From concept to delivery: a yeast-expressed recombinant protein-based COVID-19 vaccine technology suitable for global access

ABSTRACT

Introduction

The development of a yeast-expressed recombinant protein-based vaccine technology co-developed with LMIC vaccine producers and suitable as a COVID-19 vaccine for global access is described. The proof-of-concept for developing a SARS-CoV-2 spike protein receptor-binding domain (RBD) antigen as a yeast-derived recombinant protein vaccine technology is described.

Areas Covered

Genetic Engineering: The strategy is presented for the design and genetic modification used during cloning and expression in the yeast system. Process and Assay Development: A summary is presented of how a scalable, reproducible, and robust production process for the recombinant protein COVID-19 vaccine antigen was developed. Formulation and Pre-clinical Strategy: We report on the pre-clinical and formulation strategy used for the proof-of-concept evaluation of the SARS-CoV-2 RBD vaccine antigen. Technology Transfer and Partnerships: The process used for the technology transfer and co-development with LMIC vaccine producers is described. Clinical Development and Delivery: The approach used by LMIC developers to establish the industrial process, clinical development, and deployment is described.

Expert Opinion

Highlighted is an alternative model for developing new vaccines for emerging infectious diseases of pandemic importance starting with an academic institution directly transferring their technology to LMIC vaccine producers without the involvement of multinational pharma companies.

KEYWORDS:

• SARS
• SARS-CoV-2
• COVID-19
• vaccine
• recombinant protein
• microbial fermentation
• CORBEVAX
• INDOVAC

1. Introduction

The COVID-19 pandemic highlighted profound vaccine inequities, resulting in delayed or absent COVID-19 vaccine access for low- and middle-income countries (LMICs) [1,2]. The causes remain under investigation but among them were public policies that failed to adequately build a vaccine ecosystem committed to enhanced support for local LMIC vaccine production and manufacture [3]. In some cases, however, this was achieved either through transfer of new technologies such as adenovirus or particle platforms from multinational pharma companies to LMIC producers, or through collaborations between product development partnerships (PDPs) (comprised of NGOs, academic institutions, etc.) and LMIC producers [4]. In the second case, local production relied on developing and producing at-scale COVID-19 vaccines that employ indigenous technologies.

Texas Children’s Hospital Center for Vaccine Development (Texas Children’s CVD) at Baylor College of Medicine is a 22-year-old nonprofit and academic health center-based PDP with a track record of developing and testing new vaccines for poverty-related neglected and emerging tropical infectious diseases, including hookworm infection, schistosomiasis, leishmaniasis, and Chagas disease [5,6]. Based at the Texas Medical Center in Houston, TX, Texas Children’s CVD builds its vaccine technologies to be compatible with the technologies at the producers and manufacturers based in LMICs. This includes recombinant protein production through microbial fermentation in yeast, a process already used to make recombinant hepatitis B vaccine in multiple nations [7,8]. Here, we report on the development pathway taken to develop a COVID-19 vaccine technology by working primarily with LMIC vaccine producers. A summary of the achievements and challenges met to bring this vaccine technology from concept to delivery is presented where ultimately, two COVID-19 human vaccines received emergency use authorization in India and Indonesia, leading to the administration of close to 100 million doses to both children and adults [9,10].

2. The beginnings: A recombinant protein SARS vaccine

Our interest in developing new coronavirus vaccines emerged following the SARS outbreak of 2002–3 and the rapid identification of SARS-CoV as the virus etiologic agent. The high mortality rate of SARS and its ability to travel from where it first emerged in southern China to Canada and elsewhere incentivized multiple virology laboratories to pursue vaccine concepts. These included whole-inactivated virus (WIV) vaccines, as well as those produced in Venezuelan equine encephalitis virus vectors, modified vaccinia virus Ankara, adenovirus, and other viral vectors, each expressing selected genes encoding specific SARS-CoV antigens. The major vaccine candidates included the spike (S) and nucleocapsid (NP) proteins. Studies conducted at the New York Blood Center (NYBC) revealed the receptor-binding domain (RBD) of the S protein offered advantages over the full-length protein as a vaccine target [11]. Among them was the ability of the RBD to be delivered either in a viral vector or as a recombinant subunit vaccine and inducing high levels of virus-specific neutralizing antibodies, together with the observation that convalescent sera from SARS patients contained neutralizing antibodies specific to the RBD [11]. In addition, when delivered as a subunit vaccine, the RBD induced protection against virus-challenge infections in rodent models, with significantly reduced or absent eosinophilic immunopathology or immune enhancement, which was noted to occur with WIVs, selected virus-vectored vaccines, or even the full-length S protein [11]. It was further noted that the virus-neutralizing antibodies induced by RBD vaccines in rodents could cross-neutralize other coronaviruses.

Through support from the National Institute of Allergy and Infectious Diseases of the US National Institutes of Health (NIAID-NIH), Texas Children’s CVD, together with the NYBC, the University of Texas Medical Branch in Galveston, and Walter Reed Army Institute of Research (WRAIR), created a consortium to accelerate a SARS vaccine. It was based on the concept of employing the RBD expressed in yeast (Pichia pastoris X33) as the lead candidate antigen [11–14]. Texas Children’s CVD selected two related SARS RBD antigens RBD193 and RBD 219 on alum adjuvants for further study in a mouse model [12]. The highest levels of virus-neutralizing antibodies were achieved through the expression of RBD219-N1 by deleting the first asparagine responsible for the N-linked glycosylation [10]. Mouse polyclonal antibodies against RBD219-N1 neutralized at high titers both live virus in Vero E6 cells as well as pseudovirus constructs [12]. Following selection of this antigen for further development, the yield of the RBD219-N1 was optimized in a bioreactor and produced without the need for a polyhistidine tag and through a process suitable for current good manufacturing practice (cGMP), toxicology testing, and clinical testing [13]. When formulated with aluminum hydroxide, the RBD219-N1 antigen was shown to fully protect Balb/c mice against a SARS-CoV lethal challenge infection compared to negative controls [14]. Moreover, the aluminum hydroxide (Alhydrogel®) formulation of the antigen was shown to be superior to other adjuvants in terms of both inducing specific antibodies and minimizing eosinophilic immune enhancement [14]. It was further noted that formulating the RBD219-N1 antigen at a ratio of 1:25 with Alhydrogel® induced the highest virus-neutralizing antibodies and afforded complete protection that resulted in non-detectable virus loads, together with no or minimal eosinophilic immune enhancement [14]. Ultimately, this protein antigen was produced under cGMP at WRAIR and remains stable at −80°C since 2016. Lack of commercial interest in the antigen or SARS vaccinations has thwarted further product development or clinical testing.

3. Pivoting to SARS-CoV-2 (COVID-19)

When COVID-19 emerged and the sequence of the SARS-CoV-2 was revealed publicly in 2020, there was interest in repurposing SARS RBD219-N1 as a potential heterologous vaccine antigen to prevent SARS-CoV-2 infection [15]. SARS RBD219-N1 was determined to share 75% identity and 83% similarity with the corresponding SARS-CoV-2 antigen, with some monoclonal antibodies developed to the SARS-CoV RBD cross-reacting to SARS-CoV-2 antigens [15]. Diminished enthusiasm from the scientific community and policymakers for this concept arose from concerns that immunization with a heterologous antigen might pose a theoretical risk for inducing immune enhancement, especially when formulated on a Th2-inducing adjuvant such as alum or aluminum hydroxide. However, such concerns failed to incorporate experimental observations that eosinophilic enhancement linked to coronavirus immunizations was more associated with Th-17 host responses, and aluminum adjuvants reduce immune enhancement in rodent models of virus challenge infections [16–18]. Despite concurrence from an independent NIH study panel regarding these latter observations [19], the SARS RBD219-N1 antigen remains untested in human clinical trials as a potential COVID-19 vaccine antigen.

In lieu of pursuing SARS RBD219-N1 as a COVID-19 vaccine, Texas Children’s CVD developed a new RBD specific for SARS-CoV-2 and COVID-19. The enthusiasm for this approach was based on past successes for the yeast-derived RBD for experimental SARS infections [11–14,20], together with the reality that antigen production methods through microbial fermentation in yeast are widely employed in LMICs to produce recombinant protein hepatitis B vaccines [21]. Therefore, our rationale was that successful development of a SARS-CoV-2 RBD vaccine antigen could accelerate vaccine production locally in many LMICs and help in global efforts to achieve COVID-19 vaccine equity.

To begin, three different constructs based on the ancestral SARS-CoV RBD219 were genetically engineered to match the original SARS-CoV-2 (Wuhan) sequence (Figure 1). They included a wild-type molecule and an N1 deleted version [22]. As with the ancestral SARS antigen, the deletion of the N-terminal asparagine (N331 of the Spike protein) prevented hyperglycosylation that could interfere with both immunogenicity and production consistency. However, in addition, we found that a cysteine residue in the 538-position caused the formation of oligomers, and aggregation through intermolecular disulfide bonds. Substituting an alanine residue for the cysteine corrected this problem [22]. Therefore, a new antigen known as SARS-CoV-2 RBD219-N1C1 offered multiple advantages compared to the wild-type molecule, including higher yields in fermentation, absence of hyperglycosylation and oligomer formation, as well as reduced aggregation [22]. The new molecule was also amenable to purification with high yields and purity. Moreover, subsequent studies found that the protein antigen binds to the angiotensin-converting enzyme-2 (ACE-2) receptor (considered the major host receptor of coronavirus in human tissues), while antibodies generated against the protein in mice exhibited strong virus-neutralization in a pseudovirus assay with IC50 values in the 1:1,000 to 1:10,000 range [23]. When formulated on alum, or alum together with a second immunostimulant such as a CpG1826 oligonucleotide (Invivogen, San Diego, USA), the SARS-CoV-2 RBD219-N1C1 antigen induces a balanced Th1/Th2 cellular immune response with high levels of cross-neutralizing antibodies against pseudovirus (comprised of a non-replicating lentiviral particle with SARS-CoV-2 spike) variants corresponding to alpha (B.1.17), Beta (B.1.351), Delta (B.1.617.2), and BA.1 Omicron (B.1.1.529) variants [24]. A similar molecule in a CpG formulation also induced protective antibodies in aged mice [25].

Figure 1. COVID-19 vaccine development pathway. The stages of vaccine development from the laboratory to the manufacturer. 01. Genetic engineering. 02. Process and assay development. 03. Formulation and pre-clinical strategy. 04. Technology transfer and partnerships. 05. Industrial manufacturing and clinical development.

Scale-up fermentation and process development of the SARS-CoV-2 RBD219-N1C1 antigen in P. pastoris was conducted at the 1-L and 5-L levels and found to be associated with high yields (>400 mg/L of fermentation supernatant) and target product recovery after purification through combined hydrophobic-interaction, size-exclusion, and ion-exchange chromatography was ~140 mg/L [26]. The purified molecule was analyzed and characterized by multiple biophysical and other methods, including dynamic light scattering, HPLC, and ACE-2 affinity [26].

In parallel, the RBD219-WT antigen, identical to RBD219-N1C1, but missing the two mutations of that construct, was evaluated for protective immunity in a non-human primate SARS-CoV-2 challenge model, in collaboration with the Yerkes National Primate Research Center at Emory University [27]. Since the WT antigen was the first one available for animal testing, we chose it for the non-human primate studies in order to avoid delays in our development program. For these studies, the antigen was formulated on aluminum hydroxide together with an immunostimulatory alum-adsorbed TLR-7/8-agonist, 3 M–052 (provided by 3 M Corporate Research and formulated at Access to Advanced Health Institute, Seattle, USA) [27]. After inducing high levels of virus-neutralizing antibodies together with Th1-biased CD4+ and cytotoxic CD8+ T cell responses, the vaccine-induced significant protection as evidenced by a reduction in virus titers in nasal secretions as well as bronchoalveolar lavage (BAL) samples. In addition, reduced lung pathology compared to controls was observed [27]. Anti-RBD antibodies inversely correlated not only with virus nucleic acid levels in BAL fluid but also with the amount of viral mRNA in nasal secretions. Protective immunity was further associated with RBD-specific plasma cells in draining lymph nodes and other relevant immunological findings [27].

4. Target product profile and technology transfer

Based on evidence for its protection in preclinical animal studies, and the anticipated ease of production and large-scale purification, in addition to low-cost and simple refrigeration storage, for the vaccine formulation, the SARS-CoV-2 RBD219-N1C1 antigen gained international attention as a vaccine candidate for technology transfer to LMIC producers in India, Indonesia, Bangladesh, and Botswana, respectively. In the cases of India and Indonesia, the antigen was formulated with aluminum hydroxide and CpG1018, a TLR-9 agonist, from Dynavax Technologies (Emeryville, CA) for toxicology evaluation and clinical testing and to request emergency use authorization.

In India, Biological E led the production and clinical testing of the SARS-CoV-2 RBD219-N1C1 vaccine which was branded as CORBEVAX [28–30]. In phase 1 and 2 studies, healthy adults receiving two doses (28 days apart) exhibited only a low incidence of adverse events, with no grade-3 or serious events observed. Humoral antibody and cellular responses increased in a dose-dependent manner, and antibody titers remained significantly elevated with minimal change in anti-RBD IgG and neutralizing antibody titers for at least 6 months following the second dose [28]. Neutralizing antibodies matched or exceeded those of convalescent plasma from recovered COVID-19 patients or those reported from clinical trials of other COVID-19 vaccines [28]. In consultation with the Drug Controller General of India (DCGI), India’s national regulatory authority, Biological E conducted a phase 3 superiority study comparing CORBEVAX with COVISHIELD, the AstraZeneca COVID-19 vaccine produced by the Serum Institute of India [29]. This was a prospective, single-blinded, randomized, active control study across 20 sites in India among healthy adults aged 18–80 years. Both neutralizing antibodies (versus both the original Wuhan virus and Delta variant) and interferon-gamma-secreting peripheral blood mononuclear cells were higher in the CORBEVAX recipients [29]. Levels of the former were equivalent to 90% protection (versus symptomatic infection caused by the original Wuhan virus lineage) based on correlations noted for other comparison COVID-19 vaccines such as the Moderna mRNA vaccine [29]. Similar findings were noted among children and adolescents between 5 and 18 years of age [30], and neither in adults nor children were severe or serious adverse events observed, except for two events in adults – a case of dengue fever and a femur fracture – each unrelated to the vaccine [29,30]. On this basis, CORBEVAX was authorized for emergency use in India in December 2021, with the first doses of the two-dose vaccine administered to adolescents 12–14 years of age in March 2022. As of 5 December 2022, more than 73 million doses of the vaccine have been administered, with more than 41 million children having received the first dose and more than 32 million children having received their second dose [9]. CORBEBVAX has since been approved for children aged 5–11 years, and also as a heterologous booster for adults who would have previously received either COVISHIELD or COVAXIN, India’s WIV produced by Bharat Biotech.

A second version of our RBD vaccine protein technology, where the antigen and production cell bank were slightly modified through the removal of a sequence of 16 amino acids at the C-terminus [31], was transferred to BioFarma in Indonesia. This version recently underwent phase 3 clinical trials for immunogenicity and effectiveness, and emergency use authorization of this vaccine, named IndoVac, was obtained in the fall of 2022. Since the yeast-derived antigen is cloned and produced without animal or human cells or derivatives, it was designated as halal and is thus suitable for use or export in Muslim-majority countries and may also be effective in reducing vaccine hesitancy among those that avoid animal-derived medicines for ethical reasons. Figure 1 summarizes the development pathway taken to develop a COVID-19 vaccine technology by working primarily with the LMIC vaccine producers as described above.

5. Conclusions

CORBEVAX and IndoVac represent the first two COVID-19 vaccines specifically intended for scale-up and production in LMICs [8]. The World Health Organization is working with the LMIC producers to complete its evaluation for global emergency use authorization and prequalification of CORBEVAX and IndoVac. Currently, new-generation RBD-based vaccine technologies adapted at Texas Children’s Center for Vaccine Development to the emerging omicron variants have been transferred to the LMIC producers and are currently undergoing evaluation either as monovalent or bivalent vaccines.

6. Expert opinion

A noted feature of these yeast-produced vaccines is that they represent a vaccine development model for new or emerging pathogens, which were produced indigenously, with locally available technologies in an LMIC setting, and with the technical support of an academically anchored nonprofit PDP, prior to licensure or emergency authorization. Previously, the meningococcal A vaccine for Africa known as MenAfriVac was among the first vaccines developed through a similar model, a partnership between the Serum Institute of India and PATH, a Seattle-based PDP [32]. During the COVID-19 pandemic, several new WIV vaccines were developed and produced in China and India, respectively, as were recombinant vaccines for Cuba and Taiwan (although for much smaller populations) [7,33]. Such LMIC-led activities provide proof-of-concept that it is feasible to develop from scratch or in collaboration with PDPs – nonprofits that use industry practices to make vaccines, drugs, or diagnostics [6] – new vaccines for pandemic threats. Growing this sector allows the global vaccine ecosystem to expand beyond the multinational pharma companies and become faster and nimbler in protecting vulnerable populations living in poverty. LMIC producers can respond more effectively to emerging threats when they can rely on existing infrastructure and can work within an existing regulatory framework that does not require time-consuming de novo evaluations. This not only addresses vaccine equity but also reduces vaccine hesitancy fueled by an increasing distrust arising from the perceived denial of access to COVID-19 vaccines from the US and Europe during the first 2 years of the pandemic [1–4].

Areas for improvement include better addressing financial and infrastructure support for LMIC producers, with efforts focused on expanding their portfolio of technologies while maintaining quality control and safety standards. In parallel, there is an urgency to continue building capacities for national regulatory authorities in LMICs. In regard to COVID-19, areas for further research and future studies include adapting CORBEVAX, IndoVac, or other COVID-19 vaccines developed and produced in LMICs for future variants of concern. Studies are in progress between the Texas Children’s CVD and Biological E to develop and produce a yeast-recombinant protein corresponding to the RBD of the Omicron BA.4 and BA.5 subvariants spreading globally at the time of this writing [34] and to examine this antigen as a potential booster for CORBEVAX, or to create a bivalent booster together with the original Wuhan RBD antigen. There is also the potential to build on the current platforms to develop and produce universal coronavirus vaccines. For example, we are pursuing a multivalent recombinant protein vaccine comprising the RBDs from SARS, SARS-CoV-2, and potentially SARS-CoV-2 omicron or one of the other coronaviruses found among bats and other animal reservoirs in East Asia [35].

In addition, the Texas Children’s CVD RBD antigen is being evaluated as a component of an oral/mucosal COVID-19 vaccine, developed by MigVax in Israel [36]. Such vaccines could be scaled and produced rapidly at low cost and therefore are highly suitable for LMICs. Given how COVID-19 is now the third major serious coronavirus infection to emerge in the 21st century, we should anticipate the possibility that additional SARS-like coronavirus infections will emerge.

Article highlights

• The COVID-19 pandemic revealed profound vaccine inequalities due in part to the absence of new vaccines developed and produced by manufacturers in low- and middle-income countries (LMICs).

• Texas Children’s CVD works with LMIC vaccine producers to develop recombinant protein vaccines through microbial fermentation in yeast. So far, more than 73 million doses have been administered for primary immunizations in children, as well as several million adult booster doses.

• Proof-of-concept for the protective immunity elicited by yeast-derived spike protein receptor-binding domain (RBD) antigens for both severe acute respiratory syndrome (SARS) and COVID-19 caused by the SARS-CoV-2 coronavirus was obtained in animal challenge models.

• The modifications for each RBD improve immunogenicity and production efficiency and reduce costs.

• The SARS-CoV-2 RBD219-N1C1 antigen for COVID-19 was technology transferred to Biological E in India, where it was produced at an industrial scale and formulated with aluminum hydroxide and CpG1018 (Dynavax Technologies) to produce CORBEVAX.

• In phase 1-2 clinical trials, CORBEVAX was shown to be safe and immunogenic and was authorized for emergency use following a phase 3 superiority study with a comparator COVID-19 vaccine produced in India.

• In Indonesia, the vaccine producer BioFarma developed IndoVac, a similar RBD COVID-19 vaccine with Texas Children’s CVD technology, where it has received emergency use authorization. IndoVac takes advantage of its vegan properties to designate it as halal for Muslim-majority countries.

• The activities reported here highlight an alternative for developing new vaccines for emerging infections outside of the multinational pharma companies.

• Now, this approach is being tapped to create variant-specific booster shots, possibly including one to prevent the Omicron BA.5 subvariant, as well as universal coronavirus vaccines.

Declaration of interest

The team of scientists at Texas Children’s Hospital Center for Vaccine Development including its co-directors, Professors Peter Hotez and Maria Elena Bottazzi, are co-inventors of a COVID-19 recombinant protein vaccine technology owned by Baylor College of Medicine (BCM) that was recently licensed by BCM non-exclusively and with no patent restrictions to several companies committed to advance vaccines for low- and middle-income countries. The co-inventors are not involved in license negotiations conducted by BCM. Similar to other research universities, a long-standing BCM policy provides its faculty and staff, who make discoveries that result in a commercial license, a share of any royalty income. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or material discussed in the manuscript

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was funded by Fifth Generation, Inc, JPB Foundation, NIH/NIAID (AI14087201) and the Robert J. Kleberg, Jr. and Helen C. Kleberg Foundation, as well as philanthropic funds received by and intramural funding from Texas Children’s Hospital Center for Vaccine Development at Baylor College of Medicine.