Thermostable Vaccines in Veterinary Medicine: State of the Art and Opportunities to Be Seized

The COVID-19 pandemic has highlighted the weakness of the vaccine supply chain, and the lack of thermostable formulations is one of its major limitations. This study presents evidence from peer-reviewed literature on the development of thermostable vaccines for veterinary use. A systematic review and meta-analysis were performed to evaluate the immunogenicity and/or the efficacy/effectiveness of thermostable vaccines against infectious diseases. The selected studies (n = 78) assessed the vaccine’s heat stability under different temperature conditions and over different periods. Only one study assessed the exposure of the vaccine to freezing temperatures. Two field studies provided robust evidence on the immunogenicity of commercial vaccines stored at temperatures far in excess of the manufacturer’s recommended cold-chain conditions. The drying process was the most-used method to improve the vaccine’s thermostability, along with the use of different stabilizers. The pooled vaccine efficacy was estimated to be high (VE = 69%), highlighting the importance of vaccination in reducing the economic losses due to the disease impact. These findings provide evidence on the needs and benefits of developing a portfolio of heat- and freeze-stable veterinary vaccines to unleash the true potential of immunization as an essential component of improved animal health and welfare, reduce the burden of certain zoonotic events and thus contribute to economic resilience worldwide.


Introduction
Global vaccine availability and equity is a goal advocated by global leaders and by 170 Nobel Laureates [1]. Nevertheless, the current COVID-19 pandemic has highlighted that the global vaccine coverage is highly inequitable and skewed, with a high vaccine uptake concentrated in selected countries, predominantly the G7 and European ones [2]. Recently, the G20 Summit has underlined the urgent need to intensify efforts to enhance timely, global, and equitable access to safe, effective, and affordable COVID-19 vaccines [3]. In fact, logistical and supply chain system failures have slowed the vaccine availability and have hampered the global efforts to up-scale COVID-19 vaccination coverage. The lack of thermostability has been proven to be one of the major barriers limiting the worldwide distribution of these products [4]. Indeed, the race to develop efficacious SARS-CoV-2 vaccines has resulted in the first available commercial vaccine products to have storage and delivery requirements of temperatures between +2 • C and −70 • C, depending on the product [5][6][7][8].
It is surprising that in 2021 the vast majority of vaccines for human and animal diseases are still dependent on cold-chain systems to ensure their potency throughout production, shipment, storage, and administration. In both human and animal health, vaccines resistant to damage by heat and freezing could have great economic and health benefits. Heat-freezestable vaccines could help to reduce vaccine wastage and prevent the consequences of administering ineffective vaccines [9]. For these reasons, thermostable vaccines have been named a priority research area in the World Health Organisation's Global Vaccine Action Plan 2011-2020 [10]. Nevertheless, their development and production is not always a prime concern for vaccine developers, industries, and funding entities [2].
Vaccination is an effective preventive measure against infectious diseases. The main objective of livestock vaccines is to improve animal health, reducing the economic losses associated with disease occurrences [11]. The use of vaccines is recognised as an important management option during outbreaks, as it helps to control the spread of infection and reduce the need for the large-scale culling of at-risk animals [12]. Vaccines are also essential to sustain the commercial exchange of animal products between countries. Vaccines have been developed for 53% (63/117) of the OIE listed diseases (Appendix A: Table A1) [13,14], while the production of vaccines has been historically reported by members to the World Organisation for Animal Health (OIE) for 68 diseases. When considering these data, it should be noted that only the laboratories under national veterinary services are requested to provide information to the OIE on the vaccines produced (e.g., vaccines produced by private firms are not reported to the OIE) (Appendix A: Table A2).
Most vaccines require continuous storage at 2-8 • C from manufacturing through to administration, requiring a cold-chain system for their transportation [15]. Vaccination campaigns for several OIE-listed diseases (e.g., Foot-and-mouth disease (FMD) and Rabies) are highly encouraged in endemically infected countries to combat disease outbreaks and reduce their economic burden [16]. These are generally low-and middle-income countries which do not have widespread access to a stable supply of electricity or an effective coldchain system for vaccines. Considering this, thermal stability is a critical issue for most of the available vaccines against animal diseases of international concern.
Similarly, half of the supplied vaccines for human use are wasted as a result of inadequate cold-chain capacities [15]. It has been estimated that this loss accounts for about 80% of the total cost of vaccination programs, which is roughly $200-$300 million per year [17]. In the worst circumstances, the damages may remain undetected, increasing the chance that vaccines with reduced potency are administered, exposing the recipients to a higher risk of becoming infected or even ill [18]. There are no such studies for veterinary vaccines, but we can assume similar figures. For these reasons, it would seem reasonable to invest in solutions that can address the core fragilities which are embedded in most vaccines that are on the market today. Indeed, we have previous experiences which underscore the importance of having heat-stable vaccines.
To date, Rinderpest in cattle, and smallpox in humans, are the only diseases that have been officially eradicated. For both diseases, indispensable to the success of the eradication was the adequate supply of heat-stable and potent vaccines [19,20]. The benefits of developing thermostable vaccines for humans were reviewed by several studies [9,21]. Additionally, the economic impact of their use was estimated in different case studies in developing countries. For instance, Lee et al. [22] developed a computational model to simulate the effects of making some vaccines thermostable in Niger. They showed that even a single thermostable vaccine would free significant cold storage space for other vaccines, thus alleviating supply chain bottlenecks. In Benin, another study showed that replacing different existing vaccines with thermostable formulations would save medical costs and productivity losses, even with a price two-to-three times higher than the non-thermostable product [23]. Although no study evaluating the economic impact of thermostable vaccines for veterinary use has been carried out, it is reasonable to assume that it would be significant, especially considering how livestock plays an important role in the economy of developing countries, contributing to the livelihoods of about 1.7 billion people [24].
The potential impact of making certain formulations thermostable appears evident when looking at the figures of vaccines commercialized by private companies and authorized by the agencies responsible for the evaluation and supervision of medicines.
A vaccine that did not require cold temperatures to be transported and stored would eliminate the costs of maintaining the cold-chain and would address equity issues linked to the unavailability of a reliable electricity supply. The positive impact would be also seen in high-income countries, as thermostable vaccines would be easier and cheaper to store. For example, Porphyre et al. [25] identified the importance of sufficient strategic supplies of vaccines to control FMD outbreaks in Scotland. The easy distribution and storage of thermostable vaccines would greatly influence delivery rates and, thus, the reaction timing for controlling outbreaks in livestock. This is particularly true when considering highly contagious diseases, such as FMD [26]. The general consensus is that vaccination is one of the essential tools to respond to outbreaks of livestock diseases which cannot be controlled by stamping-out policies. In these cases, vaccination is also considered the control option that provides the largest economic benefits [27,28].
As an unsurprising starting point, it should be mentioned that the characteristics of thermostable vaccines are not clearly and specifically defined. The World Health Organization (WHO) encourages the production of thermostable vaccines, considering them, in general terms, as heat-and freeze-stable formulations which can be stored for extended periods of time above 8 • C, as well as not being damaged by freezing temperatures (<0 • C) [29]. The OIE, which sets the standards for the production and quality control of biological products for veterinary use across the globe [30], uses the word 'thermotolerant' to describe the ability of a vaccine to retain a level of infectivity after exposure to heat (Glossary of terms of the OIE Terrestrial Manual https://www.oie.int/app/uploads/2021/03/mailing-oct-2014.pdf, accessed on 25 November 2021). However, it does not provide a clear definition of thermotolerance or thermostability in terms of its shelf-life and its recommended stability, with reference to temperature ranges. Moreover, the Food and Agriculture Organization of the United Nations (FAO) and the Pan American Health Organization (PAHO), which have high-level scientific and technical expertise from around the world in dealing with priority health issues, do not outline a standard for thermostable vaccines [31,32]. The lack of a standard, as well as a unified definition, from the international organisations involved in the fight against human and animal diseases at a global level, contributes to the hinderance in the production of thermostable formulations.
Today, given the evidence of the inequitable access to vaccines, supply chain challenges, and the continuing rise in new cases of COVID-19, particularly in low-and middle-income countries, the world has a perfect opportunity to identify bottlenecks and to reprioritize research. The transformative power of the COVID-19 pandemic calls for major advancements in vaccine development and manufacturing, which would empower decision makers and the scientific community to unleash the full potential of vaccines and immunization. Considering the above, the objective of this study is to gather, assess, and present evidence from the peer-reviewed literature on thermostable vaccines developed for animal diseases and providing examples of their value, as well as discussing their impact on disease prevention and control.

Objectives
This systematic review and meta-analysis focus on a selection of animal infectious diseases with the objective of answering the following guiding questions:

1.
What type of thermostable vaccines have been developed for veterinary use? 2.
What are the characteristics of these thermostable vaccines? 3.
How immunogenic and effective are these thermostable vaccines?

Eligibility Criteria
The inclusion criteria are: (1) the clinical and field trials evaluating the immunogenicity and/or the efficacy/effectiveness of thermostable vaccine formulations developed against animal infectious diseases (only bacterial and viral diseases); (2) studies testing thermostable vaccines in natural hosts; (3) articles published in peer-reviewed journals after 1990; and (4) an English language full text. Experimental studies using laboratory animals (e.g., mice) and in vitro studies are excluded.

Information Sources
PubMed, CAB Abstracts, and Web of Science databases were used to perform two separate literature searches: a broad search on thermostable vaccines, and a specific search on DNA vaccines, which are the new-generation vaccines that are considered heat-stable on account of their structural character [11,33,34]. The first search was done using general keywords and was integrated by screening the reference lists of the identified eligible studies. For the search on DNA vaccines, the terms used to label articles (MeSHterms or Subject category) were implemented, and only the titles of the first 100 returns (sorted by relevance) from each database were retrieved, since the timeframe for this study only allowed for a rapid assessment. The decision of performing two separate searches was for the following reasons:

•
Authors may not specify that DNA vaccines do not need the cold-chain, a thermostability is an intrinsic characteristic of these vaccines. Thus, the computerized search would not be able to retrieve the manuscripts if it only used general keywords; • The use of a unique complex search strategy, combining multiple different terms, would not be an efficient way to identify relevant articles.
The last search was done on 8 September 2021. Details on the search strategies are provided in Table 1.

Data Collection Process and Data
Two data extraction sheets were created in Microsoft Excel, version 2017. In the first database, the following information for each study was recorded: the authors, year of publication, target agent, type of agent (bacterium/virus), animal species, country, product name, vaccine type, strain, market availability (locally produced, commercially available, or experimentally developed), thermostability characteristics, route of administration, type of study (clinical or field trial), assessment (objective), test used, main results, and comments. If the data was not provided, 'N.A.' (NOT AVAILABLE) was written. If some information was difficult to extract, a comment was written to that cell. The second database was created to retrieve quantitative data from clinical and field trials assessing the vaccine efficacy/effectiveness after its challenge with the infectious organism. The vaccine efficacy was measured in the clinical trial, as well as how well the vaccine performed in controlled settings. On the other hand, the vaccine effectiveness was defined as the measure of how well the vaccine works in the real world and was measured in the field trials. Vaccine efficacy/effectiveness can be computed by estimating the incidence rate of the disease among vaccinated and unvaccinated groups and determining the percentage of reduction in the incidence rate of the disease among vaccinated animals, compared to unvaccinated animals (1-risk ratio) [35,36]. To build this database we only considered the studies on diseases which are severe and sudden in onset (acute conditions leading to death), while studies assessing the morbidity rate were excluded. The following information was retrieved: the number of deaths after challenging in the vaccinated group, the number of survivals after challenging in the vaccinated group, the number of deaths after challenging in the control group, the number of survivals after challenging in the control group, the challenge time (days post-vaccination, dpv), the relative percent of survival (RPS)/days post-challenge (most of the studies computed the relative percentage of survival (RPS) from the cumulative mortalities in the vaccinated group (Mvac) and unvaccinated control (Munvac): RPS = [1 − (Mvac/Munvac)] × 100%). If a single study had data for more than one experimental group, then those studies were considered as separate studies according to the number of the vaccinated groups under investigation. All authors checked the quality of the data extracted. Any disagreement in the results were resolved by discussion within the team.

Risk of Bias (Quality Assessment)
To minimize the risk of bias in individual studies, anything that could have potentially affected the interpretation of the study was written in the comments section of the data extraction sheets.

Method of Analysis
Results were summarised with text descriptions, tables, and waffle graphs. A metaanalysis with a random-effects model was performed, using quantitative data from studies on fatal diseases. The pooled risk ratio (RR) was calculated, along with the corresponding 95% CI, to report the vaccine efficacy (VE). Studies with less than 10 animals per group were excluded. The analysis was done with the 'meta' and 'metafor' packages in R software version 4.1.1. [37]. The inverse variance index (I 2 ) was used to quantify heterogeneity, indicating the I 2 values of 25%, 50%, and 75% as low, moderate, and high heterogeneity, respectively [38]. Outliers were investigated using the Baujat and diagnostic plots [39,40]. The potential publication bias was assessed by the examination of the funnel plot. Considering that the asymmetry observed in a funnel plot may be also due to the correlation between the log of RR and its SE, the presence of a small study effect was tested with the Peters' test for binary outcomes [41]. Subgroup analyses, using mixed effect models, were performed to identify possible sources of heterogeneity related to the animal species and the type of agent.

Study Selection
The first literature search identified a total of 1,655 studies. After the duplicates (n = 758) were removed, the titles and abstracts of the remaining studies (n = 897) were screened for relevance, and 149 articles were further evaluated for eligibility based on the inclusion criteria. Out of them, 40 were included in the qualitative synthesis, along with three articles retrieved with the screening of the reference lists. Finally, 10 articles were included in the meta-analysis ( Figure 1A). Considering the articles on DNA vaccines (n = 300), 31 duplicates were removed, and the titles and abstracts of the remaining articles (n = 269) were screened for relevance. Seventy-six articles were assessed for eligibility. Out of them, 35 were included in the qualitative synthesis, and 18 were included in the meta-analysis ( Figure 1B).

Study Characteristics
A total of 78 studies, published between 1990 and 2021, were included in this systematic review (Tables 2 and 3).        • Two studies performing both clinical and field trials (one using vaccinated and control groups, and one with all animals vaccinated); • Thirteen studies performing field trials (eight using vaccinated and control groups, and five with all animals vaccinated); • Sixty-three studies performing clinical trials (60 using vaccinated and control groups, and three with all animals vaccinated).
Out of the 43 articles retrieved with the broad literature search, 13 studies (these include only the studies that explicitly state that the freeze-drying process was used for the vaccine development) implemented a lyophilization (freeze-drying process) to obtain thermostability [47,49,50,52,[65][66][67]70,74,76,79,85]. An alternative drying process was applied by Lv et al. [64] and Smith et al. [77], who used the vaporization method (foam-drying) to preserve the live attenuated vaccines against the porcine reproductive and respiratory syndrome virus (PRRSV) and the rabies virus (RABV), respectively, while Dulal et al. [51] successfully used the sugar-membrane technology to thermostabilize an adenovirus-vectored vaccine against the Rift Valley fever virus.
Thermostability characteristics were not reported in 23 studies (  [80]. The remaining studies assessed the heat stability under different temperature conditions and over different time periods, from 3-4 days at 40 • C [47] to 25 • C for 12 months [64]. Details on each study are provided in Table 2. All the articles on DNA vaccines did not provide information on thermal stability. Nevertheless, some of them mentioned, in the introduction section, that DNA vaccines do not require the maintenance of a cold-chain as they are thermostable (e.g., Bande et al. [88]). Interestingly, only one retrieved study assessed the exposure of the vaccine to freezing temperatures [68].
Considering the objective of the study, 27 works aimed to assess humoral immunity, estimating the antibody titres after vaccine administration, and five articles evaluated both humoral and cell-mediated immunities. In the remaining studies (n = 46), animals were challenged with an infectious disease organism, evaluating the humoral immunity, cell-mediated immunity, clinical signs, histopathological changes, or survival rates post-challenge.
Only a few authors reported an insufficient immune response after vaccination. In particular, Rahman et al. [72] described a partial seroconversion in goats after the vaccination against PPRV, and Bunning et al. [89] reported a failure of the oral vaccination with a DNA vaccine against the West Nile virus (WNV) in the American crow.

Risk of Bias (Quality) Assessment
Overall, no relevant comments that could have affected the outcomes of the studies included were identified.

Synthesis of Results
Twenty-eight studies, comprising of 60 vaccinated groups, were included in the metaanalysis (Table 4).  The animals in the trials included avian species (n = 12) and fish (n = 16), while the target agents were the virus (n = 18) and the bacteria (n = 10). As shown in Table 4, the RPS was lower than 50% in 10/60 vaccinated groups. These include vaccinated groups from studies investigating the suitability and efficacy of different administration routes [43,89,119], strains [71], or doses [105]. It is important to consider that all these studies have at least one vaccinated group with the RPS > 50%.
The pooled RR was 0.31 (95% CI: 0.25-0.38), resulting in a vaccine efficacy (VE) of 69%. A vaccine efficacy of 69% indicates a 69% reduction in the death rate among the vaccinated groups. Effect estimates and confidence intervals are presented in the forest plot ( Figure 3). The heterogeneity was significantly high, being I 2 = 95 (95% CI: 93-98), with a p-value < 0.0001. One study was detected as influential, with an individual RR of 0.07 (93% of VE) (Appendix A: Figures A1 and A2) [90]. Although the removal of this study would reduce the amount of heterogeneity and increase the precision of the estimated average outcome, we decided to keep it in the quantitative synthesis as it has one of the largest sample sizes (100 vaccinated animals and 100 control animals) and a high-quality study design. The inspection of the funnel plot shows some asymmetry (Appendix A: Figure A3). Nevertheless, the Peters' test p-value was 0.27; therefore, the hypothesis of the symmetry of the funnel plot was accepted. A meta-analysis was not performed to evaluate the vaccine effectiveness due to the small number of field trials retrieved.

Summary of Evidence
This study represents the first systematic review and meta-analysis on the current state of thermostable vaccines against a selection of animal infectious diseases, providing a quantitative measure of their efficacy against death (VE = 69%).
Most of the studies included are on vaccines against avian diseases, and, in particular, against NDV [42][43][44][45][46][52][53][54][55][56][57][58]67,[69][70][71]73,75,76,78,79,[82][83][84]. Developing a thermostable vaccine for Newcastle disease (ND) was considered a priority for non-governmental organisations (NGOs) and studies were funded to evaluate the effectiveness and economic viabilities of the vaccination in developing countries. Strong encouragement and support were provided by the FAO and the International Atomic Energy Agency (IAEA) to reduce the burden of the disease and improve the welfare of rural households [122,123]. In this context, the key success of the vaccination against NDV was the development of thermostable vaccines by the Australian Centre for International Agriculture Research (ACIAR) [124].
A similar situation can be observed for the vaccines developed against PPRV, for which progress has been driven by the PPR Global Control and Eradication Strategy (GCES) launched by the FAO and OIE [125]. The first thermostable vaccine (Nigeria 75/1 PPR strain) against this highly contagious disease has recently received the regulatory approvals required to be produced and commercially distributed in Nepal [126]. Along with Rinderpest, ND and PPR represent perfect examples of high-impact diseases which have benefitted from the support and incentives of NGOs by implementing a vaccination campaign with thermostable products.
Interestingly, and as a first step towards addressing the issue, some field studies provided robust evidence on the immunogenicity of commercial vaccines stored at temperatures far in excess of the manufacturer's recommended cold-chain conditions [46,61]. Their results raise several questions, such as: (i) why manufacturers do not test for thermostability during vaccine development; (ii) why they do not include such information on the products labels; and (iii) how many other vaccines currently on the market could be stored outside the cold-chain, and for how long, whilst retaining equivalent potency? These studies provide preliminary evidence that some commercial vaccines might be used successfully, following a period of non-optimal storage in remote areas, regardless of the manufacturer's recommendations.
If we look back at recent and past history, our literature search highlights that the freeze-drying process is a valuable method to obtain vaccine thermostability [47,49,50,52,[65][66][67]70,74,76,79,85]. An improved freeze-drying process was used to develop a thermostable Rinderpest vaccine (Thermovax), which was an essential tool for eradicating the disease in remote pastoral areas [127]. In this study we identified dried formulations (freeze-or foam-dried) for vaccines against NDV (e.g., [52]), bovine ephemeral fever virus (BEFV) [74], classical swine fever virus (CSFV) [85], rabies [77], and PPR [47], highlighting that the drying process is a useful technique to improve the thermostability of vaccines against diverse diseases in several species. However, it is worth mentioning that the drying process alone is not able to confer a long-term stability in the formulations. There are other ways that have been used to enhance the shelf life of the products at ambient temperatures. For instance, the freeze-dried vaccine against CSFV was stabilized with a buffer composed by trehalose, glycine, thiourea, and phosphate [85]. Other examples of stabilizers retrieved from this review include: (i) lactalbumin, hydrolysate, and sucrose for the Rinderpest vaccine [65] (ii) the methylglucoside for the vaccine against bovine ephemeral fever (BEF), (iii) and a formulation composed of trehalose, tryptone, and other protectants for the vaccine against the porcine reproductive and respiratory syndrome virus (PRRSV) [64].
Lyophilized vaccines are more stable prior to their reconstitution in the liquid form, while their potency is known to decline once reconstituted. In addition, not all vaccines can be lyophilized and, thus, there have been efforts to increase the stability of vaccines in liquid form. For instance, the stability of liquid vaccines can be achieved by optimizing the properties of the solvent (e.g., buffer, pH, and salt concentrations), and low-cost and safe excipients (e.g., glycerol) could provide freeze protection to vaccines with aluminum hydroxide, as an adjuvant to freeze damage [9]. Modern technologies have also become a key strategy to develop thermostable products. In this sense, Tan et al. [78] designed a thermostable recombinant NDV candidate vaccine against NDV and the infectious bronchitis virus (IBV), which was stable in the liquid form at 25 • C for 16 days. Similarly, Murr et al. [68] developed a recombinant NDV vector vaccine against PPR which was stable in the liquid form at −80 • C, −20 • C, 4 • C, 21 • C, and 37 • C for seven days.
Oral vaccinations are easy to implement and avoids stress in animals. Some thermostable vaccines have been developed with this route of administration in mind. The vaccine is incorporated into the feed during production, or it may be coated with pellets or encapsulated. Oral vaccines are particularly suitable for use in wild animals. In this sense, Smith et al. [77] developed a promising thermostable RABV vaccine using a foam drying process, highlighting the potential of this technique to produce a vaccine for oral use. The failure of the oral vaccination in the research by Bunning et al. [89] could have been due to the inactivation of the vaccines within the avian gastrointestinal tract. Oral vaccination was implemented in 16 other studies. These include articles on ND, using water and feed as vaccine carriers (e.g., [42,43,76,83]). The disadvantages of this route of administration are related to the large dose required to induce a uniform and long-lasting protection. For this reason, ND vaccines administered by eye-drops or treated feed have better performance than using water or untreated feed [42,43,57]. Additionally, oral vaccines may have an additional cost for the encapsulation, which may be necessary to avoid their degradation in the gastrointestinal environment prior to absorption [33].
Although a large number of trials using heat-stable vaccines was retrieved, very few peer-reviewed analyses exist on freeze-stable formulations. This finding shows how most efforts were directed to prevent vaccine deterioration and overcome the difficulty of maintaining the cold-chain in developing countries, which generally have high ambient temperatures. It is important to consider that although heat stability is perceived as a greater concern [128], conditions leading to freeze exposure occur, and may have an impact on the long-term stability of the vaccines, especially of those with aluminum adjuvants [9]. Damage due to freezing is likely in low-and middle-income countries, where cheap domestic refrigerators and cold boxes are used for storing and transporting vaccines. In particular, the poor performance of these refrigerators may lead to regular negative excursions, with potential damages to the vaccines during their storage [129]. Likewise, placing the vaccines with ice or gel packs inside portable containers may cause freeze damage to the vials too close to the ice and gel packs during their transportation [21]. With the exception of one article [68], this systematic review failed to identify studies in which the evaluation of the effect of freezing on vaccine potencies were assessed and, therefore, precluded identifying products fitting the definition of thermostable vaccines provided by the WHO [29]. Unsurprisingly, the information on heat stability and environmental temperatures, as provided by the authors, was reported heterogeneously in terms of different temperatures and periods of time (in ranges of days (e.g., Murr et al. [68]), weeks (e.g., Tu et al. [79]), or months (e.g., Dulal et al. [51]). Additionally, some authors defined the vaccines as thermostable only by performing a heat-treatment test in the lab (e.g., 56 • C for 60 min [58,73]). This diversity among benchmarks between the studies highlights the urgency to define standards when it comes to environmental or the freeze stability of vaccines.
With regards to the search on DNA vaccines, most of the articles retrieved were on vaccines that were experimentally developed. Although many DNA vaccine candidates have been evaluated with promising results in various animal species, it has been estimated, by a recent review, that only five DNA vaccines have been approved and licensed for veterinary use [130]. These include: • Three against viral diseases; • Two for fish (one against infectious hematopoietic necrosis virus (IHNV), and one against salmon alphavirus subtype 3); • One for horses against WNV, but used also in several avian species; • One to treat cancer melanoma in dogs; • One growth hormone-releasing hormone (GHRH) gene therapy for swine.
Conversely, no DNA vaccines have been licensed for human use to date [33,130]. DNA vaccination involves immunization with a plasmid encoding a gene of the pathogen. The production of DNA vaccines is cheaper than other types of vaccines. They are able to act in the presence of maternal antibodies, are temperature stable, and are safe to transport, which is especially important for remote areas [33]. Despite these advantages, some concerns have been raised, as DNA vaccines have failed to produce measurable antibodies, even if the host got protected, suggesting a major role of cellular immune responses. Another important concern is related to the potential deleterious effects following the integration into the host chromosome [131]. These issues, along with the cost of GMP (good manufacturing practices) grades, large-scale manufacturing restrains the commercial availability of DNA vaccines.
In the majority of the articles on DNA vaccines, both humoral and cellular immune responses were assessed, obtaining promising results on the production of a variety of immune modulators, cytokines, and co-stimulatory molecules (e.g., [102]). DNA vaccines have received particular attention in the field of aquaculture. They are safe for fish since they do not contain an oil adjuvant that can cause peritonitis, but also for the consumer, as the fish are consumed months after vaccination and the quantity of DNA used is very small [33]. In this work, 16 out of 35 studies on DNA vaccines were carried out on fish species in China. Since China is a major player in global aquaculture, contributing to roughly 61% of the total production [132], it is not surprising that researchers from China conducted extensive research on DNA vaccines against different diseases impacting aquaculture. In fish, the RPS, post-challenge, in the groups vaccinated ranged from 20% for the vaccine against Mycobacterium marinum developed by Pasnik and Smith [105], to 92% for the vaccine against Vibrio alginolyticus developed by Cai et al. [90] and the vaccine against IHNV and the infectious pancreatic necrosis virus (IPNV) developed by Xu et al. [113]. It should also be considered that the immune efficiency varies based on the immunization routes, doses, and times of DNA immunization. In fact, Pasnik and Smith [105] reported a higher protection (RPS: 80-90%) for the same vaccine administered at a higher dose, and a lower RPS at lower vaccine dose (RPS: 0%). Our search also retrieved a great number of studies on DNA vaccines against avian diseases. Promising results have been obtained in avian species, with an RPS, post-challenge, ranging from 44% for the vaccine against WNV in the American crow [89] to 100% for the vaccine against novel duck reovirus (NDR) in ducks [120]. However, Bunning et al. [89] showed that the response to the DNA vaccines depended on the inclusion of an adjuvant (RPS: 60%) and the route of administration, as none of the birds receiving the oral microencapsulated DNA vaccine against WNV developed antibodies, and none of them survived post-challenge (RPS: 0%).
The VE, in terms of protection against death, is an objective measure to aggregate data on different vaccines. Indeed, numbers or rates of death are the most used measure for comparing the impact of different diseases in epidemiology [133]. In vaccine trials, challenging humans with dangerous pathogens is ethically unacceptable. Conversely, the evaluation of veterinary vaccines mainly relies on challenge studies. This is important to consider as serological studies may not always provide a good measure of efficacy [134]. For all these reasons, the pooled estimate of the VE was provided in terms of the reduced risk of death. The protection of thermostable vaccines against fatal diseases was estimated to be high (VE = 69%), highlighting the benefits of vaccination to reduce the economic losses (direct deaths) due to the disease impact. The heterogeneity between studies was high.
Developing a portfolio of thermostable vaccines would not only help with improving access to vaccines in parts of the world where cold-chain capacity is lacking, overcoming a major supply-chain hurdle to the rollout of successful vaccination campaigns for humans and animals, but it would also greatly benefit the environment by reducing the great consumption of energy required to sustain the cold-chain. On top of the overall energy consumption of an increased number of refrigeration units, maintaining ultracold temperatures requires the use of hydrofluorocarbon gases, which are known to have a very heavy carbon footprint [135]. An additional benefit can be obtained by investing in thermostable products that can aid eradication programs, such as "differentiating infected from vaccinated animals" DIVA-vaccines, such as the ones presented in this review, developed by Verardi et al. [80], Daouam et al. [49], Dulal et al. [51], and Murr et al. [68]. These types of vaccines are promising for the effective disease control during outbreaks, and eradication programs in disease-endemic regions [136].

Limitations
The current study should be interpreted within the context of its limitations. Firstly, it does not provide a complete overview of the licensed thermostable vaccines for veterinary medicine. Instead, it aims to synthetize the peer-reviewed articles on thermostable vaccines developed against a selection of animal diseases. The target is not only to include the commercial vaccines, but also the vaccines experimentally developed, which are promising candidates. Additionally, only studies testing for VE, and the protection of the target hosts, were included, while in vitro studies, or studies testing the vaccines on non-natural hosts, were excluded. The intent was to retrieve an adequate number of studies to summarize the evidence on the efficacy of thermostable vaccines, rather than describe the progress made in vaccine technology. Some successful technologies that produce vaccine thermostability may not have been included in this study because the peer-reviewed articles were on vaccines tested under laboratory conditions. Secondly, the search on DNA vaccines was intentionally limited by sorting for relevance and extracting the first 100 records from each bibliographic database. The screening of all the papers would have allowed us to retrieve a higher number of articles, which would have compromised the time efficiency of our search. Indeed, such an approach would have been unfeasible, given the growing number of peer-reviewed articles on DNA vaccine candidates for animal species [137]. Moreover, in this case, the aim was not to provide a comprehensive overview on DNA vaccines for veterinary use, which has been reviewed by several narrative reviews (e.g., Fomsgaard and Liu [130]). Instead, this study aims to highlight some applications of these vaccines, which have intrinsic thermostability characteristics.
Thirdly, considering that the methodology to assess the immunogenicity, durability of immunity, and the safety profile is specific to each disease, comparisons on the humoral and cell-mediated immunities elicited by the vaccines were not made. The outcomes of interest for veterinary vaccines consider the livestock profitability and vary according to the disease. In the articles extracted, the outcomes ranged from the evaluation of specific disease symptoms (in cases of non-acute diseases) (e.g., Murr et al. [68]) to mortality. These different conditions could not have been compared or pooled.
Fourthly, the heterogeneity of the included studies is likely to be due to the different diseases against which the vaccines have been developed. Because of the small number of articles for each disease, the heterogeneity was not investigated using a subgroup analysis according to the disease. Additionally, other factors influencing the performance of vaccines were not assessed. These include the age and sex of the animals, the level and time of the challenge (pathogen factors), the dose, and the route of vaccine administration. Despite these limitations, it is important to consider that most of the individual estimates show the same direction of effect (RR < 1), highlighting the significant protection conferred by the vaccination.

Conclusions
This study presents the first condensed evidence from peer-reviewed literature on the current availability of thermostable vaccines for veterinary use. Over the years diverse methods have been implemented to develop and improve vaccine thermostability. Moreover, the efficacy of these formulations has been proved for several animal diseases, with an overall risk of death, in vaccinated animals, that is reduced by nearly 70% compared with unvaccinated controls. Although we were not able to identify the exact percentage of thermostable formulations, many articles cited in this review stated that most of vaccines on the market are still dependent on cold-chain systems, stressing the importance of enhancing their stability (e.g., [9,18,127]). The recent COVID-19 pandemic has highlighted the difficulties in transporting and storing non-thermostable vaccine formulations, especially for low-income countries, highlighting the necessity to improve the distribution and storage of vaccines to adequately respond to the current and future pandemics. In this regard, the reevaluation of vaccine research and development, manufacturing, and supply-chain management strategies are essential to produce vaccines that are heat-and freeze-stable to make vaccinations widely available to anyone globally, regardless of cold-chain capacity. We suggest that each novel vaccine candidate should be evaluated for its thermostability along with its safety, immunogenicity, and protective efficacy before it is licensed for use. The shelf life of existing products should be investigated, by default, under non-cold-chain conditions, coupled with efforts to boost their thermostability. We also strongly encourage regulatory agencies to adopt a standard definition of vaccine heat-and freeze-stability requirements to be used for the development of new generation vaccines both for human and for veterinary use.
As a final point, we would like to invite funding agencies and donors who support vaccine research to reflect and consider on the added value that having more stable products would bring to their philanthropic efforts both in human and veterinary medicine.

Conflicts of Interest:
The authors declare no conflict of interest. Table A1. OIE listed diseases for which vaccines have been developed [13,14].    The most produced type is the inactivated vaccine (available for 56/67 diseases), followed by the live attenuated vaccine (available for 49/67 diseases), recombinant vector vaccine (available for 10/67 diseases), conjugate vaccine (available for 5/67 diseases), subunit vaccine (available for 4/67 diseases), and DNA vaccine (available for 2/67 diseases) (supplementary material: Table A2). Generally, live-attenuated vaccines are more heat sensitive to potency loss during storage and distribution, thus requiring particular attention to maintain the cold chain. Conversely, inactivated and subunit vaccines can be particularly freeze sensitive, while DNA vaccines are very stable and do not require a cold chain. Figure A1. The Baujat Plot of roe deer studies. Study number 5 [90] could be an outlier which may distort the effect size estimate, as well as its precision. Figure A2. Influence analysis identifies study number 5 [90] as potential outlier.