As SARS-CoV-2 spreads around the world, pandemic preparedness is a key topic, not only discussed by the biopharmaceutical industry. Despite a “flattening of the curve” by confinement and social distancing measures, one thing has become blatantly clear: The only way to stop the global humanitarian and economic crisis is a vaccine – one we do not yet have. The WHO currently lists 76 companies and academic institutions with promising vaccine candidates, six of which have already entered Phase 1 of clinical trials. Optimistic forecasts suggest, a vaccine could be available in 12-18 months [i]. But even if the trials succeed, there are many barriers before global immunization is possible.
Despite increasing efficiency and speed of drug discovery, as well as accelerated tracks towards regulatory approval, one essential question remains: Is the industry prepared to meet the challenges of immediate and large-scale manufacturing of a novel vaccine to guarantee widespread and cost-efficient availability?
Delivering billions of doses of a new vaccine within a few months of its discovery is a tough challenge. The large number of doses needed in developing and emerging countries, makes it even more complex. A high cost per shot would prohibit access for large parts of the world’s population. Similarly, to the way it was done for HPV and pneumococcal vaccines, it might be the best option to agree upon multiple vaccines, each effective, but with different prices for different markets [ii]. This approach might also facilitate pandemic-scale production, as it mitigates the need for gigantic manufacturing capabilities for one single product by distributing this load. Still, the planning and construction of these expensive high-tech facilities in the shortest amount of time remains extremely challenging. Economic considerations add to the complexity.
Though it is not clear, which vaccine candidate(s) will be successful, companies must start building facilities long before regulatory approval to avoid losing time. In case a company’s candidate fails during trials, or does not make it to market in time, the enormous costs for the facility may never pay off. It therefore might be wise to design these plants as multi-product facilities, following a generic platform approach, that can be easily fitted to fluctuating demand for different vaccines. After all, the current pandemic is likely not to be the last, and it is not foreseeable, how long it will take and how quick the global demand for a vaccine may diminish. The SARS virus for example disappeared in 2003 after less than a year, and interest in a vaccine quickly vanished.
Re-using existing facilities might seem promising, but this approach clearly depends on the vaccine candidate(s), that lead the race. Additionally, the production of a novel vaccine in foreign facilities would have to be balanced against the one for other vaccines. If billions of people need a SARS-CoV-2 vaccine, manufacturing of vaccines against influenza, measles, mumps and many other diseases cannot suddenly be brought to a halt.
On a global scale, influenza vaccines account for the largest production capacities. In 2016, the total global capacity was estimated to be 1.5 billion doses per season [iii]. Luckily, companies are already used to quickly scaling up production in times of high demand to up to 6.2 billion pandemic doses (as in 2006 [iii]). While reusing these capacities might seem promising, most of the influenza vaccines rely on old-fashioned approaches: live attenuated vaccines or inactivated viruses. Most SARS-CoV-2 vaccine candidates follow completely different strategies and would not be able to make use of these capacities.
Traditional virus vaccines are based on the principle that weakened or killed pathogens or pathogen components sensitize the immune system and thus enable an active immunization against the infectious agent. These technologies are well established and have been around since Edward Jenner.
Whole virus vaccines use the entire virus particle, either in an attenuated state or fully inactivated, and look back on a very successful history. Live attenuated vaccines have led to remarkable success, up to eradication, in the fight against measles, mumps, rubella, smallpox or yellow fever. However, they are not recommended for people with weakened immune systems and pose difficulties in storage and worldwide delivery, as they must be kept cool. Only two of the current SARS-CoV-2 candidates base on live attenuated viruses. More candidates rely on inactivated versions of the SARS-CoV-2 virus, e.g. those by Sinovac (China, Phase 1) and Beijing/Wuhan Institutes of Biological Products (China, Phase 1). Inactivated vaccines are for example used to protect against influenza, Hepatits A, polio or rabies.
Both these vaccine types are grown in either in embryonated hens’ eggs or on a cell culture substrate. Subsequently, they are extracted, attenuated or inactivated (using heat, chemicals or radiation) and purified. In comparison to more recent approaches, traditional whole-virus vaccines do not require extensive purification. But every new vaccine requires its own unique production process, that is long, complex and costly. Bespoke facilities must be built years in advance and can cost up to USD 700 million [iv].
In particular for above mentioned influenza vaccines, so called “plattform approaches” exist – defined production systems, that usually also suit for manufacturing a group of related products. Such a platform process can serve as a template for further processes, and thereby lead to an acceleration of process development, regulatory approval and manufacturing. The demand for influenza vaccines shows clear seasonal trends. Therefore, the production might be reduced easily off-season to produce a SARS-CoV-2 vaccine. Unfortunately, influenza and corona viruses differ significantly, and it is not yet clear, in which way these production plants could be re-used.
Additionally, human influenza viruses are handled and cultured under biosafety level 2. Producing and purifying whole SARS-CoV-2 viruses at high concentrations requires at current time facilities with biosafety level 3 certification and these are scarce. Despite the best intentions, the restriction to BSL-3 may poses a counterproductive hurdle, making it even harder to develop and deploy a SARS-CoV-2 vaccine [v].
A more up-to-date approach utilizes protein subunits, that contain only the pathogens’ antigens, needed to trigger a protective immune response. Many influenza and tetanus vaccines work like this and the vast majority of candidates use this approach. Novavax (US), the University of Queensland (Australia) and many more pursue subunit vaccines to fight COVID-19.
These vaccines do not necessarily require growing large amounts of virus but can be produced as a recombinant protein e.g. in E. coli cells. After fermentation, the cells are disrupted, and the subunits are released. Centrifugation or depth filtration is used to remove host cell proteins. Subsequently, several downstream process steps, typically ion exchange chromatography, are utilized to purify the target from further contaminants.
In comparison to growing the virus in eggs or cell cultures, producing recombinant proteins is costly and complex. Like whole virus vaccines, every new vaccine requires its own production process and facilities must be built years in advance. Furthermore, many subunit vaccines need adjuvants to boost the immune response. Depending on the type of adjuvant these might get scarce during vaccine production on a pandemic scale.
Several promising vaccine candidates rely on an even more recent approach: viral vectors. These are typically harmless viruses, in whose genome the “building plans” for virus antigens, like the SARS-CoV-2’s spike protein, are included. Viral vectors offer promising platform approaches, that have proven successful already for a large number of diseases, including influenza, Zika and Ebola. Non-replicating viral vectors basing on the Adenovirus are being pursued e.g. by CanSino Biological Inc./Beijing Institute of Biotechnology (China, Phase 1) and University of Oxford (UK, funded by CEPI). The same approach had also been used for SARS-CoV-1 and is being investigated with two Phase-1-studies for MERS-CoV. Additionally, several replicating vectors are examined, relying e.g. on the Measles or the Horsepox vector.
Although viral vectors are tailored to a specific virus, their production follows a platform approach, enabling the scalable and cost-efficient manufacturing of viral vector systems. The Adenovirus for example is usually cultured in cell lines. After fermentation and infection, the virus is released, and the original DNA is degraded to facilitate DNA removal. Large-scale clarification from cells is accomplished by deepth filtration and/or centrifugation. Subsequently, the virus is purified from low-molecular-weight proteins and fragmented DNA by means of anion exchange chromatography or potentially ultrafiltration. Remaining impurities closely related to the target are removed by a second chromatographic polishing step. The final step is typically a membrane filtration.
As mentioned before, platform approaches can be utilized, with the individual process steps adjusted to the actual setup. It is to be assumed, that existing global manufacturing capacities for Adenovirus vectors are very scarce, as only few vaccines relying on Adenovirus vectors are in development and to the best of the author’s knowledge none is approved yet. Ramping up global production to a pandemic scale is further complicated by distinct scale-dependencies of all downstream steps, which need to be addressed in order to deliver a safe product.
Nucleic acid vaccines have emerged as a promising alternative to conventional approaches and many hopes lie on the current DNA and messenger-RNA vaccine candidates. The idea behind them is to introduce some of the virus’ genetic material directly into human cells. DNA or mRNA is injected into the patient’s muscle, where their cells start building the virus protein needed to trigger an immune response.
Moderna (US, Phase 1) were the first to start human trials for their mRNA vaccine candidate in late March 2020. The same approach is pursued by CureVac (Germany, funded by CEPI, currently in pre-clinical phase) and by BioNTech (Germany, Phase 1). Inovio (US, Phase 1) follows a similar approach but utilizing DNA instead of mRNA.
In comparison to the approaches, DNA and mRNA platforms rely on very simple and standardized production processes. They are constructed directly from the genetic sequence of the virus protein, which is much easier, more rapid and less risky than working with the pathogen itself. The downstream process following synthesis is significantly less complex and may even be accomplished by size exclusion chromatography or filtration. A factory designed for DNA or mRNA vaccines could produce all vaccines of the respective kind. In the case of mRNA vaccines, such a facility does not even have to be big: One shot is dosed on micrograms and does not require any special delivery device, but standard needles and syringes [vi]. CureVac states that already today, it has the facilities to produce more than 1g of product every two weeks and up to 400 million doses a year [vii].
But these vaccines are a novelty and have yet to be approved for human use. Even if several of them are already in clinical phases, it is to be assumed, that the current global production capacity is respectively little so far. Scaling them up should, however, be comparatively easy, and it is likely this will enable their widespread availability.
According to Brian Kelley [viii], the fastest process development strategy for monoclonal antibodies at pandemic pace precludes optimization or evaluation of process performance at pilot scale. For the sake of speed, companies should directly proceed to production following cell line selection. The same imperative likely holds for vaccines.
Scaling up a complex downstream process from lab to production, typically requires an enormous experimental workload. To mitigate time and cost involved with this, digital twins of downstream processes can be utilized to enable a virtual process transfer. The downstream process is simulated by means of mechanistic models, which describe the behavior of the process as a function of system parameters, like the dimension of the chromatography column or its packing quality. Differences in scale and their influence on the process behavior can easily be replicated by the model. By doing so, they can be used to investigate whether a process fits into an existing production facility very early and at lowest effort and predict requirements for new production facilities.
A computer-guided approach to downstream development also offers fundamental benefits all over the product development cycle. Various success stories for virus-like particles and vaccines have proven, that it may not only accelerate process development and reduce cost, they also show improved production output by completely utilizing the potential of single- and multi-product facilities.
It is to be assumed, that a virtual downstream development approach as pursued by GoSilico enables the fastest and cheapest scale-up and facility fitting, as well as the most economic design of multi-product facilities. All these benefits might make a significant contribution towards faster global immunization against SARS-CoV-2.
Vast financial means are needed for companies to scale-up manufacturing capacity, even if these facilities might ultimately not return a profit. Imperatively, vaccine development should not be driven by economic and risk considerations of individual companies or states, but by coordinated, international, and intergovernmental action.
The leading role towards this goal is being played by the Coalition for Epidemic Preparedness Innovations (CEPI). As a global alliance, between public, private, non-profit and civil society organizations, it finances and coordinates vaccine development and deployment. It is CEPI’s ambition to forward at least three SARS-CoV-2 vaccines to regulatory submission for general and emergency use within the next 12-18 months. According to them, this will require an investment of at least US$2 billion, financing Phase 1 clinical trials of up to eight vaccine candidates, Phase 2 and 3 trials for up to six candidates, completion of regulatory and quality requirements for at least three vaccines and finally enhancement of global manufacturing capacities for three vaccines [ix]. At the current time, CEPI funds 6 approaches, from which two are in Phase 1, and four still in pre-clinical research.
It is a tough challenge to deliver billions of doses of a new vaccine within a few months of its discovery. It is unlikely, that a single company with its vaccine candidate will be able to tackle this alone. Multiple candidates must therefore be supported financially to advance them to approval.
Production capacities must be built already during clinical phases. As it is not yet clear, how long the market will be there for this new vaccine, such a facility should ideally follow a platform approach and on the longer term enable multi-product production.
In order to tackle the billion dose challenge, it may be very beneficial to re-use existing facilities. For most of the approaches being pursued, this will be hindered by very specific approaches or biosafety concerns. Only nucleic acids, viral vectors and subunit vaccines might be produced in existing facilities, but this will require well-planned facility fitting. Virtual process transfer and facility fitting as enabled by GoSilico offer a promising approach to achieve this.
As things stand at the moment, the industry is very likely not yet prepared to immediately produce a novel vaccine on a pandemic scale and at lowest costs. But as outlined in this article, platform approaches, novel mRNA/DNA technologies and digital twins of downstream processes are encouraging, that there is great hope we at least have the right tools to this.