疫苗的安全性评价
来源： 浏览量：300 更新日期：2010年1月11日
Non-clinical safety evaluation of novel vaccines and adjuvants: new products, new strategies
Frank R. Brennan and Gordon Douganb
aHuntingdon Life Sciences, Woolley Road, Alconbury, Huntingdon, Cambridgeshire PE28 4HS, UK
bDepartment of Biological Sciences, Centre for Molecular Microbiology and Infection, Imperial College of Science and Technology, London SW7 2AZ, UK
Abstract
Advances in molecular biology and biotechnology, coupled with an increased understanding of disease processes and mechanisms of protective immunity have facilitated the development of new rationally-designed vaccines utilising recombinant proteins, naked DNA, live vectors, genetically-modified toxins and whole dendritic and tumour cells for both prophylaxis and therapy of a wide range of indications. These new vaccine technologies coupled with novel adjuvants, delivery systems, formulations, dosing routes and regimes present many unique and difficult challenges in demonstrating product safety and efficacy to support clinical testing. This paper aims to review these novel vaccine and adjuvant technologies and to highlight the key safety issues potentially associated with them. Approaches taken to demonstrate vaccine safety by assessing systemic and local toxicity, biodistribution and persistence, immunogenicity and immunotoxicity, reproductive toxicology, safety pharmacology and genotoxicity within the current regulatory framework are presented.
Keywords: Vaccine; Adjuvant; Non-clinical safety testing
1. Introduction
Vaccines are amongst the most widely used pharmaceuticals, providing significant benefits to human health through the eradication of smallpox, the conferring of protection against serious diseases such as polio, measles, whooping cough, diphtheria and hepatitis and through their involvement in the development of new treatments for cancer, malaria, HIV, autoimmunity, allergy and other diseases. Generally vaccines have an excellent safety record, with millions of doses being safely administered to many millions of people. Serious adverse events are comparatively rare. However, there are well-documented cases of vaccine-induced toxicity. For example, vaccination with a formalin-inactivated respiratory syncytial virus (RSV) vaccine prior to infection with RSV resulted in extensive lung pathology and death of some children [1]. More recently a simian-human reassortant rotavirus vaccine was associated with intussusception and mortality of vaccinated infants [2]. These cases, coupled with technological advances that have led to the generation of novel vaccine technologies with novel mechanism of actions, have emphasised the need for a thorough safety evaluation of new vaccines in an attempt to predict adverse events prior to clinical testing.
For many vaccines, such as inactive and protein subunit vaccines, it is the immune response, i.e. the vaccine-specific antibodies and/or T cells that are the likely cause of toxicity rather than the vaccine antigens themselves. For example, toxicity could result from molecular mimicry between a vaccine antigen and an self antigen [3] such that the vaccine-induced effectors target host tissues, a major concern in the development of vaccines for diseases such as Groups A and B Streptococcus (GAS, GBS), Lyme disease and group B meningococcus [4], [5] and [6]. Priming of inappropriate immune responses, e.g. vaccine-specific TH2 cells rather than TH1 cells is believed to explain the toxicity associated with the formalin-inactivated RSV vaccine [7].
The likelihood of many non-living vaccine antigens being directly toxic is relatively small considering the low dose levels and infrequency of dosing used in most vaccination protocols. However, for a number of the new vaccine and adjuvant technologies, there is a greater potential, at least in theory, for the products themselves to be toxic. This is in part because they represent different classes of compounds with novel pharmacological properties or, in the case of live vaccines, are attenuated by novel genetic mechanisms. For example, the insertional mutagenesis of DNA vaccines, the reversion to virulence of live attenuated bacteria and viruses and the pro-inflammatory effects of immunomodulatory adjuvants are potential safety concerns associated with these novel vaccine products (to name but a few). The consequence of general immune stimulation and immune modulation/deviation by vaccines and some adjuvants could indirectly lead to some products inducing/exacerbating hypersensitivity and autoimmune disease reactions. Other components within a formulation such as excipients, preservatives, impurities and contaminants also have the potential to be toxic.
Preclinical safety studies with vaccines therefore aim to identify these possible causes of toxicity prior to undertaking clinical studies. Such studies might be designed to demonstrate the safety and efficacy of the vaccine components and the absence of immunotoxicity, define safe dosing regimes for clinical trials, characterise the immune response and how it relates to any observed toxicity, identify immunological and toxicological safety parameters for clinical monitoring and in some cases, determine the fate of the vaccine and how it relates to toxicity and efficacy.
This paper aims to present an overview of the scientific and regulatory considerations in the preclinical safety testing of a range of novel vaccine technologies and adjuvants prior to undertaking clinical trials. For the purpose of this paper, a vaccine is defined as any medicinal product that elicits or modifies an intended and specific immune response for the prophylaxis, or therapy of a disease or condition. A brief overview of novel vaccine and adjuvant technologies in development are presented below to highlight the diversity of the products being developed and the wide array of theoretical safety concerns associated with them which are subsequently discussed. More detailed reviews can be found elsewhere [8] and [9].
2. Current vaccine technologies
Traditional vaccines comprising live attenuated microorganisms, inactivated whole microorganisms, or split or subunit preparations for prophylaxis of infectious bacterial and viral disease agents represent the vast majority of licensed vaccines. However, progress in molecular biology, biotechnology and biomanufacturing, coupled with an increased understanding of disease processes and mechanisms of protective immunity have allowed the development of rationally-designed vaccines for both prophylaxis and therapy of a wide range of indications. These include infectious disease, chronic viral disease, cancer, allergy, contraception, addiction, obesity, Alzheimer‘s disease and bioterrorism [8] and [9]. These novel vaccine technologies include synthetic peptide and recombinant protein vaccines [10], carrier-coupled conjugate vaccines [11], genetically-attenuated bacterial and viral pathogens [12] and [13], live vectors for heterologous proteins and DNA [14] and [15], naked nucleic acid (DNA) vaccines [16], virus-like particles [17] and [18], genetically detoxified toxins [18] and [19], antigens expressed in transgenic plants [20] and modified dendritic and tumour cell vaccines [21]. Many of these new vaccines are biotechnology-derived, i.e. produced by recombinant DNA technology, however, only a few recombinant vaccines such as hepatitis B surface (HBsAg) and outer surface protein A (OspA) have been licensed for human use. The vast majority of marketed vaccines are for the prevention of infectious disease indications in children and comprise live attenuated viruses and bacteria, purified bacterial or viral components, toxoids and purified or chemically-synthesised carbohydrates produced by conventional (non-recombinant DNA) methods. Indeed, many of the vaccines currently in late stage clinical trials are based on the conventional vaccines (e.g. new combinations of antigens). This present lack of progress of some of these biotechnology-derived vaccines perhaps highlights the complexity of these new products and their manufacturing processes and controls and the fact that they are targeting severe, often incurable diseases such as cancer, chronic viral disease, autoimmune disease and allergy whose aetiology and mechanisms of protective immunity are less well understood. It may also reflect a lack of experience in the regulatory authorities that handle the licensing of these products. However, scientific constraints, such as the poor immunogenicity of recombinant subunit vaccines compared to whole organisms, the over attenuation of live viral and bacterial vaccines, the low of expression of DNA vaccines and the difficulty of therapeutic vaccines to modulate ongoing immune responses in patients clearly play a significant role.
3. Adjuvants and immunomodulators
Adjuvants are being developed to both increase the immunogenicity of inactive non-replicating vaccines, enhance responses in low responder population such as the elderly and the immunosuppressed, improve immunogenicity of vaccines delivered by the mucosal route and to modulate inappropriate immune responses thereby increasing protective immunity (reviewed in [9], [22] and [23]). The use of adjuvant can also reduce the amount of antigen required giving an economic benefit also. Currently only aluminium salts and MF-59 (as part of an inactive flu vaccine) [24] are licensed for human use, although others such as monophosphoryl lipid A (MPL) [25], the saponin derivative QS-21 [26], as well as combinations of these adjuvants [27] are in late stage clinical testing. MPL is licensed in some countries as part of the allergy vaccine Pollinex Quattro [28]. Others such as incomplete Freunds adjuvant, Mycobacterium vaccae, cytokine/growth factors, CpG motifs, syntex adjuvant formulation/2 (SAF/2), muramyl tripeptide (MTP) and liposomes [29], [30], [31], [32], [33] and [34] are in clinical development and many more are at the pre-clinical testing stage (Table 1).
Table 1. Selected adjuvants (including carriers and vehicles) licensed or in preclinical/clinical development
Adjuvant class
Examples
Gel-type
Aluminium hydroxide (alum), aluminium/calcium phosphate
Microbial
Bacterial DNA, BCG, inactivated Mycobacterium vaccae, Bordetella pertussis, MPL, genetically-attenuated CT, LT and subunits; muramyl di- and tripeptide (MDP/MTP) and derivatives, streptococcal cell wall, mycobacterial cell wall skeleton and MPL (Detox); Tetanus toxoid (TT), Diphtheria toxoid (DT) and other bacterial protein carriers of T cell help; live viral and bacterial delivery vectors
Emulsions
Incomplete freunds adjuvant, MF-59, SAF
Particulate
Immune-stimulating complexes (ISCOMs), liposomes, virosomes, PLGA microspheres, chitosan, QS-21
Synthetic
Non-ionic block copolymers, MDP/MTP analogues, polyphosphazene, polynucleotides, CpG motifs
Human protein-based immunomodulators
Interferon-gamma (IFN-γ), Interleukin (IL)-1, IL-2, IL-4, IL-12, IL-15 IL-18; granulocyte-macrophage colony-stimulating factor (GM-CSF), C3d, costimulatory (CD80; CD86) and MHC molecules for incorporation into live vectors and cells
The development of new adjuvants is hampered by issues of local and systemic reactogenicity, lack of surrogate markers of protection, and a reluctance of the regulatory authorities and manufacturers to move away from the use of alum and related salts. Adjuvants act via a number of mechanism including providing a slow or controlled release of antigen, facilitating mucosal delivery by reducing proteolytic digestion and targeting mucosal associated lymphoid tissues (MALT) such as Peyer‘s patches, activating/modulating immune cells such as dendritic cells and lymphocytes, facilitating MHC Class I presentation for optimal cytotoxic T lymphocyte (CTL) responses, facilitating dendritic cell or macrophage uptake or B cell targeting, enhancing long-term memory and providing a general inflammatory stimulus [9], [22] and [23]. These adjuvants can be produced separately and simply mixed with the vaccine or the vaccine antigen. Alternatively, some of the protein and DNA based adjuvants (e.g. cytokines, costimulatory molecules, complement components, CpG motifs) can be genetically incorporated into, e.g. naked DNA vaccines, tumour or dendritic cells or viruses [35], [36] and [37] to enhance immunogenicity or immumodulation. While all these adjuvants show promise, further work is needed to better define the mechanisms of adjuvant action. Comparisons between different adjuvants derived from in vitro studies, or from studies using adjuvants in rodents or other animals, are often only partially predictive of safety, adjuvant effects, or vaccine efficacy in humans [34].
These new vaccine and adjuvant technologies, coupled with the use of novel production and delivery systems/devices, novel routes of delivery, immunisation regimes such as prime boost strategies, therapeutic immunisation/immunomodulation and oral tolerance induction [30], [38], [39], [40], [41], [42], [43] and [44] pose unique regulatory challenges that need to be addressed prior to testing in humans.
4. General strategies to demonstrate safety of novel vaccines and adjuvants
As with all biologicals, non-clinical safety programmes for vaccines are very much product-specific and designed on a case-by-case basis [45]. Non-clinical safety testing programmes must be rationally-designed with a strong scientific understanding of the product, including its method of manufacture, purity, sequence, structure, pharmacological and immunological effects and intended clinical use. In addition, more prescriptive ‘off the-shelf’ testing programmes designed for new chemical entities (NCEs) may be required for novel adjuvants, excipients or preservatives that are either synthetic, non-human derived or are derivatives of compounds known to be intrinsically toxic. Protein-based adjuvants, e.g. cytokines like Interleukin-2 (IL-2) and IL-12, however, would require a more product-specific biologics-type evaluation. Regulatory awareness, coupled with data from in silico, in vitro and in vivo studies can all contribute to the overall safety and efficacy assessment of vaccines as discussed in the following selections.
5. Regulatory guidance for vaccine safety testing
A starting point in the design of a non-clinical safety testing programme is to refer to the numerous guidelines published by the major regulatory bodies such as the European Medicines Evaluation Agency (EMEA) and the US Food and Drug Administration (FDA). A firm understanding of the regulatory requirements for vaccines and related products ensures that the most sensitive and regulatory compliant test systems are utilised to optimise the chances of gaining regulatory approval for clinical testing or marketing authorisation in the shortest time-frame.
The key guidelines relating to the safety testing of vaccines are shown in Table 2. A general guidance for the preclinical safety testing of vaccines has been published by the Committee for Proprietary Medicinal Products (CPMP) at the EMEA: “Notes for Guidance on Preclinical Pharmacological and Toxicological Testing of Vaccines” (CPMP/SWP/465/95) (Table 2). Some of the principles outlined in the ICHS6 guidance document “Preclinical Safety Evaluation of Biotechnology-Derived Pharmaceuticals (CPMP/ICH/302/95)” can also be applied to the safety testing of vaccines, particularly recombinant protein and peptide-based vaccines. Further guidance is available specifically for DNA vaccines and gene transfer-based vaccines. Guidelines pertaining to somatic cell therapies are also useful in designing programmes for whole cell cancer vaccines (based on dendritic cells or tumour cells). Separate guidances are also available for certain products such as influenza virus vaccines and smallpox vaccines. There are also guidelines for the testing of combined vaccines and for the clinical testing of vaccines. The CPMP has established a Vaccine Expert Group (VEG) whose mandate covers a number of vaccine product-related issues including the development of novel adjuvants, and is akin to the Office of Vaccine Research and Review (OVRR) at the FDA. Indeed this group has recently published a draft guidance document for the safety testing of novel adjuvants: “Guideline on Adjuvants in Vaccines” (CPMP/VEG/17/03/04v5/consultation).
Table 2. Guidelines relevant to the safety testing of vaccines and adjuvants
Guideline
Relevant to
Notes for Guidance on Preclinical Pharmacological and Toxicological Testing of Vaccines (CPMP/SWP/465/95).
All vaccines
Guideline on Adjuvants in Vaccines (CPMP/VEG/17/03/2004v5/Consultation)
Adjuvanted vaccines
Guidance for Industry. Considerations for Reproductive Toxicity Studies for Preventative Vaccines for Infectious Disease Indications (CBER, FDA, 2000 (draft)).
Vaccines for pregnant women and women of child-bearing potential
Notes for Guidance on Pharmaceutical and Biological Aspects of Combined Vaccines (CPMP/BWP/477/98).
Combination vaccines
Points to Consider on Plasmid DNA Vaccines for Preventative Infectious Disease Indications (CBER, FDA, 1996).
DNA vaccines
Guidance for Industry. Guidance for Human Somatic Cell Therapy and Gene Therapy (CBER, FDA, 1998).
Viral vector and cell-based vaccines
Note for Guidance on the Quality, Preclinical and Clinical Aspects of Gene Transfer Medicinal Products (CPMP/BWP/3088/99).
Viral vector and DNA vaccines
Points to Consider on Human Somatic Cell Therapy (CPMP/BWP/41450/98).
Cell based vaccines
Points to Consider on the Development of Live Attenuated Influenza Vaccines (CPMP/BWP/2289/01)
Influenza vaccines
Note for Guidance on the Development of Vaccinia Virus Based Vaccines Against Smallpox (CPMP/1100/02)
Smallpox vaccines
ICH Document S6: Preclinical Safety Evaluation of Biotechnology-Derived Pharmaceuticals (CPMP/ICH/302/95).
All vaccines (and other biologics)
Points to Consider in the Production and Testing of New Drugs and Biologicals Produced by Recombinant DNA Technology (CBER, FDA, 1985).
Recombinant protein/peptide vaccines
WHO guidelines on Nonclinical Evaluation of Vaccines (WHO/BS/03.1969)
All vaccines
Much useful information can also be found in the European Public Assessment Reports (EPARs), Summary Basis of Approval (SBA) documents and Summaries of Products Characteristics (SPC) published for licensed products (although admittedly these are for conventional non-biotechnology-derived vaccines) and in other documents relating to these products published by the EMEA and FDA. Pharmacopoeal and WHO monographs available for certain types of vaccines, can also provide useful information. Communication with the regulators throughout the preclinical and clinical development programme is essential to identify and implement appropriate strategies to demonstrate safety and efficacy of a new product.
6. In silico studies
In silico studies (bioinformatics) might be useful in predicting human toxicity of novel vaccines and in designing non-clinical safety studies. For a recombinant protein vaccine, e.g. a proteomics-derived outer surface protein, a first step might be to perform a sequence alignment search of the vaccine DNA and protein sequence with DNA and protein sequences in databases such as SWISSPROT, TrEMBL or GenBank. Programmes such as Compsim, Blastimer, Conservatrix or Patent-Blast could be used to determine if the vaccine antigen (e.g. a bacterial or viral protein) is expressed in humans or has identity or homology to human proteins/epitopes or allergens and therefore may have the potential to induce autoimmunity or allergy [46] and [47]. However, it should be noted that many autoantigens contain conformational B cell epitopes not detected by theses strategies. Identifying MHC classes I and II binding ligands using computer-driven predictive methods (e.g. pattern-matching algorithms such as Epicure and Epimatrix) in conjunction with T cell assays to identify naturally processed putative T cell epitopes from disease targets (immuno-informatics) might assist in predicting the immunogenicity and allergenicity of vaccines during rational vaccine design [48] and [49].
7. In vitro studies
In vitro human and animal cells and cell lines can be used to demonstrate particular aspects of product safety and efficacy. For example, demonstrating the inability of a novel live attenuated Salmonella enterica serovar Typhi (S. Typhi) vaccine to replicate in human macrophages, a reservoir for virulent S. typhi in humans, compared to the wild type strain or the licensed S. typhi Ty21a vaccine (Vivotif?) would be key in supporting safety of the strain [14]. Likewise, demonstrating replication deficiency of live attenuated virus vaccines and vectors can be performed in vitro with human and animal cell lines. The lack of toxicity/activity, e.g. cAMP production by Y1 adrenal cells in response to genetically-modified adjuvants such as those based on cholera toxin (CT) and Escherichia coli heat-labile toxin (LT) is key in supporting attenuation of these toxins [50]. As well as demonstrating safety, in vitro studies are also important in demonstrating pharmacological activity/efficacy of vaccines and adjuvants, which not only confirm potency but might assist in the choice of species for the toxicity studies. For example, the ability of modified LPS-based adjuvants such as MPL or CpG-containing adjuvants to effect human and animal lymphocyte or dendritic cell activation, proliferation and cytokine production can be performed in vitro [51] and [52]. The cellular tropism and replication competence and nuclear localisation of a gene therapy-based viral vaccine, the transformation efficiency and protein expression of a DNA vaccine and the pharmacological effect of an immunomodulatory cytokine (e.g. IL-12, IL-2, interferon-gamma (IFN-γ)) can be investigated in vitro [53], [54] and [55]. These in vitro studies have even greater significance if no relevant animal models exist whereby they become the key data to support product safety.
8. In vivo studies
8.1. Species selection
Prior to performing toxicology studies in animals, one must first determine the relevant species (most sensitive in detecting toxicity). For vaccines, this is based on the immunogenicity of the vaccine, the comparability of the immune response to that predicted in humans, e.g. antibody- or CTL-based or both and, in the case of live vaccines, the susceptibility of the animal species to infection by the wild type organism. For some vaccines, e.g. cancer vaccines, the presence of an animal homologue of the human protein target (e.g. a tumour-associated antigen present on normal cells) might dictate the choice of species since this might allow the assessment of potential autoimmunity. Of course it is always critically important to look for an in vitro test to replace any in vivo assay and all investigators are encouraged to attempt to develop in vitro approaches. However, in vivo testing is a requirement by law for the licensing of all vaccines and in vivo models remain the best method to test a vaccine candidate in a complex living system with all interactions and functions intact.
Safety assessment programmes for vaccines normally use a single species. Rodents and rabbits are the most commonly used species for most types of vaccines, although monkey may be relevant if the vaccine lacks immunogenicity or mounts an inappropriate response in lower species, or in the case of live attenuated vaccines, if the organism is host-restricted. Vaccine-specific immune responses, in particular CTL responses, are usually best characterised in mice and reagents are readily available for immune response monitoring. Inbred mice (e.g. BALB/c or C57BL/6) are often used for the toxicity studies since the CTL response to a vaccine as well as efficacy may already have been characterised in inbred mice. Mice are also the common species used for biodistribution and toxicity studies with gene transfer-based vaccines (e.g. DNA vaccines, viral vectors, etc.). Rabbits are often used for vaccines designed to stimulate antibody responses. Rabbits are a medium sized animal in which usually a whole human dose can be administered easily and for which a large body of background data exists. In some cases the pharmacological activity of the adjuvant, e.g. a human cytokine such as IL-12 will have to be considered when choosing a species also. As mentioned above, cell-based biological activity assays with cytokine adjuvants can also be used to confirm relevant pharmacological activity in a species.
In some cases, there might be no relevant animal model, e.g. a live attenuated vaccine based on a bacteria or virus that does not infect animals (e.g. S. typhi), a short peptide vaccine that binds only human MHC, a human cytokine adjuvant that does not have activity in any animal species or in the case of most cell-based vaccines, the problem of rejection of xenogeneic human cells (e.g. dendritic cells) in animals. In these cases the use of the ‘parallel reagent’ or animal homologue is used. For example, mouse IL-12 would be used in mice in place of human IL-12 and antigen-pulsed or virus-transfected mouse dendritic cells would be used to mimic the response with human dendritic cells. In the case of an S. typhi strain with attenuating gene deletions, which does not infect mice, the use of a mouse-infective S. enterica serovar Typhimurium strain harbouring the same deletions to assess attenuation in mice would be considered [14]. However, since these studies do not use the actual product and the purity and pharmacological activity of the homologue is likely to be different from the clinical product, extrapolation of any finding to man is difficult and so caution should be adopted. Studies in transgenic animals could also be considered. For example, mice transgenic for human MHC and CD4/CD8 molecules might allow the immunogenicity/efficacy testing of HLA-restricted peptide vaccines [56] and [57]. Mice transgenic for a human cell surface receptor might allow safety/efficacy testing of a live attenuated vaccine [58]. However, in many cases, very little background data would exist for these models making it difficult to put any unexpected finding into perspective. If there is no relevant species then a short term (14-day) toxicology study in rodents is advocated in the ICHS6 guideline (Table 2).
8.2. Non-GLP studies to support safety and efficacy of vaccines
Following identification of a vaccine candidate, animal studies might be designed to identify a relevant toxicology species for the formal pivotal Good Laboratory Practice (GLP)- toxicology studies (as described above), characterise the nature and duration of the immune response to a vaccine and demonstrate the requirement for an adjuvant. Such studies might also define the relationship between dose levels and immunogenicity and how this contributes to safety and efficacy in animal models of infectious agent, tumour, allergen or autoantigen challenge in both a prophylactic and therapeutic setting. For live attenuated organisms [14] and [59] and for vaccines that will target specific populations such as the immunosuppressed [60] or the paediatric population [61], toxicity and immunogenicity studies in immunosuppressed, neonatal or pregnant animals may be considered.
Vaccine antigen-specific antibody responses (class and isotypes) can be measured by immunoassay. CTL responses can be measured by a range of in vitro techniques including standard 51Cr-release assays to measure direct CTL lysis of target cells, enzyme-linked immune spot-forming cell assays (ELISPOTs) to detect cytokine release from antigen-activated CTLs or antigen-specific tetramer staining by flow cytometry to quantify the number of antigen-specific CTLs [56] and [62]. Antigen-specific T helper cell responses can be examined in T cell proliferation assays (3H-thymidine incorporation) and by measuring cytokines levels (IFN-γ, IL-2, IL-4, IL-5) in cell supernatants using commercial ELISA kits (if available) to determine T helper subset bias [62].
For mucosal vaccines, vaccine specific sIgA in mucosal secretions and T cell responses in the target mucosa (e.g. the Peyer‘s patches and lamina propria for orally administered vaccines) and at distant mucosal sites should be measured [39].
For vaccines that use a prime-boost dosing regime, e.g. with virus and naked DNA [63] it is important to demonstrate the enhanced immunogenicity of this strategy compared to the use of the single components. For vectored vaccines, the immune response to the vector (virus, bacteria, protein carrier) should be determined and the effect of this immune response on priming and boosting with the vaccine determined since carrier-mediated suppression could affect repeat dosing in the clinic [64], although this does not hold true for all antigens [65]. For combination vaccines, the immune response to each component and any toxicological consequences should be evaluated.
The CPMP vaccine guideline (CPMP/SWP/465/95) states that immunogenicity alone is not sufficient and that protection against challenge should be demonstrated if a relevant model is available. This mainly relates to vaccines for infectious disease/cancer, however, demonstration of efficacy of therapeutic vaccines for other indications such as allergy and autoimmunity would be expected using the available animal models for these diseases [66]. However, in some cases, e.g. with HBsAg vaccines where titres of HBsAg-specific antibody required to provide protection in humans is known then challenge studies are not required.
In vitro studies to demonstrate vaccine–specific effector mechanisms can also be performed to support the in vivo efficacy data. For example, vaccine-specific antibody can be assessed for its ability to augment opsonophagocytosis or complement-mediated lysis of bacteria, virus or toxin neutralisation, tumour killing or inhibit the pharmacological activity of a particular cell-based response. T cells from vaccinated animals can be isolated and vaccine-specific T helper cell and CTL responses characterised in vitro, e.g. for a particular cytokine profile (e.g. quantifying relative levels of IFN-γ and IL-4 to determine T helper cell bias) or for the ability of the CTLs to kill autologous tumour cells or virus-infected cells. Further studies might be designed to confirm the need for an adjuvant and to identify a suitable adjuvant, delivery device and formulation and to determine the optimal route and dosing regime to suit the clinical indication.
Many of these types of studies would be performed in-house, e.g. in the non-GLP environment of a biotechnology or pharmaceutical company or academic laboratory prior to performing GLP safety studies. Laboratory scale material that may or may not be well-characterised or in the proposed clinical formulation might be used. These in-house studies, although giving a gross indication of vaccine safety, often do not include detailed histopathology, haematology, clinical chemistry or urinalysis measurements which are part of regulatory GLP-compliant toxicity studies. This pharmacological and toxicological data however, would still have value and would be included in the regulatory submission to support the formal GLP safety studies. The value of these studies is increased if quality systems are in place to control and monitor the personnel, the experimental work and the technical reports, even if GLP is not claimed.
8.3. Formal GLP compliant toxicology studies
The information generated in in-house safety, immunogenicity and efficacy studies will allow the rational design of formal toxicology studies, which are frequently outsourced to an external Contract Research Organisation (CRO) and performed according to GLP.
The design of toxicity studies (dose levels, dosing regime, route of delivery, duration of study) is largely dependent on the intended clinical regime. For vaccines, the intramuscular route is the most common although the subcutaneous, intradermal, transcutaneous, oral and intranasal routes are also being utilised [38], [39] and [40]. It might not be possible to follow the clinical regime in animal studies due to limitation of dosing volumes, use of delivery devices and the use of non-standard routes of delivery. For vaccines with wide boosting intervals in humans, dosing intervals might be shortened for practical reasons (provided comparable immunogenicity can be demonstrated).
The following sections discuss the components of a vaccine safety testing programme and the reasons for their inclusion or exclusion. The discussion focuses predominantly on studies with the final vaccine formulation and hence for adjuvanted vaccines, the vaccine antigen and adjuvant are tested together. As mentioned previously, for novel chemical-based adjuvants, a classical toxicological programme with the adjuvant alone comprising acute and repeat dose toxicity studies in two species, pharmacokinetic (PK) and tissue distribution studies, e.g. with radiolabelled compound, assessment of genotoxicity using in vitro assays and determining the potential for hypersenstivity, anaphylaxis and pyrogenicity is required as outlined in the recent CPMP guideline on testing of adjuvants (Table 2).
Well-characterised material produced according to current Good Manufacturing Practice (cGMP), representative of the clinical batch and in the proposed clinical formulation should be used as early as possible in the toxicology programme. All pivotal safety studies should be performed in compliance with GLP using qualified/validated assays where possible.
8.3.1. Single (acute) dose toxicity studies
Single dose studies using the clinical route are generally designed to define the relationship between dose levels and immunogenicity and measure only gross effects of toxicity such as mortality, clinical signs, body weight and macroscopic examination. Although recommended in the CPMP guideline, these studies are not always performed for vaccines destined for repeat dosing in the clinic, particularly if there is no concern for toxicity. However, more detailed GLP compliant single dose toxicity studies might be performed for a vaccine if it is for single dose delivery (where the single dose study is the pivotal study), if a novel potentially toxic adjuvant, excipient or preservative or other novel component is used, if the relevant toxicology species is the monkey (where they are required prior to conducting repeat dose studies for ethical reasons), if the vaccine has a known theoretical risk of toxicity, e.g. autoimmunity (following in silico and in vitro studies) or is to be administered via an uncommon route where local and systemic effects might not be predicted. Single dose studies in mice and guinea pigs (abnormal toxicity test) using the intraperitoneal route with the final vialed formulation is a batch release test for vaccines to assess the intrinsic (direct) toxicity however this is considered a part of the quality determination rather than the non-clinical safety evaluation.
8.3.2. Repeat dose studies
Most vaccines are for repeat dosing and hence GLP-compliant repeat dose toxicity studies are usually required to support safety. One full human clinical dose should be tested or, when this is not practical, the highest feasible dose. Scientifically, it would make sense to also test the most immunogenic dose level (if this differs from the human dose) which may be higher or lower than the clinical dose since immune responses to vaccines often do not follow a linear relationship. However, in practice, this is rarely done. Where an adjuvant is used, groups receiving vaccine and adjuvant as well as antigen and/or adjuvant only groups should be included. The use of an antigen only group is useful in detecting adverse finding related to the mixture of adjuvant and antigen. The adjuvant only group is less critical if a full toxicological package already exists for the adjuvant alone. Control groups should receive vehicle. A recovery period should be included to examine reversibility of any potential toxic effects (see below).
Routine analysis may include appearance of injection site (local tolerance), electrocardiograpy (ECG; safety pharmacology assessment), ophthalmoscopy, clinical chemistry, haematology and urinalysis, macroscopic examination and histology of all major organs. Measurement of relevant pharmacodynamic markers (immune responses) should also be included to demonstrate that the product is having the intended pharmacological effect (discussed below).
8.3.3. Pharmacokinetics (PK) and pharmacodynamics (PD)
For biologics such as monoclonal antibodies and recombinant proteins as well as small molecule drugs, knowledge of the PK (serum half-life, maximum serum concentration, area under the curve (AUC; total exposure)) and pharmacological response is important in choosing clinical dose levels and regimes as well as in assessing safety margins and the relevance of extrapolating toxicology data to humans. However, determining the plasma concentrations of antigens is not generally performed for inactive, non-replicating vaccines such as protein/peptide-based vaccines or inactive whole cell vaccines. The plasma concentration of antigens does not correlate with pharmacological activity/immunogenicity and there is often a lack of dose response relationships in immune responses, e.g. high doses can lead to tolerance rather than increased immune responsiveness. However, for certain protein-based adjuvants such as mutant CT and LT administered intranasally, PK and tissue distribution data (e.g. with radiolabelled compound) are important in the safety assessment of these products as they may accumulate in, and direct vaccine antigens to, brain neurons when delivered by this route [67]. For many vaccines, a measurement of the pharmacodynamics (PD), i.e. the immune response/immunopharmacology is used as proof of exposure to a vaccine and is generally used to set dose levels and regimes for non-replicating inactive vaccines which themselves are unlikely to be toxic at the low doses administered. Defining the immunopharmacology of vaccines is discussed later in this review. For vaccines with the potential to be toxic, e.g. live attenuated organisms, modified toxins, etc., the maximal tolerated dose, as well as the immunopharmacology are considered in the setting of dose levels for the clinic.
8.3.4. Biodistribution, persistence and excretion
For live attenuated bacterial and viral vectors (e.g. Salmonella, Shigella, poliovirus, rotavirus, adenovirus, Vaccinia) it is important to examine the presence and persistence of the organisms in blood and tissues as well as excretion of orally administered organisms in the faeces to correlate the presence of the organisms with any observed toxicity. This is usually done by plating of blood, urine, tissue and faecal homogenates onto selective culture media (bacteria) or permissive cells (viral plaque assay) and/or by using other techniques such as PCR and immunofluorescence. Demonstrating rapid clearance from the tissues supports the attenuation of the organism [14] and [68]. Lack of persistent shedding in faeces is desirable for these products [12], not only to demonstrate a lack of replication in the gut but if the vaccine is a Genetically-Modified Organism (GMO), then its release into the environment and survival could lead to entry into the food chain and to exposure to susceptible individuals such as pregnant women, the immunosuppressed, etc.
For viral and bacterial DNA vectored vaccines and naked DNA vaccines, the potential for the DNA to recombine with endogenous host DNA sequences and integrate into the host cell chromosome is a key safety consideration. The likelihood of this is increased if widespread non-target tissue distribution and long-term persistence of the viral vector or DNA transgene is observed. Integration of the DNA into the germ cells (sperm and ova) of the gonads could result in transmission of the gene to the offspring and effects on fertility and reproductive function. Integration could also result in insertional mutagenesis and the activation of endogenous oncogenes or inactivation of tumour suppressor genes leading to tumour formation. This threat is now a reality in humans [69]. Long term-persistence of the expressed viral proteins or antigen/adjuvant transgenes (e.g. IL-12 adjuvant) could lead to inflammation and autoimmunity due to over-stimulation of the immune system or transmission of a vaccine transgene to the young could lead to tolerance/immunotoxicity. Single dose biodistribution/persistence studies are performed using both the intravenous route and proposed clinical route. The intravenous route represents systemic dissemination of the vector, a worst case scenario. Biodistribution of the DNA to a wide range of tissues is examined by PCR. If wide-spread tissue distribution is observed and/or a persistent PCR signal is seen, e.g. in the gonads, then further investigation is necessary. In situ hybridisation would monitor transgene cellular localisation, RT-PCR would determine mRNA levels in tissues and protein expression could be assessed by immunohistochemistry or Western blotting. Whether the persistent DNA is extra-chromosomal or integrated into genomic DNA can be determined using separative electrophoresis techniques to isolate genomic DNA and performing single-sided PCR techniques to test for covalent linkage of the transgene [70]. For whole cell vaccines, such as modified tumour cells or pulsed dendritic cells or lymphocytes, the biodistribution of the cells could be assessed within toxicity studies by staining of peripheral blood or lymphoid organ cells with specific mAbs by flow cytometry or immunohistochemistry [71]. However, since the number of injected cells is relatively low it is usually very difficult to detect the cells distant from the injection site by these methods and to discriminate between injected and endogenous dendritic cells. Separate studies using radiolabelled dendritic cells [72] or dendritic cells transfected with the gene for green flourescent protein (GFP) [71] have been used to assess distribution of injected dendritic cells and how this relates to pharmacological effects.
8.3.5. Immunopharmacology
Vaccines are designed to specifically stimulate humoral and/or cell-mediated immunity. The immune response to these products should be fully characterised, since in many cases it is the immune response to a vaccine or adjuvant (and/or local reactogenicity) that is most likely the cause of any toxicity rather than the product itself. Many of the detailed immunological analyses are performed in the ‘in-house’ non-GLP studies described earlier. However, it is still important to define the primary pharmacodynamic response to the vaccine, i.e. Ab or CTL responses, systemic or mucosal responses within the regulatory GLP compliant toxicity studies to demonstrate that the vaccine is having the expected immunopharmacological effect thereby validating the study.
It is important to demonstrate the reversibility of any immunopharmacological and/or immunotoxicological effects on cessation of treatment with the product. Ideally, the recovery period following treatment with an immunomodulatory product should be long enough for the immunopharmacological effects to have peaked and be on the decline to ensure that any possible immunotoxic effects are detected. However, in reality an arbitrary period of, e.g. 4 weeks is often chosen due to time constraints.
Haematological assessment (relative number of blood cell types), presence of acute phase proteins and histology of the lymphoid tissues (spleen, lymph nodes) may reveal gross effects on the immune system during toxicity studies. Histology and immunohistochemistry studies can detect antibody and complexes bound to tissues, assess the integrity of lymphoid organs (thymus, spleen, lymph nodes) and determine the presence of particular cell types contributing to any observed pathology [73]. Peripheral blood and lymphoid tissue leukocyte subsets can be examined by flow cytometry for effects on relative numbers of T cells (CD4 and CD8), B cells, macrophages and NK cells. Stimulation of the immune system can be assessed by analysing activation markers (e.g. CD25, CD45 isoforms) on leukocytes by flow cytometry or by conventional histology or immunohistochemistry techniques or by measuring levels of cytokines or soluble receptors in sera (reviewed in [74]).
8.3.6. Immunotoxicology
Immunotoxicity of vaccines is rare but when it occurs it is often due to the generation of a vaccine-specific immune response that is inappropriate in either nature, magnitude or specificity. Potential adverse effects related to immune system interaction of vaccines may be detected during repeated dose toxicity studies. These effects can often be investigated/predicted using the in silico, in vitro and in vivo studies discussed earlier as well as using models specifically designed to assess particular aspects of toxicity. Approaches to assess the immunotoxicity of chemical drugs such as measuring antibody responses to keyhole limpet haemocynanin (KLH) and sheep red blood cells [75], measuring T cell using the local lymph node assay (LLNA) [76] or by measuring delayed-type hypersensitivity (DTH) to microbial antigens [77] as well as assays to assess general NK cell function [78] are generally inappropriate for testing the immunotoxicity of vaccines since these assays are designed to assess antigen non-specific immunosuppression. These assays, however may be applicable to assess novel small molecules that are part of the vaccine formulation.
Diseases such as allergy/hypersensitivity (e.g. anaphylactic reactions, vasculitis) and autoimmunity (SLE, arthritis, MS, diabetes) are immune-mediated. Since vaccines directly interact with the immune system, there is the possibility that these diseases could be induced by or exacerbated by vaccination. However, the incidence is likely to be very low and dependent on other factors such as the dose and timing of vaccination, age of the vaccinees, and concurrent infections [66]). Vaccines and adjuvants could influence the regulation and activation of T helper cells subsets, levels of cytokines such as IFN-γ, IL-4 IL-5, IgE levels and immune complex formation which all may play a role in the risk of developing allergy or autoimmunity. Animal models have not be standarized and validated to predict the risk of allergy and autoimmunity associated with vaccines. However, a number of existing models used to assess the toxicity of other products (NCEs, immunotherapeutics) could be refined for vaccine evaluation such as guinea pig models of anaphylaxis, the local lymph node assay, as well as in vivo models of autoimmune disease [66].
A further consideration is whether the immune system is more susceptible to perturbation during its ontological development that may not be seen if immunotoxicity data is only acquired in adults. If this premise holds true (currently there is no strong data to support this), then developmental immunotoxicity testing of vaccines and adjuvants might be important (discussed later).
8.3.7. Local tolerance
Local reactogenicity may be induced as a result of the pharmacologic action of the biopharmaceutical or due to other non-‘active’ components of the formulation (diluents, stabilisers, preservatives, etc.). Visual inspection of the injection site after dosing as well as histology of the injection site or mucosa (e.g. for oral/nasal vaccines) on termination would be performed within the toxicity studies and often this is adequate assessment of local tolerance, provided the proposed clinical formulation is used. However, separate studies in rabbits are sometimes preferred by the regulators for parenteral vaccines. Where local effects are observed, further studies should be performed to examine for the presence/persistence of the vaccine antigen (protein, virus, DNA, cells) or adjuvant at the injection site and draining lymph nodes by immunohistochemistry and its relation to local reactogenicity.
8.3.8. Safety pharmacology
Safety pharmacology (secondary pharmacodynamics) studies classically focus on the cardiovascular system (CVS), respiratory system (RS) and central nervous system (CNS). The CPMP guideline on vaccine testing (Table 2) states that “the potential for undesirable pharmacological activities, e.g. on the circulatory and respiratory systems should be considered for new vaccines”. The likelihood of observing such effects with most types of vaccines is low. Safety pharmacology data can often be obtained within the main toxicity studies, e.g. ECG measurements, body temperature in primate studies or visual assessment of locomotor function and respiration in rodents. In addition, vaccine-specific safety pharmacology (immunopharmacology) measurements should be made as discussed earlier.
If no effects are noted and no pathology is observed in the heart, lung or CNS tissues then separate safety pharmacology studies should not be necessary, although consultation with the regulators is advised. Separate, more extensive safety pharmacology studies might be considered if product-specific issues deem them necessary, e.g. if a vaccine antigen/adjuvant is known to have cardiotoxic effects, e.g. pertussis toxin [79], has been shown to bind or accumulate in the heart, lung or CNS (e.g. mutant CT in neurons) or if vaccine-specific antibody or T cells cross-react with these tissues.
The core battery of tests (rat respiration study, dog cardiovascular study and mouse Irwin Study) routinely performed for NCEs is not appropriate for vaccines and hence modified protocols have been developed and used by some of the larger vaccine development companies. In a combined cardiovascular and respiratory study, anaesthetised rats are injected with a single dose of vaccine using the clinical route and formulation. Further rats are injected intravenously with increasing high dose levels of unadjuvanted antigen as a worst-case scenario where high levels of the antigen reach the systemic circulation. Blood pressure, heart rate, ECG, respiration rate and excursion are measured shortly after dosing thereby assessing immediate effects of vaccine administration such as direct toxicity of the antigen, release of cytokines, complement activation. A second set of animals given the clinical regime are left until the immune response peaks when the same safety pharmacology parameters are measured, thereby assessing the effect of the immune response.
8.3.9. Reproductive toxicology and teratology studies
Reproductive toxicology and teratology studies provide information on potential effects on fertility, in utero development and subsequent offspring [80]. The risks of reproductive toxicity with vaccines were not considered to be of any major concern since vaccines are given infrequently at low doses, many (conventional) vaccines are given during childhood and pregnant women are seldom vaccinated. Indeed there is no evidence that any of the licensed vaccines have had any reproductive toxic effects and hence very few vaccines have been subjected to reproductive toxicity testing prior to human administration. The development of new vaccine and adjuvant technologies with novel mechanisms of action and a greater understanding of immune mechanisms and how they relate to a successful pregnancy [81] and [82] have led scientists and regulators to envisage theoretical reproductive toxic effects of vaccines.
Vaccines are unique medicinal products in that even single doses (e.g. of a live attenuated or toxoid vaccine) given in childhood could lead to a prolonged pharmacodynamic effect (i.e. life-long immunity) through the generation of immune memory. In this way, the infection of pregnant women and their offspring is prevented. For other vaccines where immunity is short-lived (e.g. subunit vaccine in alum adjuvant) or needs to be boosted, vaccination of adolescent and adult women as well as prospective mothers and pregnant women is required both to protect the female and to generate immunity to protect the foetus.
This has meant that vaccination programmes for the prophylaxis of disease caused by, e.g. influenza, hepatitis, Group B meningitis, RSV, rubella, tetanus, Group B streptococcus and bioterrorism agents (smallpox, anthrax, plague, etc.), as well as therapeutic vaccines for cancer, chronic viral infection, allergy and addiction may often include women of child-bearing potential as well as pregnant women [83], [84] and [85]. However, the placental transfer of antibodies from mother to foetus via the specific transporter FcRn, important in transferring maternal immunity to the offspring could also lead to the transfer of cross-reactive autopathogenic antibodies that bind to embryonic or foetal tissues [86].
Immunological recognition of pregnancy is important for the maintenance of gestation. Vaccines and adjuvants could have immunomodulatory effects which may not only effect whether successful gestation is achieved [81], but might also affect immune system development in the foetus. The balance of TH1/TH2 responses is important in a successful pregnancy, with maternal responses being biased towards humoral immunity (TH2) and away from cell-mediated immunity (TH1) [82]. Hence any vaccine/adjuvant that alters this balance, e.g. therapeutic vaccines for allergy that induce a TH2 to TH1 switch, could in theory have adverse effect on embryo-foetal development. Cytokines, NK cells and gamma delta T cells of maternal origin are all thought to be involved in processes such as foetal recognition, placental development and in the regulation of gene expression during organogenesis [81]. Hence any effects of vaccination on these cells and mediators may affect development of the foetus.
These concerns have led the FDA to publish new guidelines specific to the reproductive toxicity testing of vaccines (Table 2). If pregnant women or women of child-bearing potential will be treated with the vaccine then reproductive toxicology and development studies will be necessary prior to large scale Phase 3 trials or regulatory submission [87]. Even if a vaccine is for single dosing in childhood, it may be that the B and T cell response persists for years through the action of memory cells and the response is boosted following exposure to the organism during pregnancy. Hence reproductive toxicity studies may still be required. No other medicinal product has such as prolonged pharmacodynamic effect. Such studies are not usually required early in the development programme as a pre-requisite for Phase I/II studies, which are performed in healthy male adults or women on effective contraception, provided there is no pathology of the reproductive organs within the toxicity studies.
For vaccines, where the likely toxicity is due to the immune responses and not the vaccine antigen itself (with exceptions such as DNA vaccines and live attenuated vaccines), modified protocols have been developed to comply with the FDA guidance. The choice of species is dependent on both the immunogenicity of the vaccine in that species and in the relative rate and timing of maternal transfer of vaccine-specific antibody to the offspring before and after birth (dependent on the physiology of the placenta). The rat, rabbit and mouse are the most frequently used species for these studies and most vaccines are immunogenic in these species. The rabbit is probably the better species since, like humans, much of the antibody is transferred pre-natally unlike the rodent where 90% of antibody is transferred post-natally in the milk. However, the rat is the most commonly used species on the basis of experience and practicality of performing post-natal studies. In some cases, however, only primates show a relevant response and the ethical, technical and logistical problems in performing developmental studies in primates are obvious [88]. If no appropriate species can be found, the FDA still expects studies in a non-reactive species.
In a combined embryo foetal/post-natal development study, females are immunised a few weeks before mating (ideally following the same dosing regimen used in the repeat dose toxicity study) and then boosted prior to mating and during the embryo-foetal and post-natal periods. This ensures that antibodies are present throughout the whole gestation period and are produced in the dams milk during weaning (early post-natal period). Maternal antibody transport is determined in maternal, foetal and neonatal serum and maternal milk. One group of pregnant females is submitted to caesarian section at the end of gestation and used for routine foetal examinations whilst another group is allowed to litter and post-natal development of the pups followed up to weaning. These types of studies could be extended to include assessment of the immune system (developmental immunotoxicity testing) in the offspring at 6–8 weeks to assess the potential for long-lasting, permanent changes (reviewed in [89]). It should be pointed out that these protocols are generally designed for vaccines whose primary pharmacodynamic effects is the generation of specific antibody rather than specific cellular responses. However, maternal T cell responses can similarly be optimised using the same protocol design.
For gene transfer based vaccines, eliminating the possibility of the gene persisting in the gonads and being inserted into the germline ensures that some elements of concern about reproductive toxicology have already been addressed for these products. However fertility, embryo-foetal and perinatal studies may be required depending on the clinical usage.
8.3.10. Genotoxicity
Classical studies designed to detect genotoxins which exert effect via direct interaction with target cell DNA are not applicable for most vaccines antigens. However genotoxicity studies should be considered if a novel chemical linker is included e.g. as part of a KLH conjugated recombinant protein vaccine product or if a novel chemical adjuvant, stabilizer or preservative are used in the formulation. In this case a battery of tests is used to assess the potential for gene mutation, chromosome aberrations and primary DNA damage. Studies might include the in vitro bacterial reverse mutation (Ames), mammalian cell (forward) mutation and in vitro chromosome aberration tests as well as the in vivo rodent micronucleus test.
9. Conclusions
Historically, vaccines have proven to be one of the safest types of medicinal product. However, new classes of vaccines and adjuvants have been developed with novel safety concerns that require careful investigation prior to human clinical testing. In developed countries, the public (rather unrealistically) expect zero risk from vaccination and the risk/benefit ratio is often scrutinised on an individual level. Hence safety prediction of these new vaccines and adjuvants will depend on the selection of the most relevant and sensitive in vitro and in vivo test systems, requiring a strong scientific understanding of the product. However, it is recognised that non-clinical studies will not detect all adverse events since many of these are rare and only seen in clinical trials involving thousands of patients. We can, however, learn from the clinical data and develop more sensitive models to try and predict these rare events. This review has highlighted that there is no prescriptive “one package fits all” approach that can be applied to the safety testing of these new vaccine products which need to be evaluated on a case-by-case basis. Both non-clinical and clinical experience over the coming years will help to determine if the largely theoretical safety concerns associated with these new vaccines and adjuvants are justified.
Acknowledgement
The authors would like to thank Dr. Francois Verdier, Aventis Pasteur, for his advice in the preparation of this manuscript.
References
|