, Berma M. Kinsey3, 4, Reetakshi Arora4, 5, Muthu Ramakrishnan3, 4 and Thomas R. Kosten6, 7
(1)
Departments of Medicine, Pathology and Immunology, and Molecular Virology and Microbiology, Center for Translational Research in Inflammatory Diseases, Baylor College of Medicine, Houston, TX, USA
(2)
Allergy, Immunology, and Rheumatology Service, Veterans Affairs Medical Center, Bldg. 109, Rm. 234, 2002 Holcombe Blvd., Houston, TX 77030, USA
(3)
Department of Medicine, Baylor College of Medicine, 2002 Holcombe, Houston, TX, USA
(4)
Research Service, Veterans Administration Medical Center, 2002 Holcombe, Houston, TX, USA
(5)
Department of Psychiatry, Baylor College of Medicine, 2002 Holcombe, Houston, TX, USA
(6)
Psychiatry Service, Veterans Administration Medical Center, 2002 Holcombe, Houston, TX, USA
(7)
Departments of Medicine, Psychiatry & Neuroscience, Baylor College of Medicine, 2002 Holcombe, Houston, TX, USA
16.1 Introduction
While all vaccines require the immune system to respond with effector molecules and cells in order to protect the body against a threat, most typically a toxin or microorganism, substance abuse vaccines have a special requirement that the effector molecules (antibodies in this case) be especially abundant (Orson et al. 2008). Although antibodies against tetanus toxin, for example, can be completely protective in the 1–2 μg/mL range (Stevens and Saxon 1979), antibodies against cocaine, methamphetamine, or morphine must be substantially higher. To bind a sufficient fraction of cocaine molecules to block or slow entry into the brain and thus inhibit the basic pharmacological effects of typical recreational doses of the drug, at least 40 μg/mL of anti-cocaine antibodies must be in circulation (Orson et al. 2009). Furthermore, the antibodies need to be of sufficient binding affinity with a rapid on-rate in order to sequester the drug in the bloodstream and reduce its entry rate into the brain (Ramakrishnan et al. 2012). As a result, it is essential to optimize substance abuse vaccine responses, and adjuvants will be a necessary component of any drug vaccine formulation.
16.2 Protein Vaccines Are Generally Combinations of Antigen and Adjuvant
Most individual proteins are perceived by the immune system as relatively bland molecules from the perspective of the immune system (Hawiger et al. 2001), though virtually any foreign or modified native protein may eventually be recognized as such with at least low-level reactions if administered in purified form and in sufficient quantity, e.g., insulins (Witters et al. 1977). It is only with costimulation of additional immune system signaling pathways that the immune system produces robust reactions to these molecules. However, some conjugate molecules already in clinical use have costimulatory properties. In some cases, these properties are very well explored, e.g., the outer membrane protein complex (OMPC) of Neisseria meningitidis. For OMPC, a major part of the costimulation comes from a retained partially hidden lipopolysaccharide (LPS) (Donnelly et al. 1990). Although humans are too sensitive to LPS for it to be used directly, the exposed portions of LPS in OMPC are stimulatory without being toxic and, as a result, have been very useful for pneumococcal polysaccharide and Haemophilus influenzae B polysaccharide vaccines (Heath 1998; Zangwill et al. 2003). More commonly, this costimulation can come through added chemicals, specific inflammatory proteins (e.g., cytokines), or in some cases from components associated with the antigens, e.g., lipids (Bublin et al. 2014) or in a few cases from structures of the protein itself (e.g., flagellin, which will be discussed in more detail later). Furthermore the direction and strength of these immune responses are guided by the types and quantities of the signaling molecules elicited by these co-stimuli. In recent years the multiplicity of signals and the complex interactions of these signals have become more appreciated and then continue to be intensely investigated, as reflected in numerous detailed reviews (Brito and O’Hagan 2014; Buonaguro et al. 2011; Harandi et al. 2009; Wilson-Welder et al. 2009).
16.3 Costimulatory Pathways: TLRs, RIHs, NLPs, Inflammasome, Cytokines, and Chemokines
As the innate immune system has become more thoroughly investigate, the numerous interactions of the signaling pathways in the component cells of this system as well as within the adaptive immune system cells have gradually been uncovered. Toll-like receptors (TLRs) comprise one of the most prominent costimulatory systems for immunity, and these signaling molecules have been under intensive investigation in the past few decades (Iwasaki and Medzhitov 2004; Kawasaki and Kawai 2014; Qian and Cao 2013). Other pathways, including retinoic acid-inducible gene-like helicases (RIHs) (Hall et al. 2011), nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs) (Motta et al. 2015), and inflammasomes (Lamkanfi and Dixit 2014), have also become more clearly elucidated, in addition to the better know cytokine and chemokine immunoregulatory pathways. In every immune response, it is the balance of all these signals that finally determines the resulting immune activities, and thus the fundamental properties of each receptor, its intracellular pathway components, and the cellular and genetic responses to the signaling are all highly variable with genetic differences on both a population and individual basis (Chen et al. 2011; Dhiman et al. 2008; Ovsyannikova et al. 2012; Ovsyannikova et al. 2011).
16.4 Individual Adjuvants in or Near Clinical Use
16.4.1 Alum
The adjuvant that was first established as clinically usable was alum, and it continues to be the adjuvant used for most clinical protein-based vaccines and a few polysaccharide conjugate vaccines (Kumru et al. 2014). Although its actual functions were not understood and misattributed to a simple “depot” function for many years, recent work has illuminated the complex effects of the variable amorphous suspension of crystalline material that is alum (aluminum hydroxide, aluminum phosphate, a mixture of the two, or in one case an aluminum hydroxyphosphate sulfate) (Hem and Hogenesch 2007). Protein-based antigens are generally adsorbed to the aluminum hydroxide crystals (Johnston et al. 2002) or aluminum phosphate amorphous precipitate structures (Ramesh et al. 2007) before administration. However, under physiological conditions, proteins generally are fairly rapidly desorbed from the alum particles (Hansen et al. 2007), taken up by antigen-presenting cells and transported to appropriate immune tissue locations. In fact, ablation of the injection site on the same day as the injection (eliminating any depot effect contribution) had no negative effect on the magnitude or quality of the immune response in certain mouse experiments (Hutchison et al. 2012).
If there is not a significant depot effect from alum, then what are the mechanisms of adjuvant action for alum? There have been numerous recent investigations into this question in recent years, and it has become clear that there are many contributing factors, rather than a single mechanism. This has been thoroughly reviewed recently (Brito and O’Hagan 2014), but in brief, these effects include inflammatory responses at the injection site with cell infiltration (Calabro et al. 2011; Kool et al. 2008; McKee et al. 2008), cytokine production (McKee et al. 2009; Mosca et al. 2008), and increased antigen uptake by antigen-presenting cells (Gupta 1998) and transport from the injection site to lymph nodes (Calabro et al. 2011). One of the more interesting aspects of these recent studies is the demonstration of the production of IL-1b, via activation of the inflammasome due to the crystalline structure of alum. Since a similar process occurs in the inflammation that occurs from uric acid crystals in clinical gout and calcium pyrophosphate crystals in pseudogout (Martinon et al. 2006), the inflammatory and febrile reactions that can occur with vaccines, especially in children, are now more clearly understood. In any case, it is quite obvious now that there is no single mechanism by which alum potentiates immune responses, but rather a complex series of events that, as a result, can vary considerably between individuals.
16.4.2 MF59
This adjuvant was the first non-alum adjuvant to be approved for human use (1997, in Europe), though only with influenza vaccines. It is a squalene emulsion in water adjuvant that has an extensive safety record thus far (O’Hagan et al. 2013), although as with even long-established vaccines, there have been claims of injury from squalene in vaccines (Asa et al. 2000), later shown to be unsubstantiated (Phillips et al. 2009). The mechanism of action of MF59 has some similarities to alum, in that cell infiltration, cell activation, enhanced antigen uptake, and cytokine and chemokine production are induced, and there is no substantive “depot” function (O’Hagan et al. 2012). Furthermore, there is no association of the antigen with the emulsion droplets, at least in influenza vaccines (O’Hagan et al. 2012).
16.4.3 ASO3 and ASO4
ASO3 is an adjuvant composed of squalene and alpha tocopherol that was approved for use with the flu vaccine for the pandemic H1N1 influenza A virus in 2009 (Morel et al. 2011). Unfortunately, this was found to be associated with a rare incidence of narcolepsy among recipients (Nohynek et al. 2012), unlike any other vaccines of related composition (e.g., MF59). The difference appears to be the alpha tocopherol, which can elicit nonspecific immune activation (Morel et al. 2011), unlike squalene alone. AS04 has a different composition and immunostimulants, using alum-adsorbed monophosphoryl lipid A, a TLR4 agonist (Mata-Haro et al. 2007; Vandepapeliere et al. 2008). This adjuvant was approved for inclusion in a human papillomavirus vaccine formulation that has been shown to be very effective (Garcon et al. 2011).
16.4.4 Adjuvants in Current Development
There are numerous agents in evaluation as potential vaccine adjuvants at early stages of evaluation and development (Brito and O’Hagan 2014). Some of the adjuvants are potentially attractive for substance abuse vaccine conjugates, but others have more promise in other applications. For example, CpG oligonucleotides are especially useful in directing immunity toward a Th1-biased response (Krieg 2000), that will be valuable for eliciting T-cell-mediated responses to cancers (He et al. 2014) and certain parasites (Mullen et al. 2008; Wu et al. 2010), or redirecting responses away from allergy (Beeh et al. 2013). Although this redirection can be very valuable for certain vaccine targets, Th1 responses in general have lower total antibody responses than do Th2 responses and the main problem for anti-substance abuse vaccines at this juncture is the total level of specific IgG antibodies that have sufficient affinity for binding drugs in the proper concentration range (Orson et al. 2014; Ramakrishnan et al. 2012).
One of the new adjuvants that has shown considerable promise for substance abuse vaccines is the TLR5 agonist derived from bacterial flagellin. The intact protein itself is an effective TLR5 stimulant, but antibodies against the intact specific protein can inhibit its function as an adjuvant (Terron-Exposito et al. 2012). In addition, injection of high doses of flagellin administered systemically were reported by some groups to have some toxic effects (Xiao et al. 2014). Recent studies from numerous labs, however, have shown that modifying the protein by truncation of the hypervariable domain, fusion to antigens, or conjugation to drugs can enhance activity and in some cases minimize the induction of neutralizing antibodies and toxicity.
Flagellin was recognized as a potential tool for enhancing responses against antigens other than itself in the previous decade (Honko and Mizel 2005; Mizel and Bates 2010), and numerous experimental vaccines have been developed against various targets, e.g., influenza (Treanor et al. 2010), but also including cocaine (Lockner et al. 2015). The latter studies showed a number of important findings. First, haptenization of the recombinant flagellin (FliC) did not reduce TLR5 activation until hapten densities exceeded ~10 haptens per molecule. Second, doses of conjugated FliC were not toxic (as measured by failure of weight gain in the mice) until the dosage was 50 μg or higher. Third, antibody production was limited in low dosing (10 μg/dose, but at higher doses (50 μg/dose and 100 μg/dose); robust antibody production was observed, with modest enhancement of responses (~40 %) at the highest dose with added alum, compared with alum alone and a KLH conjugate. However, a reduction in antibody production was observed with addition of monophosphoryl A to the 50 μg FliC conjugate dose.
Attachment of antigens and immunogenic epitopes has been thought to be required for adjuvant activity of flagellin (Huleatt et al. 2007); however our lab has shown that when combined with alum, enhancement of antibody production against a conjugate drug vaccine beyond optimized alum alone clearly occurs (Table 16.1). It is likely that delivery to the same or nearby antigen-presenting cells occurs due to co-adsorption of both the conjugate vaccine and the flagellin molecule.
Table 16.1
Alum plus flagellin or Entolimod increases anti-cocaine IgG
Formulation | μg flagellin or Entolimod | %a |
|---|---|---|
Alum | 0 | 100 |
Alum + flagellin | 10 | 149 |
Alum + Entolimod | 1 | 254 |
A more recent advance utilizes entolimod (previously termed CBLB502, developed by Cleveland Biolabs Inc, CBLI (Burdelya et al. 2008), a truncated and pharmacologically optimized derivative form of salmonella flagellin. As an adjuvant for a drug conjugate vaccine, coadministration of Entolimod with alum results in higher responses than alum alone or alum plus intact flagellin (even at tenfold lower dosing, Table 16.1). Entolimod was developed as a treatment for radiation poisoning due to its immunostimulatory function and in fact is the only TLR5 agonist that has been tested in humans, as it is currently in early-stage clinical development as a prospective immunotherapy drug (http://clinicaltrials.gov/show/NCT01527136 ). Adjuvant properties have also been demonstrated in animal models against tumors (Burdelya et al. 2013) and viruses (Hossain et al. 2014).
Due to internal truncation of D2-D3 hypervariable flagellin domains, entolimod (combined with Alum) showed maximal adjuvant activity at low doses (≤1 mg / injection), while inducing undesirable anti-flagellin antibodies only at higher doses (> 3 mg / injection) (Vadim Mett, Cleveland Biolabs, personal communication). As a result, such formulations should be able to be readministered in booster doses to maintain antidrug antibody levels for more prolonged periods of time.
16.4.5 Adjuvant Combinations
Recent experiments from many laboratories have shown that combinations of adjuvants are often much more effective than single adjuvants in stimulating robust immune responses. However, achieving this result is not merely a simple process of adding optimal concentrations of each adjuvant. As the understanding of immunoregulatory signaling has developed, the reasons underlying the challenges of multiple costimulation have become more apparent. The TLR signaling pathways were studied singly and in pairs in an extensive study (Makela et al. 2009), demonstrating that while additive or synergistic stimulation of cytokine production was achievable in some combinations, others resulted in a reduction in cytokine production. Given the number of combinations studied, it is understandable that dose-response curves for each agonist for each combination could not be examined, but this is precisely the problem facing optimization of costimulation. The number of combinations with different dosings is very prohibitive to study in a simple systematic fashion. Even stimulation of TLR4 alone is complex, due to the MyD88 and TRIF pathways being both activated (Shen et al. 2008) for maximal cytokine production, although this may not necessarily elicit maximal antibody, and there are additional recently described inhibitory components in TLR4 function involved as well (Daringer et al. 2015). However, individual TLR4 agonists do not activate both pathways equivalently. For example, MPL dominantly activates the TRIF pathway (Mata-Haro et al. 2007), while E6020 strongly activates both pathways. In any case, it is clear that there are differential effects of these pathways on B-cell differentiation and function (Yanagibashi et al. 2015).
In our laboratory, we have found that combinations of co-stimuli can have significantly different effects depending on the doses and specific combinations that are used. For example, the addition of alum to the TLR4 agonist monophosphoryl lipid A (MPL) markedly reduced antibody production compared to stimulation by either adjuvant alone (Orson, unpublished), as also observed with other conjugate vaccines (Pravetoni et al. 2014), although in other formulations of alum and MPL (ASO4) have shown no inhibition (and no enhancement) of antibody responses with alum (Didierlaurent et al. 2009). On the other hand, addition of alum to other selected agents markedly enhances production of antibody, e.g., the TLR9 agonist CPG (Mullen et al. 2008) and small molecule agonists of TLR7 (Wu et al. 2014). Similarly, a TLR5 agonist (flagellin) markedly increases antibody production at intermediate doses, but combining E6020 and flagellin with alum results in a substantial decrease in antibody production (Table 16.2).
Table 16.2
Combination of alum and flagellin derivative (Entolimod)
Formulation | μg E6020 | μg Entolimod | %a |
|---|---|---|---|
Alum | 0 | 0 | 100 |
Alum + E6020 | 3 | 0 | 259 |
Alum + Entolimod | 0 | 1 | 254 |
Aum + both | 3 | 1 | 209 |
16.5 Humanized Mice as a Tool for Adjuvant Evaluation
It remains unclear whether humanized mice (HM, immunodeficient mice stably engrafted with human hematopoietic stem cells) will provide an avenue to break through the problems with evaluating these complex systems in the mouse or other animal models which often do not accurately reflect the subsequent clinical responses and problems that may develop in human testing or widespread use. As noted above, even with human screening of adjuvants, promising data in animals has all too frequently resulted in disappointing results or unacceptable side effects. Certainly, the HM is not the answer to all such problems, but may provide some guidance by permitting relative comparisons on immune responses from replicates in individual donor HM and measurements of cytokine and other mediator production in the in vivo context.
With any model system for developing vaccines, but in particular with HM (Akkina 2013), it is important that responses be sufficiently robust and consistent in individual animals. With the NOD-Rag1nullIL2rγnull mouse, a strain of relatively long-lived immunodeficient radioresistant nonobese diabetic mice (NOD), efficient engraftment with human hematopoietic stem cells occurs since these mice lack the common IL2rγ chain cytokine receptor component blocking natural killer cell development. There are no murine T cells or B cells due to the Rag1 gene deletion, and since human antigen-presenting cells are required for effective antigen presentation to human lymphocytes, the full function of immune activation for vaccines is present. Our laboratory has shown that features essential for vaccine evaluation are present in the NOD-Rag1nullIL2rγnull mouse (Orson et al. 2014). These include consistent responses to single vaccine formulations for replicate mice of individual donors, distinct differences between groups of mice engrafted with different donor stem cells to the same formulations, and critically significant differences between different formulations in groups of mice all engrafted with stem cells from a single donor.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
