasma Derivatives in the Division of Hematology: Doug Frazier, Dot Scott, and Dov Golding.
Thanks.
[Applause.]
DR. FRUCHT: We're going to take a break now and reconvene at 10:15.
[Off the record.]
DR. FRUCHT: I should just mention that Dr. Collier's slides aren't available today, but they will be available on the WEB.
Also, if folks are looking for index cards, there will be extra index cards in the front, if you have written questions.
Well, Dr. Collier really needs no introduction, especially to a group like this. But I thought I'd highlight a few of his many accomplishments.
He's a pioneer in the field of diphtheria toxin research, an authored numerous landmark publications in this field. Fortunately for our field, he expanded his research investigations to other bacterial species, including bacillus anthracis. In large part due to the accomplishments of his laboratory, we now have a much better understanding of the three-dimensional interactions of anthrax toxin with its target molecules in the cell. And this is the basis for developing potential therapeutics.
Among his many other honors, Dr. Collier is a member of the National Academy of Sciences. He is currently the Maude and Lillian Pressley Professor of Microbiology and Molecular Genetics at Harvard Medical School. Dr. Collier.
Novel Inhibitors of Anthrax Toxin
DR. COLLIER: Thank you very much, David.
It's really a pleasure to be here, and it's an honor to be invited to participate in this meeting.
I thought what I would do this morning is to expand on a couple of the inhibitors that Steve Leppla mentioned in his opening talk--inhibitors of the anthrax toxin that we've been involved with.
So--let's see here. [Slide.]
So, just to remind you of the current model of the way the anthrax toxin components interact, assemble into the cell surface--assemble into toxic complexes at the cell surface--I won't go through this in detail, but I will tell you that the two types of inhibitors I'm going to focus on are, first of all, the dominant negative inhibitor, which is a mutant form of the protective antigen that acts to block the conversion of the heptameric pre-pore that's assembled from PA63 of the cell surface in the pore stage. And I'll detail that as we go along.
And the second type of inhibitor is a soluble form of one of the two receptors that Steve Leppla mentioned: the CMG2 soluble form created by genetically truncating the molecule and eliminating the transmembrane component of that part of that receptor, giving you a soluble form that can bind to PA. [Slide.]
So, the concept of the dominant negative inhibitor is detailed here further. The idea is if you have a form of PA that itself is unable to go to undergo the conversion from the pre-pore to the pore stage, you can envision subsets of mutants of that class that might be dominantly negative, in the sense that they would co-assemble with wild-type PA during the normal assembly process at the cell surface. And then getting down all the way here to the step of conversion of the pre-pore to the pore would dominantly inhibit the ability of the wild-type parts of that heptamer to undergo that conversion, and therefore would block the entry of both EF and LF into the cytosol, and therefore block all toxicity. [Slide.]
So, the basic mechanism of pore formation--the current model is illustrated here. The pore-forming domain of PA is Domain 2, and I--sorry--I think a slide got left out here--Domain 2, that was, in any case, shown earlier, I think, in Steve's talk.
And during the crystallographic determination of the structure of PA, it was found that there was a loop region up about halfway along the height of the Domain 2 that, first of all, was not seen in the heptameric structure--nor the monomer, for that matter. The loop seemed to have properties that suggested it might be able to form an amptipathic beta barrel that would span the membrane in this fashion--similar to what had been found by Eric Groh with the staphylococcal alpha toxin. [Slide.]
And so this loop--according to the model, then, this loop region, in order to form the transmembrane beta barrel would have to be relocated down to the base of the heptamer, and that would imply a major conformational change in the pore-forming domain--Domain 2. And the way that is envisioned to happen is that Domain 2 is built in the form of a Greek key motif--as illustrated here--and one can imagine, then, that if you stripped out these two flanking beta strands--2-beta-2 and 2-beta-3--from the Domain 2, that would allow the loop to be relocated down in this fashion. [Slide.]
Now, one of the manifestations of the conversion of the pre-pore to the pore is illustrated here. In the pre-pore form, if one exposes the structure to the denaturing detergent SDS, the subunits will fall apart into--or the heptamer will fall apart into individual subunits. After pore formation occurs, the structure is resistant to SDS, and hence on SDS polyacrylamite gels one sees a very high molecular weight olicamer here formed that we believe corresponds to the pore form. [Slide.]
In the course of studies a few years back, a post-doctoral fellow in the lab, Bret Sellman, was performing directed mutagenesis over here on the opposite side of Domain 2, from the pore-forming loop, in these loops here--and came across some sites which mutation absolutely blocked the activity of PA; and specifically blocked its ability to convert from the pre-pore to the pore form. Three of these sites are shown here: this lysing 397, and this loop--aspartic acid 425, and phenylalanine 427 in this loop here.
It turned out that these two mutations--either of them--was dominantly negative. [Slide.]
The K397 mutations--mutations of K397--blocked pore formation, but the mutations themselves were not dominant.
We've made several combinations of these--of mutations at these sites, and settled on, early on, a combination of D427K and K397D as being a double mutant that had very high dominant negative activity. And that's sort of become the working standard that we've carried forward. [Slide.]
Along the lines of--in the theme of in vitro assays of this session, I just wanted to mention that sort of our standard bread-and-butter assay that we use is based upon--not upon using either EF or LF per se as the effector molecule but, rather, what we do, based in part on work that was done in Steve Leppla's lab as well, is to take the N terminal domain of LF--and that we call "LFN." Catalytic machinery is "C terminal." We eliminate the catalytic machinery and replace it by the catalytic domain of diphtheria toxin. And, as most of you know, the catalytic domain of diphtheria toxin blocks protein synthesis in, essentially, every cell that it gets into but 80 pure eboscylating EF2. And it gives us a very nice effector molecule--at least nice for investigative purposes in the laboratory. We call that LFNDTA--diphtheria toxin A, chain-linked to LFN. And we look for the inhibition of protein synthesis as our standard bread-and-butter laboratory assay for translocation. [Slide.]
So here is--this slide illustrates some of the properties of the negative mutants at these sites--at two of these sites. I mentioned the K397 and D425A. So we're looking at the K397 and D425A mutations.
And we see here, in this slide, that the binding of these mutant forms of PA to cells is unaffected by the mutations. [Slide.]
The translocation, however, is drastically affected. And here we're looking at an assay where we assembled the complexes at the cell surface, using a radio-labeled ligand--LFN radio-labeled ligand--and then acidify the medium and look at translocation across the plasma membrane, after pyridically degrading anything that's left at the cell surface. And so you can see that that step is drastically affected here--in fact, it's essentially completely dead, these molecules are. [Slide.]
This panel shows that the ligamerization occurs normally, fostered by LFN, to compete off PA20 from nicked PA. And the central panel here shows the effect on LFNDTA, inhibition of protein synthesis. This is wild-type PA, and these are the two mutants. So either of these two mutations, then has strong effects on translocation, specifically. [Slide.]
This is just--again, along the lines of illustrating some of the assays that we used for cell permeabilization by PA, the conversion of the pre-pore to the pore, we load cells with radioactive rubidium, and then look at the release of that into the medium upon acidification of the medium. [Slide.]
And here's the illustration of the assay I just descried on using a radio-labeled ligand at the cell surface, looking at translocation across the plasma membrane in response to low PA. [Slide.]
This is an assay showing the dominant negative character of the double mutant here, in comparison with a non-cleavable PA mutant that is a much weaker inhibitor of toxin action. So what we've done here is to set up a combination of wild-type PA and LFNDTA that would inhibit protein synthesis about 90 percent, and then titrate into that mixture the double mutant, or this non-cleavable mutant.
And as you can see here, by the time you get to a one-to-one ratio of the double to the wild-type PA, you've almost completely inhibited toxin action; whereas, the non-cleavable mutant, which will compete for the receptor, and inhibit toxin in that way, it's an extremely weak inhibitor. [Slide.]
We've gone through--scanned through--the entire PA63 molecule in collaboration with Rod Tweeten and Jimmy Ballard, looking for other sites where mutation would create a defective PA. And we found a number of sites heavily concentrated in Domain 2, and the dominant negative ones--those that we found--we found a few more of those, besides the ones in this loop that are distributed in these two beta strands here.
How do the dominant negative mutants work? [Slide.]
Well, these sites where the residues are mutated are in that pre-pore they're solvated. They're not in contact with any other part of the PA63. But they've obviously got to be recognized at some point. And we think that what's happening is that if we envision these sites as composing one site that may be recognized by another site in PA63, and that we can envision some sort of a rotational model here where sites 1 and 2 in the pre-pore are not in contact with each other, but in response to pH, would come in contact--perhaps site 1 of one subunit with site 2 of the next--and that a dominant negative mutation would simply interrupt that link. And this has got to be a highly concerted process, the conversion of the pre-pore to the pore. And if all seven subunits do not work in concert, the whole thing won't work.
So, that's the basis--that's our current thinking, then of the way this whole thing works. [Slide.]
Then the original studies on the Fischer 334 rat that we did a few years back. We combined 40 mcg of PA and 8 mcg of LF and the animal became moribund in about 90 minutes. And if you add as little as 10 mcg of either the double mutant or the F427A mutant, the animals showed no symptoms and survive indefinitely. [Slide.]
Recently--so I've done no more in animal work beyond this, but the company that's licensed this technology--Pharmathene--has conducted a spore-challenge model recently in rabbits that has given interesting results and favorable results to the whole thing.
And these are groups of six rabbits that were injected either with a high dose or a low dose and challenged with a whopping dose of spore--some 7,000 times the LF50. And as you can see here, with the high dose, then the animals--five our of six animals survived--well--indefinitely, through day 85 here.
If there are any questions regarding these data I'll refer you to Sol Layermand, who's in the audience.
So the dominant negative inhibitors them, in summary, combine with wild-type PA and dominantly inhibit pore formation and translocation. We think that as little as one dominant negative PA molecule can inactivate, then, up to six molecules of wild-type PA, plus up to three molecules of LF and/or EF that are bound in that complex.
Interestingly, DNA--the dominant negative PA retains immunogenicity. We haven't seen any diminution in immunogenicity in the tests that we've done so far. And so potentially, therefore, the dominant negative PA represents a combination--potentially--of a therapeutic antitoxin and a vaccine in one molecule. [Slide.]
I'm still struck by the fact that you can take a single mutation, or double mutation, and convert a toxin subunit into a potential inhibitor of toxin action, and potentially, a vaccine. There are many ligameric pore-forming toxins that assemble at the cell surface, or outside cells. And so potentially, this approach is generalizable to some other systems. [Slide.]
Okay, finally, in the last five minutes I want to go one and tell you about one of the cellular receptors for PA. As Steve mentioned, there are two of them known: ATR artemate, and CMG2. And these are both single-pass membrane proteins that have a von Willebrand A domain; about 60 percent identity between the two examples here. And they both have a MIDAS motif, which is a metal ion-dependent adhesion site. That turns out to be important in the interaction with PA. [Slide.]
Recently, Borden Lacy in my laboratory has determined the crystallographic structure of the extracellular von Willebrand domain of CMG2, which is illustrated here. I won't go into detail except to tell you that the MIDAS motif is up here, and there's a magnesium atom right there that we believe interacts with an aspartic acid in Domain 4 of PA to form part of the binding affinity.
The affinity is tight. The CMG2--soluble CMG2 binds in a one-to-one ratio with monomeric PA, and a seven-to-one ratio with the heptamer. So there appears to be no stearic inhibition to the interaction of the CMG2 with each of the monomers of PA--heptameric PA. [Slide.]
So, this summarizes what we know about the interaction. PA binds one CMG2 von Willebrand A domain at saturation. That's an AD molecule. The pre-pore binds seven at saturation--incredibly high affinity. KD is sub-nanomolar. It's roughly 200 picomolar. Whopping high affinity. The off rate--the rate of dissociation of the complex once it's formed is also extremely slow--on the order of a day. So once this CMG2 latches onto PA, basically you've locked it up for a very long time. So therefore it's a potential inhibitor of toxin action, in fact it has been shown to be so in in vitro systems. [Slide.]
So, finally, I just thought I would mention a few other potential approaches in inhibiting anthrax toxin action. But these, basically, have all been mentioned and discussed in greater detail. Steve mentioned the polyvalent inhibitors that we were involved in developing, and that work is being carried forward by a former post-doc in my lab, Jeremy Mogridge. We're not working on this anymore.
One thing that's not, I think, widely known is a project that's going on in collaboration with John Young--and, by the way, the CMG2, I should have mentioned--CMG2 was cloned in John Young's lab, and the ATR also was cloned in John Young's lab; the CMG2 by Heather Scobey. And we've been working collaboratively with John Young and his people for many years now.
And as an extension of that collaboration, we're involved in an NIAID-sponsored program project that's headed by Maryann Manchester, and involving John Young and Jack Johnson and a number of other people in southern California. And the idea is to take certain viruses--a plant viruses, Kalpi mosaic virus and an insect virus, flockhouse virus--very well characterized small viruses, and substitute certain peptides or even domains on the capsid protein--one of the capsid proteins of these viruses, and perhaps creating what we call "molecular sponges" with bioactive peptides that might, for example, attach to PA and suck it out of the serum, or various other ways that one can envision that this technology might be applied to anthrax and many other infectious diseases. [Slide.]
And then, finally, Steve Leppla already mentioned this--the hexa-D-arginine--and there are inhibitors of LF action that have been developed by Ben Turk and Lou Cantley's lab. And others--and there are a number of others, as well, that have come out--active site inhibitors of EF and LF that Steve already mentioned. And the one that's being developed at Merck by Jeff Hermes and his colleagues is also very exciting. [Slide.]
So, finally, I'll conclude--and these are the people that have worked on anthrax toxin in my lab over the years. And I want to highlight the efforts of Bret Sellman, who isolated the first dominant negative mutations,