This talk is going to be focusing on the use of ISCOMATRIX® in both prophylactic and therapeutic vaccines.
(Click on image for a larger version)
I want to first couch this in the context of the challenges that we all face in the development, in particular, of therapeutic vaccines.
The most successful vaccines have been those that have been used in the prophylactic setting. They have had a significant public health impact, and they have tended to be based either on whole or on part components of organisms. But in more recent times there have also been some successful prophylactic vaccines based on recombinant protein strategies.
That has left the potential for the development of vaccines for the therapy of infectious disease and cancer, but to date there has been no successful therapeutic vaccine that has been a player in the commercial sense.
There have been many adjuvants evaluated as part of the development of a therapeutic vaccine, but very few that have been approved for human use. So there is an unmet medical need for the development of an adjuvant system that is compatible in humans and can be used in a successful therapeutic vaccine setting.
(Click on image for a larger version)
At CSL there was an extensive evaluation of possible adjuvants about 15 years ago, and they identified the ISCOM® technology as one of the major platforms that would be worth evaluating for development in a vaccine sense. They licensed the technology, and they developed an ISCOMATRIX® adjuvant platform which is actually different from ISCOM’s. It has basically been developed to an industrial stage of manufacture. It has regulatory acceptance now; there are several clinical trial experiences demonstrating that it is safe and quite immunogenic in humans, and we have been working on understanding the mechanisms by which it works, to further understand how it can be used in a therapeutic context.
(Click on image for a larger version)
What is ISCOMATRIX®? It is basically a saponin-based adjuvant that is composed of phospholipid, cholesterol and certain fractions of saponin. It is capable of inducing very profound antibody and also cellular responses, in particular both CD4 and CD8 T cell responses. And it forms 40-nanometre particles which are, interestingly, very similar to the size of a virus.
In some ways we think of it as a dumb virus to the immune response. We think that the particle size is important, but some of the immunomodulatory properties of what is composed within the ISCOMATRIX® cage-like structure are also important to the way that it functions.
(Click on image for a larger version)
The key features are, firstly, that it has broad immune responses. It is safe and well tolerated in humans, we have clinical experience with a range of antigens now, and there are also non-clinical safety and toxicology packages that we are developing, together with our partners. It has regulatory acceptance, it is well defined and characterised, and there is a strong IP portfolio there, in terms of going forward into the clinic.
(Click on image for a larger version)
The benefits are that it has both antigen delivery and immunomodulatory properties; as I mentioned, it has that specific particle size; there has been substantial improved tolerability to the saponin component as a result of understanding which fractions are immunogenic and removing the fractions that are more reactogenic; and it is able to be formulated with a wide range of recombinant antigens.
(Click on image for a larger version)
If we look at other adjuvant systems that are also being evaluated in the clinic, we see that the one thing that ISCOMATRIX® has is that it combines both the immunostimulatory and the delivery components. The other adjuvant systems have either one or the other, but tend not to have both in a very dominant form. We think that that is quite an important reason for some of the immune responses we see in patients.
(Click on image for a larger version)
Some of these antigen delivery aspects include the ability to release antigens into specific compartments within the antigen presenting cells, such as the dendritic cell (DC), which allows presentation into the class I pathway for generation of cytotoxic T lymphocytes. But also what we are finding is that there are all sorts of cascades of cytokine profiles induced in vivo which are important in the recruitment and activation of certain types of immune effector cells, and again which facilitate the induction of adaptive immunity.
(Click on image for a larger version)
We view it as a system that integrates both the innate and the adaptive immune system by harnessing DCs. I will present data that shows that ISCOMATRIX® adjuvant delivers antigen to the DCs and enhances class I presentation via an exogenous route, which is termed cross-presentation. There is prolonged presentation in the draining lymph node, which we find to be advantageous in identifying the right T cell clones for expansion to the vaccine. And these immunomodulatory properties seem to be important in then conditioning the immune response in an appropriate way.
 |  |
| (Click on images for larger versions) | |
We have adopted work in the animal model to understand how ISCOMATRIX® works, and we do this by basically injecting ISCOMATRIX® with a model antigen – in this case ovalbumin. We then look in the draining nodes or the spleen and we measure antigen presentation from the DCs ex vivo by presenting to transgenic T cells specific for the antigen.
What we find is that after one injection of OVA-ISCOMATRIX® all of the antigen presentation is going on in the lymph node; there is very little going on in the spleen. If we boost the response, there is a lot of activity in the spleen. But in this early phase, as you would predict, a lot of activation and presentation is going on in the draining node.
(Click on image for a larger version)
We examined the DCs in this site, and the first thing to note is that the numbers of DCs in the node draining the ISCOMATRIX® injection are increased in comparison with the contralateral node.
Then when we look at those DCs and ask whether they are presenting peptides from the vaccine itself, we find that there is a dramatic presentation within the first 12 hours after injection. This is dramatic and surprising in terms of how quickly we can get peptide on the surface of the DC.
In comparison to OVA alone, it is about a 100-fold difference in the number of DCs that are engaged and presenting, but I think the most beneficial thing is that this is prolonged. For up to about three days we get continuous presentation of the DCs that have received the vaccine in the node.
So which DCs are doing this?
(Click on image for a larger version)
As you may or may not know, there are different subsets of DCs, both in the draining node and also in the spleen. The resident DCs can be identified based on the expression of the CD8α chain, but there are also some CD8- DCs in the lymph node.
And then there are what we call the migratory DCs, which traffic into the node from the injection site. These can be either Langerhans cells or dermal DCs, and you can distinguish these based on certain phenotypic markers.
(Click on image for a larger version)
When we actually do that and have a look at who is doing what at the various time points, what we find is that in the early stages, where we are seeing this big peak of antigen presentation by the DCs in the node (as shown in the graph at the top right of this slide), it is all happening by means of the resident DCs within the node, suggesting that the vaccine is getting directly to the node, as opposed to being trafficked there by other DCs.
So this distinguishes ISCOMATRIX® from being a depo adjuvant to actually being a delivery type of adjuvant. If anything, you could look at it as an intracellular depo of high concentration of antigen into the APC, as opposed to at the injection site.
The interesting thing is that in the later stages, where we are getting this prolonged presentation, we are now getting an increased predominance of the migratory DCs facilitating most of the antigen presentation. So there are actually two waves of activation and antigen presentation going on. The first is within the node, and the second is cells trafficking to the node from the injection site, and that takes at least 24 hours to occur.
The net outcome is this prolonged presentation of antigen to T cells.
(Click on image for a larger version)
Just to summarise: the ISCOMATRIX® adjuvant allows efficient cross-presentation not just of the resident DCs, which are very specialised at cross-presenting – in other words, taking an exogenous antigen and translocating it into the class I pathway, which they do very specifically – but also of the migratory ones. These don’t do it very well but ISCOMATRIX® adjuvant enables them to do it, so you are harnessing a much larger pool of APCs for the generation of T cell responses.
It also enhances class I cross-presentation up to 100-fold, and prolongs this presentation for up to three days. I will present some data showing that this is recapitulated in the human system as well.
(Click on image for a larger version)
Now, what is the role played by DCs? Which DCs do it? And is there a need for T helper responses?
(Click on image for a larger version)
The next thing to notice here is that if we look at the various DCs in the node – the CD8 positives and also the plasmacytoid DCs, which are the interferon-alpha producing cells in response to viral infections – after ISCOMATRIX® injection, what we find is that they are actually activated in response to various cell surface markers. So they upregulate class II MHC, they upregulate CD86, and the plasmacytoid DCs also upregulate the early activation marker CD69. So we think this is probably in response to various cytokines that are induced by ISCOMATRIX® and subsequently cause maturation of the DCs.
(Click on image for a larger version)
The interesting thing is that this is almost as good as, if not better than, what LPS does. If you inject LPS into mice and look at DC activation, you find that ISCOMATRIX® can do at least as well as LPS in terms of activation of DCs.
(Click on image for a larger version)
We then looked at T cell response readout. What we did was to immunise mice in a prime-boost strategy, day 0, and then seven days later we gave them the boost; seven days after that we then looked in the spleen, stimulated the T cells just for four hours in an ex vivo readout assay, and looked at CD8 versus interferon-gamma.
What you see here is basically the antigen-specific T cells in the top readout, and the percentage that we get there is a sort of measure of T cell responses.
(Click on image for a larger version)
We asked the question: are DCs necessary in this response? In order to answer this, we used the dip. tox. chimeric mice. Essentially, what you do is lethally irradiate a mouse and reconstitute it with bone marrow from mice that are transgenic for the diphtheria toxin receptor under the CD11c promoter. So basically all the reconstituted DCs are susceptible to diphtheria toxin when you put it into the water of the animals, and once these mice settle down you can ablate the DCs and actually look at what consequence that has when you immunise them with ISCOMATRIX®.
Shown here, we have an untreated mouse and all the DCs, and after treatment with diphtheria toxin you pretty much remove all the DCs in the site.
The interesting thing to point out is that the host Langerhans cells are resistant to the lethal irradiation and they are still present. So they are still capable of participating in the response.
(Click on image for a larger version)
When we do this, we find that by depleting the DCs but actually keeping intact the Langerhans cells, the B cells and macrophages, we still completely ablate the ability of ISCOMATRIX® to generate T cell immunity, suggesting that DCs are critical and the other APC populations don’t play as prominent a role and are not able to rescue the response.
(Click on image for a larger version)
In terms of T help requirements, another surprising thing for us was that if you immunise class II knockout mice which are deficient in CD4 T cells, you get normal CD8 T cell responses in response to the ISCOMATRIX® vaccine.
However, if you then look at these mice 28 days later and try to re-boost them – in other words, ask the question of what happens to the memory response – there is a need for CD4 help in the maintenance of memory, even with an ISCOMATRIX® generated immune response, but there isn’t a need for CD4 helper responses in the generation of the effector cells during the early stages of the vaccination.
(Click on image for a larger version)
This is also seen in CD40 knockout mice, where you get normal CD8 responses in terms of the effector phase. If you look at the memory phase, you see that there is a requirement for CD4 help there.
(Click on image for a larger version)
Antibody responses generated with ISCOMATRIX® adjuvant are completely dependent on CD4 help. There is a defect in class II knockouts in terms of antibody responses to an ISCOMATRIX® vaccine.
(Click on image for a larger version)
Let me summarise. ISCOMATRIX® vaccines require DCs for CTL induction, but Langerhans cells, B cells and macrophages are less required and are not able to compensate. They are potent activators of DCs in vivo. And the need for CD4 help in an ISCOMATRIX®-generated T cell response really is reliant on the maintenance of memory but is not necessary for the generation of the effectors in the early stage. We are now exploring the link between the innate immune response and the adaptive. I won’t talk too much about that today except to say that there is a link between the innate and the adaptive immune response in the way that ISCOMATRIX® works.
(Click on image for a larger version)
ISCOMATRIX® vaccines enhance presentation of antigen and enable cross-presentation of even non-specialised DCs, and this is prolonged; and they integrate the innate and the adaptive immunity.
(Click on image for a larger version)
So how does this translate into humans?
(Click on image for a larger version)
I will explain the key points that we found from the human study.
ISCOMATRIX® adjuvant does access several DC populations in the human system. There is rapid antigen delivery into the cytosol that allows access to the class I pathway, just as we found in the mouse system. Processing can occur in a proteasome-independent fashion, which is actually a surprise to us. It also occurs in a proteasome-dependent fashion, but I will point that out a little bit more clearly. It generates tumour-relevant epitopes – in other words, the T cells that are generated will see the peptide that is expressed on a tumour cell. And it results in efficient and prolonged presentation of CD8 T cells in the human system, just as we saw in the mouse.
(Click on image for a larger version)
Just to give everyone a little bit of an overview of class I cross-presentation: basically exogenous antigen is getting into the class I pathway. Usually exogenous antigen gets presented in class II MHC to CD4 T helper cells, but DCs have this unique capacity to take an exogenous antigen, release it into the cytosol, which is where viruses would normally be uncoding themselves and getting into the class I pathway for presentation, and the protein is then available for proteasomal cleavage, as well as other peptidases and protease complexes to be shuttled into the endoplasmic reticulum through the TAP transporter, to load onto empty class I and get to the surface for presentation to CD8s.
(Click on image for a larger version)
We can try and track down where the checkpoints are, by using various chemical inhibitors. Some of them are very specific for certain proteases and others are a bit more pleiotrophic in where they block, but it gives us a bit of an understanding.
(Click on image for a larger version)
We did the work using NY-ESO-1, which is a cancer/testes antigen that was being evaluated at the Ludwig Institute as part of an ISCOMATRIX® cancer vaccine. It is expressed frequently in tumours and also in germ cells in the testes. It is highly immunogenic, spontaneously eliciting antibody and T cell responses in patients with NY-ESO-1-positive tumours. There are quite well-described epitopes that are restricted to both class I and class II MHC. And there are human CD4 and CD8 T cell lines that were generated at the Ludwig and could be used as readout for presentation of peptides from NY-ESO-1 ISCOMATRIX® vaccines.
(Click on image for a larger version)
When we did that, we found that if you look at monocyte-derived DCs, if you load them with peptides, the sort of response that you get is shown on this slide, in terms of gamma production from an HLA-A2 specific CTL line. Protein-loaded DCs are very inefficient in allowing class I presentation to occur. Cross-presentation is, in other words, very inefficient if you just use protein alone.
What was interesting was that ISCOMATRIX®-formulated NY-ESO-1 was very efficiently presented onto class I. The other thing to point out is that an immune complex form of NY-ESO-1 was also quite efficient, and it is probably the more physiological way that cross-presentation might naturally occur in vivo, in terms of an antibody immune complex. But one can view ISCOMATRIX® as a sort of a chemical trick to get cross-presentation to occur, and it happens quite efficiently.
(Click on image for a larger version)
The other revelation was that if you do a pulse-chase experiment of human DCs – in other words, pulse them with peptide, wash them, put them back into culture and then harvest them over time and have a look at how much peptide is still on their surface, as a measure of what their potential for T cell stimulation is – you find that peptide-pulsed DCs rapidly lose peptide, and their ability to stimulate T cells and interferon-gamma bursts is rapidly lost. So by 48 hours the DC has no peptide that is recognisable by the T cell to cause it to make interferon-gamma.
If you pulse and chase a DC with protein alone, cross-presentation is very inefficient. It takes about 24 hours and only a minority of the T cells can see anything on the surface of the DC.
With an ISCOMATRIX®-formulated antigen, within four hours you are getting quite profound peptide-MHC complexes on the surface that the CD8 T cell can see, and this is prolonged for up to three days. So if you consider the outer point in the vertical red box here, you see there is a huge differential between formulating the antigen with an ISCOMATRIX® adjuvant and, say, peptide-loading approaches. And what we see in the horizontal red box is very reminiscent of what we found in the mouse system, in terms of a prolonged presentation in vivo.
(Click on image for a larger version)
To look at what else is required: uptake of an ISCOMATRIX® vaccine seems to be an active process. So it is not the saponin punching holes in the outer membrane and getting into the cell; it requires active actin polymerisation and phagocytosis, and you can actually block the ability of ISCOMATRIX® – and even immune complex – by blocking actin polymerisation.
(Click on image for a larger version)
Looking within the endosomal compartment, we know that acidification is important for how ISCOMATRIX® can generate these class I epitopes.
(Click on image for a larger version)
If you block proton pump inhibitors, you completely shut down the ability of ISCOMATRIX® to generate class I epitopes.
(Click on image for a larger version)
Interestingly, other lysosomal enzymes and calpain inhibitors seem to play some role, although not complete, so you get about a 50 per cent reduction in the ability to generate the class I epitopes.
(Click on image for a larger version)
The surprising thing for us was that, unlike the immune complex, which is very sensitive to the proteasome inhibitor in the way that it can generate class I epitopes, ISCOMATRIX®-formulated antigen doesn’t appear to require the proteasome to generate at least the A2 epitope.
(Click on image for a larger version)
We have evidence that with other epitopes within the same protein, the proteasome is important, so I think it defines the rules by which these protease complexes identify the correct amino acid sequence to start cleaving.
(Click on image for a larger version)
I think it is fair to say that ISCOMATRIX® enables both the proteasome-dependent and independent protease complexes to access the protein.
(Click on image for a larger version)
Now, we were trying to work out what this proteasome-independent complex might be, and it emerged in the literature around the same time that tripeptidyl peptidase II was important in generating class I epitopes independently of the proteasome, and in some cases in tandem with the proteasome.
(Click on image for a larger version)
So we brought in some of the inhibitors for this – AAF-CMK and butabindide oxalate – and we found that these two inhibitors don’t affect immune complex generated class I epitope presentation, which we know is proteasome-dependent, but they do block the ability of ISCOMATRIX® to generate class I epitopes.
So it looks as if TPPII, at least, is relevant for the way the A2 epitope is generated.
(Click on image for a larger version)
The story at the moment is that, in most pathways, ISCOMATRIX® and an immune complex formulation seem to parallel each other till they get to the cytosol, and for some epitopes a proteasome-independent pathway is what ISCOMATRIX® is enabling, to handle the protein. Everything else seems to be the same.
(Click on image for a larger version)
Cross-presentation is protein formulation dependent; ISCOMATRIX® can certainly allow cross-presentation to occur. This is enhanced and prolonged for up to three days, both in the mouse and in the human system, as far as we can tell. And the ISCOMATRIX®-mediated cross-presentation is an active process that requires acidification of the endosomal compartments and, in some cases, may be proteasome-independent.
(Click on image for a larger version)
In terms of human DCs that can cross-present, the mono-DCs (monocyte-derived DCs) can, and the CD1c circulating blood DCs can. Plasmacytoid DCs, B cells and macrophages don’t cross-present to CD8 T cells, even with ISCOMATRIX® adjuvant.
(Click on image for a larger version)
How does this happen in terms of uptake and tracking?
(Click on image for a larger version)
The way we approached this was to use OVA that was tagged with Alexa488, a non-quenchable fluorescent tag, and followed OVA in human monocyte derived DCs. If you use soluble antigen alone, you can see that within the first 10 minutes a lot of the antigen is in very distinct compartments, as shown on this slide, in proximity to the cell surface membrane. These have actually been identified as endosomal compartments, by using two-colour staining.
(Click on image for a larger version)
What is interesting is that ISCOMATRIX® very rapidly translocates this into the cytosol. So not only do you see these distinct compartments but you see a lot of cytosolic blush. This translocation is explaining the rapidity with which we are getting class I epitope generation – within four hours we were almost getting maximal class I epitope-MHC complexes on the surface of the DC.
(Click on image for a larger version)
Just to summarise this (I have not shown all the data, because of time): ISCOMATRIX® adjuvant traffics protein into very early and late endosomal compartments, and translocates the protein into the cytosol; and this is really the prerequisite for getting it onto class I MHC for presentation to CD8 T cell responses.
(Click on image for a larger version)
ISCOMATRIX® adjuvant targets and conditions multiple DC populations in vivo, delivers proteins into the cytosol for cross-presentation, and generates tumour-relevant T cell effectors of broad specificity.
(Click on image for a larger version)
I will show you a little bit about the broad specificity of the T cell responses in the human system.
(Click on image for a larger version)
We have administered ISCOMATRIX® adjuvant now to thousands of subjects, from the healthy to the chronic infectious disease setting to cancer; with a large range of antigens, including viral, recombinant proteins, cancer and infectious disease settings; and also in therapeutic and prophylactic vaccine programs, and up to, I think, nine or 10 doses. In all of those cases it has proven to be safe and well tolerated.
(Click on image for a larger version)
It has also been shown to be quite immunogenic in humans. We have robust neutralising antibody responses, and broad CD4 and CD8 T cell responses that are restricted to multiple class II and class I alleles – which basically covers almost 90 per cent of the patient population that we have looked at – and these T cell responses are long lived.
(Click on image for a larger version)
I will show you the data supporting what I have just said.
(Click on image for a larger version)
In terms of a NY-ESO-1 ISCOMATRIX® vaccine trial, this was a trial that was set up to be monthly injections of a NY-ESO-1 ISCOMATRIX® vaccine in a dose escalation setting and comparing it with protein alone. When you look at antibody responses, essentially what you find is that the highest dose of NY-ESO-1 ISCOMATRIX® generated the most robust antibody titres to NY-ESO-1, and that protein alone at the highest dose, in the absence of adjuvant, was actually a poor inducer of antibody responses.
(Click on image for a larger version)
To summarise the antibody responses: essentially 100 per cent of the individuals that received the vaccine generated quite high antibody titres to the NY-ESO-1 protein, as compared with a quarter of the patients receiving the protein alone at the same protein dose.
(Click on image for a larger version)
We also looked at DTH responses in this trial, and again the highest vaccine dose generated the most profound DTH responses between pre-vaccination and post-study samples. And again protein alone was not particularly dramatic in generation of DTH measurements.
(Click on image for a larger version)
In some of these cases we biopsied the DTH site and found there was clear lymphocytic infiltrate, with the presence of some CD8 T cell responses and also quite a predominant CD4 T cell infiltrate.
(Click on image for a larger version)
Looking at the A2-restricted T cell response, which was the initial measure in this trial, we found that only a third of the patients generated the A2-CD8 T cell response. We were perplexed by this, and disappointed in some ways, in that only a third of the patients were generating T cell responses, but essentially we were just focusing on this one epitope as a surrogate for immunity to the whole vaccine, and this was a recombinant protein.
(Click on image for a larger version)
So the guys at the Ludwig then started looking at a much broader response, using overlapping peptide sets. What we found, essentially, was that this is really the tip of the iceberg. When you started looking at overlapping peptide sets (18-mers and 13-mers and 11-mers overlapping the full-length of NY-ESO-1) in an ex vivo readout assay about half of those patients generated CD8 and CD4 T cell responses. And if you expanded those T cells in vitro to amplify the response, you actually saw that almost 90 per cent of these individuals had T cell responses to the vaccine.
(Click on image for a larger version)
This slide presents an example of some of the responses, highlighting the region for the A2 epitope. Looking at the overlapping peptide sets and interferon-gamma CD8 responses by intracellular cytokine staining, you find that there are hot spots all around the place, distinct from the A2 region.
(Click on image for a larger version)
This is seen with the CD4 responses as well – so not a restricted DR4 response but also responses to other regions of the full-length protein.
(Click on image for a larger version)
Basically, this slide contains a protein map showing that very few responders were showing responses to the A2 region, and the hot spot for CD8 responses was actually in the region between 60 and 110 amino acid sequence.
(Click on image for a larger version)
As to the other thing that was looked at, some of the patients that finished on the first study were then eligible, up to two to three years later, to go on a second ISCOMATRIX® study. Their response to various epitopes in the first study is shown on this slide, and they showed quite broad responses in the first study; 862 days later, before they went on to the second study, they had a pre-bleed sample taken to look at responses to various epitopes that were induced in the first study. And with three out of the four epitopes that we measured, they still had an intact T cell response to that vaccine. (We lost a response to one particular epitope.) We were quite pleased with this. It was seen in about three out of five patients where this was looked at.
(Click on image for a larger version)
To summarise the human studies: we had frequent CD8 T cell responses at the highest vaccine dose, with a broad range of antigen-specific, both CD4 and CD8 T cell responses detected in the majority of (healthy cancer and HIV-infected) subjects that we have looked at, and also against multiple class I and class II alleles.
Essentially, the more you look, the more you find, and that was probably one of the major lessons for the Ludwig group: not to place all your bets on using just one surrogate epitope as the measure when you are using a recombinant protein approach.
(Click on image for a larger version)
In overall conclusion to this talk: the science is trying to underpin the mechanisms by which the adjuvant is working; we have quite a robust series of development activities that are trying to optimise the way we immunise and vaccinate and apply the adjuvant in the clinical setting, and are able to support all the safety and regulatory side of things; and this is identifying that ISCOMATRIX® adjuvant could be quite an integral part of novel vaccines to both prevent and treat human diseases.
(Click on image for a larger version)
None of this work would have been possible without the large team of people involved at several institutes. Nick Wilson, who was a postdoc fellow from the CRC at CSL was instrumental in generating a lot of the mouse data, in collaboration with Gabrielle Belz. Neil Robson, at the Ludwig, and also Max Schnurr, who was at the Ludwig and is now at Munich, generated a lot of the human DC data and evaluation of trafficking and proteasome complex work. Jonathan Cebon provided the clinical trial data. And a very large body of researchers have contributed to this work.
Discussion
Question: Eugene, that was an absolutely beautiful presentation. Certainly for someone like me who had not caught up with the ISCOMATRIX® story for 10 years, this was just marvellous.
The 40-nanometre size seems to keep popping up, in terms of optimal immunogenicity of a formulation. What is so magic about that size – not 10, not 60 – for a 100 micron DC? What is it about 40 nm?
Eugene Maraskovsky: I am not quite sure what the answer is. The link we are thinking of is that evolutionarily speaking it is a virus-like size, and it might be that these cells have evolved to be able to either phagocytose or endocytose. And particle size is one of the components by which an APC senses pathogens, including all the pattern recognition receptors on its surface and intracellularly and thereabouts.
Question (cont.): In terms of your cartoon showing progress from the outside to the inside, there is not one point where we can say, ‘Right, that’s where the 40 nm really counts, versus these other sizes’?
Eugene Maraskovsky: I’m not sure. It is not something we have really been focusing on directly, so I would not be the expert to be answering the question. But there are several other groups that are focusing on the importance of particle size in terms of optimal activation of the DC, and they may have more information on that than I do.
Question: I am wondering about the route of delivery. Can you direct the ISCOMATRIX®, say, to get trapped in the lung or in the liver? I am wondering what happens if you can deliver it to antigen-presenting cells in the lung. Do you actually get enhanced therapies for lung infections? And for the liver, does it get into the Kupfer cells, and can you then actually direct it to liver infections?
Eugene Maraskovsky: We certainly haven’t tried the liver. We have tried to focus it mainly in a parenteral vaccine route – so, intramuscular, subcutaneous. I know in mice it is probably not a good idea to go systemically with ISCOMATRIX®, which can be quite potent. So that hasn’t been an approach we have even evaluated. We have steered away from that and tried to keep to localised administration of the adjuvant.
Perhaps Debbie Drane from CSL would like to talk about the lung delivery.
Debbie Drane: We have done quite a bit with lung delivery in sheep, actually, and it looks quite promising. But that is early work, being done with the University of Melbourne.
Question: In relation to the process of delivery, the antigen is obviously entering the endosome and then it gets out of the endosome into the cytosol. Is that due to the disruption of the endosome by components of the ISCOMATRIX® or is there an antigen process? And once it actually gets into the cytosol, what is the mechanism that underpins the longevity of the presentation? Is it simply the gradual release from the ISCOMATRIX® of protein, or is it actually affecting in some way the underlying presentation processes?
Eugene Maraskovsky: That is a very good question. It’s one we are trying to explore, but it is not so easy to dissect those out. The working hypothesis we have – and that is all I have at the moment, because we don’t have the definitive data – is that it might be at that point that the saponin is having its activity, in terms of punching holes in the endosomal membranes or at least compromising them to a point where you can get antigen access to the cytosol, as opposed to cell surface. (I don’t think that that is where it does its job.)
It depends on soluble antigen versus associated antigen. In a soluble antigen we don’t have good control of whether ISCOMATRIX® is still part of the vaccine. In theory, antigen can go one way and ISCOMATRIX® can go the other way, and there is no need for one to be inhibiting the other. We think that in many cases it might be that when ISCOMATRIX® does get into the cytosolic compartment with the antigen, it has certain effects on some of the players in the cytosol that may skew who accesses the protein and whether certain epitopes are accessible or not.
I think the intracellular depo of antigen is probably one of the reasons why you are getting this prolonged presentation. So, rather than having slow-release antigen at the injection site, in the way that some of the other adjuvants, like alum, would be working, we think we are loading a lot of antigen into the cell in a very efficient way and then it is probably being released in a slow fashion, maybe because the ISCOMATRIX® is preventing it all from being catabolised. So you are getting a continuous feeding of quite a decent amount of antigen that can keep getting sent up to the surface and being presented as a T cell response.
But again we don’t have all the data-based evidence for that. It is just a working hypothesis at the moment.
Question: My question relates to the fact that you didn’t mention in your series of adjuvants, when you were comparing them, GlaxoSmithKline’s family of ASO2 adjuvants, which in point of fact have very close similarities to ISCOMATRIX®. And GlaxoSmithKline, in their preclinical work, anyway, leading up the successful malaria vaccine candidate, were insistent that without the monophosphoryl lipid A it doesn’t work. Now, they had the QS21 as a saponin, similar to yours; they had the cage-like structures, a bit similar to yours; but they had to insert the monophosphoryl lipid A, and you have not. First, is that going to bring you down when you get to some of the tougher clinical situations, like malaria or cancer? And second, if not, why not?
Eugene Maraskovsky: Why we haven’t used monophosphoryl lipid A?
Question (cont.): Why you seem not to require it.
Eugene Maraskovsky: I think we don’t require it because the composition of the phospholipid, cholesterol and saponin not only generates the cage-like structure but, I think, also ameliorates some of the more toxic effects of the saponin. So, in terms of the side effects, we don’t see what you would see with a naked saponin or a QS21, in terms of reactogenicity.
That is not to say that we don’t need other components in harder settings, like trying to deal with solid cancer that is an established disease, or a very smart viral chronic infectious disease. I think philosophically I would be of the approach that this is a very good beginning for the therapeutic targeting of vaccines, but in certain settings it won’t be enough. It has certain components that we want – very strong antigen delivery and immunomodulatory effects – but I think we could all benefit from the addition of other components to this. Essentially, what I am saying is that I think the road to successful therapy of existing cancer will be in a combination setting.
So we are not saying that this is going to cure everything, but it is a very good beginning. We understand the process, and we are very aware of the further challenges of why tumours escape the immune response. It may still not be enough for that setting, but it is a good start.
Question: Good adjuvants, as you would know, lead to upregulation of costimulatory molecules. You showed that ISCOMATRIX® led to upregulation of B7, like LPS. That raises the issue: is it stimulating the DC via Toll-like receptors? And if so, which ones?
Eugene Maraskovsky: We have looked at that. It doesn’t appear to be a Toll-like receptor agonist in the classical sense. We are exploring that a bit further, in terms of whether some of the downstream pathways of the TLRs are engaged, and that is still work in progress at the moment. We suspect in part it is due to cytokines that are induced by ISCOMATRIX® in situ and are secondarily having impact on the DC in terms of maturation. So we are looking at various cytokine knockout mice. For example, TNF seems to be an important player, so if you immunise TNF knockout mice you lose some of the impact on generation of CD8 T cell responses. But it is not the only player.
So, in answer, it is not a classic TLR agonist, but there appear to be secondary effects on DC maturation.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
