A modular protein subunit vaccine candidate produced in yeast confers protection against SARS-CoV-2 in non-human primates

TLDR: In this article, the SARS-CoV-2 receptor binding domain (RBD) and hepatitis B surface antigen virus-like particles (VLPs) were used for preclinical testing in cynomolgus macaques.

Abstract: Vaccines against SARS-CoV-2 have been distributed at massive scale in developed countries, and have been effective at preventing COVID-19. Access to vaccines is limited, however, in low- and middle-income countries (LMICs) due to insufficient supply, high costs, and cold storage requirements. New vaccines that can be produced in existing manufacturing facilities in LMICs, can be manufactured at low cost, and use widely available, proven, safe adjuvants like alum, would improve global immunity against SARS-CoV-2. One such protein subunit vaccine is produced by the Serum Institute of India Pvt. Ltd. and is currently in clinical testing. Two protein components, the SARS-CoV-2 receptor binding domain (RBD) and hepatitis B surface antigen virus-like particles (VLPs), are each produced in yeast, which would enable a low-cost, high-volume manufacturing process. Here, we describe the design and preclinical testing of the RBD-VLP vaccine in cynomolgus macaques. We observed titers of neutralizing antibodies (>10 4 ) above the range of protection for other licensed vaccines in non-human primates. Interestingly, addition of a second adjuvant (CpG1018) appeared to improve the cellular response while reducing the humoral response. We challenged animals with SARS-CoV-2, and observed a ~3.4 and ~2.9 log 10 reduction in median viral loads in bronchoalveolar lavage and nasal mucosa, respectively, compared to sham controls. These results inform the design and formulation of current clinical COVID-19 vaccine candidates like the one described here, and future designs of RBD-based vaccines against variants of SARS-CoV-2 or other betacoronaviruses.

Figures

Fig. 1 Design and analysis of the RBD-VLP drug product A) Schematic of protein expression and conjugation. B) Reduced SDS-PAGE analysis of the formulated RBD-VLP vaccine samples. Alum-bound protein antigen (with and without CpG) was separated by centrifugation and desorbed from the alum using an elution buffer combined with heat treatment prior to SDS-PAGE.

Fig. 3 Challenge with SARS-CoV-2 A,C) Peak levels of SARS-CoV-2 sgRNA after challenge for each animal from nasal swabs (A) or bronchoalveolar lavage (C). Statistical significance was determined by a KolmogorovSmirnov test. N.S. = not significant (p > 0.1). Black bars represent median values. Dotted gray line represents limit of detection. B,D) Levels of sgRNA after challenge from nasal swabs (B) or bronchoalveolar lavage (D). Each thin line represents one animal. Thick black lines represent median values. E,F) Correlation of RBD-specific antibody titer and pseudovirus neutralizing antibody titer from week 5 animal sera with peak sgRNA levels in nasal swabs (E) and

Fig. 2 Humoral and cellular immune response to the RBD-VLP vaccine A) Design of the non-human primate study in cynomolgus macaques. B) Titers of RBD-specific antibody binding in animal sera. C) Titers of SARS-CoV-2 pseudovirus neutralizing antibody in animal sera. D) Expression of IFNγ from cells stimulated with S1 protein peptides. Statistical significance was determined by a Kolmogorov-Smirnov test. N.S. = not significant (p > 0.1). Black bars represent median values. Dotted gray line represents limit of detection.

Full text

1
A modular protein subunit vaccine candidate produced in yeast confers protection against
SARS-CoV-2 in non-human primates
Neil C. Dalvie
1,2
^, Lisa H. Tostanoski
3
^, Sergio A Rodriguez-Aponte
2,4
^, Kawaljit Kaur
5
, Sakshi
Bajoria
5
, Ozan S. Kumru
5
, Amanda J. Martinot
3,6
, Abishek Chandrashekar
3
, Katherine
McMahan
3
, Noe B. Mercado
3
, Jingyou Yu
3
, Aiquan Chang
3,7
, Victoria M. Giffin
3
, Felix
Nampanya
3
, Shivani Patel
3
, Lesley Bowman
8
, Christopher A. Naranjo
2
, Dongsoo Yun
2
, Zach
Flinchbaugh
9
, Laurent Pessaint
9
, Renita Brown
9
, Jason Velasco
9
, Elyse Teow
9
, Anthony Cook
9
,
Hanne Andersen
9
, Mark G. Lewis
9
, Danielle L. Camp
2
, Judith Maxwell Silverman
10
, Harry
Kleanthous
11
, Sangeeta B. Joshi
5
,
David B. Volkin
5
, Sumi Biswas
11
, J. Christopher Love
1,2,12
,
Dan H. Barouch
3,7,12,13
1
Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge,
Massachusetts 02139, USA
2
The Koch Institute for Integrative Cancer Research, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139, USA
3
Center for Virology and Vaccine Research, Beth Israel Deaconess Medical Center, Harvard
Medical School, Boston, MA, USA
4
Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge,
Massachusetts 02139, USA
5
Department of Pharmaceutical Chemistry, Vaccine Analytics and Formulation Center,
University of Kansas, Lawrence, Kansas, 66047, USA
6
Departments of Infectious Diseases and Global Health and Comparative Pathobiology,
Cummings School of Veterinary Medicine, Tufts University, North Grafton, MA 01536, USA
7
Harvard Medical School, Boston, MA 02115, USA
8
SpyBiotech Limited, Oxford Business Park North, Oxford, OX4 2JZ, United Kingdom
9
Bioqual, Rockville, MD 20852, USA
10
Bill & Melinda Gates Medical Research Institute, Cambridge, MA 02139, USA
11
Bill&Melinda Gates Foundation, Seattle, WA 98109, USA
12
Ragon Institute of MGH, MIT, Harvard, Cambridge, MA 02139, USA
13
Massachusetts Consortium on Pathogen Readiness, Boston, MA 02115, USA
^Contributed equally
Correspondence to: dbarouch@bidmc.harvard.edu, clove@mit.edu
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted July 14, 2021. ; https://doi.org/10.1101/2021.07.13.452251doi: bioRxiv preprint

2
Abstract
Vaccines against SARS-CoV-2 have been distributed at massive scale in developed
countries, and have been effective at preventing COVID-19. Access to vaccines is limited,
however, in low- and middle-income countries (LMICs) due to insufficient supply, high costs,
and cold storage requirements. New vaccines that can be produced in existing manufacturing
facilities in LMICs, can be manufactured at low cost, and use widely available, proven, safe
adjuvants like alum, would improve global immunity against SARS-CoV-2. One such protein
subunit vaccine is produced by the Serum Institute of India Pvt. Ltd. and is currently in clinical
testing. Two protein components, the SARS-CoV-2 receptor binding domain (RBD) and
hepatitis B surface antigen virus-like particles (VLPs), are each produced in yeast, which would
enable a low-cost, high-volume manufacturing process. Here, we describe the design and
preclinical testing of the RBD-VLP vaccine in cynomolgus macaques. We observed titers of
neutralizing antibodies (>10
4
) above the range of protection for other licensed vaccines in non-
human primates. Interestingly, addition of a second adjuvant (CpG1018) appeared to improve the
cellular response while reducing the humoral response. We challenged animals with SARS-CoV-
2, and observed a ~3.4 and ~2.9 log
10
reduction in median viral loads in bronchoalveolar lavage
and nasal mucosa, respectively, compared to sham controls. These results inform the design and
formulation of current clinical COVID-19 vaccine candidates like the one described here, and
future designs of RBD-based vaccines against variants of SARS-CoV-2 or other
betacoronaviruses.
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted July 14, 2021. ; https://doi.org/10.1101/2021.07.13.452251doi: bioRxiv preprint

3
Introduction
Prophylactic vaccination is effective in eliciting protective immunity against SARS-CoV-
2 and preventing coronavirus disease 2019 (COVID-19) (1). Multiple vaccines have now been
distributed at large scale in many countries, and have resulted in a lower incidence of infection
and severe disease caused by SARS-CoV-2 (2, 3). Access to vaccines remains limited, however,
in low- and middle-income countries (LMICs), where infectious variants of SARS-CoV-2
continue to emerge in large scale outbreaks (4). In addition to financial and logistical support
from developed countries and health organizations, vaccines produced by local manufacturers
could enable the lowest costs for interventions in these countries and potentially minimize the
infrastructure required for their distribution (5–7). Protein subunit vaccines are a promising
solution because they can be manufactured using existing large-scale microbial fermentation
facilities in LMICs (8), typically do not require frozen storage and distribution, and are safe and
effective when used with adjuvants (9, 10). Here, we describe the design and immunogenicity of
a modular protein subunit vaccine, comprising a SARS-CoV-2 spike protein subunit receptor
binding domain (RBD) displayed on a Hepatitis B virus-like particle (VLP) that is constructed
using a covalent peptide-mediated linkage (SpyTag/SpyCatcher) . Both of these vaccine
components are currently produced by microbial fermentation at a large-scale manufacturing
facility in India. We show that this vaccine candidate elicits a strong immune response in
cynomolgus macaques and protects against SARS-CoV-2 challenge. Based on these promising
data, this vaccine candidate is currently being tested in clinical trials (ANZCTR Registration
number ACTRN12620000817943).
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted July 14, 2021. ; https://doi.org/10.1101/2021.07.13.452251doi: bioRxiv preprint

4
Results
Design of an accessible protein subunit vaccine
We sought to design a protein subunit vaccine that would be both suitably immunogenic
and simple to manufacture for affordable distribution in LMICs. Multiple protein vaccines based
on the trimeric SARS-CoV-2 spike protein have demonstrated efficacy, but are manufactured in
insect or mammalian cells, which are difficult to transfer to existing facilities in LMICs (11, 12).
The receptor binding domain (RBD) of the spike protein has been proposed as an alternative to
the full spike protein because it has been shown to elicit multiple potent neutralizing antibodies
directed at multiple epitopes (13–15), and can be manufactured in microbial systems like the
biotechnological yeast Komagataella phaffii (Pichia pastoris) (16, 17). Formulations comprising
only monomeric RBD (Wuhan-Hu-1) and adjuvant tested to date have required three doses to
elicit potent neutralizing responses in humans (18, 19). While further optimization of such
formulations could improve these designs, we and others have demonstrated that multimeric
display of RBD on VLPs can be highly immunogenic (20–24).
Here, we selected Hepatitis B surface antigen (HBsAg) virus-like particles (VLPs) as a
nanoparticle on which to display the RBD. This choice leverages extensive experience with a
previously tested commercial product, GeneVac-B, that is manufactured at low cost and
distributed in LMICs for prevention of Hepatitis B (25). As a model for our design here, we
referenced previously reported designs with this core nanoparticle decorated with a malarial
subunit antigen (26). A polypeptide-based system (SpyTag/SpyCatcher) allowed for covalent
linkage of the antigen to the VLP by a transpeptidation reaction. This general design can increase
antigen-specific antibody titers in mice, and the responses elicited in prior examples were
unaffected by the presence of pre-existing antibodies against HBsAg (27). The modularity of the
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted July 14, 2021. ; https://doi.org/10.1101/2021.07.13.452251doi: bioRxiv preprint

5
SpyTag/SpyCatcher system allows each component of the final particle to be expressed and
purified independently to maximize yields and quality.
We adapted this approach to make a vaccine candidate for COVID-19. We genetically
fused the SpyTag peptide onto the SARS-CoV-2 RBD. This fusion protein was manufactured in
an engineered strain of K. phaffii (28). The RBD-SpyTag and HBsAg-SpyCatcher VLPs were
each purified separately and then conjugated in a GMP process to produce the RBD-VLP antigen
(Fig. 1A). In this study, the RBD-VLP antigen was formulated with two adjuvants: 1) aluminum
hydroxide (alum) and 2) alum combined with CpG1018—a potent commercial TLR9 agonizing
adjuvant known to elicit Th1-like responses (29). Analysis of the formulated vaccine drug
product by SDS-PAGE showed only small fractions (<20%) of unconjugated HBsAg-VLP and
RBD, and complete adsorption of the RBD-VLP antigen onto the alum adjuvant (Fig. 1B). We
detected CpG1018 in both the unbound and bound to alum fractions.
We also performed additional analytics on unformulated RBD-VLP antigen to confirm
antigenicity and nanoparticle formation. The RBD-VLP antigen exhibited strong binding to the
human ACE2 receptor and a known neutralizing antibody CR3022 by biolayer interferometry
(Fig. S1A). The large difference in signal observed in this analysis between the RBD-VLP and
soluble monomeric RBD-SpyTag confirms the multivalency of the RBD conjugated on the VLP.
We also confirmed formation of nanoparticles by electron microscopy (EM) (Fig. S1B-C). These
analytics confirmed the structural attributes of the conjugated RBD-VLPs used for non-clinical
evaluations here.
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted July 14, 2021. ; https://doi.org/10.1101/2021.07.13.452251doi: bioRxiv preprint

Frequently asked questions

Q1. What are the contributions in "A modular protein subunit vaccine candidate produced in yeast confers protection against sars-cov-2 in non-human primates" ?

In this paper, a modular protein subunit vaccine was proposed to prevent SARS-CoV-2 and prevent coronavirus disease 2019 ( COVID-19 ). 

Q2. What was the RNA used to generate the standard?

To generate a standard, a fragment of the subgenomic E gene was synthesized and cloned into a pcDNA3.1+ expression plasmid using restriction site cloning (Integrated DNA Techonologies). 

Q3. How many g of each protein was loaded on the gel?

Approximately 0.1 µg of each sample was loaded on 4-12% Bis-Tris gel (Invitrogen, NP3022) run for 50 min at 150 V in 1X MES-SDS running buffer (Invitrogen, NP0002). 

Q4. What is the effectiveness of a vaccine against SARS-CoV-2?

Prophylactic vaccination is effective in eliciting protective immunity against SARS-CoV-2 and preventing coronavirus disease 2019 (COVID-19) (1). 

Q5. What is the effect of CpG1018 on the antigen stability of the RBD?

Reduced dosing of CpG1018 could lessen the impact on the stability of the RBD antigen while still eliciting balanced Th1 and Th2 cellular responses (36). 

Q6. What are the main reasons why vaccines are inaccessible in southeast Asia?

Vaccines based on mRNA have been shown to be effective against emerging variants (40, 41), but remain largely inaccessible in southeast Asia and Africa due to high costs and cold chain requirements. 

Q7. What is the definition of the titers?

SARS-CoV-2 neutralization titers were defined as the sample dilution at which a 50% reduction in relative light unit (RLU) was observed relative to the average of the virus control wells. 

Q8. What was the effect of the alum and CpG1018 vaccines?

The authors observed significant IFNγ expression in cells from both the alum formulation and the alum and CpG1018 formulation after two doses. 

Q9. How long after infection were the lysates lysed?

48 h after infection, cells were lysed in Steady-Glo Luciferase Assay Reagent (Promega) according to the manufacturer’s instructions. 

Q10. How many copies of sgRNA were used per mL of BAL fluid?

C for 1 s and 60º C for 20 s. Standard curves were used to calculate sgRNA copies per mL of BAL fluid or per swab; the quantitative assay sensitivity was 50 copies per mL or per swab. 

Q11. What is the effect of the conjugated RBD-VLP vaccine?

The formulation with alum and CpG1018 at the concentrations used here appears to have altered the antigenicity of the RBD-VLP antigen, potentially leading to a reduced humoral immune response. 

Q12. What was the process used to produce the SpyCatcher nanoparticles?

Production of unformulated drug substance Hepatitis B Surface Antigen (HBsAg) SpyCatcher nanoparticles were produced in a GMP process at the Serum Institute of India. 

Q13. What is the role of the RBD-VLP vaccine booster?

There is growing interest, therefore, in the utility of RBD-based vaccine boosters to provide immunity against emerging variants of SARS-CoV-2 that may escape antibody responses induced by vaccination against the Wuhan-Hu-1 or D614G spike protein. 

Q14. What was the sgRNA score for each lung lobe?

Two lung lobes (1 section from the right and left caudal lung lobes) were assessed and scored (1-4) for each of the following lesions:1) Interstitial inflammation and septal thickening 2) Eosinophilic interstitial infiltrate 3) Neutrophilic interstitial infiltrate 4) Hyaline membranes 

Q15. What is the way to prevent COVID-19?

This modular design of protein antigens paired with existing large-scale manufacturing capacity in yeast demonstrates a vaccine platform that could keep pace with an evolving COVID-19 pandemic, and provide better access to COVID-19 vaccines around the globe. 

Q16. How many cycles of 95oC were used?

Reactions were carried out in duplicate for samples and standards on the QuantStudio 6 and 7 Flex Real-Time PCR Systems (Applied Biosystems) with the following thermal cycling conditions: initial denaturation at 95ºC for 20 s, followed be 45 cycles of 95º 

Q17. What was the effect of the alum vaccine on the cynomolgu?

The authors observed ~30% less binding to human ACE2 for the alum and CpG1018 formulation compared to the formulation with alum only (competitive ELISA) (Fig. S2C). 

Q18. What is the role of the vaccine candidate in preventing COVID-19?

The authors show that this vaccine candidate elicits a strong immune response in cynomolgus macaques and protects against SARS-CoV-2 challenge. 

Q19. What is the effect of the RBD-VLP vaccine on the upper respiratory tract?

Both formulations of the RBD-VLP vaccine significantly reduced the levels of detected sgRNA in the upper respiratory tract (Fig. 3A), and no sgRNA was detected after day 4 post challenge from the group immunized with RBD-VLP formulated with alum (Fig. 3B). 

Q20. What is the reason for the difference in the antigenicity of the RBD?

The authors attributed this difference to a reduction in antigenicity of the RBD; inclusion of CpG may impact the structural integrity of the RBD, leading to altered antigenicity. 

Q21. How long before the vaccine was formulated?

Another multimeric RBD vaccine has reported robust humoral responses in nonhuman primates with the same dose of CpG1018 (35), but in that instance, the vaccine was formulated 30 minutes prior to injection.