6th United Kingdom-Ireland Controlled Release Society Symposium on:
"Novel Vaccine Formulations and Delivery Systems"

6-7 January, 2000, Trinity College, Dublin

Meeting Report

The discovery of new antigens for old and new diseases is still at the forefront of research in the main vaccine companies and in academia, but this aspect is only part of a complex set of inter-dependent disciplines that comprise vaccinology. Successful immunisation firstly requires interplay between biologists and chemists working in discovery, delivery, formulation and toxicology. Secondly, due largely to economic factors such as the high costs of needles and syringes, cold chain maintenance and labour, the distribution of expensive injected vaccines to the Developing World simply does not occur for many vaccines. Notwithstanding those issues, even in Developed countries when effective injected vaccines become available, patient compliance still remains a serious problem. For example, in adults over the age of 65 or in healthcare workers, the take-up of the influenza vaccine or and hepatitis B vaccines is rarely above 65% respectively. This scenario can be radically improved by replacement of painful and costly injections by new vaccine delivery methods. Non-injected vaccination will lead to better overall compliance, lower overall costs and, if cold chain issues are solved, better vaccine distribution. For any particular non-injected vaccine, the minimum requirement is that the efficacy is similar to the existing injected formulation. Moreover, since oral and nasal vaccine delivery can lead to mucosal immunity, this offers a strong supporting scientific rationale for preventing the attack by 90% of pathogens which invade either via the intestine or via the respiratory tract, the majority for which injected vaccination is ineffective. Many of these issues were discussed at the 6th UK-Ireland Controlled Release Symposium on Novel Vaccine Formulation and Delivery Systems (6-7 January 2000, Dublin, Ireland).



Delivery strategies presented at the conference ranged from oral and nasal to transcutaneous immunisation using combinations of particulate, live-vaccine and vector methods with or without appropriate adjuvants. Specific topics were introduced in a stimulating keynote presentation by Bob Langer (MIT).

Throughout his career, Langer and his collaborators at MIT have been a source of innovative drug delivery research, which has then been advanced by other academics or placed in start-up companies for development. For example, he described his 1979 research on biodegradable poly-DL-lactide-co-glycolide (PLG) microspheres as injectable vaccine delivery vehicles. Bovine serum albumen (BSA) was successfully entrapped in PLG and a single injection to mice produced equivalent antibodies to two injections of alum-adsorbed BSA. Subsequently Langer examined the potential of staggering antigen release from blends of PLG particles in an attempt to mimic booster immunisation injections. In 50:50 lactide: glycolide, in vitro release of antigen was between day 15 and 35, while for the ratio 75:25 the release rate was day 30-52. This work helped to stimulate the renewal of interest in mucosal PLG vaccine delivery by SRI and others in the 90s. While antigens in PLG still have great potential at least by injection, it is only recently that attempts to stabilise and protect antigens from moisture-induced aggregation has become a major research area in this endeavour. For example, it was shown by Langer's group that tetanus toxoid (TT) aggregated and lost essential lysine at only 65% relative humidity and that extreme pHs damage free thiols and disulphides in tetanus toxoid (TT), all features that occur for TT in PLG after 9 days storage. The problem of lysine loss and production of damaging isopeptide formation can be solved in part by succinylation and d-alkylation of TT. He described other work with TT in PLG by Maria Alonso at MIT in which stability of TT at the solvent interface was maintained better by immersing the TT in protective non-ionic surfactant oil in the particle core.

Of the research originating at MIT that Langer viewed as having vaccine delivery potential, he believed that stabilised polymeric liposomes (OrasomesTM) offer promise for oral vaccination. He cited the study by Chen et al (1996) in which radiolabelled particle uptake by the Peyer's patches of the intestine was increased from 3.2% to 10.6%, indirect evidence that targeting particles to M cells should be possible. On the inhalation side, he was excited by the potential of porous spray-dried particles for pulmonary vaccine delivery since they had a density (1g/cc) and size (low µm) that promoted deep lung deposition in a non-aggregated format. As a first step, aerosolised insulin in porous particles made by Advanced Inhalation Inc. was now in the clinic. Other technology originating at MIT with potential for non-injected vaccination as sonophoresis (ultrasound) across the skin. Experiments were described in which large peptide molecules were delivered transdermally at a critical frequency of 1 MHz, the mechanism being induction of structural changes in the skin through generation of gaseous bubbles (cavitation). On the adjuvant side, he felt that the polyphosphazine (PCPP) copolymers now being developed by Avant still held promise. Such tyrosine-containing polymers had already shown themselves to be able to stimulate higher titres than alum-adsorbed antigens upon injection, but their potential for non-injected vaccination was relatively unknown.


Oral vaccine delivery

Oral vaccination using particulate was discussed by David Brayden (Elan, Dublin) and Rob Brey (Endorex, Chicago) in back-to-back presentations representing the Joint Venture, Innovaccines. Results were shown from injected and oral immunisation of antigens in PLG in which the most pertinent result was successful protection against a pertussis aerosol challenge in orally immunised mice. It was emphasised however that oral doses need to be 50-100 times those of injections and that the free pertussis antigens were highly immunogenic. Echoing the work of Langer, the Innovaccines work showed that antigen stability, premature release, poor uptake by Peyer's patch M cells and then incomplete release were the major additional issues for oral over injected immunisation with PLG particles. In injected PLG vaccine studies, studies described by Brayden proved that standard solvent evaporated 1-3 µm diameter particles produced cell mediated Th1 immunity in mice by the intra-peritoneal (i.p.) and intramuscular (i.m.) routes. In marked contrast, smaller 400nm nanoparticles made by a coacervation process demonstrated stronger humoral (Th2) immunity when administered by these injected routes. This work has relevance in that injected particulates may be tailored to suit vaccination against intracellular pathogens or allergens (Th1) or against extracellular pathogens and parasites (Th2).

Immunology questions may well be addressed for oral vaccines in PLG by incorporating stabilisers and adjuvants and by attaching M cell targeting agents. Serious issues remain unresolved such as the prevention of antigen release from PLG in advance of the 3-4 hours required to reach the human M cell, and then the ability to release all the antigen out of the particle within a short time-scale within the Peyer's patch. In Ed Lavelle's work (Rowett Institute, Aberdeen) sponsored by West Technologies (Nottingham), preliminary data suggested that enteric polymeric coatings such as carboxy methyl ethyl cellulose may be sufficient to prevent antigen being degraded and / or detached from polymeric lamellar substrate PLG particles (PLSP) in gastric fluid. Such combination of technologies may assist in keeping antigens associated with the polymer in the intestine, thus increasing the chances of reaching M cells. In a formulation variation on PLG, Lavelle used polylactic acid / polyethylene glycol (PLA/PLG) blends in a successful attempt to get highly loaded particles. These 'hi-load' formulations appear to maintain antigens in a more stable state than in either PLG or PLA alone and, moreover, gave more sustained antibody titres to antigens after sub-cutaneous (s.c.) injection. Finally, they showed that polyvinyl alcohol (PVA) and surfactants used in synthetic process for PLG could also modulate the immune response in addition to their role as excipients. Bringing all these strands together for successful mucosal immunisation will be the next step.

OrasomesTM as initially described by Langer may currently offer a simpler solution at least to the premature release question since these modified liposomes retain their cargo in the presence of bile, detergents and pH extremes. Data was shown by Rob Brey confirming that TT co-entrapped with the adjuvant Monophosphoryl Lipid A in OrasomesTM gave acceptable serum antibody titres in mice after two and three oral immunisations with 50 µg doses of antigen. This is still a high dose by comparison with the 1-3 µg TT required for intra-nasal (i.n.) and injected immunisation in orasomes, but it is still more encouraging than literature for TT formulations in which doses of up to 250µg are regularly cited for bicarbonate-treated mice. Importantly, using confocal microscopy Brey showed that small 150 nm orasomes-TM conjugated with a mouse M cell-specific lectin, UEA-1, demonstrated 15 fold enhancement in binding to mouse M cells in in situ intestinal instillation experiments. Current studies are aimed at translating this increased M cell binding to increased immune responses for TT in the targeted particle.

Particles containing sub-unit antigens are not the only approach being taken in oral vaccination. Annick Mercenier (Pasteur Institute, Lille) described the use of lactic acid bacteria (LAB) as live vectors for oral vaccine delivery in research carried out by an EU 4th Framework Consortium. Lactic acid pro-biotic bacteria are non-pathogenic and non-invasive and live as commensals in the human intestine. With intrinsic adjuvant properties allied to their resistance to gastric acid, the main issues for LAB are the quality of the immune response to co-administered antigens, the stability of the strains and the practicality of the formulation process. With the lactococcus lactis MG1363 strain capable of secreting 50mg/l protein and of presenting antigen on the surface, co-formulated tetanus toxoid fragment C (TTFC) has been produced at a level of 10 µg. In mice this formulation stimulated serum IgG2a antibody against TTFC following sub-cutaneous (s.c.) and i.n. administration. After three oral doses two weeks apart, albeit in the presence of bicarbonate buffering, strong serum IgG1 and IgA were seen, which were accompanied by splenic cell production of IL-2 and IL-6 and by protection against a TT challenge. Current studies are using pneumococcal capsular polysaccharide in LAB and the group has also advanced to vaginal administration of LAB to monkeys. Particular discussion points of this work include comparisons with live attenuated salmonella vectors as regards efficiency of immunisation for co-administered antigens, responses in immuno-compromised people, and the feasibility of repeat dosing of the vector.


Transcutaneous immunisation (TCI)

Greg Glenn of IOMAI discussed the potential of using vaccines on the skin with ADP-ribosylating adjuvants. Reviewing studies of over 50 antigens co-administered with adjuvants to the shaved skin of mice, Glenn provided evidence that the Langerhans cells (LC) in the lower epidermis and dermis were the key mediators of the immune response. Confocal microscopy and FACS had shown that cholera toxin adjuvant (CT) placed on the skin led to physical changes in the LCs, MHC Class 2 up-regulation and detection of the co-stimulatory molecule, B7-2. It seems that just a few LCs in the right cell cycle stage are required to migrate to the draining lymph nodes in the dermis, a site where CD4+ T cells have also been located upon activation by TCI. There is evidence that the technology may have broad potential as cytotoxic T cells (CTL) have been detected in response to peptide antigens from Yersina pestis and from HIV, which were co-administered to mouse skin with CT. Furthermore, serum, lung wash and stool IgA has been detected in mice treated with CT on the skin and this raises the intriguing possibility that mucosal immunity can be stimulated at distant sites. Early published work with TCI showed that CT could illicit prime and boost immune responses to both itself and to co-administered diphtheria toxoid (DT) when 100 µg of each agent placed on the skin of immobilised mice for several hours. The IOMAI group has now shown that doses can be reduced to 10 µg in the case of heat labile enterotoxin from E. coli (LT) co-administered with TT and moreover, that skin exposure time can be reduced to 15 minutes in hydrating conditions. In discussion of the actual range of adjuvants, additional effective agents include LT mutants, CpG adjuvants and purified (but not recombinant) cholera toxin B sub-unit. While ADP-ribosylation is clearly an important the adjuvant effect as shown by the excellent responses to CTB only when contaminated with wild type CT, the success with LT mutants prove that the ribosylation is not obligatory and this has implications in respect of ultimate safety for the adjuvants used in this technology.

Glenn described two successful Phase 1 studies in man with LT in a 2 x 2 gauze patch overlaid with Tegaderm® and a saran wrap. 500 µg of LT applied for 6 hours was not toxic and three doses given 3 weeks apart stimulated serum and stool IgG and IgA in all subjects. Boosting was seen on the second and third immunisations and up to a 23-fold increase in serum anti-LT IgG was detected in some subjects. Current studies are proceeding to properly formulate antigens and adjuvants in prototype patches being made by Elan in the joint venture called Xairo and ongoing work aims to prove the concept that co-administered antigens can produce immune responses when applied to the skin with an appropriate adjuvant in man.


Nasal vaccine delivery with adjuvants

The presentations by Lisbeth Illum (West Technologies, Nottingham), Oya Alpar (Aston University, Birmingham) and Rino Rappuoli (Chiron Biocine, Siena) had a common thread showing that nasal immunisation with an appropriate adjuvant or delivery system can achieve titres in mice similar to alum-adsorbed antigens given i.m. So far however, with the exception of Aviron's live cold-adapted live flu vaccine, data in man for sub-unit vaccines has not shown quite such an encouraging relationship between the routes.

Illum described an ongoing human trial with flu antigens in range of 7.5-15 µg given i.n. with chitosan in three doses, given at least a month apart. The test protocol was either i.n prime followed by i.n. boosts or i.m. prime followed by i.n. boosts. Preliminary data shows that in contrast to their mice data, responses in both groups to i.n. administration were less than for the control i.m prime followed by i.m. boosts. On the positive side, the nasal route was well tolerated by subjects and in both i.n groups the hemagglutinen inhibition (HI) titres were over the 4-fold increase threshold level known to correlate with protection in 18/23 subjects. Chitosan powders appear to be more effective than chitosan solutions in the case of guinea pig nasal immunisation with the diphtheria toxoid mutant, CRM 197, but some of the data on chitosan efficacy in this model was confounded by positive responses to CRM 197 alone in s.c primed animals. On a weight-for-weight basis chitosan is effective nasally in a ratio of 140:1 in respect of 20 µg quantities of CRM 197. In a discussion on the mechanism of chitosan, Illum argued that the rationale is based on a combination of chitosan uptake in particulate format by nasal M cells, the opening of nasal epithelial tight junctions and the stimulation of local cytokine production. Such arguments may well also be the basis in part of the encouraging data with PLG and PLA microspheres as nasal vaccine delivery systems as described by Oya Alpar. Successful nasal immunisation of mice with several antigens either as sub-units or as DNA vaccines was reported for small amounts material adsorbed to PLG and to cationic nanoparticles, the smaller particles being more effective.

In Rino Rappuoli's presentation, the participants were encouraged to look beyond mice for proof-of-concept for nasal immunisation as in his experience they tended to give false positives for adjuvants and particle delivery efficacy, or in his words, "mice lie!" In a review of the development of the LT and CT mutants as non-toxic mucosal adjuvants, he described how the toxicity of the parent toxins had been eliminated in some recombinants by reducing the A1 sub-unit activity or by removing its activity altogether in others. Toxicity data in the Y1 adrenal cell viability assay and in the rabbit fluid secretion models showed clearly that the mutants were far safer than the wild type toxins. Importantly, crucial experiments attempting to show that the Chiron mutants have no demonstrable penetration of the nasal olfactory bulb are being carried out at present. The experiments showing that LTK63, LTR72 and CT106 are effective nasal adjuvants for a range of antigens including acellular pertussis and pneumococcal conjugate antigens were described in mice. Evidence was presented showing that i.n. of LTR72 mutant toxin polarised the immune response towards Th2, while LTK63 polarised the response towards Th1. In an interesting comparison of nasal with oral, it was shown that a 2-log increase in dose to achieve a 2 log less response was typical for oral delivery versus nasal. Consequently, to account for some literature demonstrating oral immunisation effectiveness at low doses, he suggested that it should be viewed sceptically since it was very easy to accidentally contaminate the nasal passageway using gavage methods in mice.


Advanced adjuvants

The saponin QS-21 (Aquila, Cambridge, MA) and the micro-emulsion MF59 (Chiron, Emeryville, CA) were the most advanced of the adjuvants presented at the conference. MF59 is approved in Italy for influenza vaccination by injection. In attempting to follow alum salts as the next approved adjuvant in the US, the emphasis is as much on safety and tolerability as much as efficacy given that the formulations will be given to healthy individuals except in the case of therapeutic vaccination. There is no indication however, that any new adjuvant is about to be approved by the FDA.

Charlotte Read-Kensil (Aquila) described progress in man with the tri-terpene glycoside adjuvant, QS-21. The saponin is already approved in an animal vaccine, Leucogen ®, which prevents feline leukemia virus infection. It has also been given by injection to 2000 people in doses of 25-100 µg in 50 Phase 1 / 2 trials and in one Phase 3 trial for a melanoma vaccine. In these studies the material was filter-sterilised and stored as lyophilised formulations, important features for safety and stability for injections. Transient site pain was reported for i.m injections with QS-21, but pain had been reduced by either changing the antigen or by using excipients such as Tween-80 and cyclodextrins in the formulation. Inclusion of other excipients may, however, lead to further safety requirements by the regulatory authorities. In terms of immunology, QS-21 has shown synergy in injections with other delivery systems and adjuvants such as PLG, MPL ®, alum and IL-12. The agent is very effective by injection at generating CTL production and gives a mixed Th1:Th2 type response. Less work has been done by non-injected routes: i.n. vaccination with flu antigens and QS-21 (5-15 µg) in mice showed a 10 fold increase in protection over solution controls, but the data was not as good as s.c immunisation with free flu antigens. No evidence of toxicity was seen in rabbits exposed i.n. to 100 µg of QS-21, the maximum dose so far used in man. DNA vaccine i.n. studies with QS-21 and HIV gp160 DNA are underway and yet to be reported. Orally, increased titres were seen to TT co-administered with QS-21 in bicarbonate-treated mice dosed weekly for three immunisations, but the dose of TT was extremely high and impractical. The quickest way to approval for QS-21 as an adjuvant in man will be to give it with a therapeutic vaccine for cancer or with antigens with potential for therapeutic vaccination in Alzheimer's disease (Aquila Press Release, January 18th, 2000).


Facing up to delivery challenges

A realistic appraisal of the benchmark for non-injected vaccination was provided by Rino Rappuoli, (Figure 1). Since Chiron were the only actual vaccine company participating, his comments were particularly illuminating. On an arbitrary log scale of efficacy in man, he argued that the target benchmark for efficacy, 5, was that of responses achieved with alum or MF-59 emulsion adjuvants given by injection. Currently, in his view nasal vaccination in man was at a level of +1, just over the minimum requirement of zero, but still 4 logs lower than for the real requirement. All other routes and technologies were below zero and therefore huge log increases in efficacy were required for all the non-injected routes with delivery systems and / or adjuvants. Yet Rappuoli believed that the targets required for non-injected vaccination in man could be achieved by the efforts of some companies and academics now working in the field.

In summary, the conference provided a useful snapshot of some of the key technologies that offer hope for the realisation of non-injected vaccine delivery. What is certain is that five years ago a conference entirely devoted to this theme would not have been a viable proposition and the high attendance at the meeting reflected this increasing interest. Despite the significant hurdles that must be overcome, researchers in non-injected vaccine delivery can be reasonably optimistic. Unlike the case of oral peptide delivery in which demonstrable delivery of set quantities to the post-hepatic blood circulation must be achieved for a sustained period, modest antigen delivery to a few key antigen presenting cells may adequately trigger the self-perpetuating systemic and mucosal immune systems.

Figure 1. Estimate of benchmark challenge for successful non-injected vaccine delivery in man. On a log scale, +5 is the estimated efficacy for injected adjuvants, alum salts and the microemulsion, MF59. Zero is the minimum required for efficacy, but +5 is what is actually required. Nasal vaccination is the most promising alternative route, but is still 4 logs lower than required. The hurdles for all other routes are slightly higher. Modified originally from Rino Rappuoli (unpublished).


David J. Brayden

Copyright©2000 UKCRS

ukcrs - Home Page

PharmWeb® - Copyright©1994-2003. All rights reserved