SLIDES & TRANSCRIPTS
Tuesday, February 1, 2000

Classic and Alternative Drug Resistance Pathways: ABC Transporters
Alan List, MD

DR. WILLMAN: That was terrific, Scott. Thank you so much, and we are going to move into the second phase of our plenary session to take what we have learned now about transcriptional regulation and signal transduction pathways and apoptosis into other pathways and how these might be actually manipulated in the context of new therapies. Scott alluded to the growing number of drug resistance transporters, and I would like to bring Alan List to the stage, who all of you know is from the Arizona Cancer Center and has been a real leader in the use of therapeutic modulation of drug resistance pathways in leukemia.

DR. LIST: As Cheryl mentioned, we are going to move from some of the more elegant discussions we had in the morning back to the clinic now and talk about something that has been an area of intense clinical and translational research for the last decade and that is multidrug resistance. What we refer to as multidrug resistance is an in vitro phenomenon that was first described over two decades ago by June Biedler and other investigators and is the development of resistance to a wide range of structurally and functionally unrelated drugs following exposure of tumor cells to a single anticancer agent.

This was associated, we found later, most commonly with overexpression of a specific gene mdr1 and its translational product P-glycoprotein. It is the recognition that P-glycoprotein's function can be inhibited by other competitive inhibitors, pharmacologic agents, that made this a very attractive area to investigate in acute myeloid leukemia.

TOP

But as Dr. Kaufmann has mentioned and Dr. Willman mentioned, there are now known to be more than 50 different genes in the family of ATP-binding transmembrane gene transporters that have similar, at least developmental relationships to mdr1 or P-glycoprotein. I have listed just a sampling of some of those genes here, and in addition to MDR1, certainly the MRP genes, including what we consider MRP6 which is the same as MOAT-b; BCRP, the breast cancer resistant protein described by Doug Ross who is here today (identical to a gene identified as MXR, for mitoxanthrone resistance gene), and ABC1, 2 and 3. ABC2 has been associated with extrusion of estramustine. ABC1, and to some degree ABC2, may be involved in the translocation of interleukin-1 as well.

So there is a host of different transporters, some of which may have clinical relevance in acute myeloid leukemia.

TOP

Just to broaden things a little bit further is to categorize some of the potential mechanisms that can contribute to what we consider a multidrug resistance phenotype in vitro.

There are a number of those, some of which we have just heard about: that is, altering a product threshold for BCL2 family members and so on; autocrine growth signals and cell adhesion molecules which we didn't have a chance to hear about yet today; the transport membrane transporters that we are talking about: P-glycoprotein, the MRPs, the CRP, intracellular entrapment that can occur from the major vault protein, again MRP or the TAP gene; drug detoxification; altered nuclear targets and so on.

The complexity that we have heard for apoptosis generation obviously extends when it comes to multidrug resistance as well.

TOP

Just to narrow down, to create a list of some of those that we know have prognostic relevance in AML, certainly the one that we are going to focus on the most this morning is MDR1. MRP1, in particular, also has some prognostic relevance, particularly when it is expressed in concert with MDR1, and some of the best work that has been done is by Jean Pierre Marie's laboratory. What we call the major vault protein or lung resistance protein is controversial, but recent work indicates that this does contribute to nuclear translocation of drug and drug resistance. The problem that we have had is demonstrating consistent prognostic relevance in AML and some of that relates to the method of analysis.

Also autonomous growth is one that is a relatively older biologic feature, but associated with drug resistance and, of course, the BCL-2 family members as well.

TOP

The P-glycoprotein has been the one that has been the major focus of investigation for the last decade, and there is a very good reason for that. It is associated with a number of the prognostic factors that we recognize in AML including advanced age. What we see in the laboratory as far as decreased intracellular accumulation, you can demonstrate in clinical samples as well.

We are looking at anthracycline retention associated with an adverse cytogenetic pattern, a CD34 surface phenotype, cytologic dysplasia, secondary AML and most importantly, it has been linked to a lower frequency of induction response as well as an inferior disease-free survival particularly in de novo AML but also to some degree in secondary AML as well.

Based upon what we have done in the past which was to identify prognostic relevance, it was a very appropriate target for investigation in the clinic.

TOP

When first addressing pharmacologic inhibitors of P-glycoprotein transport, there were a number of what we consider first-generation compounds which were drugs that are currently available that were found in the laboratory to inhibit in a competitive fashion the export function of P-glycoprotein.

I summarized these for you here. The ones that have made it for study extensively in AML have been quinine, generally by the French groups, and cyclosporin A. This gave rise to a number of second generation compounds, and this list is by no means completely inclusive obviously, but it highlights some of those that are entering clinical trials or are targeted to enter clinical trials; the cyclosporine analogue PSC-833 has been the most investigated in AML. The Lilly compound is now in trials in acute myeloid leukemia, the Vertex compounds and so on. Although we selected these second generation compounds based upon greater activity in vitro against P-glycoprotein and apparent greater affinity for P-GP, like any kind of pharmacologic agent, they have effects on other targets as well.

Although cyclosporin A and PSC-833 inhibit P-glycoprotein, they also have some modest inhibition of MRP1. The Vertex compounds and MS-209 both can inhibit MRP1 as well.

Although we believe these are good targets for P-glycoprotein, when they get into the clinical setting, we start to identify all their additional potential targets.

TOP

Now when considering all the trials that have been done using P-GP inhibitors, I think it is important to remember all the potential variables that will impact the outcome of any of these trials using drug resistance modifiers.

Certainly one of the more obvious variables is to make sure that we are achieving in vivo what concentrations are believed to be effective. There are newer attempts now to try to demonstrate in vivo that modulation can be demonstrated, to ensure that the ratio of the modulator and the anticancer drug exposure are optimized. I will show you some data later on that is showing that perhaps a longer exposure of the anticancer drug to the modulator may be more optimal as far as overcoming drug resistance.

Certainly we want to select anticancer drugs whose activity is limited in a drug resistance model by a high affinity for P-glycoprotein. So in other words, we want to take an agent such as the anthracyclines, daunomycin being a good example. There is a very high affinity substrate for P-glycoprotein. The resistance in vivo then might be impeded by P-GP expression.

We also have to take into account that some of the modulators, particularly the cyclosporines, can change the AUC of the antineoplastic and account for that when evaluating outcome. We have to ensure that there is an equal frequency and comparable distribution of the MDR1 phenotype in each of the arms of such studies and also ensure that there is functional P-glycoprotein. I have shown you a very long list and a growing list of potential non-P-GP mechanisms of resistance. These certainly can impact the outcome of the modulator trial and need to be considered.

TOP

There are two major trials that were done looking at cyclosporin A which is where I would like to begin and spend most of my discussion as far as illustration of the clinical outcomes.

The MRC trial, which was headed by Alan Burnett, looked at a diverse group of patients with relapse and refractory and elderly AML, using an Ara-C, daunomycin, etoposide combination, three different time sequential induction regimens with and without the addition of cyclosporin A. This is a negative trial as far as looking at induction outcome, but the majority of the patients received a dosage of cyclosporine that did not achieve blood levels expected to be effective which is 5 milligrams per kilogram per day.

The SWOG trial is the one I want to focus on the most. This uses a very much higher dose of cyclosporin A which at least achieves blood levels that we can demonstrate in ex vivo models to be effective and by concentrations obtained in the plasma are known to be effective as well, and it did have some important clinical benefits.

TOP

This was a randomized trial based upon the initial work that we did at Arizona looking at a regimen including high-dose cytarabine followed sequentially by daunomycin given by continuous infusion. I will address that more in detail later on, but suffice it to say the continuous infusion was selected not because of concern about changes in the AUC of the anthracycline but simply because of what we knew preclinically that to optimize overcoming drug resistance we need prolonged exposure to the anticancer drug target as well as the modulator.

Patients were randomized to receive either this combination or the combination with cyclosporine administered concurrently with daunomycin at the concentration I mentioned.

Eligibility included patients with high-risk AML either in relapse, primary refractory, secondary AML or patients with RAEB-T with an age cut off of 70 or younger. Patients who did remit then were eligible to receive one course of consolidation using the same regimen but with an abbreviated course of Ara-C for 3 days rather than 5 days. All patients were stratified prospectively based upon age and status.

TOP

These are the patient demographics, and I know that most of you have seen this data before. Relapsed AML accounted for the majority of patients -- approximately 60 percent of the patients in both arms of the study -- and secondary AML and RAEB-T accounted for anywhere between 20 to 25 percent of the patients on the study.

TOP

The regimen was actually very well tolerated which I think is very important. When we come to some of the other trials that have been performed with PSC-833, the most common toxicity that was found was hyperbilirubinemia which was reversible. This was generally not believed to be a true hepatic toxicity because of the relative absence of a transaminase elevation and also some degree of nausea but no noticeable increase in the frequency of significant mucositis.

TOP

When analyzing the induction outcome, what you will see is the most striking change was a reduction in frequency of resistance to induction therapy, 48 percent in the control arm compared to 31 percent in the treatment arm using cyclosporin A and that was significant.

What was more problematic was demonstrating a difference in complete response rate, and you can see it was 44 percent with cyclosporine and 35 percent with daunomycin. This was not significant but trended in that direction.

What we found was that when patients were entering remission and some of them living for a year or even longer that they did not retain what we consider a true complete response based upon platelet counts maintained above 100,000 although they may be leukemia free, but what it did give us was some index of activity of the combination.

TOP

What is most impressive from this study is the relapse and survival. I apologize I haven't been able to update this with a new PowerPoint file but this data has been reanalyzed in the last several months, and the differences really still persist.

Looking at relapse free survival, we can see here that at 2 years it was 44 percent on the cyclosporine arm. We know now with 3 years' follow-up we have about a 35 percent disease-free survival. That compared to, again, only 5 percent in the control arm of the study.

TOP

As far as I know, this is one of the first trials in relapse and refractory AML that showed a difference in overall survival. This difference in overall survival still persists at the recent re-analysis. Approximately 22 percent of patients remain alive in the cyclosporine arm compared to 9 percent of the patients in the control arm.

TOP

That brings us back to the variables that we talked about earlier. Did we achieve this simply by modulating P-glycoprotein function or was this an impact of other parameters with cyclosporine administration? Indeed, with cyclosporine administration, as we anticipated, there was significant change in systemic exposure with daunomycin. If we look at the steady state levels of daunomycin, you can see that in the control arm compared to cyclosporine the cyclosporine increased the level roughly twofold; with daunomycinol the primary metabolite it increases approximately four-fold when looking at steady state concentrations.

These drop off fairly quickly when we compare the day 10 levels which is 24 hours after cyclosporine termination and indeed we were able to achieve levels of cyclosporine that were active. Not only did we reach the target concentration of roughly 1600 nanograms per ml but also in a surrogate biologic assay they sensitized the cell line that was P-GP positive five-to-six-fold.

TOP

If the effect that we saw as far as benefit on outcome with cyclosporine modulation related simply to a change in exposure to daunomycin what we would expect to see in both arms of the study is an improvement in induction outcome and other response criteria with increasing daunomycin exposure. If we compare both arms of the study, what we see on the control arm of the study shown here on the left is with increasing daunomycin to steady concentrations, we did not see an improvement in CR rate nor did we see a decrease in the frequency of resistant disease. However, if we look at the interaction when cyclosporine is added, we see a trend for increasing CR and a trend for decreasing resistant disease, and that interaction is certainly consistent.

When we look at daunomycin at all levels, the same can also be seen. In the control arm there is no improvement in CR. With higher levels there is no trend for reduced induction resistance. However, with the addition of cyclosporine we see a clear trend towards improvement in CR, and decreasing resistant disease; again, that interaction is statistically significant.

TOP

If we look at that same relationship between daunomycin steady state levels and outcome based upon survival analysis, we see the same trend. In fact, seen here in the orange is the highest level, greater than 30 nanograms per ml of daunomycin compared to the lowest level shown here in yellow looking at the control arm. If you had higher daunomycin steady state levels, you actually had a shorter survival compared to the lowest levels, with intermediate levels falling in between. When we look at that with the cyclosporine arm, the higher your daunomycin levels, the better the outcome and again each of these levels was better in comparison to the control.

TOP

When you look at relapse-free survival, the same trend also persists but obviously the numbers are smaller. We are looking here at the impact on cytogenetic groups. What you can see here is the patients with favorable cytogenetics shown in yellow who had an improved outcome with cyclosporine. Patients who had intermediate cytogenetics also showed an improved outcome with cyclosporine, and if you look at 2-year survival with poor risk, we also did better in the poor risk subset of patients with an unfavorable karyotype with the addition of cyclosporin A. So all karyotype groups appear to be impacted by this.

TOP

So what is the impact of P-glycoprotein expression and did we impact our target? The answer is yes with a few caveats, and I am sorry that I don't have this in color, but this shows you the overall survival based upon expression of P-glycoprotein by MRK16 expression. Again, this is based upon looking at all populations of blast cells. For the patients who were P-glycoprotein positive based upon a D value greater than .20 who did not receive cyclosporin A as shown here in the dashed lines, you can see that the survival curve is here; those patients who were P-GP positive and received cyclosporine are shown here. The median survival for these groups is 4 months in the control arm versus 12 months with cyclosporin A, about a threefold increase in median survival. If we look at the impact in those patients who are P-GP negative as shown here for the control arm with the solid line, you can see the curve here. For those patients who were P-GP negative with cyclosporine, the dotted dashed line shows the overall survival. Median survival was no different, and in fact, 6 months is the median survival for both of those P-GP negative groups.

What does seem to separate though at 2 and 3 years, and this still looks the same on the re-analysis, is that the two groups receiving cyclosporin A whether you are P-GP positive or negative seemed to come together and appeared to be superior to the control groups whether they are P-GP positive or negative at 3 years, suggesting that we had an impact on what we consider P-GP negative leukemias based upon bulk leukemia population analysis.

TOP

What about PSC-833 which is the most extensively studied of these second generation compounds in AML? There are four trials that have now been completed, much of which has been highlighted at the recent ASH meeting.

The ECOG trial was targeting relapsed/refractory AML using the MEC regimen, that is mitoxantrone, etoposide and Ara-C. This was closed prematurely because of failing to demonstrate activity in an early analysis.

The CALGB trial targeted patients with elderly AML using daunomycin, etoposide and Ara-C as the regimen that was modulated. That was also closed prematurely, this time for toxicity.

The HOVON study looked at elderly AML patients with daunomycin and Ara-C. That has been completed and found to be an overall negative study. However, when focusing just on the P-GP positive patients, they show an impact. There is a positive impact of the administration of PSC on induction outcome. Also another European trial which was looking at an induction with MEC simply to gain complete remission before going to transplant in relapsed/refractory patients, that has also been completed and found also to be a negative study.

TOP

Obviously, it is important now to look at what are the differences between these trials and the cyclosporine trial. What can we learn from this to address further investigations? There are a number of differences between this trial and the SWOG trial. Obviously it is unusual as far as a schedule of using sequential Ara-C followed by daunomycin.

That was not done because we thought it was a specific biologic feature we wanted to exploit, but simply because we were concerned about CNS toxicity with cyclosporine administered with Ara-C. However, there is preclinical data to show that sequential administration of Ara-C followed by daunomycin in vitro is synergistic compared to an additive fashion of giving them concurrently. It has never been adequately tested. It has been tested in a French trial but not in a sequential fashion looking at day 1, 2, 3 daunomycin versus day 5, 6, 7, showing no difference. Perhaps that is one of the differences for the SWOG trial.

Not all of these trials looked at P-GP modulation during induction and consolidation as we did in the SWOG study. Some of these were designed just to answer an induction question which may be more difficult to look at in these high-risk patients.

Some of these trials incorporated VP16 with the intent that they were going to have two targets to modulate for P-glycoprotein function. The problem with this is that etoposide is actually a very low affinity substrate for P-glycoprotein. When we add an additional drug to the regimen, the AUC is going to be modulated by the modulator. We will increase toxicity but not potentially impact the outcome as far as improving the benefit of the drug. Also because of the change in AUC that was anticipated in the anticancer agents, the drugs were decreased in dosage. So there is the potential since about one-third of the patients will not have a change in AUC that we are underdosing the anthracycline if that were the target drug in many of the patients.

Obviously, there are potentials for altered targets of cyclosporin A compared to PSC-833. We don't know of any to date that can account for this, and so we assume that that is not the answer to these differences in results.

What I think is probably as important is prolonged versus rapid infusion of the antineoplastic and there are several lines of investigation to suggest that that may have clinical relevance. This are some older data from some of our Scandinavian colleagues published in the 1980s looking at changes not only in plasma pharmacokinetics but cellular pharmacodynamics of the anthracyclines when administered to patients over varying time periods.

TOP

As you can see here, this is the result of the pharmacodynamics of plasma. This is looking at pharmacodynamics within the blast population of these AML patients and you can see with a short infusion, as you would expect, the Cmax in the plasma decreases with no corresponding changes in AUC or changes in daunomycin Cmax and no change obviously in daunomycin or AUC. But if you look just at the blast cell pharmacodynamics focusing on the AUC inside the cells, is that we see an increasing AUC of daunomycin with prolonged exposure, roughly a three-to-four-fold increase in drug accumulation compared to the short exposure which appears to be superior to what we expect with the short administration.

TOP

What about when it comes to overcoming multidrug resistance? This is an example from data generated in our lab but the same thing has been shown with other modulators by investigators at the NCI, Arizona and so on. This is looking at the impact of exposure to the anthracycline in a multidrug resistant cell line, in this case the K562 cell line, with a fixed control of the exposure time of PSC-833 using two micromolar. What you see is, looking at the IC50 shown here with a 1-hour exposure to the anthracycline and 24-hour exposure to PSC, we get roughly a nine-fold sensitization in leukemia cell kill.

If we prolong the daunomycin exposure and leave the PSC time fixed, we can increase that over 20-fold, a 220-fold sensitization. If we prolong the PSC exposure to 48 hours, there is no advantage, and I can tell you that with increasing daunomycin exposure to 48 hours it adds no additional benefit as well, suggesting that the duration of exposure may be important for efficient modulation of P-GP function.

TOP

As I mentioned earlier, VP16 appears to be a relatively low affinity substrate for P-GP. This is looking at photoaffinity labeling of membranes that are P-GP positive and you can see here, shown in the circles is vinblastine and triangles for verapamil. This is etoposide which shows the lowest affinity for P-glycoprotein in this model.

TOP

I thought I might finish with two more slides. What would I consider the important areas for additional clinical investigation when it comes to P-GP and drug resistance modifiers? Let us go back to what I mentioned earlier and compare the results of the other trials. Certainly we know that selecting an agent, an anticancer drug, the avidity for P-glycoprotein impairs its activity. It is very important, and sticking with the anthracyclines may be very important in this regard. We have to account for the change in AUC and some of the newer second generation modulators don't have that impact on AUC change. But perhaps that is a benefit when we are impeding cell defense and, also the duration of drug exposure may be important as well.

TOP

My questions that remain are: does the variability in systemic drug exposure to the antineoplastics indeed impact the outcome when you effectively modulate P-glycoprotein function? That still needs to be looked at and addressed in a randomized fashion in future trials.

Is prolonged infusion of the antineoplastic important to improve outcome with the drug resistance modifier?

Should we be targeting de novo patients versus high risk? I think the likelihood is perhaps de novo should be our target despite the prevalence of P-GP expression being higher in relapsed or secondary AML patients? What is the relevance of the MDR phenotypes? Should we be targeting the AML progenitor? Is this really the cell that we need to be able to define the phenotype of rather than the bulk leukemic population? What is the impact that we are having with these treatments on altering the mechanisms of drug resistance?

TOP

I want to finish with one slide which I know may be difficult to see but these are some data that are unpublished that actually were generated from the EORTC which I think at least gives us an idea of where should we be targeting this type of therapy.

This is trying to identify the predictive value of P-glycoprotein expression in different categories of AML, patients that either presented with de novo AML, a late relapse defined as relapse occurring beyond a year, or those patients who had an early relapse within 6 months or were primary refractory. In fact, what you see here in the orange is showing the predictive value of P-glycoprotein for induction outcome. You can see P-GP in de novo AML has a very high predictive value for induction outcome. The overall prevalence of detection in a bulk population is relatively low, 20 to 25 percent.

If we look at the late relapses, those beyond a year, the predictive value decreases. The frequency of

P-GP detection actually increases, but it is those patients that we identify that are the hardest ones to treat. Those patients with the early relapse and the primary refractory are the ones that had the poorest predictive value for P-glycoprotein although they had the highest frequency of P-glycoprotein expression.

My suspicion is this is probably where we should be targeting our treatment to have the best impact on overall survival, and I will stop there.

DR. LOWENBERG: Bob Lowenberg. In fact, Alan, the MDR model is an example of a mechanism where there is really a bell curve of experimental evidence that would predict that this would provide a clinical effective approach and the clinical results of accumulating studies are so far not very positive. I mean the SWOG study is an exception. You stressed the point of drug dose levels in the body and just to clarify the point that you made, did you state that in the SWOG study the cyclosporine was active in that subset of patients that had achieved high dosages of daunomycin plasma levels? Because if that is the case, I think that is perhaps the most likely explanation for understanding failure so far, even if there are complex mechanisms of drug resistance, even if there are other interfering factors.

I mean one would predict that it should work in P-GP positive patients at least to some extent.

DR. LIST: The answer is yes, but let me clarify this a little bit more. It worked at all steady state daunomycin levels, but the outcome as far as looking at induction outcome and overall relapse-free survival was better as we saw the daunomycin levels increasing, but the opposite was true in the control arm, and that is the important distinction here.

If this is simply due to daunomycin exposure, we would expect the same trend in the control arm, and in fact, people died faster on the control arm with higher daunomycin exposures whereas you had a better outcome at all daunomycin steady state levels with cyclosporine, but the higher your level, you improved your outcome even more. That makes sense to me when you are thinking about P-GP as a cell defense inhibitor as what it is.

If we block the major mechanism to extrude the drug from the cell, then the higher the exposure of the daunomycin the better the outcome, and that makes sense to me.

So the answer is yes, but it was seen at all daunomycin steady state levels.

DR. LOWENBERG: So maybe to add one point you referred to the HOVON/MRC study that overall was a negative study, but if you look at subgroups the P-GP negative patients did do worse on PSC. The P-GP positive patients did slightly better with PSC, again indicating the contradictory effects that may be achieved in vivo.

DR. LIST: And I think it is important. Bob, that is a very good point, but it is important to point out that these are very different studies. I think there certainly are differences with administering the drug, at least daunomycin as a rapid infusion, as far as toxicity profile as opposed to those patients that are getting instantaneous infusion and also to remember that the doses were cut on that as well.

DR. LOWENBERG: I think the development of these types of drugs is entering a critical phase. Is it going to be discontinued or is it going to proceed? Therefore, this is something we should probably address in the meeting during the afternoon, how to go on.

DR. BERMAN: Ellin Berman from Sloan-Kettering. Alan, can you comment on Kathy Scoto's recent publication showing that in patients undergoing thoracic surgery for metastatic sarcoma, exposure to doxorubicin actually provoked MDR expression and that maybe our initial analyses of patients with acute leukemia that they are iMDR negativei is not accurate, and in fact that these patients can be inducible. Is there a relationship between those patients perhaps in your studies that looked at that?

DR. LIST: I think that is a very good point. A number of investigators have shown that P-GP is rapidly inducible. Within an hour of exposure even to Ara-C, you can induce it.

It is a very important point. I think that may be another explanation for the apparent benefit when you look at survival at 2 and 3 years. Even if we disregard the idea of identifying the P-GP phenotype in the progenitor population, perhaps these patients were having up regulation of P-GP and then benefiting by the time cyclosporine was added. I think it is a very important point but a hard one really to get at when you get to the clinic. Perhaps it can be addressed in some of the other studies, but it is a very good explanation, I think, and a very plausible explanation for what we saw.

DR. VAN DER JAGT: Richard van der Jagt, Ottawa. Alan, perhaps I missed it, but did you get a sense from your data whether there is a clear relationship as the daunomycin levels increased you got increased toxicity in that population, or have you had a chance to look at that?

DR. LIST: You know, I have asked Ken Kopecky, our SWOG biostatistician, to look at that. The problem is people die from so many different causes. Then you try to break down them into some of these subgroups. The thing to remember, even though we looked at about 150 patients, we had adequate levels to do this in only approximately three-quarters of the people in the trial. When you do a subgroup analysis, it is hard to tell any common recurring theme for causes of death in the higher group or the control arm. It is a good question, but I can tell you that we cannot answer it.

DR. GREVER: Mike Grever, Ohio State. One of the concerns I have is the determination of the concept is not very impressive, and a lot of these studies have been as you indicated. So one of the concerns I have is if we take the traditional approach to the patient with the maximum tolerable dose, we are putting all of our eggs into the assessment basket. I really think the timed dose didn't change. I would be not in favor of doing things the way we have always done them. I think it is time to put at least a few of these complex diagrams to the test. I think that we may not know the distal end point, but I think it is very important to try to assess them in a systematic way. Phase I studies are nice because they allow you to do dose response assessment. There have been very few forms of malignancy that have been cured with monotherapy and so one of the concerns we have is to the end points because we are always evaluating the concept, and it is complicated. It is hard to believe patients only have one drug resistance. I think the real challenge here is the pharmacodynamic end points with combination use.

DR. LIST: I certainly wouldn't disagree with you at all, Mike. In fact, we all tend to like to do things like we have done them for years, and we want to give our anthracyclines as a rapid infusion and certainly managed care would like us to do that as well, but if we are going to develop a new approach based on biology then I think you have to apply it and test it based upon the way it worked in the lab. If a prolonged infusion appears better, certainly test the point clinically as well.

DR. ESTEY: Eli Estey, MD Anderson. You mentioned that one of the problems with the PSC study was the avidity of the PSC for VP16 was relatively low. Certainly for Ara-C I guess it is even less. So that raises the question why include Ara-C in these studies. I mean in principle you have shown that when the cyclosporine is there the level matters, and in principle if you omitted the Ara-C could you possibly give more, infuse more daunorubicin possibly in combination with VP16? There are certainly patients, and I think this is something that is going to come up all day, where there is no evidence that they benefit from Ara-C at all, high-dose Ara-C or whatever. So would that be a possible avenue to go to omit the Ara-C and therefore you could give perhaps more infusional daunorubicin?

DR. LIST: I would say yes and no to that, Eli. I mean from a conceptional standpoint, yes. You are right. Why put it in there, but the reality is if we believe that the people who are going to benefit from this the most are the pre-marrow transplant patients, I think none of us would feel comfortable omitting Ara-C in that situations. That is really what I think we need to test.

DR. ESTEY: Even with a complex karyotype patient, you think?

DR. LIST: If we can know that ahead of time, perhaps we can try that then.

DR. LARSON: I think we will have just one last comment from Peter and then we must move on.

DR. WIERNIK: Peter Wiernik, New York. Just a comment that is a little off the subject, but it seems to me that in APL where you have no P-GP you have the best data for a dose-response curve to levels of daunorubicin which is consistent with what you observed in that study.

DR. LIST: I think it is a great model to illustrate that.

TOP

TOP