(Click on image for a larger version)
The most important slide, of course, is the acknowledgements. I acknowledge my own group at the University of Melbourne and various other groups at the University, including Lorrie Brown and members of the Doherty laboratory; John Walker, at Pfizer Animal Health; Mike Good and Mike Batzloff, at QIMR; Eric Gowans and Dominic Wall; Richard Roden, at the Johns Hopkins University, with Hannah Alphs; and Berma Kinsey and Tom Kostner at Baylor, in Texas. I will be talking about the work that we have been doing with most or all of these people, except for Eric’s work which was presented yesterday.
(Click on image for a larger version)
I am going to tell you how we target epitope-based vaccines to dendritic cells (DCs), how we target protein-based vaccines to DCs and, if we have time, how we are now targeting DNA-based vaccines to DCs. And this is all based on quite simple lipid structures, one in particular.
(Click on image for a larger version)
So what do we want to do? We want to take this very quiescent DC and turn it into a very angry cell which will then kick off immune responses, either antibody or CMI (cell-mediated immune) responses.
(Click on image for a larger version)
How have we done that? In the first place, about 15 years ago we started with epitope-based vaccines, very simple structures of a contiguously synthesised T helper epitope with a target epitope which could be either an antibody-inducing epitope or a T cell inducing epitope. And because we are basically chemists we were able to play with the geometry of these structures that we were assembling, these vaccine structures.
We were interested in examining branched structures and we were also interested in looking at multiple copies of the epitope within the simple vaccine structure.
(Click on image for a larger version)
One of the first experiments was to look at the way that these various simple epitope-based vaccines were recognised by DCs. What you see here is the ability of linear or branch structures to turn on, in this case, a CD4 T helper response. And what we find is that branched structures are much more efficient at stimulating DCs to activate the T cells. So it looks as if DCs like odd-looking things. They are poised to recognise unusual-looking structures.
Also, branched structures are more stable to proteolysis. What we are looking at here in the blue is the stability of these branch structures to proteolytic enzymes. So DCs like odd-looking things, and those odd-looking things are also more stable to proteolytic digestion. So they are more stable than simple linear structures.
(Click on image for a larger version)
What else is attractive to DCs apart from geometry? Well, danger signals, of course, and one of the most popular is Pam3Cys. We have all heard about this from Günther Jung’s work from 15 or even 20 years ago. There is another variety, Pam2Cys, which is only two palmitic acid residues attached to a glycerol moiety and then a cysteine residue.
So what we have done is to take our simple epitope-based vaccine structure and put this danger signal either in the centre of the molecule or at the end of the molecule. Many people have done the sort of thing shown at the bottom of this slide, but when we put the lipid in the centre we found that our structure was much more soluble and therefore amenable to manufacture than it was when we put it at the end of the molecule. It also turns out that this more soluble structure is more immunogenic than our linear structure.
So we have a simple epitope-based structure – you see here our target epitope and our helper – and it is targeting DCs through the Toll-like receptor using this very simple lipid moiety. This is a very simple synthesis. We can make 30 or 40 residue peptides overnight; another day and we have got the lipid on the end of it. And we can make this in 100 mg quantities. So it is scalable, and it also lends itself to very rigorous quality control by HPLC and mass spectrometry.
(Click on image for a larger version)
The uptake of these lipopeptide structures by DCs is shown in this slide. In the top left-hand corner at A we have a lipopeptide. These are murine DCs, although we have also done this with human DCs and we get a very similar result. Below it you see a contour map of one of those cells, showing the amount of fluorescently labelled lipopeptide that is taken up. You can see that there is very good uptake of the fluoresceinated lipopeptide by DCs.
In the series A–E the time is shown. You can see that there is uptake within five minutes, in 20 minutes you get even more, and again the contour maps are shown here. In four hours we have loading into vesicle type structures, and even at two hours, and again, as Eugene Maraskovsky was showing with ISCOMATRIX®, we seem to get a blush of material which is not associated with these intercellular vesicles but distributed throughout the cytoplasm.
(Click on image for a larger version)
Uptake is shown more quantitatively in this slide, where we can see that the percentage of DCs which are taking up the lipopeptide is 60 per cent, within 10 minutes. Uptake increases to about 80 per cent. It is also concentration-dependent, a biphasic type response and is saturatable, which indicates that it is a receptor-mediated reaction.
(Click on image for a larger version)
What do these lipopeptides do to DCs? If you take a population of mouse DCs, then some of those spontaneously mature. We are looking here at MHC class II expression as a measure of maturation. If you add non-lipidated peptide there is very little effect. Shown next is what LPS does to DCs, and then what our lipopeptide does. LPS and this lipidated structure Pam2Cys have similar effects.
(Click on image for a larger version)
The mechanism is that it works through TLR2, which is complexed with either TLR1 or TLR6, depending on the acylation status of the lipid moiety that we have in the peptide.
(Click on image for a larger version)
Does it work for us? It does. We have taken HEK293 cells and transfected them with TLR2, or left them untreated, and we find that only those cells which are transfected with TLR2 are activated..
TLR2, I might add, is known to be an endocytic receptor, so you are not only targeting the DCs through this receptor, you are also loading the DCs through this receptor, and that was supported by the data I showed you using fluorescence microscopy.
(Click on image for a larger version)
Our antibody responses using lipopeptides are T helper dependent. We have used MyD88 knockouts, MHC class II deficient mice and also CD40 deficient mice. In none of those cases do we get antibody produced, but when they are present we get antibodies produced in (as I am showing here) two different strains of mice. So our antibody responses with these T helper antibody-epitope structures with lipid in the middle are T helper dependent.
(Click on image for a larger version)
So does it work, and if it does, what antigens have we used? Most of our work has been done with LHRH, which is a reproductive hormone – an example of a self antigen. The point I want to make here is that we can break tolerance to self antigens.
We have also recently, with Tom Kostner and Berma Kinsey at Baylor, been examining drugs of abuse, particularly cocaine, amphetamine and nicotine; Group A Strep, with our colleagues at QIMR; influenza in my own laboratory and with Lorena Brown; HPV (human papilloma virus) with Richard Roden; and also some work with two tumour models, with Andreas Suhrbier in QIMR. This is all published work, so I won’t be talking in detail about all of the results.
(Click on image for a larger version)
But let me convince you with some of the data that this really does work in a real-world situation.
LHRH, or GnRH, is a hormone produced by the hypothalamus. It moves through the portal system to the pituitary gland, and instructs the pituitary gland to release the follicle-stimulating and luteinising hormones which act on the gonads – whether you are a man or a woman it doesn’t matter, it’s the same hormone. If you are a giraffe, an elephant or a human it doesn’t matter, it’s the same sequence of hormone. And here it is on this slide. It is a 10-residue peptide. It is responsible for everything which makes us men and women and everything that goes on around that.
Why would you want to make a vaccine against LHRH? Well, if you are in the livestock business or if you are a vet, then it is very useful to stop your dog making a nuisance of itself with the lady dog next door, it is a way of immunocastrating dogs. You no longer have to spay or surgically castrate. If you are into animal husbandry, it is also very useful in the case of cattle and horses – horses don’t run very fast when they are heavily pregnant, so castration or control of reproduction is very important in the animal industry. You can also think of some uses for it in humans. There is evidence that anti-LHRH vaccines could be useful in the case of prostatic cancer and, indeed, in the case of some estrogen-dependent breast cancers.
(Click on image for a larger version)
This slide shows individual dogs, time in weeks, and testosterone levels. You can see here what the vaccine does to the testosterone levels in those animals. After two inoculations they hit a brick wall where they suddenly stop producing testosterone.
(Click on image for a larger version)
The abuse of cocaine and methamphetamine are major problems in our society. What we have done is to put, in this case, cocaine at the end of a T helper epitope. We have our lipid in the middle, our usual branched structure. We have done this also with amphetamine, we have done it with DNP (dinitrophenol) and other experimental, small molecules. I hasten to add that these really are very small. The moiety at the left here is the size of a single amino acid.
(Click on image for a larger version)
What we find is that we can get some very good antibody titres. Shown here is our lipidated cocaine vaccine – very high titres of antibody here. Shown also is the non-lipidated structure (same vaccine but no lipid) but administered in complete Freund’s adjuvant. You can see that we are getting the same sort of secondary antibody response as we do in CFA – primary not so good, but certainly a very good secondary antibody response. You can also see our control here.
Fourth from the left is cocaine, coupled now in the traditional way to a protein carrier and administered in CFA – not so good, nor do we get a primary response in that case. And there are our various controls.
What is startling to us is that you don’t only get an antibody response to the very small molecule.
(Click on image for a larger version)
One of the infectious diseases that we have worked with is Group A Strep (GAS). I am horrified to have to tell you that this great country of ours has got the highest incidence of rheumatic heart disease, in our indigenous people. The cause of that is the Group A Streptococcus, and Michael Good at QIMR has been working for years to assemble a vaccine against the causative bacterium GAS.
(Click on image for a larger version)
Michael asked us to try our lipopeptide approach using his target epitope in a vaccine.
One of the major criticisms of peptide-based vaccines is, ‘You are only going to get antibodies against linear sequences, linear epitopes, in which case if the target epitope possesses conformation then you won’t get meaningful antibodies.’ Michael Good and his team knew that the important part of the M protein from Group A Strep was conformational – it was a helix – so he engineered some extra residues at the flanking ends of the epitope which had a propensity to fold into a helix, and sure enough, assembled a helical peptide very successfully. It’s still quite short, only about 20 residues in length. We then incorporated this sequence into our lipopeptide-based vaccine candidate.
This is all administered intranasally – this is another point I wanted to make about lipopeptides. You can put them up the nose, you can put them in intramuscularly, subcutaneously, almost where you like. The data you are going to see now are obtained following intranasal inoculation.
The data show the average serum IgG titre against the Group A Streptococcal epitope designed by Michael Good and colleagues. You can see that the two lipidated forms of our peptide are very good at inducing antibody and that they are also opsonising antibodies.
(Click on image for a larger version)
Because the lipopeptide-based vaccines are administered intranasally, we hoped that they would induce secretory IgA. And sure enough it was: IgA was present in the saliva, again with the lipopeptide structures only, very good titres, and also fecal IgA – again only the two lipopeptides that we made are able to induce IgA.
(Click on image for a larger version)
Here are some survival curves, showing again that the lipidated forms of the peptide are the only ones which give significant survival after a lethal challenge of live bacteria.
(Click on image for a larger version)
Ian Frazer has introduced the first vaccine against HPV and cervical cancer. Richard Roden from Johns Hopkins University has a different antigen. It is not an L1-based antigen but an L2-based antigen, and he has a promising peptide from within that sequence.
When we incorporated this L2 peptide into our platform lipopeptide we found that the antibodies were reactive with multiple strains of HPV.
What we wanted to determine was to whether those antibodies were biologically effective.
(Click on image for a larger version)
This is a cutaneous challenge model and makes use of the IVIS 200 imaging system. Animals have been inoculated intranasally with the peptide vaccine and then challenged either with HPV-16 or HPV-45. In both cases the lipopeptide-based vaccine protects the animals from infection.
(Click on image for a larger version)
We also made use of an intravaginal challenge system. Mice are inoculated and then they are sensitised with a subcutaneous inoculation of progesterone and are then challenged some time later. The vaginal tracts are harvested and visualised using MAESTRO® fluorescence microscopy.
(Click on image for a larger version)
And again results similar to those in the cutaneous challenge system were found. At the left is a negative control; this is the vaginal tract which has been dissected out and splayed. In the centre is a positive control – no vaccination. And at the right are our lipopeptide-vaccinated animals, and you can see again that there is very good protection against intravaginal challenge with virus.
(Click on image for a larger version)
If you are a real immunologist, of course, you are not really interested in antibodies, you want to know about CTLs. And CTL epitope-based vaccines are important, particularly for viruses, tumours and intracellular bacteria, including TB.
(Click on image for a larger version)
Our favourite virus is influenza, and again I am going to show you some intranasal data. We have also done this subcutaneously and we get similar results.
The peptide vaccine is administered intranasally and after seven, 28 or 90 days these animals are challenged with live virus intranasally. Then after five days lungs are removed and the lung virus titre determined.
(Click on image for a larger version)
On the left-hand side you can see that the lung virus titres in the mice receiving non-lipidated virus are very high, as high as they are in the control. Lipidated vaccine decreases these titres by 99.7 per cent. (This varies between 97.5 and 99.9 per cent in multiple experiments that we have carried out.) With an unrelated lipopeptide, there is no effect.
ELISPOTs for these mice are shown in the next graph, and the chromium release assay is shown below that.
What is particularly interesting is the survival curve in the bottom left-hand panel. Again this is lipopeptide vaccination intranasally, and a challenge with a lethal PR8 strain of virus. We see complete protection. Just to remind you this is a single CTL epitope causing complete protection against a lethal PR8 virus challenge.
(Click on image for a larger version)
What about memory? We can see that three months after inoculation with the lipopeptide we are still getting very good protection from challenge with live virus. Not only do we obtain protection in the short term, we also elicit a long-term memory response.
(Click on image for a larger version)
With Stephen Turner, in Peter Doherty’s lab, we examined what happens at those two times, ie, the acute and the memory phases of the CTL response. We looked in the lung and the spleen at the CD8 cells which were specific for that CTL epitope, and the numbers of CD8 cells you get there are shown here. You can see that there are differences in the numbers of cells: in mice which have received influenza virus you get far more than in those with the lipopeptide vaccine. But remember that these numbers of cells are still sufficient to effect a 99.7 per cent protection against live challenge. So even though the numbers are much smaller than those induced by the live virus, this is still more than sufficient for protection.
But look what happens at day 31 in the memory response. In the lung, the numbers of resident memory cells that you have is the same as you induced with virus, and very high levels are also found in the spleen at day 31.
(Click on image for a larger version)
When we examined the cytokine profiles which were produced by either virus or the lipopeptide, we found quite big differences between virus-induced CD8s and the cytokines induced by the lipopeptide.
(Click on image for a larger version)
We also examined the TCRβ usage and sequence diversity in individual single CD8 cells which we isolated from mice inoculated either with influenza or with PA lipopeptide, and the Vβ usage was similar. The dominant Jβ usage is shown here, as well as the modal and the clonotypes/mouse. There are some differences, but it looks as if the sequence diversity is maintained at the single cell level, whether you use virus or the epitope-based vaccine.
(Click on image for a larger version)
To summarise this work with influenza, we do not need to induce precisely the same kind of immunity as viral infection to be effective and so you don’t have to try to mimic with your vaccine what you get by infection.
We have used conserved epitopes of influenza virus to design this vaccine, I think this is very important. We are used to having to change the influenza virus vaccine every year, just to keep up with the circulating strains of virus, but now we have an opportunity to use conserved CTL epitopes in an influenza vaccine. This avoids having to have prior knowledge of the emerging strain in order to design a vaccine strategy.
Our results suggest that a lipopeptide-based CTL epitope vaccine could be very useful as an adjunct to existing influenza vaccines.
Discussion
Question: A very interesting talk. Your comment about it not always being necessary to stimulate the same type of immune response to a virus to get a protective effect is obviously a very pertinent one, and you can think of many diseases, such as HCV or HSV, where you have got some example of immune pathology occurring as a response to the virus, and where you might need a different type of vaccine response to get protection. You mentioned HSV in your slides of studies that you have done. Can you say anything about the nature of the response you get to HSV antigens?
David Jackson: We have done a small number of experiments, with Andrew Brooks, in which we used a single CD8 epitope and put it into our lipopeptide structure. We were able to show protection, again 99 per cent protection or thereabouts, with that lipopeptide vaccine following HSV challenge. So we have done a small amount of work only, but we have shown protection against HSV. It would be nice to take the candidate gp (glycoprotein) antigen from HSV and lipidate that, to show whether we can produce a protective antibody response.
Question: Have you been able to use any of your constructs for transcutaneous immunisation, and do you know if you can activate Langerhans cells with these peptide constructs?
David Jackson: We haven’t done that. Wouldn’t it be interesting to do it! The thing has got lipid; maybe that transfers.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
