AUSTRALIAN FRONTIERS OF SCIENCE, 2005
Walter and Eliza Hall Institute of Medical Research, Melbourne, 12-13 April
New imaging techniques in detection and treatment of cancer
Associate Professor Grant McArthur, Head, Molecular Oncology and Translational Research Laboratories, Peter MacCallum Cancer Centre
What
I am going to talk to you about is the use of some newer technologies, particularly
positron emission tomography (PET) imaging to assist us in developing some of
these new anti-cancer drugs.
(Click on image for a larger version)
One of the problems that we face at the moment in saying ‘we’, I mean patients and clinicians who look after the patients, as well as the companies developing the drugs is that is that the cost of new drug development is escalating rapidly. So to bring a new drug to market in 2001 cost US$800 million, and now it is actually over US$1 billion a substantial sum of money.
(Click on image for a larger version)
Not only that, but the path to bring a new treatment to the marketplace to help our patients is a very long, time-consuming process, an average of around 14 to 15 years. It is also very inefficient. Of many promising compounds developed preclinically and in screens very, very few will get through to the marketplace. So we really need to do this more efficiently, more rapidly, at less cost and, given that a lot of the costs occur during the clinical trials phase of development rather than the preclinical development, we really have to do our clinical development a lot more efficiently.
I will put it to you today that perhaps we are just at the start of using some new imaging technologies that may be able to help us do this a lot more efficiently.
(Click on image for a larger version)
Before I give you a couple of examples of the way we are going about this, I will just go back and give you a brief overview of the so-called different phases of the clinical trials in cancer development. Traditionally in phase I clinical trials we ask one simple question: is it safe? We also ask questions about the pharmacokinetics, or delivery of the drugs. In phase II clinical trials, traditionally we have asked simply: does it reduce tumour size? And then in phase II clinical trials, we have asked the most important question: can we prolong survival of patients with cancer, or even cure some patients?
(Click on image for a larger version)
Now with the onset of some newer, perhaps more rational approaches to developing drugs, particularly these new targeted drugs such as Ricky Johnstone spoke about, I think we are changing this paradigm. In fact, we are asking some additional questions, during both phase I and phase II development. Yes, we need to know whether the drug is safe; but Tony Burgess in introducing this session mentioned the importance of knowing whether the drug is inhibiting the target. So we need to know whether our drug is modulating or inhibiting the target. Not only that, but it would be very reassuring if we could work out whether the drug modulates a biological process. We think that is important, based on our understanding of the biology of the target.
So I will give you a couple of examples today of where we have attempted to use functional imaging to help us with this process.
(Click on image for a larger version)
By doing this, hopefully, we can speed up this very time-consuming process.
(Click on image for a larger version)
Ricky introduced to you the positron emission tomography, and this is basically a very nice technique where one can get functional information about tissues in the body by using positron-emitting isotopes labelled to biologically relevant molecules in this case, fluorodeoxyglucose, the most common positron emission tomography scan used in oncology. What is done here is that patients are injected with the F18-labelled deoxyglucose, they go under a scanner, and then in tissues that take up high amounts of glucose, such as the brain, you can image the actual metabolism occurring in certain anatomical locations, including tumours.
(Click on image for a larger version)
The reason this is useful in cancer is that many cancers express very high levels of glucose transporters, so that high levels of this fluorodeoxyglucose are taken up into cancer cells, where it is phosphorylated, which makes it resistant to the normal metabolism that occurs with normal glucose in our cells. So it is quite a useful imaging technique, quite commonly used now in routine oncology.
(Click on image for a larger version)
But there are a number of other, novel tracers that can be used for imaging biological processes in patients and in preclinical models. Tracers that are currently being used at Peter MacCallum Institute and the Austin Hospital, as part of Cancer Trials Australia, include markers of cell proliferation, fluoro-L-thymidine; markers of hypoxia, fluoro-azamycin aribinoside; markers of amino acid transport; fluorocholine, which is a marker of cell membrane production; as well as other markers such as oxygen-15-labelled water, which one can use to actually measure quantitatively blood flow.
(Click on image for a larger version)
I am going to tell you two example stories today. The first one involves an inhibitor of the EGFR family of protein, receptor tyrosine kinase, the ErbB family, a family of molecules that Tony Burgess, the Chairman of this session, has worked extensively on. This is a multi-membered family of receptor tyrosine kinases some cancers have amplification of one of the family members, for example ErbB-2 in breast cancer but these molecules are commonly expressed and activated in a variety of human cancers, and have generated some interest as targets for molecular therapy. So the drug I am going to tell you about inhibits all of this family of receptor tyrosine kinases.
(Click on image for a larger version)
This family of receptor tyrosine kinases are activated by ligands, and that leads to phosphorylation of the receptor and then a whole stream of downstream signalling pathways that lead to a variety of biological effects on the cell. This includes promotion of proliferation, inhibition of apoptosis such as we heard about this morning, promotion of metastasis, and also promotion of other important biological processes in cancers, such as angiogenesis. The drug I am going to talk to you about today inhibits the auto-phosphorylation of these receptors, and that leads to inhibition of a number of these downstream biological processes.
(Click on image for a larger version)
The way we evaluate this preclinically is to use the machine shown here, a small animal PET-scanner. It can image lesions down as small as 1.5 millimetres in the animal shown here as a phantom, and it was installed in mid-2003 as a collaboration between Peter MacCallum Institute and the University of Pennsylvania. The next panel shows an image of a mouse injected with fluorine-18, which goes to very high levels in the bones. This is imaged on a conventional clinical scanner, but if we do this with a small animal scanner you can see a significant difference in resolution. You can now see the individual ribs in the animal, which you just can’t make out with a traditional scanner.
(Click on image for a larger version)
I will now run you through an example of how we have used this method of imaging to address biological processes in a cancer model. This cancer model is a xenograft model of a squamous cell carcinoma line called A431. It has been implanted into the hind limb of the mouse. Vehicle treated animals, as you can see, have uptake of the fluoroxyglucose analogue throughout the treatment period, but when treated with the Pan-ErbB kinase inhibitor there is actually a reduction in the uptake of fluoroxyglucose into the tumour.
(Click on image for a larger version)
One can quantitate that, using tumour to background ratios, and there is a reproducible reduction in the glucose uptake. One can also, at the end of the experiment, sacrifice the animals and validate that uptake by doing ex vivo counting, as shown here on the right.
(Click on image for a larger version)
So why, in this tumour model, might there be less glucose uptake occurring in cells? Well, it is known that cells that are rested in G1 have less uptake of glucose, so reduced cell proliferation would be an explanation. Interestingly, glucose uptake is also increased in cells in response to hypoxia, so possibly a reduction in glucose uptake could reflect reduced hypoxia in the tumour cells. Finally, of course, if the cells are dying they won’t be actively metabolising, so increased apoptosis is another possible explanation. So, using a variety of different PET tracers, one can address some of these questions in this model.
(Click on image for a larger version)
Shown here is uptake of a thymidine analogue into these tumours. Again you will see with vehicle-treated animals there is actually an increase in intensity of the uptake of the thymidine analogue into the tumour, but that is greatly attenuated and in fact inhibited in the ErbB inhibitor-treated mice.
(Click on image for a larger version)
And one can again quantitate this in multiple animals.
(Click on image for a larger version)
In the way we do these experiments we also look for other validations of the PET tracers, and in this particular model we also saw other markers of proliferation being inhibited, such as incorporation of bromodeoxyuridine. You can see here the effect in, first, the untreated animal and then in the treated one. So certainly there was an anti-proliferative effect of this drug in this tumour model.
(Click on image for a larger version)
So what about hypoxia? I mentioned that cells that are hypoxic take up high levels of glucose. Intriguingly, in this particular model, this time injected into the flank of the mouse, near the forelimb, with continued growth of this tumour there is increased uptake of this hypoxic PET tracer, and we have also validated this using different methods of markers of hypoxia as well ex vivo. And you can see, with continued growth, increasing labelling with the hypoxic marker. In fact, the ErbB inhibitor completely reverses this increase in label of the hypoxic marker, suggesting that not only is there an anti-proliferative effect occurring but potentially hypoxia to the tumour is being reduced.
(Click on image for a larger version)
And again one can quantitate these sorts of experiments, using tumour to background ratios.
(Click on image for a larger version)
So the glucose transport goes down and is associated with cell cycle arrest and an improvement in hypoxia. How do we believe this is actually being mediated at the molecular level? Well, in terms of the glucose transport, we found quite a significant reduction in the staining with the glucose transporter Glut1 following treatment with the drug, and that is probably the rate-limiting step that explains the reduction in glucose uptake in these cells. Precisely how hypoxia links to this we have not established in this model to date.
Importantly, though, in this particular model we don’t see much of a change in markers of apoptosis during the treatment period. So probably the changes in glucose are reflecting changes in proliferation and possibly hypoxia as well.
(Click on image for a larger version)
To summarise that: our ErbB inhibitory drug seems to improve hypoxia and reduce cell proliferation as the explanation of why the glucose uptake was reduced.
(Click on image for a larger version)
Now, in finishing up, I just want to move from an animal model into work we have done in the human, using similar PET tracers. We heard briefly from Ricky about the importance of angiogenesis in tumours, and we know that tumours can produce angiogenic factors that promote proliferation of blood vessels to nourish the tumour. One critical player in that pathway is vascular endothelial growth factor (VEGF).
(Click on image for a larger version)
What is interesting about therapies that target vascular endothelial growth factor is that they do not cause rapid shrinkage of tumours. So in our traditional drug development model, where we are relying on seeing shrinkage of tumours, you would throw these drugs out as being useless. However, if you do alternate clinical endpoints, such as time to progression of a tumour, there is no doubt that these therapies in this case, a monoclonal antibody, Avastin, or bevacizumab do prolong progression-free survival in patients with kidney cancer. But it doesn’t shrink the tumours.
So here we believe this functional imaging can be really useful, early in the development of a new drug, to tell us that we are really having a biological effect on the tumour.
(Click on image for a larger version)
I will now show you a clinical study we did that supports this idea.
This drug was another of the small-molecule protein kinase inhibitors. It inhibits the receptor for vascular endothelial growth factor but also other receptor tyrosine kinases important in tumour angiogenesis and stromal cell interactions like platelet-derived growth factor.
(Click on image for a larger version)
So we did a very complex clinical trial here in Melbourne, in the Cancer Trials Australia Group, where patients underwent multiple PET scans with FDG and then different cohorts of patients had different types of PET scans. So every patient had two types of PET scans. Some had proliferation PET scans, some had hypoxia PET scans in this case, with misonidazole and some had oxygen-15 water studies to determine whether the drug had any effect on tumour blood flow. Not only that, but some patients also underwent tumour biopsies before and after treatment with the drug.
(Click on image for a larger version)
The patients we treated had a wide range of different types of cancers.
(Click on image for a larger version)
What we found, interestingly, was evidence of biological effects of the drug in multiple tumour types, including colorectal cancers, sarcomas, melanomas, and lung and breast cancers. So many different types of tumours responded to the PET scan, and the PET scan response was much, much greater than the objective response in terms of tumour shrinkage just as you would anticipate, perhaps, for a drug that is targeting angiogenesis.
(Click on image for a larger version)
We have here an example of one of the patients treated on this study. This patient had both a deoxyglucose scan and also a thymidine proliferation scan. Here is the pre-treatment scan: a very complex tumour, uptaking FDG, in the thorax. You can see by two weeks substantial reduction in FDG uptake. And then the same thing is occurring with the proliferation marker. Around the periphery of the tumour you can see areas of uptake of thymidine, which resolves following treatment with the drug so, very clear biological effects of this drug on glucose metabolism and proliferation.
(Click on image for a larger version)
Here is another example of a patient treated on this study. This patient had the hypoxia scan. This was metastatic colorectal cancer, shown here in the chest with the abnormal FDG uptake, and that is associated with this yellow area which is uptake of the hypoxic marker. And in fact, following treatment with the anti-angiogenesis drug, the hypoxic region improved, in concert with a reduction in the FDG uptake. So that example I gave you in the animal xenograft model may also apply in humans.
Why might this be important? Well, I can remember sitting around a table in 1999 talking to Sugen when they were embarking on their anti-angiogenic therapies. We were having a heated debate about whether it was ethical to use an anti-angiogenic drug along with radiotherapy, because we know that if we make tumours hypoxic by taking their blood supply away, radiotherapy won’t work. So can we possibly do this? Well, in fact now we have evidence in the very earliest clinical trials that in fact the opposite is happening. The anti-angiogenic drug is actually improving hypoxia, not increasing it in the tumours.
(Click on image for a larger version)
Finally, one thing I will show you from this trial is that, as Tony pointed out, it is really important to know with your drug that you are inhibiting the target. This is a patient with thyroid cancer, treated on the trial. The patient had a very good PET scan response and also had some activation of the vascular endothelial growth factor receptor actually on the tumour cells themselves, and this was abolished following treatment with the drug evidence that in fact this drug was inhibiting the target that we thought it should be inhibiting.
(Click on image for a larger version)
Just one last thing to share with you from this trial, which I think is also very interesting: often it is a very difficult thing, in developing a new drug, to work out what dose of drug should be given. Based on our PET scans we found a very, very good correlation between the drug levels in the serum and the ability to obtain a PET scan response. So patients that did have a PET response had a much higher average concentration of drug in the serum than patients that did not have a response very useful information, because we therefore know that exposure to the drug may be important.
(Click on image for a larger version)
What I have done for you is to run through some examples of where we can use this functional imaging with PET scanning to potentially accelerate the phase I and phase II part of drug development.
(Click on image for a larger version)
Hopefully, what that will mean is that we can speed up the process, make it cheaper, get rid early of some non-promising drugs that are not inhibiting the target, not having biological effects, and concentrate on the drugs that are most effective.
(Click on image for a larger version)
Finally, I want to acknowledge my collaborators, particularly those in the Translational Research Group at Peter MacCallum Institute, co-headed with Rod Hicks, and also my clinical colleagues in Cancer Trials Australia, who collaborated on the clinical trial project.
