I would like to start by acknowledging the traditional owners of the land on which we meet, the Ngunnawal people, and also, on behalf of the organising committee, to thank the two Academies and the Australian government for giving us this opportunity to meet together here today.
We live 20 years longer, on average, than we did 100 years ago, largely due to the control of some infectious diseases during the 20th century. The challenge that we face in the 21st century is to recognise that it will not be quite so easy to make gains in the field as we have done over the last century.
Cancer is now the commonest cause of death in Australia, and the International Union Against Cancer recognises that over the course of the next 50 years cancer will likely become the commonest cause of death in most countries, not only in the developed world but also in the developing world. At the same time, we face the challenge of many infectious diseases that we have well controlled in this part of the world but that still have to be controlled in the parts of the world where the economies are not able to support the widespread deployment of vaccines.
What I would like to do for you this morning is just to diversify a little bit from the theme of infectious diseases and think about how we might make use of the immune system to help control cancer, and to face up to the challenge of cancer being the commonest cause of death in the developed world.
(Click on image for a larger version)
Of course, this is a story that started a very long time ago. While he perhaps didn’t realise that it was what he was doing, William Coley – through the discovery (or perhaps ‘creation’ would be the better word) of Coley’s toxins – was the first to use the immune system in the management of cancer. Indeed, he achieved remarkable results, with nearly 40 per cent of the patients he selected as suitable for treatment being, at least at some level, cured by the treatment that he gave, which consisted basically of live bacteria injected intravenously – not the sort of thing you would get past an ethics committee very easily these days, although in fact there are some treatments currently in development based exactly on that principle.
(Click on image for a larger version)
The aim of vaccination is to protect against disease with minimal side effects. We have already heard from Senator Jan McLucas that this is a story which also goes back a very long time, to the time of Jenner and before. Vaccine development then was largely empiric: you tried it to see if it worked, and if you did better than you were expecting, then you called it a success.
I would like to propose to you, just as a thought before we actually meet over the next three days, that in fact vaccine technology has not moved very much further forward than that. Nowadays we would perhaps design the vaccines a little more scientifically, but when it comes to understanding how they work, it is still a case of trying, after the event, to rationalise what we see and determine how the vaccine has worked, rather than setting out in advance to plan how it might.
(Click on image for a larger version)
Obviously, side effects were a concern even in Jenner’s time, and the cows shown here as growing out of the arms show that the thought that we were involved in cloning cells and the consequences of cloning was not a new one as a result of the stem cell debate but in fact one that had been considered for at least 200 years.
(Click on image for a larger version)
So, cancer immunotherapy: it sounds like something where you can take a tumour antigen and say, ‘Right, we’re going to attack that and try and control cancer.’ But it’s important to remember that in fact the two areas of cancer immunotherapy which we already practise are not nearly so scientifically designed as that.
The first is what was started with William Coley – Coley’s toxins – but is now routinely used in clinical practice as stimulation of the non-specific components of the immune system, the innate immune system. We use this in the treatment of bladder cancer, which we can treat by instilling live BCG organisms, and the direct use of the Toll-like receptor 7 (TLR7) agonist Imiquimod as therapy for skin cancer. So here there is no antigen involved; we are really just stimulating the natural immune responses. And of course this technology is being advanced as we speak, through the development of new drugs which will achieve the same sort of aim. Anti CTLA-4 therapy, which unfortunately has had a little bit of a knock-back in the last 48 hours with the results of a melanoma study showing no benefit is nevertheless still to be developed as a potential therapy for ovarian cancer.
Passive specific immunotherapy, where we actually transfer-in the components of the immune system rather than inducing them by vaccination has also been extremely successful, and we have a range of monoclonal antibodies available for routine use in the management of cancer.
So these two areas are already in clinical practice.
Active specific immunotherapy, which has been the goal of many people working in this area over a long period of time, has in fact proven the Cinderella area of the field, because although it has been mooted for a very long time and many trials have reached the stage where the effectiveness of therapy is being evaluated, none have yet routinely entered clinical practice.
(Click on image for a larger version)
So, thinking about the antibody side of things briefly, because I think it is important to remember that that is where our success stories actually lie at the moment, and to recognise that, we can recognise that there are a whole range of antigens on the surface of cells which the immune system can see. At the moment we are exploiting these by using monoclonal antibodies targeted at some of them to help control the cancers.
It is as well to remember, however, that we could, in principle at least, think about inducing immune responses through deliberate immunisation which might achieve the same result. The advantage of monoclonal antibodies is that they are an easily defined therapy and easily transferred to an individual who needs them; the disadvantage is that you need to keep doing it.
On this slide I have listed some antigens which are considered important in terms of immunotherapy.
(Click on image for a larger version)
But perhaps the more important side of things is shown on this next slide, when you look at the successes. These are the antibodies that are actually in routine use in clinical practice to help control cancer, with the antigens that they target listed.
There is a limited number of antigens, but they are all ones which are well recognised as being in some way helpful in controlling particular cancers, and it is notable that we are basically achieving a new one every single year since the turn of the millennium, so, hopefully, that will carry on.
I am not going to speak very much about the immunotherapeutic side of things focused on B cells, but I think it is important that we don’t forget about it.
(Click on image for a larger version)
The T cell side of things is the area which has always interested me, and indeed most tumour immunotherapy, at least at the experimental level, is directed to this area. What we are trying to do is to recognise that tumour cells express antigens which are unique to the tumour – viral antigens expressed in tumours induced by the virus – or mutated regulatory proteins (oncogenes) which express new epitopes that the immune system, or normal proteins over-expressed to immunogenic levels in the tumour. We can target these, at least in theory, through specific immunotherapy.
(Click on image for a larger version)
I am going to use my own interest in papillomavirus as a model for what we are going to discuss in the field.
It is worth remembering at the start of all this that 20 per cent of all cancer is actually due to infectious disease, and while human papillomavirus is at the top of that list, representing 5 per cent of the total global cancer burden due to infection with human papillomavirus, there are many other targets on that list that we should be picking out and attacking. We have effective vaccines against the top two listed here, papillomavirus and hepatitis B virus infection. We will hear during the course of this conference about potential vaccines against several of the others on this list.
(Click on image for a larger version)
Human papillomavirus is associated with cervical cancer. And cervical cancer, uniquely amongst the cancers listed out here, is almost 100 per cent due to infection with the virus that is responsible for it. But the whole burden of papillomavirus-associated cancer is not limited to cervical cancer. Basically, what I would like to do is to remind you of a couple of diseases that are involved with this virus, and then talk about vaccines to prevent this infection, and then vaccines that might be used as immunotherapy to treat existing infection. The first vaccines are here now and the data are really to show that the vaccines work; the immunotherapeutics are a challenge for the future, and I will show you some of our experimental data that say where the problems lie and how we might perhaps overcome them.
(Click on image for a larger version)
Cervical cancer is primarily a disease of the developing world. The countries in red on the map are the countries with the highest cervical cancer burden. China is perhaps anomalously shown in green, because the World Health Organisation figures for cervical cancer probably do not reflect the real burden of disease in that country. India is the country with the largest total number of cases, shown in pink only on the map but because more than one billion of the population live there it is then the country with the highest number of cases of cervical cancer.
Cervical cancer is in fact the second commonest cause of cancer in women worldwide, and in many countries the commonest, with a quarter of a million deaths attributable to it every year. As I have already alluded to, this is a disease which is 100 per cent due to papillomavirus. In fact, if we didn’t have human papillomaviruses we almost certainly would have no cervical cancer.
Two virus types, out of the 200 that are recognised, are together responsible for about 70 per cent of the cancer burden. That is of significance, because in fact those are the two types that are represented in the vaccines that have become available to prevent cervical cancer.
The infection itself is globally very common. At least 30 per cent of sexually active individuals will acquire an infection with one or other of those two high risk viruses at some time in their lifetime. But fortunately for us, the infections only very rarely lead to cancer. Less than 2 per cent of them will become chronic and lead to cancer.
There is a 15-year lag time, on average, between the time when you get the virus and the time when you get the cancer – which gives us a chance for intervention, both therapeutically to try and prevent that progression and also for screening for cervical cancer in those countries which can afford it.
(Click on image for a larger version)
But I thought I should put up for your consideration one other cancer which is increasing in frequency, which is also attributable to high risk papillomavirus infection. This is tonsillar epithelial cancer, which contributes about 20 per cent of the global burden of head and neck cancer.
It has always been recognised that head and neck cancer is associated particularly with smoking and excess alcohol consumption, and indeed one subset of tonsillar cancer is due to that. That is the subset which you see in older people, as shown on the right-hand side of this slide. But there is another group of tonsillar cancer which is the one which is increasing in frequency, and which is associated with high risk human papillomavirus infection in the cancers. This is a disease of younger people, younger men in particular, where the virus is present and where the risk factors of smoking and alcohol consumption are much less obvious. And this is a disease which has nearly doubled in prevalence over the last 30 years and is continuing to increase, so clearly while vaccines might help prevent one of these cancers, we would hope that it would prevent the other one as well.
(Click on image for a larger version)
The immune system does recognise that we get infected with these viruses, but it doesn’t do it very well. That is perhaps one of the reasons why a vaccine is possible against this virus – the natural immune response is poor. This slide is looking at antibody accumulating after the time that we become infected with these high risk viruses. Time zero is the first detection of virus in the cervix of a woman and, as you can see, it takes a year for half of the women that have been infected with this virus to accumulate antibody, and another four years before two-thirds of people have. So clearly the immune system doesn’t see this virus very well.
This has some practical implications. You can’t screen for infection by looking for antibody, and also we know that the antibody is not a marker for likely regression of disease. So that antibody is not the means by which we get rid of this infection, although I will show you in a moment that it seems to be sufficient to protect us against it.
(Click on image for a larger version)
We don’t really understand how we do get rid of papillomavirus infection, but I thought I would put up just a historical flashback, because that was how I actually got involved in this field some 25 years ago. While I was working at the Walter and Eliza Hall Institute we drew an association between papillomavirus infection of the anal canal and pre-cancer lesion of the anal canal in people who were immunosuppressed, and that was what got me first interested in this area. Indeed, we thought this might be an example of something which had been postulated by Macfarlane Burnet many years previously, that the immune system was actively involved in surveying for cancer and then eliminating it. We were perhaps half right with that, but more likely we were wrong, as we now recognise that HIV infection, which was responsible for immunosuppression in the men who developed anal pre-cancers, was in fact promoting persistent HPV infection. And persistent infection with papillomavirus, rather than impaired immunusurveillance, is what conveyed the increased risk of cancer.
The data here show resolution of papillomavirus infection with time. On the right-hand side in a population who are HIV-negative, half the infection is clear in a year, and 90 per cent of them over four years. But in an HIV-positive population, it takes two years for half of the infections to clear, and in fact 30 per cent of them are still persisting after four years. So this suggests very strongly that a cell-mediated immune response to the virus is important in controlling it, which is what would be predicted, if you like, by classical immunology. And it suggests very strongly that the reason that this is a cancer that was more common in immunosuppressed people was not because of immune surveillance for the cancer, but rather for the viral infection.
(Click on image for a larger version)
We recognised this further in people who were deliberately immunosuppressed to receive a renal transplant, where there was an increased relative risk – shown in yellow on the list here – of several cancers associated with papillomavirus infection, but notably no increased risk in the cancers in red on that list, in epithelial cancers which are not thought to be associated with an infection. So the immune system surveys for and controls cancers that are associated with viral infections, but possibly the evidence that it surveys for and controls cancers not associated with viral infections is rather less good.
(Click on image for a larger version)
Having said that, we know very well what is necessary to prevent papillomavirus infection. All I have said up to now is about how to get rid of it, but we know that what is necessary to prevent the infection is, very simply, antibody. As with every other vaccine that we have, a vaccine against papillomavirus infection that induces antibody in an animal model can protect against live challenge with the virus.
This study done by Bennett Jenson and his colleagues basically used canine oral papillomavirus virus-like particles to immunise dogs against their own papillomavirus, canine oral papillomavirus, and then the dogs were challenged with live virus. And live virus produced warts unless the dogs had been immunised. The red arrows on this slide point to an example, group 4, where the dogs had been effectively immunised. But, more importantly, serum taken from these animals and transferred into another animal, an animal that had never seen the virus before – and indeed the IgG fraction of serum – was able to protect those dogs in group 4 against live challenge with further virus.
So, clearly, antibody is sufficient to protect against infection with these viruses. It may not be necessary; indeed, there is some direct evidence for other viruses that antibody is not always necessary to protect against infection. But it is at least sufficient. And if it is sufficient, that allows you then to develop a vaccine which will induce neutralising antibody to help protect against infection.
(Click on image for a larger version)
That story started with Professor Harold zur Hausen and his work back in the 1970s, because he was the one that first drew the association between papillomavirus and cervical cancer and also mapped out much of our understanding of the basic virology of these viruses. That was possible because he was able to sequence these viruses, which had not been previously possible. You can’t grow papillomaviruses in tissue culture, so conventional approaches to vaccine development were not possible. Similarly, conventional studies on how the virus worked were not possible. It was only with the sequence data that it was realised that in fact there was not one or two papillomaviruses producing warts in humans, but rather a whole family of them. Harold zur Hausen thought about 20, but we now recognise 200 that are responsible for different diseases, some of which were responsible for cervical cancer, the so-called high risk viruses.
The availability of that information, plus recombinant DNA technology to allow expression of proteins, allowed the development of these virus-like particles which are the basis of the vaccines we now have. That work was done by my colleague, the late Dr Jian Zhou, who was a postdoc in my lab at the time. We came up with these virus-like particles, which resemble the virus at least to the extent that they are seen by the immune system in the same way, so that the antibodies that are raised against these virus-like particles actually recognise the virus.
The virus-like particles are made of the major capsid protein of the virus, and the surprise for us was that these major capsid proteins, if we expressed them using recombinant DNA technology, assembled themselves into the virus-like particles, and it was indeed that event that enabled the vaccines. If that did not happen, we would not have the vaccines. It was indeed quite a surprise to us. It was a bit like having a child’s building blocks, throwing them into the corner and having them turn themselves into the Sydney Opera House, because it was that much of a surprise.
At any rate, the virus-like particles produced by self-assembly are highly immunogenic virus-like particles which are the basis of the vaccines that have become available to help prevent cervical cancer. These are very conventional vaccines – virus-like particles and adjuvant induce neutralising antibody and give long-term protection.
The yellow arrow you see on the slide, joining the virus-like particles to the vaccines, is a little bit of an understatement, because it really represents the work of nearly 2000 scientists worldwide, with 60,000 women taking part in randomised placebo-controlled trials of vaccine and the expenditure of US$1 billion by the companies involved in developing the vaccine, as well as the 15 years that separate those two. Perhaps the arrow should be a bit bigger than the one that is actually there, in recognition of that.
(Click on image for a larger version)
I am going to show you some results so I have to make a disclosure of conflict of interest. Here is a nice picture of me becoming Australian of the Year a couple of years ago, just to take your mind off the fact that I am actually going to make some money off the vaccine I am going to talk about. I must stress that all the data that I am going to show you after this are data generated by other people, not by myself, so the conflict of interest is more apparent than real.
(Click on image for a larger version)
Of course, that wasn’t the only title I got that year. There was another one that was given to me – although I must stress that it was the vaccine that they were talking about with the title ‘God’s gift to women’, not me. I have had to point out to my children that to the best of my knowledge that is not a title they are likely to inherit from me.
(Click on image for a larger version)
At any rate, the data from the vaccine trials done with those virus-like particles over the last 15 years have consistently shown that, so long as you look at disease attributable to the virus types in the vaccines, the disease caused by HPV 16 and HPV 18, and so long as you select women for the studies who are not infected with these viruses when they come into the trials – both of those are very important conditions on these results – then vaccine efficacy is near 100 per cent. That of course has to be taken into consideration in the real world against the fact that the two virus types that are in these vaccines, 16 and 18, are together responsible for only 70 per cent of cervical cancer.
(Click on image for a larger version)
What that translates into in the real world is, as the green arrow on this slide shows, a reduction in cervical pre-cancer, the pre-cancer that we have to treat surgically, by about 70 per cent as a result of vaccinating a previously unexposed population, when all cervical pre-cancer is taken into consideration. This is, if you like, the real-world effectiveness of the vaccine.
But the red arrow represents residual disease due to other virus types associated with cervical cancer, and that red arrow is the reason why we have to carry on screening against cervical cancer for the foreseeable future.
Of course, in this country that green arrow translates into one operation per 1000 women per year that will not have to be done for cervical pre-cancer, once the population has been vaccinated – about 22,000 operations every single year. And that is the real public health benefit of this vaccine in a country where cervical cancer screening occurs.
In the developing world, where there is no screening, that green arrow translates into lives saved.
(Click on image for a larger version)
There is a point that I need to make about this, and to stress again. The data shown here are actually from the end of one of the big randomised placebo-controlled trials, showing vaccine efficacy and showing that basically in the real world, for a population who came into the trial without exposure to the virus previously documented, the vaccine is about 95 to 100 per cent effective, which is fine.
[SLIDE 19: Efficacy Against HPV 6, 11, 16, 18 – Related Disease by Baseline Serostatus and PCR Status] (second slide, adding red box around lower two items and obscuring the caption below the table)
But the other side of the coin is that if a population comes into the trials already infected with one of these viruses – seropositive for the virus or DNA positive, PCR positive – then as you can see from the column second from the right the efficacy is effectively zero. This is not a therapeutic vaccine. It cannot be used to treat existing infection and therefore reduce the risk of cervical cancer in the future, a point which I will return to in a moment.
(Click on image for a larger version)
So we’ve got a vaccine to prevent a very significant cancer, and I guess that’s the ‘glass half full’. We have to look at how well we are getting on with deploying it.
The vaccine is licensed in many countries worldwide, and indeed there are now very effective government-sponsored programs in some countries. Australia, I am pleased to say, leads the way in that regard, thanks to the decisions of the government, and that vaccine has been made available for schoolgirls aged 12 routinely, and then for the next couple of years for schoolgirls aged between 12 and 18, and for older women between the ages of 18 and 25, for some catch-up immunisation.
We are getting great coverage in Australia, and I can tell you, when I talk about the vaccines elsewhere in the world people are extraordinarily jealous of the effectiveness of the program for delivering vaccine in this country – although I must note that really it only works as a program when delivered through schools. We get the same coverage in the 18- to 25-year-olds as has been achieved in America, where it is roughly 20 to 50 per cent, varying by states, if it is left to the individual’s decision as to whether they take part in the vaccine program.
I think it is very important to remember that vaccines are only as effective as they can be when they are delivered. It is not just the vaccine development that is important; it is all the understanding of human behaviour that goes into whether people choose to get themselves vaccinated or not that determines the real effectiveness of a vaccine in the field.
Other countries are trying to emulate what we have done in Australia. The European Community and Canada will start this year to immunise in a similar mode, and I was pleased to find out while I was in the United States last week that one state in the United States at least, Virginia, has decided to go the same way and introduce routine programs through schools.
But the big challenge that we face is to get this vaccine where it is really needed in the developing world, and that is not something that is going to happen overnight. The Gates Foundation, along with the companies manufacturing vaccine, and WHO, are collaborating to try and develop policies for delivering these vaccines in the developing world. There is a big challenge there. The countries that most need these vaccines well exceed the threshold that would enable them to get benefit from the global vaccination scheme, because US$1000 GDP leaves for example China well out of the loop, and yet this is a country with a significant burden of cancer.
Field trials for delivery of vaccine are absolutely critical, because it is not self-evident how you would achieve immunisation three times over six months to a year with a vaccine designed to be given to 12-year-olds, when 12-year-olds in many of the countries that we are talking about no longer attend school, have no fixed address, do not know their date of birth and have no medical records. That is one of the challenges that we are undertaking in Vanuatu at the moment as a demonstration project to find out how best to deliver vaccine in that part of the world. It is not that we are checking whether the vaccine works; we already know it does. We are just trying to work out whether we can come up with delivery strategies that will achieve effective vaccination three times over the period of a year.
(Click on image for a larger version)
I said that these prophylactic vaccines are not therapeutic. Clearly, with 20 million women out in the world at the moment who are already infected with one of these viruses and are going to die of cervical cancer as a result of that infection, it would be very nice to have therapeutic vaccines that we could use to treat existing infection. Indeed, 5 per cent of ‘healthy’ Vanuatu women over 30 already have cervical pre-cancer, and these are the women who are going to die. So we really need to have something we can do to try and control this disease.
Of course, we don’t have any therapeutic vaccines at the moment, perhaps excepting Zostavax, which is not really designed to cure you of an infection but rather to ameliorate the symptoms of the reactivation of a latent infection with Herpes zoster virus.
(Click on image for a larger version)
The clinical problem for papillomavirus is that we know the infections resolve spontaneously very frequently, which implies development of cellular immunity. We also know that there are tumour-associated viral antigens expressed in the epithelial tumours that arise from HPV infection. And we know that vaccines can and indeed do, in humans, raise specific immune responses of the sort we think ought to be therapeutic, including cytotoxic T cells and CD4 cells.
Unfortunately, the bottom line is that these vaccines in practice don’t work. And we have really got to understand why that is.
(Click on image for a larger version)
I will show you one example of a clinical trial that we were involved with, and I acknowledge CSL as being the sponsor of this trial, and Stirling Edwards, sitting in the audience here.
This is a vaccine which was basically designed to produce cellular immune responses against two viral non-structural proteins of papillomavirus, and was undertaken some years ago.
(Click on image for a larger version)
Basically, what we found is that we got good immune responses. We measured antibody – I’ll not go through the details of the slide particularly – and in comparison with the placebo group, significant levels of antibody were raised as a result of vaccination in women with cervical pre-cancer.
(Click on image for a larger version)
Similarly, cell-mediated immune responses – DTH reactions and cytotoxic T cells – were raised as a result of vaccination.
(Click on image for a larger version)
But the bottom line as far as the vaccine was concerned was that the disease that these women came into the trial with, cervical intraepithelial neoplasia, persisted at the end of the study, despite the immune responses that we thought might be therapeutic. That is obviously a challenge for us, to understand why that might be.
(Click on image for a larger version)
You have to understand that these studies are based on a transplantable tumour model, which is really the one that most of the work is done with, and this transplantable tumour model has been used to pedigree every vaccine that is put into clinical trial as a therapy for cervical pre-cancer. We used the same transplantable tumour model, and our potential vaccines worked just fine against that, but the problem is that the test is rather non-sensitive, because if we use just the antigen on its own or just the adjuvant on its own, we get significant resolution of disease. In other words, this is a model which will pedigree many more vaccines than are actually likely to be effective in the field.
(Click on image for a larger version)
So a while ago we came up with an assay which we have used to try and work out a little better what actually is important for immunotherapy. Basically, the assay consists of a bit of skin transgenic for our antigen of choice, so expressing E7 protein of HPV 16, and then transplanted onto another animal, so that now a normal immune system can see a bit of epithelium expressing E7 protein.
(Click on image for a larger version)
This slide shows what it looks like when you do that, and the slightly more ragged looking graft on the right-hand side is the one where the transgene, the E7 protein, is expressed.
If you leave those grafts alone, they will turn into cancer. The cancer is driven by the E7 transgene, the protein which is involved in oncogenesis in the cervix. But if you don’t do anything to these animals, that is all that really happens. The grafts never come off. So, if you like, this is a model which is a little bit more like what goes on in the cervix of a woman with persistent infection. We have used this to understand a few of the rules about what is important for immunotherapy.
(Click on image for a larger version)
One thing we have learned is that not all antigens are equal, that the papillomavirus antigens E6 and E7 tend to be associated with persistence of the grafts – the green line at the top of the graph on the slide – whereas the red line shows a number of other antigens expressed in the same system in epithelial cells where the grafts come off spontaneously. These include antigens like ovalbumin which are, if you like, ‘good’ antigens.
We don’t understand why there is a difference between the antigens which we are really interested in and the ones which we use as model antigens, but we can use this to exploit the system to find out why some grafts come off and some don’t, and what is necessary for graft rejection.
(Click on image for a larger version)
One of the things that we learned very early on was that you need not just cytotoxic T cells but CD4 helper T cells to get effective graft rejection to occur – this in a primed animal where the animal has already learned the immune response that will lead to graft rejection because it has already had one graft and rejected it. Now if you put a second graft onto that animal, the grafts are rejected. (That is the black line on the slide.) If you then take away the CD4 cells (the red line) or the CD8 cells (the green line), you protect the graft against rejection. So you need both CD4 and CD8 cells at the effector stage in the process of getting the grafts to reject.
This is of particular significance because there was an interest for a while in developing epitope-specific therapeutic vaccines for cancer, focusing on CD8 epitopes, and if these data are generalisable to other epithelial models, then what you would have to say is that that would not be a successful strategy. You would really need to have both the CD4 and the CD8 responses to get it to work.
(Click on image for a larger version)
The second thing we learned early on was that conventional vaccines don’t work to get rid of these grafts. If you immunise with E6 or E7 protein with a good strong adjuvant, you get the right responses – you get cytotoxic T cells, you get delayed-type hypersensitivity, you get rejection of a transplantable tumour expressing E7 – but the grafts expressing E7 never come off. Yet the grafts express 10 times as much E7 as the transplantable tumour, per cell. So clearly there is something inhibiting graft rejection.
(Click on image for a larger version)
The short answer is that it is the local environment where the graft is. You can get rid of these grafts if you create an inflammatory environment. We recognised this early on, using live Listeria as a means of stimulating a local inflammatory environment.
(Click on image for a larger version)
If we take that model further, and actually deliberately test the hypothesis that you need to have inflammation for effector T cells to work, then you find that the answer is that you do. So we use the model again where we prime against the antigen by grafting and the grafts come off, we take away the effector cells, and the grafts stay put. So a second graft put on the animal survives, and it seems to survive indefinitely as a result of having been placed in a privileged environment when there were no effector cells.
If we let the T cells recover, now the graft should come off. But it doesn’t. And if we put a third graft on, we can prove that that animal has recovered its immune response, primed to the same antigen very successfully, but the second graft, which is now healed in place and still expressing antigen but without local inflammation, is able to survive.
(Click on image for a larger version)
If you then go on formally to test whether, if you take that second graft, healed in place with antigen, and inflame it deliberately with the Toll-like receptor 7 agonist Imiquimod, the grafts will now come off; this tells you that it is something local that is important. It is not that the vaccine has failed. The vaccine, if you like, has produced the immune response. Here the vaccine is a graft, but it doesn’t really matter what it is. We have got the right immune response, but to get it to work locally we have got to change the environment.
(Click on image for a larger version)
Indeed, we have used that model to go on and look at a whole range of different things which regulate the local environment, and found a whole range of interesting molecules which are, if you like, switching off the immune response locally and preventing the effector T cells from doing their job. I’ll not go through them in detail but just point out that there are a number of favourite molecules in there which are known to be immunoregulatory, including a natural inhibitor of interleukin-1, interleukin-10, TGF β, and a surprising molecule, interferon-gamma, which turns out to be a negative regulator of local T cell function in a primed animal. So what these give us is targets that we can think about attacking. With a bit of luck, if we can inhibit those inhibitors, we might have some chance of getting effector function of the vaccine-induced immune responses to actually work.
(Click on image for a larger version)
I will leave you with one last thought in this area, that when you are doing immunotherapy, whether for chronic viral infection or for tumour, the immune system has already seen the antigen. You are not getting in there to do the job the first time round, because the immune system has already made a decision about what sort of immune response it wants to make to that antigen, and has decided that it will not be an effective immune response.
So we wanted to model that, and find out what might be going on there. We modelled that by immunising again with our favourite antigen, human papillomavirus capsid antigen, to establish immunity, and then immunising with an antigen which also had a new cytotoxic T cell epitope in it, the E7 protein, to see if we could induce specific immunity to our new favourite antigen by vaccination. The question was: could we get cytotoxic T cells induced in an animal that was already primed?
(Click on image for a larger version)
The green bar basically shows the immune response you get to that novel antigen in an otherwise naïve animal – so, you are immunising for the first time with the E7 protein, fused to the virus-like particle, and you get a good cytotoxic T cell response. But the red bar shows that if you are already immune, if you have already been deliberately immunised against the carrier protein, and we then come in with exactly the same vaccine with the E7 in it, we no longer see that immune response. So, if you like, the animal, by previous exposure to the carrier antigen, has decided already that it is not going to make a new cytotoxic T cell response to the antigen you would like to have induced an immune response against.
(Click on image for a larger version)
We also know – and I’ll not go through it in great detail – that this is due to IL-10 secreted by T helper cells in response to exposure to antigen. The graph at the left basically shows that you can get an immune response to the virus-like particles whether an animal makes interleukin-10 or not. But the graph at the right shows that in an animal that can’t make interleukin-10 we get a very good cytotoxic T cell response, even if the animal is already primed to the carrier antigen – whereas if the animal can make IL-10 we don’t see that cytotoxic T cell response.
Indeed, we know that that is IL-10 secreted from CD4+CD25+ regulatory T cells in response to the carrier antigen. So these are induced regulatory T cells, antigen-specific.
This is, in a sense, bad news, because it says it will be hard to get the cytotoxic T cell responses, but it is good news too, because it gives us another clue as to how we might intervene to allow those responses to be generated by, for example, neutralising interleukin-10. And in that model, if you use an antibody to neutralise interleukin-10, you restore the ability of the animal to make cytotoxic T cells. Interestingly, there was recently a paper in the Journal of Experimental Medicine using a model virus, lymphocytic choriomeningitic virus, and showing exactly the same finding.
(Click on image for a larger version)
So we are going to have to manipulate the immune system to allow T cells to work locally if we are going to get immunotherapy to work, and to prevent the problem of prior exposure to antigen biasing the immune response away from effector function. The problem with that is that it is going to be a double-edged weapon. Basically, all of these blocks to effective immunotherapy are regulators of self-reactivity, and if we disturb them for long or overdo the disturbance, we are going to promote the generation of autoimmune responses – animals without interleukin-10 develop colitis; animals treated with interferons, or lacking negative costimulators, develop autoimmunity. Clearly, we will disturb these areas of the immune system at our peril, but they may be the only way we can get immunotherapy to work in the short term.
An example of a drug in trial at the moment is Anti CTLA-4, with which it looks as if temporarily you can switch off the regulatory T cells – but at the expense of autoimmunity attacking the pituitary and the bone marrow. So we just have to work out how best to use these therapies.
(Click on image for a larger version)
I will leave you with the conclusions for immunotherapy, and for prevention of cancer through immunisation: that it will be much easier to prevent than to treat, and that when we have got viruses that are responsible for cancer, vaccines to prevent infection with the viruses are going to be the highest priority.
Monoclonal antibody therapy is proven successful for control of some tumours, and therefore vaccines targeting B cells may also be feasible to raise the same sorts of protective immune responses.
Immunotherapy generated against tumour-specific antigens should include both CD4 and CD8 responses, and we may have to modulate the immune system in some way to allow these T cells to do the job we would like them to do. This might include, particularly, modulation if there are already existing ineffective immune responses to the disease we are trying to treat.
I haven’t talked about cellular immunotherapy, because I see this as a rather difficult area where you are customising treatment to an individual patient, but we have to point out that there have been some startling, if limited, successes with cellular immunotherapy for cancer. In melanoma, 10 to 20 per cent of advanced disease can be at least temporarily sent into remission by cellular immunotherapy based on dendritic cells primed with tumour-specific antigen. I see these studies as providing proof of principle that we have actually got some wins to score.
(Click on image for a larger version)
I would like to acknowledge a very large number of people who have been involved in the work in my lab over the years. I’ll not name them all individually or we will be here till lunchtime, but I would draw your attention specifically to a couple of people – first of all, the late Dr Jian Zhou, who was very much involved in the development of the prophylactic vaccine against papillomavirus infection, and also one of my PhD students, Rachel De Kluyver, who stuck for nearly 14 years as technician and subsequently a student with me through all of the grafting experiments that I have described to you today.
(Click on image for a larger version)
Discussion
Question: Is anything known about the cell-mediated immune responses to the VLP papillomavirus vaccines? You mentioned neutralising antibodies, but how much of a role does CMI play in this?
Ian Frazer: The question is obviously an important one, and the short answer is that we know that there is a strong helper T cell response to the vaccine after immunisation – and you can actually measure that as delayed-type hypersensitivity, as we did in one of the trials that we did. Cytotoxic T cells do not seem to be induced by vaccines adjuvanted with alum, as this vaccine is, and certainly we were unable to find any, again, in the trial that we did. That doesn’t mean they are not there, but they are certainly not there enough that we can see them in the peripheral blood of people who have been immunised.
We recognise – and I sort of alluded to this idea but didn’t actually give any evidence to support it – that these vaccines can work by means other than antibody, and indeed we recognise for the hepatitis B virus vaccine that the protection against hepatitis B lasts long after the antibody has gone from the blood, at least as measurable against the assays that we use at the moment.
So it is quite possible that cellular immune responses contribute to protection against papillomavirus infection. Antibody may be sufficient but other responses may help. We actually know that is true for natural papillomavirus infection, because the people born with immunodeficiency disorders where they cannot make B cells, cannot make antibody, are still able to cope with papillomavirus infection in the normal way. They get one attack of each papillomavirus, they become immune, and then they do not become persistently infected – or recurrently infected. This suggests that cellular immunity plays a part in controlling the infection.
Question: Ian, you mentioned in one of your trial slides that treatment with interferon had potential [inaudible] immunity [inaudible] treating very large numbers of hepatitis C patients with a high dose [inaudible]. Have people looked at combinations of those [inaudible]?
Ian Frazer: I think the short answer to that is no at the moment. The problem is that it is the business of combining therapies that puts people off. There are some very practical realities about vaccine development, and one of them is that you really have to do each bit individually before you get onto combining things. But I think that we are going to have to start addressing these issues seriously, because if anything that we have learned from our animal models has any reality at all – and I have to stress, they are animal models – if they are reflecting in some way the immune reality in humans, it is going to require a combination of therapies to overcome the problems.
Certainly we know that all of the studies that have been done for papillomavirus-associated cervical pre-cancer, using conventional vaccine approaches, induce the right sorts of immune responses where they have been looked for, but don’t seem to be therapeutically effective, so that we really do have something extra that we are going to have to do.Visit these websites
The Sir Mark Oliphant Conference Series is supported by the Australian Government under the International Science Linkages Program.
© 2008 Sir Mark Oliphant Conferences.
Website design acidgreen
