DNA vaccine technologies have developed slowly over the past 30 years. These vaccines introduce a vector containing a DNA sequence that encodes an antigen(s) selected to induce a specific immune response (
21). Early attempts revealed issues with low immunogenicity (
19,
21,
38). Additionally, initial skepticism about the approach suggested that DNA vaccines might bind to the host genome or induce autoimmune disease (
22,
39), but preclinical and clinical studies have consistently disproved this hypothesis and indicated DNA vaccines to be safe (
38). Another concern, antibiotic resistance introduced during the plasmid selection process, did remain a concern during this initial phase of development (
22), but this issue was resolved through strategic vector design (
40,
41). However, for many years, the immunogenicity of DNA vaccines failed to reach expectations (
22). Several developments during the 2010s led to greater efficacy of DNA vaccines (
22). However, no DNA vaccines had been approved for use in humans prior to the COVID-19 pandemic (
38,
42). As of 7 October 2022, 10 vaccines have been approved worldwide (
Table 1). These vaccines fall into two categories, vaccines that are vectored with a plasmid and those that are vectored with another virus.
Plasmid-vectored DNA vaccines.
Many DNA vaccines use a plasmid vector-based approach, where the sequence encoding the antigen(s) against which an immune response is sought is cultivated in a plasmid and delivered directly to an appropriate tissue (
43). Plasmids can also be designed to act as adjuvants by targeting essential regulators of pathways such as the inflammasome or simply just specific cytokines (
39,
44). The DNA itself may also stimulate the innate immune response (
21,
41). Once the plasmid brings the DNA sequence to an antigen-presenting cell (APC), the host machinery can be used to construct antigen(s) from the transported genetic material, and the body can then synthesize antibodies in response (
22). The vectors are edited to remove extra sequences (
41). These types of manufacturing advances have improved the safety and throughput of this platform (
41).
(i) Prior applications. In the 1990s and 2000s, DNA vaccines delivered via plasmids sparked significant scientific interest, leading to a large number of preclinical trials (
22). Early preclinical trials primarily focused on long-standing disease threats, including viral diseases such as rabies and parasitic diseases such as malaria, and promising results led to phase I testing of the application of this technology to human immunodeficiency virus (HIV), influenza, malaria, and other diseases of concern during this period (
22). Although they were well tolerated, these early attempts to develop vaccines were generally not very successful in inducing immunity to the target pathogen, with either limited T cell or limited neutralizing antibody responses observed (
22).
Early plasmid-vectored DNA vaccine trials targeted HIV and subsequently diseases of worldwide importance such as malaria and hepatitis B (
45). The concern with these early development projects was immunogenicity, not safety (
45). Around the turn of the millennium, a hepatitis B vaccine development program demonstrated that these vaccines can induce both antibody and cellular immune responses (
46). Prior to COVID-19, however, plasmid-vectored DNA vaccines had been approved for commercial use only in veterinary populations (
47–49). Between 2005 and 2006, several DNA vaccines were developed for nonhuman animal populations, including against viruses such as a rhabdovirus in fish (
50), porcine reproductive and respiratory syndrome virus (
51), and West Nile virus in horses (
52). Within the past 5 years, additional plasmid-vectored vaccines for immunization against viruses were developed against a herpesvirus (in mice) (
53) and an alphavirus (in fish) (
54).
(ii) Applications to COVID-19. Several plasmid-vectored DNA vaccines have been developed against COVID-19 (
Table 1). In fact, the ZyCoV-D vaccine developed by India’s Zydus Cadila is the first plasmid-vectored DNA vaccine to receive approval or to be used in human medicine (
55–57). Another plasmid-vectored DNA vaccine, INO-4800 (
58), was developed by Inovio Pharmaceuticals Technology and uses electroporation as an adjuvant. Electroporation was developed as a solution to the issue of limited immunogenicity by increasing the permeability of cell membranes by delivering electrical pulses (
59). It has been shown that electroporation can enhance vaccine efficacy (VE) up to 100-fold, as measured by increases in antigen-specific antibody titers (
60). The temporary formation of pores through electroporation facilitates the successful transportation of macromolecules into cells, allowing cells to robustly take up INO-4800 for the production of an antibody response. For INO-4800, a plasmid-vectored vaccine is delivered through intradermal injection, which is then followed by electroporation with a device known as CELLECTRA (
61). The safety of the CELLECTRA device has been studied for over 7 years, and these studies support the further development of electroporation as a safe vaccine delivery method (
59).
These vaccines therefore represent implementations of a new platform technology. In particular, they offer the advantage of a temperature-stable vaccine, facilitating worldwide administration (
62). Although an exciting development in DNA vaccines, the cost of vaccine manufacturing and electroporation may make scaling the use of this technology for prophylactic use for the general public difficult.
(iii) Trial safety and immunogenicity. The INO-4800 trials began with a phase I trial evaluating two different doses administered as a two-dose series (
61). This trial found the vaccine to be safe, with only six adverse events (AEs) reported by 39 participants, all grade 1, and effective, with all but 3 participants of 38 developing serum IgG binding titers to the SARS-CoV-2 S protein (
61). A phase II trial of 401 adults at high risk of exposure to SARS-CoV-2 similarly supported the safety and efficacy of INO-4800. Only one treatment-related AE was observed, and the vaccine was found to be associated with a significant increase in neutralizing activity (
62). Results of phase III trials are not yet available (
63–66).
Trials of ZyCoV-D have progressed further. This vaccine uses a plasmid to deliver the expression-competent spike protein and IgE signal peptides to the vaccinee (
67). During the phase I trial, vaccination with a needle versus a needle-free injection system was evaluated, and the vaccine can now be administered without a needle (
55,
56). A phase III trial enrolling over 27,000 patients found no difference in AEs between the placebo and treatment groups and estimated the efficacy of ZyCoV-D to be 66.6% (
68). It was authorized for people ages 12 and older (
57). The highly portable design offers advantages over traditional vaccines (
67), especially as the emergence of variants continues to challenge the effectiveness of vaccines. As of August 2022, ZyCoV-D has only been approved in India (
69) and is not tracked by Our World in Data (
70).
(iv) Real-world safety and effectiveness. In terms of the ability of plasmid-vectored vaccines to neutralize VOC, varying information is available. The situation for ZyCoV-D is somewhat different, as the phase III trial occurred during the Delta wave in India (
68). At present, no major press releases have addressed the vaccine’s ability to neutralize Omicron and related VOC, but reporting suggests that the manufacturers were optimistic about the vaccine in light of the Omicron variant as of late 2021 (
71).
As for INO-4800, studies have examined whether the induced immune response can neutralize existing VOC. They assessed neutralization of several VOC relative to the index strain and found no difference in neutralization between the index strain and the Gamma VOC (P.1) (
72). However, neutralization of the Alpha and Beta VOC was significantly lower (approximately two and seven times, respectively) (
72). These findings are in line with the shifts in effectiveness reported for other vaccines (
1). In addition to loss of neutralizing activity due to viral evolution, studies have also evaluated the decline in neutralizing antibodies (nAbs) induced by INO-4800 over time. Levels of nAbs remained statistically significant relative to the prevaccination baseline for 6 months (
73). Administration of a booster dose induced a significant increase of titers relative to their prebooster levels (
73). Given the timing of this trial (enrollment between April and July 2020), it is unlikely that participants were exposed to VOC associated with decreased efficacy.
In light of the emergence of VOC against which many vaccines show lower effectiveness, Inovio Pharmaceuticals began to develop a new vaccine with the goal of improving robustness against known and future VOC (
74). Known as INO-4802, this vaccine was designed to express a pan-spike immunogen (
75). Booster studies in rodents (
76) and nonhuman primates (
75) suggest that it may be more effective than INO-4800 in providing immunity to VOC such as Delta and Omicron when administered as part of a heterologous boost regimen, although boosting with INO-4800 was also very effective in increasing immunity in rhesus macaques (
75). Therefore, boosting is likely to be an important strategy for this vaccine, especially as the virus continues to evolve.
Virus-vectored DNA vaccines.
Plasmids are not the only vector that can be used to deliver sequences associated with viral antigens. Genetic material from the target virus can also be delivered using a second virus as a vector. Viral vectors have emerged as a safe and efficient method to furnish the nucleotide sequences of an antigen to the immune system (
77). The genetic content of the vector virus is often altered to prevent it from replicating, but replication-competent viruses can also be used under certain circumstances (
78). Once the plasmid or viral vector brings the DNA sequence to an APC, the host machinery can be used to construct antigen(s) from the transported genetic material, and the host can then synthesize antibodies in response (
22).
One of the early viral vectors explored was adenovirus, with serotype 5 (Ad5) being particularly effective (
22). This technology rose in popularity during the 2000s due to its being more immunogenic in humans and nonhuman primates than plasmid-vectored DNA vaccines (
22). In the 2000s, interest also arose in utilizing simian adenoviruses as vectors because of the reduced risk that human vaccine recipients would have prior exposure resulting in adaptive immunity (
22,
79), and chimpanzee adenoviruses were explored as potential vectors in the development of a vaccine against MERS-CoV (
80).
Today, various viral-vector platforms including poxviruses (
81,
82), adenoviruses (
83), and vesicular stomatitis viruses (
84,
85) are being developed. Viral vector vaccines are able to induce both an antibody and cellular response; however, the response is limited due to the immunogenicity of the viral vector used (
83,
86). An important consideration in identifying potential vectors is the immune response to the vector. Both the innate and adaptive immune responses can potentially respond to the vector, limiting the ability of the vaccine to transfer information to the immune system (
87). Different vectors are associated with different levels of reactogenicity; for example, adenoviruses elicit a much stronger innate immune response than replication-deficient adeno-associated viruses derived from parvoviruses (
87). Additionally, using a virus circulating widely in human populations as a vector presents additional challenges because vaccine recipients may already have developed an immune response to the vector (
88). Furthermore, repeated exposure to adenoviruses via virus-vectored DNA vaccines may increase reactivity to these vectors over time, presenting a challenge that will need to be considered in long-term development of these vaccines (
89,
90).
(i) Prior applications. There are several viral vector vaccines that are available for veterinary use (
22,
91), but prior to the COVID-19 pandemic, only one viral vector vaccine was approved by the U.S. Food and Drug Administration (FDA) for use in humans. This vaccine is vectored with a recombinant vesicular stomatitis virus and targeted against the Ebola virus (
92). Additionally, several phase I and phase II clinical trials for other vaccines are ongoing (
77), and the technology is currently being explored for its potential against numerous infectious diseases, including malaria (
93,
94), Ebola (
95–97), and HIV (
98,
99).
The threat of MERS and SARS initiated interest in the application of viral vector vaccines to human coronaviruses (
80), but efforts to apply this technology to these pathogens had not yet led to a successful vaccine candidate. In the mid-2000s to late 2000s, adenovirus-vectored vaccines against SARS were found to induce SARS-CoV-specific IgA in the lungs of mice (
100) but were later found to offer incomplete protection in ferret models (
101). The Gamaleya National Center of Epidemiology and Microbiology in Moscow sought to use an adenovirus platform for the development of vaccines for MERS-CoV and Ebola virus, although neither of the previous vaccines was internationally licensed (
102).
In 2017, results were published from an initial investigation of two vaccine candidates against MERS-CoV containing the MERS-CoV
S gene vectored with chimpanzee adenovirus, Oxford University #1 (ChAdOx1), a replication-deficient chimpanzee adenovirus (
103). This study reported that a candidate containing the complete S protein sequence induced a stronger neutralizing antibody response in mice than candidates vectored with modified vaccinia virus Ankara.
The candidate was pursued in additional research, and in the summer of 2020, results of two studies were published. The first reported that a single dose of ChAdOx1 MERS induced an immune response and inhibited viral replication in macaques (
104). The second reported promising results from a phase I trial that administered the vaccine to adults and measured safety, tolerability, and immune response (
105).
(ii) Application to COVID-19. While not all of the above-described results were available at the time that vaccine development programs against SARS-CoV-2 began, at least three viral vector vaccines have also been developed against SARS-CoV-2 (
Fig. 2). First, a collaboration between AstraZeneca and researchers at the University of Oxford successfully applied a viral vector approach to the development of a vaccine against SARS-CoV-2 using the replication-deficient ChAdOx1 vector modified to encode the S protein of SARS-CoV-2 (
106). In a phase I trial, the immunogenic potential of vaccine candidate ChAdOx1 nCoV-19 was demonstrated through the immune challenge of two animal models, mice and rhesus macaques (
106). In a phase I/II trial, patients receiving the ChAdOx1 nCoV-19 vaccine developed antibodies to the SARS-CoV-2 spike protein that peaked by day 28, with these levels remaining stable until a second observation at day 56 (
107).
Second, a viral vector approach was applied by Russia’s Gamaleya Research Institute of Epidemiology and Microbiology to develop Sputnik V, a replication-deficient recombinant adenovirus (rAd) vaccine that combines two adenovirus vectors, rAd26-S and rAd5-S, that express the full-length SARS-CoV-2 spike glycoprotein. These vectors are intramuscularly administered individually using two separate vaccines in a prime-boost regimen. rAd26-S is administered first, followed by rAd5-S 21 days later. Both vaccines deliver 10
11 viral particles per dose. This approach is designed to overcome any potential preexisting immunity to adenovirus in the population (
108), as some individuals may possess immunity to Ad5 (
109). Sputnik V is the only recombinant adenovirus vaccine to utilize two vectors.
Third, Janssen Pharmaceuticals, Inc., a subsidiary of Johnson & Johnson, developed a viral vector vaccine in collaboration with and funded by the United States’ Operation Warp Speed (
110,
111). The vaccine candidate JNJ-78436735, formerly known as Ad26.COV2-S, is a monovalent vaccine that is composed of a replication-deficient adenovirus serotype 26 (Ad26) vector expressing the stabilized prefusion S protein of SARS-CoV-2 (
33,
112). Unlike the other two viral vector vaccines available to date, JNJ-78436735 requires only a single dose, a characteristic that was expected to aid in global deployment (
113). JNJ-78436735 was selected from among a number of initial candidate designs (
33) and tested
in vivo in Syrian golden hamsters and rhesus macaques to assess safety and immunogenicity (
33,
113–115). The JNJ-78436735 candidate was selected for its favorable immunogenicity profile and ease of manufacturability (
33,
113–115) and was found to confer protection against SARS-CoV-2 in macaques even after 6 months (
116). The one- versus two-dose regimen was then tested in volunteers through a phase I/IIa trial (
112,
117). A major difference between this vaccine and the other two in this category is that the S protein immunogen is stabilized in its prefusion conformation, while in the Sputnik V and AstraZeneca vaccines it is not.
As of 9 October 2022, data describing the distribution of 4 virus-vectored vaccines in 201 countries are available (
Fig. 2). ChAdOx1 nCoV-19 was first approved for emergency use on 30 December 2020 in the United Kingdom (
118). Sputnik V was available soon after, and as early as January 2021, Sputnik V had been administered to 1.5 million Russians (
119) and distribution of doses to other countries within Europe, such as Belarus, Bosnia-Herzegovina, Hungary, San Marino, Serbia, and Slovakia, was begun (
120–122).
(iii) Trial estimates of safety and efficacy. The first DNA virus-vectored vaccine for which efficacy estimates became available was AstraZeneca’s ChAdOx1 nCoV-19. In December 2020, preliminary results of the phase III trial were released detailing randomized control trials conducted in the United Kingdom, Brazil, and South Africa between April and November 2020 (
12). These trials compared ChAdOx1 nCoV-19 to a control, but the design of each study varied; pooling data across studies indicated an overall efficacy of 70.4%. For Sputnik V, the phase III trial indicated an overall vaccine efficacy of 91.6% for symptomatic COVID-19 (
123). As for Janssen, the vaccine was well tolerated, and across all regions studied, it was found to be 66.9% effective after 28 days for the prevention of moderate to severe COVID-19 and to be 81.7% effective for the prevention of laboratory-confirmed severe COVID-19 (
124). There were no COVID-19-associated deaths in the vaccine group. However, the emergence of the Beta variant in the South African trial population was associated with a slightly reduced efficacy (64% 2 weeks after receipt), and all of the COVID-19-associated deaths in the trial occurred in the South African placebo cohort (
124). In February 2021, the FDA issued an emergency use authorization (EUA) for the Janssen vaccine based on interim results from the phase III trial (
125,
126).
Two of the three vaccines have faced a number of criticisms surrounding the implementation of their clinical trials. In the race to develop vaccines against SARS-CoV-2, President Vladimir Putin of Russia announced the approval of the Sputnik V vaccines on 11 August 2020 in the absence of clinical evidence (
127). A press release on 11 November 2020 indicated positive results from an interim analysis of the phase III Sputnik V trials, which reported 92% efficacy in 16,000 volunteers (
128). However, this release came only 2 days after both Pfizer and BioNTech reported that their vaccines had an efficacy over 90%, which led to significant skepticism of the Russian findings for myriad reasons, including the lack of a published protocol and the “reckless” approval of the vaccine in Russia months prior to the publication of the interim results of the phase III trial (
128,
129). Consequently, many international scientific agencies and public health bodies expressed concern that due diligence to the clinical trial process was subverted for the sake of expediency, leading many to question the safety and efficacy of Sputnik V (
127,
130,
131). Despite regulatory, safety, and efficacy concerns, preorders for 1 billion doses of the Sputnik V were reported within days of the vaccine’s approval in Russia (
127). Almost a month later, the phase I/II trial data were published (
132). It was not until February 2021, 6 months after its approval in Russia, that interim results of the phase III trial were released (
123). This publication reported a VE of 91% and a low rate of serious AEs, although there were several serious AEs that were determined not to be associated with the vaccine by an independent data monitoring committee about which little other information was released (
133).
AstraZeneca’s clinical trial also faced criticism. The trial was paused in September 2020 following a severe adverse event in one participant (
134). It was restarted soon after (
135), but it seems that the recent pause was not mentioned to the FDA during a call the morning before the story broke (
136). Additionally, individual sites within the trial employed somewhat different designs but were combined for analysis. For example, in South Africa, the trial was double-blinded, whereas in the United Kingdom and Brazil it was single-blinded, and one of the two trials carried out in the United Kingdom evaluated two dosing regimens (low dose and standard dose, both followed by standard dose). Some of the trials used a meningococcal conjugate vaccine (MenACWY) as a control, while others used saline. Data were pooled across countries for analysis, a design decision that was approved by regulators but raised some questions when higher efficacy was reported in a subgroup of patients who received a low dose followed by a standard dose. This group came about because some participants in the United Kingdom were erroneously primed with a much lower dose, which turned out to have higher efficacy than the intended dose (
137). Combining the data then led to confusion surrounding the VE, as VE varied widely among conditions (e.g., 62% VE in the standard dose group versus 90% in the group that received a low prime dose [
12]). Subsequent research, however, suggests that reducing the prime dose may, in fact, elicit a superior immune response in the long term despite a lower initial response (
138). Therefore, this error may serendipitously improve efficacy of vaccine-vectored vaccines broadly.
(iv) Real-world safety and efficacy. Following the trials, additional concerns have been raised about some of these vaccines. Within a few days to a few weeks following their first dose of the AstraZeneca vaccine, three women developed extensive venous sinus thrombosis (
139). In March 2021, administration of the vaccine was paused in several European countries while a possible link to thrombotic events was investigated (
140), as these adverse events had not been observed in clinical trials, but the European Medicines Agency (EMA) soon determined that 25 events were not related to the vaccine (
141). The following month, the United States paused administration of the Janssen vaccine for 10 days due to 15 similar AEs (
142,
143), but the EMA, U.S. Centers for Disease Control, and the FDA’s Advisory Committee on Immunization Practices again identified the events as being very rare and the benefits of the vaccine as likely to outweigh its risks (
144–147). In Denmark and Norway, population-based estimates suggested that AstraZeneca’s vaccine increased the incidence of venous thromboembolic events by 11 cases over baseline per 100,000 doses (
148). Estimates of the incidence in other Western countries have also been low (
149). In the United States, thromboembolic events following the Janssen vaccine have also been very rare (
145). Subsequently, a potential mechanism was identified: the adenovirus vector binding to platelet factor 4 (
150,
151). Because this adverse event is so rare, the risk is likely still outweighed by the risks associated with contracting COVID-19 (
152), which is also associated with thrombotic events (
143,
153). Similarly, concerns about Guillain-Barré syndrome arose in connection to the Janssen vaccine, but these events have similarly been determined to be very rare and the benefits to outweigh the risks (
147).
Given that vaccines from multiple platforms are now widely available, people at increased risk of a specific severe AE may have options to pursue vaccination with a platform that does not carry such risks. For example, a woman in the United States with a history of thromboembolic concerns might feel more comfortable with an mRNA vaccine (described below), because such AEs have not been identified in association with COVID-19 vaccination in the United States. However, within the United States, no clear framework has been established for advising patients on whether a specific vaccine may be preferable for their individual concerns now that vaccines based on three different technologies are widely available (see reference
1 for information about Novavax, which is a protein subunit vaccine).