I have deliberately left Nathalie’s name on this slide, because this is really Nathalie’s work over almost 17 years now, spearheading the research and development at GlaxoSmithKline Biologicals, particularly heading up the adjuvant system research and development.
I also have to preface this talk by saying that within GSK Biologicals we have amassed so much data that it was very hard to come up with a succinct presentation within a half-hour period, but I hope that what I can achieve today is to outline to you the strategy and the approach that GSK Biologicals has taken to rational vaccine design, in terms of utilising modern adjuvant systems.
To go back into history: it was way back in 1925 that Ramon first recognised that a variety of substances – and these were ordinary household substances, like corn starch, agar and tapioca – could increase antigen-specific antibody production when added to diphtheria or tetanus toxoids. And it was in 1926, a year later, that Glenny first identified aluminium salts as potential adjuvants. Since then, aluminium salts have featured for a large part in terms of adjuvants within the vaccines that we have used over the last 80 years.
As Sir Gus Nossal pointed out in his address, we would acknowledge that today vaccines represent one of the safest and most cost-effective medical advances that have ever been developed.
Linked to the target population]
However, we acknowledge that challenges remain, and that we require new strategies for the development of efficacious vaccines, in part linked to the specific requirements in certain target populations – for example, where long-term persistence of protection may be required, requiring higher levels of immune response or improved immune memory; and also in order to adapt to poorly responsive populations which we recognise here – for example, the very young, the very old, the immunocompromised and the chronically ill. And also we recognise, with better understanding of disease pathogenesis, that we need better targeting of the effector responses, whether it be directing them more towards humoral or cellular responses, Th1 and CD8.
Linked to the target pathogen]
There are also challenges facing vaccine development in terms of the target pathogens. We know of very complex pathogens that have developed subtle mechanisms to try to evade or subvert the immune defences, such as hepatitis C virus, HIV and human papillomavirus. But there is a good news story, obviously, on HPV.
There are also very complex, multistage parasites, like malaria, which have thwarted development of effective vaccines for many years now, and also potentially weak immunogens, such as the very purified recombinant proteins or peptides or polysaccharides – and then we face events such as pandemics, where pathogens do not give us enough time to respond by developing effective vaccines.
I will outline the GSK Biologicals adjuvant system approach.
What we have been trying to do over the last 17 years is to use adjuvant systems to overcome the limitations of classical adjuvants. They can be insufficient when complex immune mechanisms beyond just antibodies may be required for protection; or when protection needs to be extended beyond the antigen present in a vaccine, for example to deliver cross-protection; or when infection is silent or very poorly recognised by the immune system and does not allow for, for example, natural boosting; or, as we said before, in the presence of weakened immunity.
So we see the need for the right adjuvant, or adjuvant system, with the right antigen to protect against disease in the right target population, balancing at all times the risk-benefit profile ultimately to deliver a vaccine to the market.
In terms of design principle it looks simple: combine the antigen with the right adjuvant, the antigen directing the specificity of the immune response. But in the use of adjuvant systems , what we hope to achieve is to combine different adjuvants so that we can tailor the immune response to achieve the desired immune protection.
Here are some examples of the adjuvant systems that have been in development within GSK Biologicals. We have used combinations of the classical adjuvants listed here: aluminium salts but predominantly aluminium sulphate, emulsions and liposomes, together with immunostimulants such as monophosporyl lipid A, QS21, CpG and tocopherol.
MPL is a non-toxic derivative of LPS (lipopolysaccharide), which we know is a TLR4 agonist. QS21 we know directs immune responses towards the innate immunity. CpG is a TLR9 agonist. And tocopherol is also used in our adjuvant systems.
I will give you some examples from the bench to the clinic.
This is certainly non-exhaustive. I would add that today the GSK Biologicals pipeline of vaccines in development actually has a whole host of different adjuvant systems in deploy.
By way of explaining the nomenclature: ‘AS’ stands for ‘adjuvant system’ and the numbers are simply a way for us to classify internally which combination of the classical and immunostimulant we have been using. You can imagine that over 17 years the number of permutations and combinations would be quite a lot larger; however, over the last few years we have managed to hone down to a much narrower number of ASs to deploy in the pipeline.
The first of these examples, AS02, has been used in the malaria candidate vaccine development. It comprises MPL and QS21 as the immunostimulants, with the classical adjuvant being a proprietary oil and water emulsion. AS03 has been deployed in malaria and prepandemic influenza. Here the immunostimulant is tocopherol, and again we have a classical adjuvant, a proprietary oil and water emulsion. AS04 has been deployed in HPV, hepatitis B and HSV candidate (vaccine) development. Here the immunostimulant is MPL, with the classical adjuvant aluminium hydroxide.
I will run through for you some examples of what we have managed to achieve with some specific programs.
When we try to achieve additional benefits beyond the antigen content of a vaccine, I think the AS03 adjuvanted prepandemic influenza candidate is one way to exemplify this.
We all know that the medical need is certainly great – there would be a global health crisis in the event of a pandemic – and that the global population would be largely naïve toward any declared pandemic strain. Also, there are supply limitations if you just use a non-adjuvanted vaccine, in terms of using pure HA, and you require a two-dose vaccine regimen to achieve sufficient immunity.
We know that non-adjuvanted inactivated purified H5N1 vaccines are poorly immunogenic, even if they have a very high HA content, and that the aluminium salts – aluminium hydroxide, for example – provide some modest enhancement of the immune response.
How to achieve protection: a lot of influenza vaccine manufacturers are approaching this in a similar format. Prepandemic vaccination we believe offers the potential for earlier protection. However, in order to achieve that you would require the prepandemic vaccine to be able to elicit immunity against drifted strains, that is cross-protection, and to spare the required doses of antigen in order to meet global demand.
Proof of concept: human study]
Let me just take you through a proof of concept in terms of a human study involving our prepandemic influenza vaccine development.
In terms of vaccine development we looked at, obviously, the option to reduce antigen content. So in the chart at the top left of this slide you will see the blue groups representing unadjuvanted vaccines across a range of different doses, and in the brown-yellow tones the adjuvanted system vaccines, again across a number of different dose ranges.
In regard to all three CHMP (European Union Committee for Medicinal Products for Human Use) criteria, whether it be seroconversion rate, HI titre or seroprotection rate, after the second dose we met all the criteria. And specifically the criteria were met with the lowest dose of antigen being used – 3.8 micrograms.
Proof of concept: human study] (second slide)
Further studies have also shown that the 3.8 microgram HA H5N1 adjuvanted vaccine can indeed demonstrate cross-reactive humoral and cellular immunity against a substantially drifted clade 2 strain. (I haven’t got the data here to show you, but that was demonstrated using neutralisation assays which WHO utilises to diagnose avian flu infection.) It also induces a high level of functional antibodies against the clade 1 vaccine strain.
So here we see an example of an adjuvant system allowing for dose sparing and also for cross-reactivity.
In terms of safety, over 8000 subjects to date have received an AS03 adjuvanted vaccine – there have been more than 6000 subjects from 18 to 60 years of age and also we have a paediatric safety database, already comprising 2200 children. They are ongoing studies. Obviously, it is very important to be able to demonstrate the safety of the adjuvants in use as well, especially in the context of a vaccine that might need to be used in the entire population.
Just by way of example: in a recent study that was conducted across multiple sites in Asia, involving over 1000 subjects, we examined the safety and reactogenicity in an 18- to 60-year-old study population. We do see higher rates of solicited local and general symptoms, including grade 3 symptoms. Specifically, these are local injection pain, redness and swelling, and the general symptoms include arthralgia, myalgia and headaches.
There were a number of serious adverse events (SAEs) identified. However, none were deemed causally related to the vaccine.
These data have been borne out as well in a more recent study involving over 3000 subjects across Europe, with a very similar picture. What we do see in the population over 60, though, is systematically similar differences between adjuvanted and non-adjuvanted vaccine; however, these adverse event reporting rates are actually lower in the older population as compared with the younger age groups.
Overall, the reactogenicity profile of the candidate vaccine was acceptable, and the safety profile of the candidate vaccine was favourable.
Moving on to malaria, I think you all know the dire situation faced in Africa in terms of the malaria morbidity and mortality.
How to achieve protection]
Malaria is a particularly challenging pathogen in terms of how to achieve protection. The parasite itself exists in so many different forms and stages. And we know that we don’t just require antibodies to block infection, we also need T cell responses to destroy intra-hepatic parasites effectively.
In terms of the composition of the GSK malaria vaccine candidate, RTS,S: the antigen comprises a recombinant fusion protein which is derived from the circumsporozite protein. ‘RTS’ stands for ‘repeat T epitopes’, and actually the S antigen is from hepatitis B. So these are recombinant hep-B S antigens.
The adjuvant system in deployment is AS02, as previously highlighted; the immunostimulants here are MPL and QS21, in conjunction with the proprietary oil and water emulsion.
Proof of concept in human]
It has been a long haul. I think development of the malaria vaccine program first started in 1987. The adjuvant system development program in GSK actually started only around 1990, however. The first proof of concept in a challenge model was done in ’96. So in the scheme of things from bench to the clinic it is actually not that long, in terms of from when adjuvant systems were first being researched internally.
This first proof of concept study was conducted at the Walter Reed Institute, on healthy volunteers. Different groups of adjuvanted RTS,S vaccines were examined. You will see here that we examined three different formulations, involving three different adjuvant systems, and came up with quite variable results in terms of the vaccine efficacy rate, protection from infection and also delay in infection. It was clear that the AS02 formulation actually delivered the greatest results.
Proof of concept in human]
In trying to explore further why this might be the case, this is where we have got some very interesting results. On the left-hand side here, looking at the humoral immune responses we see very similar humoral immunity, in terms of total IgG, with both the AS03 and AS02 adjuvanted vaccine (but much lower immunity with the AS04 adjuvanted vaccine in this instance). This correlated also with the protection that was being seen.
However, we can try to tease out why we were seeing better responses in the AS02 study vaccine versus the AS04 vaccine. We were seeing clear differences in terms of the cell-mediated immune responses, as demonstrated by interferon-gamma ELISPOT. There was clearly a trend towards greater protective efficacy in subjects who had been vaccinated with AS02 adjuvanted vaccine, who were also demonstrating higher levels of interferon-gamma.
So the cellular response in combination with the humoral response is associated with protective efficacy.
As to where we are today, in terms of the development: obviously, the aim is to target the EPI (Expanded Program on Immunization) population, so it is in the very young, and further studies have been conducted in phase II development to identify the optimum regime and to set up effectively the infrastructure for a large-scale phase III study.
It is recognised that at this stage the objective would be to demonstrate protective efficacy against severe malaria. We acknowledge that there is partial efficacy against the pre-erythrocytic phase of malaria, but this in effect, in terms of the phase II clinical data, has shown that we see strong protection against severe disease and against fatal malaria.
To relate this to the rotavirus experience: certainly rotavirus vaccines don’t necessarily protect against infection by rotavirus but do prevent severe manifestations of rotavirus disease.
As to the safety profile: these data are an example of the safety in children and in infants. Almost 2000 subjects have been administered the RTS,S/AS02 formulation, and there has been no difference in the frequency of SAEs as compared with subjects administered the control vaccines, with the exception of, obviously, a lower incidence of malaria.
There was one SAE in this particular study, a simple febrile convulsion.
In summary, the RTS,S candidate vaccines had an acceptable safety and reactogenicity profile in subjects aged 6 weeks to 11 years, and we are about to embark on a phase III program across multiple centres in Africa that will, hopefully, bring this program even further towards reality.
Finally, I want to touch on an example where adjuvant systems have helped us overcome a complex pathogen where infection may be silent or poorly recognised by the immune system. The example here is the HPV vaccine Cervarix R which, as you know, has been registered in Australia.
How to achieve protection]
I think the need for HPV vaccines, in terms of prevention of cervical cancer, is clear and well known. However, HPV is a tricky virus. It has developed mechanisms to evade the immune system, it is a purely intraepithelial infection that also downregulates local immunity, and it is present in an environment where the mucosal immunity is limited and where the bulk of antibodies present derive from cervical transudation or exudation. We know that it remains poorly recognised by the immune system, in terms of humoral immunity. It doesn’t appear to provide natural boosters to the immune response as measured by systemic antibody titres.
Also, the mechanism of HPV protection needs to be strong and sustained over time in order to prevent virus entry at the site of infection.
In terms of vaccine efficacy, we believe that sustained neutralising antibodies would offer the best chance of success for a prophylactic vaccine. We know these are not therapeutic vaccines.
High and Sustained Antibody Levels and Seropositivity]
Data from our phase II study – actually, with an extended follow-up phase up to 5.5 years – has shown that with total antibodies against both HPV-16 and HPV-18, after a three-dose regimen we see that antibody levels at month 7 (after the third dose) start very high, and over 5.5 years remain very high, with seropositivity levels almost 100 per cent and with total antibody levels sustained at over 11-fold higher than natural infection levels.
We particularly note that there is no waning of immunity across both HPV-16 and -18 total antibodies.
That story is replicated when we look at the neutralising antibody responses out to 5.5 years: again very high seropositivity levels and well above natural infection.
We are especially encouraged by the fact that high serum antibody levels correlate particularly well with high cervical mucosal antibody levels, which we understand to be a very important protective mechanism at the local site of infection, preventing viral entry into the cells.
Translating this to efficacy data out to 6.4 years: we have seen sustained protection against HPV-16 or -18 incident, persistent infections and precancerous lesions.
As to clinical safety: a very large safety database already exists, in a very broad age range, with data up to 5.5 years post-vaccination. It summarises that Cervarix R is generally well tolerated across all age groups. We have comparable rates of unsolicited adverse events, serious adverse events and autoimmune diseases in the vaccine and control groups, a comparable safety profile in women with HPV exposure prior to vaccination, as compared with women with no previous exposure, and similar overall rates of pregnancy outcomes in vaccine and control groups.
Looking at the initial phase I/phase II study in terms of solicited adverse events with Cervarix R, we see no statistically significant differences between the general symptoms with Cervarix R and the control, which is just aluminium hydroxide.
And in the broader phase III study, involving over 18,000 subjects, where the control was hepatitis A vaccine based on the registered GSK vaccine, again we see no significant differences between groups in either the unsolicited adverse events or the serious adverse events.
In summary, in terms of strategies for the development of efficacious vaccines for infectious diseases, we are still facing challenges when complex immune mechanisms are required for protection or when additional benefits beyond the antigen content of a vaccine are required, or when infection could be silent or poorly recognised by the immune system. It is our approach with the GSK adjuvant systems that has enabled us to overcome the limitations of classical adjuvants to try and tackle these specific needs.
I have not given you any examples here, but in our pipeline we are also developing antigen-specific cancer immunotherapies which again deploy the adjuvant systems, in order to also better effect the desired immune responses.
We would really like to acknowledge all those involved in the last 16 years across GSK Biologicals in R&D, Manufacturing, Regulatory and Clinical, and also our collaborators. I am very proud to say that in Australia we have collaborated with many institutions here on HSV vaccines and HPV vaccines, and also now with our antigen-specific cancer immunotherapies. In all these vaccine programs involving adjuvant systems we have had clinical studies here in Australia as well. So, many thanks to our collaborators around the world.
Graham Mitchell (Chair): Thank you very much, Su-Peing. We now have time for some discussion.
First I must say that my reaction about this whole field of adjuvants and the study of adjuvanticity is that there were a lot of carcases in the landscape, as it were, in the old days, over the last two decades, yet listening to that talk you would have to be very positive about the dissection of the immunology, the rational approach, and now there are clinical data – it is just very encouraging. As you say, it takes a long while, but congratulations. From Eugene Maraskovsky’s talk on the scene at ISCOM, it takes a long while and is really starting to bear fruit, right now.
Question: I was hoping to see, when you showed your influenza H5N1, that you had achieved a single-dose vaccine. Are we now settling for a two-dose influenza vaccine, or is there still work ongoing to try to get that down to a single dose?
Su-Peing Ng: I think if you look at even non-pandemic influenza, just seasonal influenza, it is well recognised that anyone who has been naïve to any influenza vaccine in the past should actually receive a two-dose regimen to achieve protection. So certainly in children we already advocate two doses if it is their first exposure to influenza vaccination.
The particular circumstance is that in the event of a pandemic it would be a naïve strain to the vast majority of the global population, hence the requirement for two doses in order to achieve immune protection.
Question: I am curious: with AS02 and AS03, are you seeing any differences with regard to the reactogenicity profile based on gender, or is there a difference between children and adults?
Su-Peing Ng: Within AS03, as I previously highlighted, in the H5N1 development programs we do see, across the board, that subjects over the age of 60 have lower adverse events as compared with those between the ages of 18 and 60. However, if you then compare it within the adjuvanted group versus the non-adjuvanted group, they are still systematically higher than the non-adjuvanted group. I don’t think we have seen any systematic differences between genders. We will certainly be exploring these in greater detail as well.
Question: When I was a very young immunologist I was taught that the immunogenicity of a vaccine adjuvant and its reactogenicity were pretty much linked. Do you think that, with the newer adjuvants that you have been developing, there is a possibility of breaking that link between local reactogenicity and systemic immunogenicity? Obviously, local reactogenicity is one of the things that discourage people from coming back for their second and third shots.
Su-Peing Ng: I know that in the long road towards developing some of these adjuvants and adjuvant systems, certainly better formulation of some of the adjuvants has improved their tolerability and reactogenicity. However, I think at the end of the day we acknowledge that, in terms of a rational design approach, there will be ‘horses for courses’. For the right target disease, the right target population you may accept a different balance in risk-benefit, and the objective is really to get in place the right combination between the adjuvant and the antigen – and the disease, and the target population.
Graham Mitchell (Chair): So Polly Matzinger might be partly right, in terms of the immune system really getting used in the danger signals.
Su-Peing Ng: That’s right.
Question: In Gus Nossal’s talk he alluded to the development of needle phobia as we get a lot more vaccines, particularly for the young. Does GSK have any part of its adjuvant development program dedicated to adjuvants for non-injectable delivery?
Su-Peing Ng: Certainly we have quite a lot of preclinical and early clinical development in which we are exploring non-injectable delivery formats. But at this stage, in terms of late-stage clinical development, all the programs that we have are injectables.
Question: Just on a similar line: this is obviously a series of iterations on the basic theme and your adjuvant study. Would you like to speculate on what the next additive will be, either of an immunostimulatory nature or as a vehicle? What is out there at the moment, swimming round, so that you are saying, ‘Gee, I’d like to get that one into my mix’?
Su-Peing Ng: I think it would be fair to say that we are constantly horizon-scanning, and we would be open to looking at any of the ones out on the horizon. What we have to balance, though, is that the regulatory hurdles for licensure of new vaccines are significant, and therefore safety is one of our greatest issues in the consideration of selection of any adjuvant. That means that a great deal of preclinical work already needs to be done for quite a number of these adjuvants, otherwise it really ends up being another decade before we get to the next point.