SLIDES & TRANSCRIPTS
Tuesday, February 1, 2000

Antibody-Delivered Therapy
Irwin Bernstein, MD

DR. BERNSTEIN:Actually it is an exciting time to be here to speak to antibody directed therapy because it is more than 2 decades since the development of methods for preparing monoclonal antibodies and we are finally beginning to see the fruits of these endeavors. Some of the studies that I would like to tell you about are beginning to appear quite promising.

Antibodies have obviously been used in a variety of ways. Initially they have been used in unmodified form with the notion that antibody binding to the target cell would lead to death of that cell by interacting with host effector components, effector cells or complement, but, also in some instances antibody has been used in other diseases to block important receptor sites leading to cell death, such as Herceptin.

Because naked antibody requires interaction with limited host components, antibody has been shown at least in preclinical models to be more effective when used to deliver other agents.

For example, antibody has been used to deliver radioisotopes, shown here with an isotope that emits energy with a relatively long path length emitted over several cell diameters.By binding to some but not all the cells in an area of tumor, the scattered radiation would also kill non-targeted, perhaps antigen negative, cells. An advantage and this would be true, for example, for I131 or Y90. Also, it is possible to use other isotopes, alpha emitters that have a very short pathway that would essentially only kill the targeted cells and perhaps an immediately adjacent cell. The final way that it has been used is for the antibody to deliver a drug or a toxin. This, of course, would require that the conjugate be internalized by the cell and the drug or toxin then released from the antibody and delivered to the appropriate target area.

What I would like to do in the next few minutes is just give you examples of each of these endeavors that hopefully will serve as fruit for the ensuing discussion in the breakout groups.

TOP

Now, first with respect to naked antibody this has been mostly done for AML by Scheinberg, Jursick and their colleagues who use an anti-CD33 antibody. I will talk to you more about the specificity of anti-CD33 antibody later, but this antigen is essentially on virtually all myeloid leukemias and the expression of the antigen is virtually limited to the hematopoietic system.What they found first for patients in relapsed AML and here in a Phase II trial where twodifferent doses were tested, that in fact it was possible to observe a complete response in 2 of 15 patients, but these were patients who had relatively modest blast counts within the marrow according to David and that patients with higher AML burdens did not show responses.

This limited effectiveness suggests, as has been suggested in animal studies, that antibody might be more effective against minimal residual disease. So Scheinberg and Jurasik tested this humanized anti-CD33 antibody in patients with APL.These were patients induced into remission following treatment with ATRA and/or chemotherapy.They found whereas most of the patients were PCR positive for the translocation associated with APL before receiving antibody, a significant portion became PCR negative after receiving antibody. These patients have done quite well, virtually all remaining in remission with a follow-up at least at time of this abstract of 3 to 54 months.heir data is also suggestive compared to historical controls the potential advantage for this antibody treatment.However, certainly for this disease definitive evidence would need to come from a randomized trial.

TOP

Now, Scheinberg's group has also used antibody radiolabeled with isotopes in an amount that would be non-myeloablative. The treatment has been used in the absence of stem cell transplantation.

They have used this same anti-CD33 antibody conjugated with yttrium in a dose-escalation study and found that the highest dose caused significant myelosuppression but that they did in fact see a complete response in one of three patients.

Based on these studies, they are planning to go ahead and utilize this in combination with chemotherapy -- etoposide followed by stem cell transplantation -- in this kind of myeloablative approach I will discuss in just a moment in terms of our studies in Seattle. In addition, this group has also used an alpha emitter, bismuth 213 and has learned, number one that toxicity was quite limited in a Phase I study and the MTD not reached, while at the same time they could cause decreased marrow blasts.
Since this in a sense selectively targets the CD33 cells, they are considering whether or not to use this as a single agent, the problem being that the half life is exceedingly short of bismuth 213, less than an hour, making this approach extremely difficult.

TOP

In Seattle in studies mainly done by Dana Matthews, Fred Applebaum and Janet Eire, we used anti-CD33 antibody labeled with I131 as part of conditioning regimens in a marrow transplant regimen. What this shows here is gamma imaging following infusion of radiolabeled I131 labeled antibody.

This happens to be taken from a patient after infusion of I131 labeled anti-CD33 antibody. What we found is that although we can localize the marrow space, the isotope was rapidly lost from the marrow and did not deliver a sufficient amount of radiation. I show you this because what we learned was that the anti-CD33 antibody was internalized rapidly by the cells, degraded and then the iodine secreted, suggesting that this approach would be very useful for delivering a drug which I will get to in the last part of the talk, but Dana showed that another antigen which is expressed within the hematopoietic system, CD45 which is stable on the surface following antibody binding, was a useful target. The isotope could be localized and would be retained within the marrow.

TOP

It is part of the Phase I study in which patients first underwent a biodistribution study.

She was able to show that substantially greater doses of radiation per millicurie of I131 attached to the antibody could be delivered to marrow or spleen as compared to critical normal organs, liver, lung or kidney and that use of anti-CD45 antibodies was, in our hands, superior to use of anti-CD33 antibody.

With the notion that this could target the radiation to sites where leukemia may reside, it suggested that this was a way to increase the therapy as far as the conditioning regimen prior to a transplant without increasing toxicity.

Now, in a Phase I study Dana was able to show that this approach was quite feasible and she could administer up to 10.5 Gr to the normal organs receiving the highest dose, the liver. As this therapy was very well tolerated, Dana and Fred became very enthusiastic and before completion of this study began to apply this approach

TOP

in a Phase I-II trial of patients with less advanced disease. In this study patients with AML in first remission who would receive an HLA matched marrow from a related donor were entered. As I just showed you, these patients first underwent a biodistribution study where antibody was labeled with a trace amount of I131.

For those patients in whom you could show that greater amounts of radiation would be delivered to marrow and spleen compared to the normal organs, these patients then received the antibody labeled with an amount of iodine designed to deliver a predetermined dose to liver which for most patients in this study was 5.25 Gray. They then received busulfan, cytoxan and methotrexate/cyclosporine as prophylaxis.

TOP

What she found was that virtually all the patients showed this favorable biodistribution, that is, greater amounts of radiation to spleen and marrow compared to normal organs.

Of 24 patients treated, most were treated at 5.25 Gray to liver with higher doses, of course, to marrow and spleen. Four of them received a lower dose as we started at a lower dose and did some dose seeking here. These patients received in addition to busulfan/cytoxan, an average of 10 Gray to marrow and 28 Gray to spleen.

As you can see, this was not a particularly favorable or overly unfavorable group of patients based on cytogenetics. What was encouraging is that, thus far, there have been four transplant-related deaths of these 24 patients and only two relapses, one at 8 and one at 10 months

TOP

providing a disease-free survival curve that has been encouraging. Dana is now extending the study in a multi-institutional trial on the West Coast involving Stanford and City of Hope to accrue additional data. Assuming that this data holds she will hopefully plan a multi-institutional Phase II trial to definitively test out this approach.

TOP

Now, the last approach that I want to talk to you about is the use of antibody for targeting a drug. I will talk to you about the experience that we have had in Seattle in collaboration with Wyeth-Ayerst. What we have done there is use anti-CD33 antibody, one that we prepared in Seattle. I already told you CD33 is mainly expressed in the hematopoietic system and is expressed by the vast majority of AMLs.

In addition, when you conjugate the antibody to the drug, it will effectively kill virtually all target cells. Then one needs to be concerned whether or not you are killing normal stem cells as well as leukemic stem cells.

What this shows here for normal hematopoiesis with a pluripotent stem cell giving rise to myeloid precursors, granulocyte, monocyte precursors, red cell precursors and platelet precursors, and all lymphoid cells, that the CD33 antigen is on committed myeloid progenitors but is not on the stem cells, suggesting that if one depletes these cells that the stem cells would eventually repopulate normal hematopoiesis.

Now, in studies that we performed early on, that I don't have time to go into, those data suggested that for at least some patients with AML the CD33 negative precursors were predominantly or completely normal in origin. This prompted us to seek collaborators who would conjugate the anti-CD33 antibody with a drug or a toxin. I must say that subsequent studies by John Dick and others have suggested at least for some and perhaps all patients that the primitive precursors are involved in AML as well. Fortunately, however, we were not aware of that data at that time and this still remains a controversial area.

TOP

Nonetheless, working with Lois Hinman and Phil Hammond at the then Lederle Laboratories and now Wyeth-Ayerst, the anti-CD33 antibody which was P676 was humanized and conjugated with calicheamicin, a member of the ene-diene family which is a minor DNA binder that causes double stranded DNA breaks and apoptosis of the target cells.

TOP

In preclinical studies it was possible to show that this conjugate would selectively inhibit the growth of CD33 expressing cell lines including HL60 and would inhibit the in vivo growth of a xenograft of HL60 cells in nude mice.

In addition, significantly, it was able to inhibit the growth of leukemic colony forming cells obtained from patients in vitro.

Based on this, a Phase I trial was initiated, run mainly by Eric Sievers and Fred Applebaum in Seattle and in collaboration with Steve Forman at City of Hope.

TOP

In this dose escalation trial, patients received up to three doses at 2-week intervals. These were patients with white counts less than 30,000 with refractory or relapsed AML.

TOP

The important points here, first of all, since this was a Phase I toxicity trial, are that toxicity was remarkably limited in that it mainly consisted of fever and chills.

Patients did experience hematopoietic toxicity presumably in part from leukemia, but also in the patients who received three doses, prolonged myelosuppression was seen in some. In the subsequent study only two doses were given.

There were also transient hepatic abnormalities seen, but significantly there was no significant CNS, cardiac, or renal toxicity, and in fact, the MTD for non-hematopoietic toxicity was not reached. The study stopped because, as I will show you, saturation of target sites was achieved.

In additionantibody responses against the conjugate proved not to be a problem. It was seen in two patients. One of these patients -- really upon retreatment that this patient responded later relapsed and eventually retreated with developed an antibody.

This graph illustrates some points I would like tomake from this trial.What is shown here on the Y axis is the percentage of target CD33 sites of circulating blasts that are saturated by the infused conjugate, and what this shows is that for most of the patients the sites were essentially saturated. In fact for patients receiving the highest doses of the conjugate, these sites were virtually highly saturated in virtually all the patients.

The second thing you can see here is that a portion of patients responded, shown in red, and these are patients in whom there was disappearance of morphologically detectable leukemia in the marrow, some of whom ended up recovering their counts and having a complete response.

TOP

What is shown here is the proportion of patients that is graphed on the X axis, and the results of dye efflux studies as a measure of drug resistance. What this is showing is that within the group of patients with high saturation and relatively low dye efflux a significant portion of patients was, indeed, able to respond.

TOP

Overall, the encouraging results of this Phase I trial led to establishment of a multicenter trial by Wyeth. In this trial patients with CD33 positive AML in their first untreated relapse following a remission of at least 6 months were entered. They received a dose that was the highest one administered on the Phase I study, 9 milligrams per meter squared, but as I mentioned only two doses 2 weeks apart.

The results of 59 patients were reported by Eric Sievers and what this showed is that the patients did experience myelosuppression and some transient liver abnormalities, but again, the toxicity was remarkably limited in that the patients, while experiencing myelosuppression, had infections, but they did not have the severe mucositis and many of the other symptoms associated with conventional chemotherapy.

Remissions were obtained in about 20 of these 59 patients or about 34 percent. These remissions, however, were defined as disappearance of leukemia, recovery of neutrophil count, and that the patients were platelet transfusion independent.

The reason for that is that many of these patients showed delayed platelet recovery, and they did not achieve a platelet count of 100,000. Although the reason for this is not known, we suspect that following ablation of CD33 positive precursors, given that limited numbers of stem cells may be re-establishing hematopoiesis, there may then be a not unexpected delay in platelet recovery. Overall, these results continue promising with the anti-CD33 antibody conjugate, termed CMA676. I might add also that at ASCO additional patients will be reported by Eric Seivers which shows a continuation of the response rate that we presented at ASH.

So to summarize, I think the examples that I have given you suggest that there may be limited effectiveness of non-modified or naked antibody such as anti-CD33 antibody, but this may be an area worth exploring in situations of minimal residual disease.

Radiolabeled antibody appears promising particularly in conjunction with stem cell transplantation and in particular, the study of anti-CD45 labeled antibody may be one approaching definitive Phase III evaluation.

Thirdly, antibody drug or toxin conjugates may be another promising area, and certainly at least the anti-CD33 calicheamicin conjugate appears effective and needs to be tested in a variety of ways to determine its optimal use.

Taken together, I think we conclude that further studies of targeted therapy of AML are certainly warranted, including the development of new antibody reagents by specific antibodies, new antibodies that are radiolabeled in different ways, and certainly additional antibody drug conjugates or antibody toxin conjugates. I should finally add that although my charge was to speak about antibody-directed therapy, we have to remember that antibody is simply a ligand used to target a substance to the leukemia cells, but certainly one can envision use of many other ligands that bind with high affinity such as hematopoietic growth factors. Art Frankel, for example, is testing out GM-CSF conjugated with diphtheria toxin, and this may, also provide a promising approach. So I think we have a great deal to talk about later at the workshop about how to extend these studies.

TOP