The trials and tribulations of peptide vaccines development.
Learning from past errors.
Dante J. Marciani, Sc.D., Ph.D.
Vaccination or immunization against pathogenic agents has been one of the most successful developments to prevent disease in the history of medicine. Indeed, vaccination goes back as early as the XV century, when in China immunization against smallpox was carried out by inhaling the dry scabs from those affected by the disease. Subsequent developments by Jenner, Pasteur, and others, stablished vaccination as an effective method to prevent some infectious diseases. Moreover, as the understanding of immune protection against infectious agents increased, new refinements to vaccine development were used; like the introduction of the sub-unit vaccines, which frequently avoids immune responses leading to allergies and even autoimmunity. Indeed, the developments that allowed the creation of sub-unit vaccines, also benefited basic research, when the use of antibodies against specific antigens started to be used in biochemistry, microbiology, and others areas. An approach that lead to the development of the monoclonal antibodies by Milstein and Köhler, for which they received the Nobel Prize in Medicine and Physiology in 1984. Hence, antibodies have become an important tool in research; i.e. the initial results derived from basic research have frequently encouraged the development of vaccines against infectious agents and cancer. Hence, due to their relevance in basic research, it is not surprising that sub-unit vaccines, composed of an antigen plus an adjuvant, are the most commonly used. In fact, the antigen directs the immune system to induce a specific response against itself, while the adjuvant decides the nature and magnitude of the immune response against that antigen.
In fact, vaccine development until two decades ago was sort of an empirical process, where the adjuvant component was frequently a mixture of ill-defined immune stimulators, like the complete Freund’s adjuvant (CFA), which contains whole bacteria. A situation that changed with the discovery of innate immunity, its receptors and ligands; suddenly, a series on well-defined compounds, typically derived from pathogens, were found to induce an immune response by binding to specific cell receptors called Toll-like receptors, or TLRs. This type of response, i.e. innate immunity, has been well-conserved during evolution, with an additional refinement in vertebrates, where innate immunity triggers a more specific, long-term one, called acquired immunity. But, different TLRs induced different kinds of acquired immune responses. Also, TLRs can interact with each other, a situation that may lead to synergic effects, which result in a better immune response, or in other cases in an antagonistic effect, leading to inhibition of the immune response. A result of the induction of immunity via TLRs, is that the acquired response is a pro-inflammatory Th1/Th17 immunity; immunities that are effective in fighting pathogens, but detrimental when they are directed against self-antigens, leading to a new type of increasing diseases, known as autoimmune conditions.
The increasing popularity of antibodies as protective agents and research tools, has created a frequently chaotic situation, where vaccines or immunization strategies have induced an undesirable response, rather than the beneficial one expected. A consequence of a variety of errors made during the design of those “vaccines,” which may include the immunogen or antigen, as well as the adjuvant, has been the significant number of vaccine failures at the development phase or even at the clinical phase. Mistakes that could have been prevented by a critical evaluation of the vaccine’s different components and a rational approach to their design. Here, we would discuss some of those errors, which have led in many occasions to unexpected and unexplained failures, which frequently have been blamed on the scientific bases supporting a vaccine, rather than the defective design of those vaccines.
The adjuvant
A widespread believe in vaccine development, is the wrong assumption that all adjuvants are created equal; which may have been acceptable more than twenty years ago, before the discovery of innate immunity its receptors and ligands. Indeed, it is common to see papers where an antigen after being found to be immunogenic by using CFA, is then combined with well-defined adjuvants, like QS-21, monophosphoryl lipid A (MPL), and others, expecting to have a similar immune response. A situation that usually results in disappointments and reaching the wrong conclusions, as well as causing delays in the development of critically needed vaccines and frequently abandoning their development. Unfortunately, a frequent explanation for these disappointments, has been to blame the science rather than the poor vaccine design. Indeed, several vaccines that when tested in animals using CFA, induced a strong immune response as determined by in vitro assays, failed in vivo testing. A situation that may be worsened by the fact that in some diseases, the CFA induced immune response may potentiate the disease that it is supposed to prevent and/or treat. A finding that usually led to the wrong conclusion that a vaccine against that disease is impossible; a speculation frequently proven to be wrong, when an adjuvant that induces the right immune response is used to develop an effective vaccine. Hence, while adjuvants like CFA may be practical to elicit the production antibodies against an antigen, it may not be advisable to consider those initial methods as stepping stones to develop a vaccine.
To complicate the issue more, the fact is that there are three types of acquired immunity, i.e. Th1 and Th17, which are pro-inflammatory, and Th2 that is anti-inflammatory. Yet, the most common immune responses are Th1 and Th2, with the strongly inflammatory Th17 being less common, but more damaging, as many autoimmune diseases show. Paradoxically, in many papers it is indicated that the optimal immune response must have both pro-and anti-inflammatory immunities, which conveys the impression that a pro-inflammatory immunity, Th1 or Th17, may exist alone and without the anti-inflammatory Th2 immunity, the latter being responsible for the humoral immune response that induces antibody production. Thus, many articles strive to reach that optimal adjuvant that induces both immunities rather than only the pro-inflammatory one. The fact is that there is no adjuvant or agent that can induce a sole pro-inflammatory immunity, Th1 or Th17, since both of them are always followed by the anti-inflammatory Th2 immunity; a relatively newcomer from the evolutionary point of view. Indeed, the milder Th2 immunity always follows the Th1 and/or Th17 immunities, likely as a repair mechanism to ameliorate the damage in the host caused by the pro-inflammatory immunities. Of interest is that different from Th1 and Th17 immunities, Th2 can exist by itself by inhibiting the pro-inflammatory immunities; a property that offers potential therapeutic uses in autoimmunity.
Hence, as a result of the advances in immunology, derived from the use of genetic as well as biochemistry, medicinal chemistry and other methods to study the immune response, the adjuvant component has evolved from being a “little dirty secret” to a collection of structurally well-defined agonists that by interacting with the appropriate cell-receptors, promote a useful immune response. Interestingly, as indicated before, different immune stimulatory ligands interact with specific receptors to induce an immune response following distinctive pathways; pathways that may be synergistic or antagonistic, depending of the adjuvants. Therefore, a deliverate adjuvants’ selection, can increase and modify the immune response well beyond the limits of that attained with single adjuvants; a strategy that is seldom exploited in vaccine development. Hence, the rational design of effective vaccines, requires a knowledge of the mechanisms of action (MOA) of the different adjuvants, in order to select those that induce an optimal response. Actually, the MOA for several innate immunity ligands, e.g. MPL, CpG oligonucleotides and others ligands, acting on antigen-presenting-cells (APCs), have been elucidated. A more complicated situation is that of the MOA for saponin-based adjuvants like QS-21, where this glycoside acts independently in a coordinate but differential manner, on both T-cells and APCs, to induce a strong pro-inflammatory response. However, while QS-21 is a potent pro-inflammatory adjuvant, a potential drawback is its instability, that results in a change in its immune stimulatory properties, from being a pro-inflammatory to a sole anti-inflammatory adjuvant. A severe problem in vaccines against pathogens or cancer, where a pro-inflammatory immunity is crucial for their efficacy. Saponin-based adjuvants, because of its physicochemical properties and MOA, are sensitive to a vaccine’s formulation; in fact, the addition of non-ionic detergents to QS-21, results in a significant increase in its Th1 immune stimulatory activity. An increase that results in damaging inflammatory effects, like those observed during the AN-1792 vaccine clinical studies against Alzheimer’s disease, which led to terminating this vaccine’s studies. Hence, a rational vaccine design must consider besides the adjuvant and antigen, the nature of the excipients and the possibility that unexpected interactions may lead to unwanted side effects.
Several strategies have been tried to achieve stabilization of QS-21, most of them using formulations in which this glycoside is sequestered, limiting the exposure of its ester bond between its fucosyl residue and fatty acids to the aqueous environment, to prevent this way its hydrolysis. A protection that can be somewhat achieved by incorporating QS-21 into liposomes, this way removing the ester bond from contact with water; while this approach reduces the hydrolytic process, it does not abolish it, and its effectiveness may be affected by excipients, temperature and other factors. Another strategy has been to replace the labile ester bond by a stable covalent bond, an approach that unfortunately failed to recognize the role of the different chemical structures from QS-21. As a consequence, the final products while reminiscent of QS-21, lack some chemical groups essential for immune stimulation. While these new derivatives induce an antibody immune response, it is unknown which type of immunity, pro- or anti-inflammatory, is being induced; a critical requirement in vaccine design. Another alternative to achieve a stable immune stimulatory glycoside, has been to replace the original acyl group with a new lipophilic chain linked by a stable covalent bond, at a different location in the glycoside, but preserving the groups critical for adjuvanticity. This new compound, GPI-0100, induces a pro-inflammatory immunity, albeit at higher doses than QS-21, but it is stable in an aqueous environment, regardless of the temperature, pH and other factors. Moreover, the immune stimulatory properties of GPI-0100, may be enhanced by changing the vaccines’ formulations. Thus, some effective options have been developed to overcome the instability of the adjuvant QS-21, while retaining most of its unique immune stimulatory properties.
An adjuvant commonly used is alum, traditionally presumed to be a sole Th2 adjuvant, is the oldest adjuvant allowed in human vaccines. While its MOA was a mystery for many decades, it has been elucidated; a complex process where uric acid plays the role of an endogenous danger signal, which triggers a complex cascade of events leading to an immune response. Although alum induced response has the characteristics of Th2 immunity, it also elicits some responses associated with pro-inflammatory immunity, e.g. complement activation, induction of IFN-γ production and monocytes activation, all responses that indicate that alum may not be a sole Th2 adjuvant. In fact, that alum has been successfully used for years in infectious disease vaccines where a pro-inflammatory immunity is desirable, points to a potentially not too obvious pro-inflammatory activity. But, alum in contrast to adjuvants like QS-21, does not elicit the production of high affinity or avidity antibodies, characteristic of an effective immune response; which may explain the alum-containing vaccines’ poor performance in the elderly, a classic example being the influenza vaccine. Hence, while safe, alum would be a dubious first choice in view of the new and more effective adjuvants currently available. Hence, it is relevant to identify the type of immune response desired, the need for a long-term memory, production of cytotoxic T-lymphocytes (CTLs), antibody isotype, and other factors, in order to select the best suited adjuvant or combination of adjuvants to fulfill those needs.
The immunogen – T-dependent antigens
The other vaccine’s critical component is the immunogen, i.e. the antigen, which is supposed to induce an immune response directed against itself, this way providing the antigen-specific response required from vaccines. Hence, it is evident that immunogens are very diverse and their number quite large. Broadly speaking, immunogens can be classified into two groups: T-dependent and T-independent antigens, depending on the need for helper T-cells; the first group is composed of protein antigens while the second is made up of non-protein ones, like polysaccharides and nucleic acids, with lipids being a special group. The best-characterized group is the T-dependent made up of proteins; hence, the present article will be focused on protein and peptide antigens, with a subsequent article dealing with T-independent antigens.
Proteins, because of their complex structures have a large variety of antigenic determinants or epitopes, which are protein regions recognized by the immune system, i.e. B- and T-cells. Thus, epitopes may be linear, i.e. peptide amino acid sequences, or conformational, which are composed of various discontinuous amino acid sequences which form a 3-dimensional structure, the same that is recognized by the immune system. Actually, conformational epitopes play a crucial role in the interactions with various cellular receptors; hence, these epitopes are important targets for neutralizing antibodies, which by blocking the interactions of these 3D-epitopes with their receptors, interfere with infectivity and processes initiated by stimulatory compounds like growth factors and cytokines. In contrast, the linear epitopes are usually T-cell epitopes, that induce T-cell mediated immunity with secretion of inflammatory cytokines and CTL production. Yet, the initial T-cell immunity will be always followed by an antibody response against that linear epitope. Therefore, it is evident that conformational epitopes would include both B- and T-cell epitopes, to allow the polypeptide chain’s proper folding, in order to achieve the right conformation. Yet, these epitopes may be also formed by discontinuous amino acid residues from a linear polypeptide chain; like in the case where an epitope is formed by contiguous amino acid residues on the surface of an α-helix. In this case, the contiguous amino acids on the coiled peptide, form a linear epitope, amino acids that are surrounded by a diversity of amino acid residues. Hence, it is apparent that peptides have a significant potential to be used as effective immunogens; but, in general the attempts to develop effective vaccines based on peptides have delivered mostly disappointing results.
A close examination of the principles on which peptide vaccines are founded, shows a series of systematic errors and wrong assumptions that have contributed to those failures. Peptide vaccines, due to their size and lack of T-helper epitopes, elicit a weak immunity with a poor immunological memory; hence, it has been customary to conjugate the relevant peptides to a carrier protein, like serum albumin, tetanus toxoid and KLH, to provide the necessary T-helper epitopes. However, these proteins cause a phenomenon known as “carrier-induced epitope-specific suppression” or CIESS, where the specific anti-peptide antibody response is suppressed, because the carrier protein induces an antibody response that competes with the anti-peptide response. A situation that results in a weak immune response that frequently is different from the expected anti-peptide. Hence, a new trend is the use of well-defined T-helper epitopes, derived from other proteins, or an artificially designed one, like PADRE. Yet, the carrier problem is often exacerbated by the use of certain cross-linkers, to covalently link the peptide to the carrier protein. For instance, one of the most common linkers used are the heterobifunctional ones carrying both maleimidyl and succinimidyl groups, popular because of their commercial availability and their easy reactivity with the thiol group from cysteine, an amino acid residue added at one of the ends of a peptide of interest. However, due to the high immunogenicity of the maleimide group, the immune response is often directed against that chemical group rather than the peptide, a situation that leads to the suppression of the antipeptide immune response and the researcher’s emotional distress. An unusual situation, since this undesirable effect of the maleimide group has been known for over two decades. Yet, these undesirable effects are not limited to maleimide, but they include other groups with aromatic character, like benzyl and some heterocyclic aromatic compounds.
Paradoxically, many immunogenic preparations, including some vaccines undergoing clinical studies, has the maleimide group in their conjugated antigens in combination with carrier proteins like KLH. Constructs that may interfere with the production of an effective immune response against the targeted peptide. Thus, the CIESS phenomenon may explain some of the peptide vaccines’ poor clinical results and why their progress has been so far, limited at best, despite of their promise. Still, these negative results are not limited to T-dependent antigens, i.e. proteins and peptides, but they also extend to T-independent antigens, a topic that will be the subject of a future article.
The excipients
By definition, an excipient must be inert, i.e. it should not alter the properties of the adjuvant or the immunogen; a situation that frequently is not fulfilled in vaccine development, which leads to unsatisfactory results. A typical example would be that of the AN-1792 vaccine to treat Alzheimer’s disease; a vaccine formulated with the potent pro-inflammatory adjuvant QS-21. Although the selection of that adjuvant to treat a neurodegenerative proteopathy was unwise, the patients receiving the initial vaccine formulation did not show any side-effects in Phase I. But, during the Phase II of such clinical studies and using a vaccine formulation that was modified by addition of a non-ionic detergent, many patients had meningoencephalitis, a damaging brain inflammation; a development that lead to termination of the study. Evidently, those designing this vaccine, were unaware that QS-21 was a potent pro-inflammatory adjuvant, and that the formulation used during Phase I was being altered by the addition of a non-ionic detergent. An apparently innocuous change in formulation, which induced a large increase in the pro-inflammatory activity of QS-21; an alteration that would have discovered in Phase I, if the protocols governing clinical studies would have been followed. Regrettably, those damaging effects due to shortcomings in the vaccine formulation, caused significant damage to that vaccine’s development. Yet, there are cases where an excipient, accidentally may contribute in a positive or negative way to a vaccine’s efficacy.
A group of excipients that it has been assumed to be inert and stable, is the one composed of various non-ionic detergents, like the detergent used in the AN-1792 vaccine; compounds that are usually used to prevent aggregation of the vaccines’ antigens. Nonetheless, several infectious disease vaccines are formulated with oil-water emulsions that contain non-ionic detergents, as adjuvants. But, a concern that is usually ignored, is that non-ionic detergents oxidize with production of aldehyde groups, which are immunologically active and that can deliver alternative co-stimulatory signals to T-cells, leading to their activation. Therefore, it is likely that those aldehyde groups by delivering a co-stimulatory signal, may contribute to the efficacy of vaccine requiring a pro-inflammatory immunity; yet, it may cause damaging side-effects in cases where the immune response must be an anti-inflammatory one. While it is apparent that those working in vaccine development are largely unaware of the implications of these oxidation products on vaccines’ performance, scientists working with biologics are well aware of those effects. In fact, many therapeutic proteins, like monoclonal antibodies cytokines and other products, suffer changes induced by the reaction of aldehydes with the proteins’ amino groups.
Hence, during sub-unit vaccine development all of the possible interactions between the different components must be considered, to prevent undesirable and/or unexpected results, which may confuse the subject. But, these comments should also apply to the development of the immunogenic formulas used in basic research; particularly because frequently those initial findings progress to become therapeutic products.
Relevant publications
Akira, S. et al. Toll-like receptors: critical proteins linking innate and acquired immunity. Nat. Immunol. 2001. 2:675-680
Allen JE, and Wynn TA. Evolution of Th2 immunity: A rapid repair response to tissue destructive pathogens. PLoS Pathog. 2011. 7(5): e1002003
Bergmann-Leitner ES and Leitner WW. (2014) Adjuvants in the driver’s seat: how magnitude, type, fine specificity and longevity of immune responses are driven by distinct classes of immune potentiators. Vaccines 2014. 2:252–296.
Boeckler C, et al. Immunogenicity of new heterobifunctional cross-linking reagents used in the conjugation of synthetic peptides to liposomes. J Immunol Methods. 1996. 191:1-10.
del Guercio MF, et al. Potent immunogenic short linear peptide constructs composed of B cell epitopes and PAN DR T helper epitopes (PADRE) for antibody responses in vivo. Vaccine 1997. 15, 441–448.
Genito CJ, et al. Liposomes containing monophosphoryl lipidA and QS-21 serve as an effective adjuvant for soluble circumsporozoite protein malaria vaccine FMP013. Vaccine 2017. 35:3865-3874
Ghosh M, et al. Carrier protein influences immunodominance of a known epitope: implication in peptide vaccine design. Vaccine, 2013. 31:4682-4688
Jegerlehner A, et al. Carrier induced epitopic suppression of antibody responses induced by virus-like particles is a dynamic phenomenon caused by carrier-specific antibodies. Vaccine. 2010. 28:5503-5512.
Kooijman S, et al. Novel identified aluminum hydroxide-induced pathways prove monocyte activation and pro-inflammatory preparedness. J Proteomics. 2018. 175:144-155
Kornbluth RS and Stone GW. Immunostimulatory combinations: designing the next generation of vaccine adjuvants. J Leukoc Biol. 2006. 80:1084-1102.
Li W, et al. Peptide vaccine: Progress and challenges. Vaccines, 2014. 2:515-536
Maggio ET. Polysorbates, peroxides, protein aggregation, and immunogenicity – a growing concern. J Excipients and Food Chem. 2012. 3:45-53
Marciani DJ, et al. Fractionation, structural studies, and immunological characterization of the semi-synthetic Quillaja saponins derivative GPI-0100. Vaccine. 2003. 21:3961-3971
Marciani DJ. Vaccine adjuvants: role and mechanisms of action in vaccine immunogenicity. Drug Discov. Today. 2003. 8:934–945.
Marciani DJ. A retrospective analysis of the Alzheimer’s disease vaccine progress – The critical need for new development strategies. J. Neurochem. 2016. 137:687-700.
Marciani, DJ. Elucidating the mechanisms of action of saponin derived adjuvants. Trends Pharmacol Sci. 2018. 39:573-585
Rhodes J. et al. Therapeutic potentiation of the immune system by costimulatory Schiff-base-forming drugs. Nature. 1995. 377:71-75.
Skwarczynski M and Toth I. Peptide-based synthetic vaccines. Chem Sci. 2016. 7:842-854
Yamaguchi Y. et al. Guidance for peptide vaccines for the treatment of cancer. Cancer Sci. 2014. 105:924-931