Why Does Treatment Fail? by Carmen Allegra MD

SLIDES & TRANSCRIPTS
Wednesday, February 16, 2000

Why Does Treatment Fail?
Carmen Allegra, MD

Slide 1:

Dr. TEPPER:I want to start off and ask Carmen Allegra from the Medicine Branch to give us the first talk.

Dr. ALLEGRA: So, Joel, are you going to tell them whether I am one of the people who work in GI cancer or one of the other two?

Dr. TEPPER: They will have to figure it out.

Dr. ALLEGRA: Okay. Thank you. I thought that I would take a fairly broad look at some of the mechanisms whereby cells become insensitive or are insensitive, why therapy doesn't work, but I thought that to be able to do this effectively we really have to ask the question as to how therapy works or what we know,

Slide 2:

what we believe as to how therapy works. I thought I would discuss a little bit about kinds of an emerging links between carcinogenesis and drug sensitivity, and I will speak about this towards the end of the discussion. I thought I would start out a little bit with some of the fundamental steps involved in drug-induced toxicity and show you some examples of some of these major determinants.

I would like to look at this fairly broadly. So, I am going to kind of click through a fair amount of these determinants for a number of these determinants fairly quickly, and if we need to discuss that, we will save it for the discussion section for more detail.

Slide 3:

Basically this is the therapeutic pathway, and what we are hoping to accomplish is to injure a cell and to have this result ultimately in cell death. Obviously there are a number of things that have to happen in between, but almost all of our therapy, as you know, is directed towards damaging the DNA. It is the hope, one, that we could inflict enough damage on the DNA and two, the cell will responds to this damage in a way that invokes cell death pathways and the cell ultimately succumbs.

As you can see there are a number of events that have to happen, and I have divided these into upstream events and downstream affecters. For the upstream events, there are a number of issues that we have spent probably decades studying, and these include the pharmacokinetics of the drugs, drug activation. We probably have spent too little time exploring tumor physiology B how the reagent enters the cell as membrane transport and then how it is metabolized, both in terms of its anabolism and catabolism. Finally, where we have spent the vast majority of our preclinical efforts is in the target interaction, and that target might be an enzyme or it might be DNA, but ultimately for the agents that we commonly use, certainly for colorectal cancer, what we are hoping to accomplish is DNA damage. Once we inflict that damage the question is what happens? How does the cell respond to that damage, and we hope in a way that it is one that invokes the cell death programs. P53 really plays I believe a central role in this pathway. The damage has to be recognized initially and then either through p53 dependent or perhaps not pathways a cell death program is invoked and ultimately the cell succumbs, hopefully.

There are lots of issues that control whether or not a cell death program is turned on. One of these is the family, the BCL family genes, which are modifiers of the cell death pathway program. The BCL family is huge, and I will show you some examples of how these factors impact cell death.

NF kappa B, which is a survival factor, I will show you some recent interesting data on and of kappa B and its role in cellular resistance, and I would like to discuss another means to access the cell death pathways and that is through the TNF family of receptors. I will focus some of my talk on the trail ligand specifically a little bit later on.

Slide 4:

I thought I would focus on 5-FU. That is my favorite drug, and it is one that we use all the time for all the people in this room. As you know, there are a number of upstream determinants of sensitivity to 5-FU, and a lot of these have already been discussed. I won't belabor these points this morning, but it has to be anabolized. Obviously it is a pro drug, that needs to be anabolized inside the cell. There has to be the availability of phosphate donors to get it into the nucleotide state.

One of the issues that you heard already discussed at great length in the last session is the level of TS and its important role in determining whether or not a cell responds to 5-FU or not, and I would also include what happens to TS levels after it is exposed to an inhibitor like 5-fluorouracil and its potential importance as a mechanism of insensitivity.

Slide 5:

We are also concerned about the levels of intracellular folate polyglutamates. That is why we use leukovorin in our therapies now, the activity of catabolic enzymes, the extent of nucleoside salvage which could also potentially impact sensitivity, the level of incorporation of RNA and the DNA by both fluoropyrimidines, as well as false incorporation of DUTP where the levels expand enormously with the inhibition of thymidylate synthase and finally, the activity of uracil DNA glycosylates which is responsible for clipping out these misincorporated uracils and results in DNA damage if there isn't a thymidine available to replace it, which there often is not because you have inhibited one of the major pathways for making thymidine, namely thymidylate synthase.

Slide 6:

These are data that have already been shown by Len and I think Peter, so I won't go through these, but this is to illustrate to you the importance of levels of thymidylate synthase in terms of predicting response.

This is an early trial of Leichman's where they looked at patients with metastatic advanced colorectal carcinoma and found with low levels of thymidylate synthase there was a relatively high rate of response, whereas with high levels there is a low rate of response, and the survival was better in those who had responses to these.

Slide 7:

. Similarly here is a study that was published by Lenz and colleagues in 1998, and this is again using the assay that Peter described for you, the RTPCR assay as opposed to the immunohistochemical assay. This was in 36 patients with advanced disease once again. Patients with low TS had a higher response and a better survival,

Slide 8:

and as you can see in this plot those who responded tended to have low levels but obviously not everyone with low levels responds to therapy, thus the role of potentially other markers to help us to with more precision determine who will and who won't respond to therapy.

Slide 9:

I wanted to spend just a few minutes on some of the activities in our group. I don't want to be too parochial, but I will spend a few minutes on some of our lab work that is ongoing now.

As I suggested, when one exposes a cell to a thymidylate synthase inhibitor like fluorouracil the levels of the enzyme are not static. The levels of the enzyme increase, and this has been shown by a number of individuals in preclinical systems. We were able to show this in cutaneous tumors taken from patients with breast cancer. This was probably in the late eighties when we did these studies, more than a decade ago.

Slide 10:

We went ahead and investigated the intracellular or the biochemical mechanism through which the cell increases acutely its levels of thymidylate synthase, reasoning that this was an important mechanism of insensitivity and that if we could control this inductive process we may be able to improve our therapy, and we came up with the following model. The messenger RNA of thymidylate synthase is shown here. There are two binding sites on the message and when the protein is in its unoccupied form, it is able to bind to the message and regulate the translational efficiency of that message to keep the levels of protein low.

When one exposes a cell or a person to 5-FU and occupies the enzyme, the enzyme is no longer able to bind to the message, and the message then is translated at a very rapid rate and the levels of protein in the cell increase thus explaining the clinical observation that when you expose a person, give a person 5-FU, the levels of the enzyme markedly increase.

What we are trying to do is to identify a small molecule which can take the place of the protein, bind specifically to the message and inhibit its translation.

Slide 11:

To do that, the first thing we needed to do was to model the protein RNA interaction and, using a series of overlapping peptides, we were able to define an area on the protein, and the protein is a homodimer as shown here where the RNA specifically binds, and it forms this sort of saddle between the two homodimeric subunits where this RNA could potentially bind. This is shown in a space-filling diagram. Protein is shown at the bottom.

Slide 12:

RNA is shown in purple, and the arginines are shown in blue. These are the presumptive contact points between the two. The job now is using three-dimensional libraries of compounds to identify small molecules that can take the place of the protein and inhibit relatively specifically the translation of the message. That is ongoing work, kind of where we are at right now.

Slide 13:

I would like to now move to downstream determinants of drug sensitivity and as I showed you there are manifold ways in which a cell can develop in sensitivity through manipulation of these very complex downstream recognition of DNA damage and then ultimately initiation of cell death pathways. I will focus on just a few of them as shown here, p53, the BCL family, NF kappa B and finally, the trail.

Slide 14:

This is a diagram that came out of a recent edition of Oncogene, and it shows really the central location of p53 in responding to DNA damage, in this case the ionizing radiation, but obviously to any of our agents that ultimately result in DNA damage. It has a number of functions, and I don't want to go into all of these, but it certainly may lead to an apoptotic response and it certainly has direct effects on a number of very important modifiers of the apoptotic response, namely the BCL family members. As this shows, there is an induction of the pro-apoptotic BCL family member bax. There is a decrease in the survival gene bcl-2 and interestingly it also induces CD95 or fas. There is some very nice work that came out of Janet Halton's lab recently that has shown that fas elevation is probably very important in the mechanism of death due to a thymidine-less state which is what we create when we treat a cell with a fluoropyrimidine.

Slide 15:

There have been a number of studies looking at p53 in various cell systems and with various agents trying to determine whether wild-type p53 or mutant p53 is more associated with sensitivity or resistance, and I think you heard Dr. Fearon yesterday mention the fact that there are studies that are kind of a mixed bag all over the place, but the bulk of the literature suggests that mutant p53 is associated with diminished sensitivity, and I just thought I would show you two studies that would support this.

This is a study that came out of Sandy Markowitz' lab where they took cells, mutant p53 and then they went ahead and overexpressed wild-type p53 on an inducible promoter, and what they found is that if they transfected these cells with another mutant p53 there was no change in sensitivity to 5-FU, but two separate clones where they were able to induce and overexpress wild-type p53 were both sensitized to the effects of 5-FU. So, it is consistent with the notion that mutant p53 breeds resistance to 5-FU.

Slide 16:

This is work that came out of Burt Vogelstein's lab in the past 9 months or so, and I like this because they did an animal study where they use homologous recombination to disrupt the function of wild-type p53 in this cell line. What they show is that in the cells that contained a wild-type p53, the animals with the tumors were relatively sensitive to treatment with 5-FU, whereas in those cells where the p53 was disrupted they were somewhat less sensitive. So, again, it is consistent with the notion that resistance is associated with mutant p53. I think that is where the bulk of the literature stands today with respect to p53.

Slide 17:

I wanted to talk a little bit about bcl-2 family, multiple family members, and I am just going to focus a little bit on bcl-2 itself as well as on bax.

Slide 18:

This is a nice model which suggests the balance, it is really the balance of these multiple members of the bcl-2 family that are important to determining whether a cell goes down a cell death pathway which is shown here when you have the overexpression of pro-apoptotic members like bax, versus goes on to survival when you have the overexpression of the survival members of this family, both of which are influenced by p53 after DNA damage.

Slide 19:

John Hickman several years ago put together a very nice review on these particular factors and looked broadly at the literature and looking at a variety of different agents really representing almost every class of chemotherapeutics that we use, looking at a variety of preclinical model systems showed very nicely that in the main overexpression of bcl-2 which is anti-apoptotic results in diminished sensitivity to a whole broad class, all of these classes of chemotherapeutics, including 5-FU.

Slide 20:

What about bax? Bax is a pro-apoptotic gene and this is a group from Japan who showed very nicely in this animal model that when one transfers bax into this squamous cell carcinoma cell line in a mouse model, in this case using a gene gun to insert the gene in the tumor in vivo in the animal, and they showed very nicely that the combination of bax plus platinum resulted in a very nice inhibition of growth in this in vivo cell model, whereas the use of platinum alone or just bax alone really had very little impact on the growth of the tumor, again suggesting the importance of the bcl family members in determining sensitivity to chemotherapeutics.

Slide 21:

Finally, this is a paper that was recently published on 15 patients with CLL where the ratio of bcl to the bax was examined and what these authors found was that there was a direct relationship between this ratio of bcl to bax and sensitivity to¾ in this case¾ cytarabine but again supports the notion that the balance of these various members of this family are a very important determinant of clinical outcome.

Slide 22:

I wanted to spend a few minutes on NF kappa B. NF kappa B is a survival factor, and I have already shown you some of that in an earlier slide, and this is work that was published recently in Nature Medicine where the authors found that when one exposes cells, in this case the SN38 which is the active metabolite of CPT11 there is an overexpression, an increase in induction in the levels of NF kappa B which is shown here.

These authors also showed that they can, through transfection, transfect these cells with an inhibitor of NF kappa B, I kappa B, and in so doing eliminate the induction of NF kappa B, and as you can see on the right hand part of the slide when you transfect these cells with the inhibitor of NF kappa B you can sensitize them to the killing effects of SN38.

Slide 23:

They went a little further and showed that the same inhibitor of NF kappa B can also sensitize cells in vivo, and this is an animal model, lovo-xenograft response to CPT11. I will just focus you on this penultimate line here which shows that the tumor growth response was marked in those animals using an adenoviral vector which contained the inhibitor of NF kappa B, the I kappa B inhibitor, and treatment with CPT11 resulted in marked growth inhibition and a high apoptotic score, again supporting the notion that NF kappa B is an important determinant of sensitivity and in this case can be overcome through the use of this inhibitor.

Slide 24:

Some of our own work has shown in colon carcinoma cells, treated with either SN38 or a specific antifolate inhibitor, TS results in about a 2.5-fold induction of NF kappa B, again suggesting that this is an important reason why cells would not tend to respond to these particular agents. It is something that needs to be considered.

Slide 25:

The last thing I wanted to address was trying to get to the apoptotic machinery and turning it on through a different mechanism other than through DNA damage, and that is by using ligands to so-called "cell death receptors," and in this case the DR, the death receptor 4 and 5 are recognized by a particular ligand, a member of the TNS family of ligands, namely, trail, and when it interacts with this death receptor through its intracytoplasmic death domain turns on the caspase sequence and ultimately results in cell death.

Slide 26:

There is some very nice work that has been done in the Medicine Branch by one of our investigators whose name is Stan Lipkowitz. As shown here, Stan took a number of breast cancer cell lines and exposed these cell lines to either 5-FU alone as shown in kind of the open bars or trail alone, trail ligand alone shown in these hatched bars and then finally the combination shown in the black bars. I think what this points out is that neither 5-FU nor trail in this system at these concentrations with this timing of exposure were particularly good at inhibiting growth of these various cell lines, but the combination seemed to be impressively interesting.

Slide 27:

I wanted to sum up with a couple of thoughts. One is that a lot of the upstream dependent DNA damage which is what we depend on with our current armamentarium of therapeutics really requires the existence of balance of multiple cellular functions. What we are asking the cell to do when we expose to our cytotoxins that we use today is a very complicated task. We are asking for the drug to get there and be metabolized properly, to not be degraded, to stay there long enough to inhibit the target and then ultimately we are asking for the cell to recognize that and invoke the cell death pathway. It is a very complicated task that we are asking of our therapeutics, and it may be one of the reasons why they are not particularly effective.

There is a notion that carcinogenesis, the mere process of becoming a cancer, requires the prevention of cell death which is really a default position. If you think about a normal cell as it becomes a malignant cell it needs to incur a certain degree of DNA damage, and we know that cancer is a genetic disease, that there is instability of the genome in a cancer cell. Somehow the cell has to live with that. It has to develop a mechanism to survive despite that instability, and this to me potentially represents a house of cards because it is a complicated series of events that has to happen to enable the cell to stay alive because in the setting of genetic instability the natural tendency for the cell would be to invoke its cell death pathways and eliminate itself.

So the cell has to actually actively do something to keep itself alive, and it may represent a relatively easier way for us to incentivize the cell to death by interacting with the downstream pathways as opposed to the upstream pathways where the complications are ours to deal with.

In the downstream pathways the complications are for the cell to deal with and for us to tease and manipulate in a way that would incentivize the cell towards death.

Slide 28:

A few conclusions. One, cellular responsiveness is complex, multilevel and multifactorial. As opposed to upstream complexity which is ours to deal with, which we deal with all the time, downstream complexities may be relatively more easily manipulated for benefit because it is the cell that needs to deal with those complexities and get around them. Finally, future strategies I believe should combine up and downstream targets with agents that utilize potentially alternative pathways to the apoptotic machinery like the trail mechanism or through fas or other agents that do not depend on DNA damage.

I will go ahead and stop there.

(Applause.)