Comparison of DNA vaccines with AS03 as an adjuvant and an mRNA vaccine against SARS-CoV-2

Summary Emerging variants of SARS-CoV-2 call for frequent changes in vaccine antigens. Nucleic acid-based vaccination strategies are superior as the coding sequences can be easily altered with little impact on downstream production. mRNA vaccines, including variant-specific boosters, are approved for SARS-CoV-2. Here, we tested the efficacy of DNA vaccines against the SARS-CoV-2 Spike aided by the AS03 adjuvant using electroporation and compared their immunogenicity with an approved mRNA vaccine (mRNA-1273). DNA vaccination elicited robust humoral and cellular immune responses in C57BL/6 mice with Spike-specific antibody neutralization and T cells produced from 20 μg DNA vaccines similar to that from 0.5 μg mRNA-1273. Furthermore, a Nanoplasmid-based vector further increased the immunogenicity. Our results indicate that adjuvants are critical to the efficacy of DNA vaccines in stimulating robust immune responses against Spike, highlighting the feasibility of plasmid DNA as a rapid nucleic acid-based vaccine approach against SARS-CoV-2 and other emerging infectious diseases.

iScience Article in protein expression in the culture supernatant ( Figure 1B). The twin-Strep-tagged Spike proteins were purified using Strep-Tactin XT 4 flow kit and concentrated using Amicon centrifugal spin columns ( Figure 1C). We confirmed the monomeric state of the purified Spike protein by SDS-PAGE ( Figure 1D) and validated its expression in vitro using 293T cells by western blot ( Figure 1E). Similarly, pooled sera of C57BL/6 mice immunized with 2-doses of Spike-DNA vaccines (20 mg per mice at two week-intervals; n = 5 animal group) show enhanced binding to Spike displayed on the cell surface of 293T ( Figures 1E and 1F). This suggests that some Spike proteins are anchored onto the plasma membrane even without the transmembrane (TM). 10 Next, qRT-PCR data show that mRNA expression levels of inflammatory cytokines IL6 and TNF-a were significantly elevated by Spike-plasmid in all the transfected cell lines compared with control ( Figures 1G and 1H).
Spike-DNA vaccination leads to enhanced humoral immune response and neutralizing activity against SARS-CoV-2 in the sera of vaccinated mice We immunized C57BL/6 mice, following a schedule of prime-boost at 2-week intervals, with three types of vaccines. For the DNA groups, 2 mg (low dose) or 20 mg (high dose) of the Spike-plasmid 10 or a control plasmid with or without AS03 were delivered IM followed by EP using the ICHOR electroporation system. 11 The protein subunit group mice were IM vaccinated with either 0.5 mg or 5 mg of the purified HexaPro Spike protein with AS03. The mRNA groups received either 0.5 mg or 5 mg of mRNA-1273 IM ( Figure 2A, Table 1). Antigen expression was confirmed by measuring the anti-Spike IgG levels in mice sera before immunization and 5 days after the first vaccination ( Figure S1). Four weeks after the second vaccination, mice sera were collected to detect anti-Spike total IgG, IgG subtypes IgG1, IgG2b, and IgG2c by ELISA ( Figures 2B-2E). Immunization with 2 mg DNA resulted in a significant increase in anti-Spike IgG than the control, but lower than that with 20 mg DNA. The addition of AS03 increased the levels in each IgG category for the DNA vaccines. In either low dose or higher dose, the mRNA vaccines exhibit the best or comparable IgG levels, compared to DNA or protein vaccines with AS03 ( Figures 2B-2E). Consistent with the literature, 12 all vaccinated mice showed a Th1-bias response based on the ratio of IgG2c and IgG1 ( Figure 2F). We did not find a difference in the magnitude of IgG subtype responses between the vaccine groups. We noted a Th1-type bias in the 20 mg DNA with AS03 groups comparable to that of the 5 mg mRNA group ( Figure 2F).
To evaluate the anti-Spike antibodies binding to the Spike protein exogenously expressed on 293T cells, we tested the binding ability of serum antibodies of immunized mice by flow cytometry ( Figure S2). Pooled sera of 2 mg or 20mg DNA-vaccinated mice allowed the detection of 25% or 40% of 293T cells that expressed Spike. Sera from AS03-adjuvanted DNA vaccination increased the numbers to 33% or 61.1%. The sera from mice immunized with 0.5 mg or 5 mg mRNA-1273 showed 29% or 96.6%, compared to 31% or 54.6% from the 0.5 mg or 5 mg protein group ( Figure S2).
Next, we evaluated the functional activity of the antibodies using the cPass neutralization assay, which measures the reduction in the binding of Spike (the viral receptor binding protein) to the human ACE2 receptor by neutralizing antibodies (nAbs). The percentage of neutralization from anti-Spike sera (nAbs) of the 20 mg DNA group reached that from 0.5 mg mRNA, but less than that of the 5 mg mRNA or 5 mg protein group, which potentially had the highest levels of nAbs ( Figure 2G). A significant positive correlation between anti-Spike total IgG and the percentage of neutralization was observed among various groups ( Figure 2H; Pearson correlation coefficient r = 0.95, p < 0.010).

DNA vaccination induced cellular responses against SARS-CoV-2 Spike in splenocytes of vaccinated mice
We next evaluated the induction of systemic cytokines and cellular responses in mice at 4 weeks after the second immunization using TNFa and IFNg double ELISpot assay and flow cytometric analysis of splenocytes. Isolated splenocytes were stimulated with 5 mg/mL of the Spike peptide pool for 24 h. ELISpot  4D). There was a positive correlation between the percentages of neutralization and IFNg-producing CD4 and CD8 T cell populations among various vaccinated groups (nAb: IFNg+ CD8 T cells, p < 0.05, r = 0.73) (nAb: IFNg+ CD4 T cells, p < 0.05 r = 0.74; Figures 4E and 4F). Notably, 20 mg DNA with AS03 led to more IFNg-positive CD4 + or CD8 + T cells than 0.5 mg mRNA but fewer than 5 mg mRNA.
We next characterized the induction of systemic cytokines in response to the SARS-CoV-2 Spike antigen using ELISA. Splenocytes of vaccinated mice were stimulated with the Spike protein (5 mg/mL), and cytokine secretion in cell culture supernatants was detected 3 days post-stimulation. Three Th1 cytokines (TNFa, IL-12p40, and IFNg) and one Th2 cytokine (IL-6) were measured. All were significantly increased in the 20 mg DNA with the AS03 group at a level comparable to 0.5 mg mRNA or 5 mg protein but lower than 5 mg mRNA ( Figures 5A-5D). The level of IL-6 was an order of magnitude lower than those of Th1 cytokines. These data suggest that 20 mg DNA with AS03 led to an enhanced cellular immune response skewed to the Th1 response when exposed to the SARS-CoV-2 Spike antigen.
Nanoplasmids expressing spike elicit a stronger humoral and cell-mediated immune response than the conventional plasmid in the vaccinated mice We further improved the Spike-DNA Vaccine using the Nanoplasmid technology and compared its efficacy with the conventional plasmid (DNA: paH-based 10 ) in vitro and in vivo. Nanoplasmids are small circular DNA constructs without antibiotic-resistance genes that can be produced in large yields. The open reading frame of HexaPro Spike was cloned into the Nanoplasmid NTC9385R-eRNA41H-CpGRNA vector (Nano-DNA; Figure S3). Expi293T cells transfected with Nano-DNA produced more Spike after purification than paH-Spike (DNA) ( Figure S5A). C57BL/6 mice immunized with 5 mg Nano-DNA had elevated Spike antigen levels in their sera compared to mice with 5 mg DNA ( Figure S5B). Next, we immunized C57BL/6 mice twice, following a schedule of prime-boost at 2-week intervals with 5 mg of DNA or Nano-DNA with or without AS03 using EP. At 4 weeks post-second vaccination, the sera were collected from immunized mice to detect the anti-Spike IgG, IgA, and IgG subtypes using ELISA. Mice vaccinated with Nano-DNA showed higher levels of anti-Spike IgG, IgA, IgG1, IgG2b, and IgG2c than DNA ( Figures 6A-6E). The AS03 adjuvant further increased the antibody levels ( Figures 6A-6E). We then characterized the cellular response of systemic cytokines in response to vaccination. Flow cytometry revealed that the percentages of TNFa-and IFNg-positive cytotoxic CD8 + T cells followed the rank of Nano-DNA+AS03, Nano-DNA, DNA+AS03, and DNA ( Figures 6F and 6G). The percentages of IFNg-and TNFa-positive CD4 + T cells were elevated in the Nano-DNA+AS03 group compared to Nano-DNA alone or DNA+AS03 groups ( Figures 6H and 6I). Similar patterns were observed for the percentages of neutralization ( Figures 6J and  6K). The highest neutralization was achieved by Nano-DNA+AS03. Again, the neutralization levels were positively correlated with the numbers of IFNg-secreting CD8 + cells ( Figure 6L) or CD4 + T cells ( Figure 6M).
Next, we measured the induction of systemic cytokines TNFa and IFNg using ELISpot. Splenocytes from mice immunized with Nano-DNA+AS03 produced more TNFa-or IFNg-positive splenocytes than other  iScience Article groups ( Figures 7A-7D). These data suggest that Nano-DNA with or without AS03 show enhanced cellular immune responses than their conventional counterparts.

DISCUSSION
The past few years have observed the unparalleled successes of mRNA-based COVID-19 vaccines against SARS-CoV-2. Yet there is still a disparity in vaccine equity, access, and supply to low-and middle-income countries, as over 2 billion people remain completely unvaccinated. The anti-Spike IgG and nAb levels were correlated with the percentage of protection of the vaccines, 13 and the severity of COVID-19 inversely correlates with anti-Spike antibodies. 14 The two mRNA vaccines produced about 4-fold nAbs over the convalescent plasma and showed over 94% efficacy, much higher than conventional vaccines based on the inactivated virus. 15,16 However, mRNA vaccines need ultra-low temperatures for storage, a barrier for many low-or middle-income countries. DNA vaccines delivered via EP are thermally stable and easy to manufacture, yet their efficacies in protection are not on par with mRNA vaccines. 17 The INO-4800 DNA vaccine from Inovio Inc. elicited favorable T cell responses, but the nAb levels were only $10% of the convalescent plasma. 18 A needle-free intradermal COVID-19 DNA vaccine (ZyCov-D) developed by Cadila Healthcare Inc. had low humoral and cellular immune responses with a 1,000 mg dosage. In comparison, the 2,000 mg ZyCov-D dose group with three injections had moderate IgG, nAb, and T cell responses, with a protective efficacy of 66.6% in patients. 19 Notably, the Spike antigens from these two DNA vaccines are not explicitly stated to contain the two proline (S-2P) substitutions critical for maintaining the Spike protein in a prefusion conformation. 20,21 In this study, we used DNA plasmids expressing the newest prefusion stabilized Spike antigen (HexaPro, the 2P form plus four additional residues substituted by proline). 10,22 HexaPro Spike is more stable than S-2P. Preclinical studies have proved that HexaPro is an ideal candidate for development of new vaccines. 23,24 In human DNA vaccine clinical trials, 2,000 mg of DNA were frequently used, whereas 30 or 100 mg of mRNA were included in the two approved mRNA vaccines. Specific to mRNA-1273, doses ranging from 0.0025 to 20 mg per mouse were tested with 0.2 and 1.0 mg inducing robust pseudovirus-neutralizing activity and CD8 T+ cell responses. 25 Thus, we considered 20 mg DNA in mice are proportional to 2,000 mg in humans, comparable to 0.5 mg of mRNA in mice when scaled to 50 mg in humans. There are over 20 studies testing DNA vaccines against Spike of SARS-CoV-2 (Table S1), yet none has compared DNA with mRNA. This is the first study to simultaneously compare the immunogenicity of DNA, mRNA, and protein vaccines in mice. We demonstrate that 20 mg of DNA plus AS03 achieve IgG and neutralization at levels similar to 0.5mg of mRNA ( Figure 2, Table 1).
Adjuvants are critical for improving the quality and magnitude of immune responses. 26 There are no adjuvants in INO-4800 or ZyCov-D. We noticed that the Inovio group added a CC-chemokine receptor 10 (CCR10) as an adjuvant in their upgraded synthetic DNA vaccine against SARS-CoV-2. 27 The mRNA vaccines are mainly composed of cationic lipids like SM-102, stabilizer lipids like PEG2000-DMG, helper lipids (distearoylphosphatidylcholine), and cholesterol, each serving distinct functions. 28 Ionizable lipid SM102 facilitates the intracellular delivery of LNPs, while other helper lipids stabilize the LNP. 29 Protein antigens delivered with LNPs (without encapsulation) elicited strong T helper cell and humoral responses, indicating that LNPs within mRNA vaccines not only facilitate delivery but also possess the strong adjuvant activity and enhance the immunogenicity of protein subunit vaccines. 26 These data support that sufficient adjuvants are critical to the success of nucleic acid-based vaccination. AS03 is an oil-in-water emulsion composed of squalene, polysorbate 80, and a-tocopherol, which potently induces antibodies and increases vaccine durability, promoting heterologous cross-reactivity, and having antigen dose-sparing effect (fewer antigens used for vaccination). 30 In a study comparing different adjuvants with the same amounts of Spike antigens, AS03 exhibits superior efficacy to other classical adjuvants like AS37, CpG1018, and alum in protecting experimental animals from SARS-CoV-2 infection. 31,32 The addition of AS03 in our vaccination increased the neutralization levels significantly in the 5 mg and 20 mg DNA groups, supporting that adjuvants augment the efficacy of DNA vaccines delivered via EP.
Further, we improved the efficacy of the Spike-DNA vaccine through vector engineering. The ZyCoV-D used the pVAX1 as the vector, 19 and INO-4800 used pGX9500, a modified version of pVAX1. 18,33 Next-generation iScience Article vector designs improve antigen expression, manufacturing yield and quality, and regulatory compliance. 34 The Nanoplasmid NTC9385R-eRNA41H-CpGRNA vector was developed by Nature Technology Corporation (Lincoln, NE) for DNA vaccination. First, Nanoplasmids bypass the antibiotic selection when amplified in E. coli. They incorporate and express a 150 bp RNA-OUT antisense RNA. RNA-OUT represses the expression of a chromosomal counter-selectable marker (SacB), 35 which encodes a levansucrase, a toxin to E. coli, in the presence of sucrose. Plasmid selection is achieved in sucrose-containing media. Second, Nanoplasmids contain highly productive heat-inducible R6K origins for DNA replication. Under a fermentation process called HyperGRO, 36,37 DNA yields of up to 2,400 mg/L have been obtained with Nanoplasmid vectors. Third, Nanoplasmids have an optimized chimeric promoter-intron (CMV-HTLV-I R synthetic intron), thereby achieving significantly higher expression levels (2-to 10-fold) than pVAX1, which has no intron but has the same CMV promoter and enhancer as Nanoplasmids (and bovine growth hormone polyadenylation signal). 34,35 Nanoplasmids use a backbone <0.5 kb ($2.0 kb for pVAX1) for plasmid preparation (replication and selection). Finally, the NTC9385R-eRNA41H-CpGRNA plasmid contains two adjuvant elements. (1) eR-NA41H is a 114 bp RNA fragment engineered to induce type I interferon (IFN) production through the activation of retinoic acid-inducible gene 1 (RIG-I). 38 (2) CpG RNA is a potential agonist of alternative innate immune receptors that activate adaptive immunity. 39,40 We use the Nanoplasmid NTC9385R-eRNA41H vector to express Spike HexaPro for maximized antigen expression and immunogenicity. The final construct (Nano-DNA; 6,452 bp; with the CMV promoter and the CMV-HTLV-I R synthetic intron) is smaller than the paHbased parental vector (DNA; 8,370 bp, with the chicken b-actin promoter and a chimeric intron from chicken b-actin and rabbit b-globin) 10 ; both plasmids use the same CMV enhancer and the rabbit b-globin polyadenylation signal. Nano-DNA elicits a more robust immune response than the parental vector, which is further enhanced by the addition of AS03. Five mg of Nano-DNA exhibit an IgG level and a percentage of iScience Article neutralization comparable to 5.0 mg mRNA, albeit the assays are not performed in parallel. We noted that a previous study used an earlier version of Nanoplasmid to target Spike. 41 Yet it required 3 doses of 10-50 mg plasmid DNA. Nonetheless, our data support that adding adjuvants like AS03 to the 9th iteration of Nanoplasmids, which already express eRNA41H and CpG RNA as adjuvants, is needed to maximize the immune responses from DNA vaccination.
Vaccines that generate potent nAbs and Th1-biased T cell responses reduce the risk of antibody-dependent replication enhancement. INO-4800 T cell response was close to convalescent plasma (albeit with a small number of patients). 18,42 ZyCov-D induced minimal cellular response compared with the placebo group. 43 In our Spike-DNA vaccines, the T cell response from 20 mg DNA plus AS03 was higher than that of 0.5 mg mRNA and close to that from 5.0 mg mRNA (Table 1). Nanoplasmid DNA vaccines further improved cellular responses. T cell response plays a central role in inducing anti-tumor response. 44 In this context, DNA vaccination represents a promising strategy for eliciting adaptive immune responses to cancer. Easy delivery of multiple antigens and induction of cellular and humoral immunity without being restricted to HLA-patient type makes the DNA vaccines a promising cancer prevention and treatment strategy.
To summarize, the Spike-DNA vaccines aided by AS03 at 5-20 mg per mouse activate robust humoral and cellular responses specific to the SARS-CoV-2 Spike protein, comparable to 0.5 mg per mouse of a leading mRNA vaccine. Nanoplasmids express higher antigens than conventional plasmids, producing more robust immune responses. Together, this study underscores the importance of antigen design (i.e., 2P and HexaPro mutations), vector engineering, and adjuvant stimulation in DNA vaccines.

Limitations of the study
There are several limitations to this study. First, because of technical restrictions, we could not reliably track and quantify the exact amount of plasmid DNA uptake in mouse tissues, which may be a critical factor that impacts the expression of Spike in vivo. Second, there are several differences between the mRNA-1273 and DNA (and Nano-DNA): (1) S-2P in mRNA and HexaPro Spike in DNA; (2) the C-terminal sequence (the TM domain plus the short intracellular domain in mRNA was replaced by a trimer foldon domain from T4 bacteriophage fibritin). The HexaPro Spike is more stable than S-2P, whereas the TM domain in S-2P allows Spike to anchor better to the plasma membrane and evoke more potent neutralizing antibodies. Third, there is no live virus to challenge the animals, which could help delineate how the enhanced humoral and cell responses may impact the pathogenicity of SARS-CoV-2. Lastly, our results were generated using mouse models, which may not completely mimic the characteristics of human immune responses.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

DECLARATION OF INTERESTS
The authors declare no financial and non-financial competing interests.

INCLUSION AND DIVERSITY
We support inclusive, diverse, and equitable conduct of research.  . Mice immunized with Nano-DNA potentiate functional T cells specific to SARS-CoV-2 Spike compared to the conventional plasmid (A) C57BL/6 mice (n = 5/group) were immunized with indicated vaccine groups. Lymphocytes in the spleens were collected from each group after four weeks of second immunization, stimulated with SARS-CoV-2 Spike peptide pool for 24 h, and measured stimulated splenocytes using a TNFa and IFNg double color ELISpot assay kit. Blue spots represent TNFa; Red spots represent IFNg. (B-D) (B) Count of TNFa producing T cells (C) Count of IFNg producing T cells (D) Count of TNFa and IFNg doublepositive T cells. The bars represent the means with error bars denoting the SD of five samples. Significant differences (ns = not significantly different; *p < 0.05; **p < 0.01; ***p < 0.001; one-way ANOVA) are shown.

RNA extraction and qRT-PCR
Total RNA was extracted from TRIzol reagent (Invitrogen # 15596026) and reverse-transcribed into cDNA with iScriptä cDNA synthesis kit according to the manufacturer's instructions (Bio-Rad #1708896). Gene expression levels were quantified by qRT-PCR performed on a QuantStudio 7 Pro qRT-PCR system (Thermofisher #A43183). The qRT-PCR was performed using primers for each gene, and the results were ll OPEN ACCESS