Rendezvous with Vaccinia Virus in the Post-smallpox Era: R&D Advances

Smallpox was eradicated in less than 200 years after Edward Jenner’s practice of cowpox variolation in 1796. The forty-three years of us living free of smallpox, beginning in 1979, never truly separated us from poxviruses. The recent outbreak of monkeypox in May 2022 might well warn us of the necessity of keeping up both the scientific research and public awareness of poxviruses. One of them in particular, the vaccinia virus (VACV), has been extensively studied as a vector given its broad host range, extraordinary thermal stability, and exceptional immunogenicity. Unceasing fundamental biological research on VACV provides us with a better understanding of its genetic elements, involvement in cellular signaling pathways, and modulation of host immune responses. This enables the rational design of safer and more efficacious next-generation vectors. To address the new technological advancement within the past decade in VACV research, this review covers the studies of viral immunomodulatory genes, modifications in commonly used vectors, novel mechanisms for rapid generation and purification of recombinant virus, and several other innovative approaches to studying its biology.


Introduction
The World Health Organization (WHO) officially announced the year 1979 as the mark of smallpox eradication [1,2]. Vaccinia virus (VACV), an orthopoxvirus belonging to the family Poxviridae and subfamily Chordopoxvirinae, was widely used as a vaccine in the eradication program and decisively contributed to the accomplishment of such extraordinary achievement. However, the manufacturing process of these smallpox vaccines at the time was not in good compliance with the standards acceptable nowadays. After all, inoculating calves with live viruses and grinding their 'ripe' lesions to generate a vaccine stock is neither hygienic nor quality assured [3]. Additionally, this type of vaccine contains live viruses that may lead to multiple side effects, including fever, mild rash, eczema vaccinatum, progressive vaccinia, and encephalitis, especially in immunosuppressed or immunocompromised populations [1,[4][5][6]. Therefore, extensive efforts have been made to increase the safety profile of VACV, giving rise to a number of second-generation (ACAM2000 [7][8][9][10], Elstree-BN [11], and CCSV [12]), third-generation (Imvamune [13][14][15][16], NYVAC [17][18][19], and LC16m8 [20][21][22][23][24]), and next-generation smallpox vaccines (DNA vaccines [25][26][27][28], protein subunit vaccines [29][30][31], and T-cell epitope vaccines [32]) [33]. Despite the successful eradication of smallpox, poxviruses have not become a thing of the past. The potential threat of bioterrorism using the smallpox virus and the emergence or re-emergence of other poxvirus diseases should make it clear that the continuity of research is paramount. The recent outbreak of monkeypox in May 2022 is a perfect example, with more than 80,000 cases reported globally in less than a year, of which about 37.5% were identified in the US [34].
The double-stranded DNA genome of VACV has a considerable size of approximately 190 kb, containing over 200 open reading frames (ORFs). Many of these ORFs govern viral plaque formation as well as abortive infections. Without DOX, two constructs viP11A3L and viP11A6L, of which the essential gene is under the control of a late VACV P11 promoter, only produced individually infected cells and failed to form plaques. This was evidenced by green fluorescence confined to each infected cell and the absence of cytopathic effects in surrounding cells. The successful expression of the transgene, independent of the generation of infectious progeny, was simultaneously confirmed. However, in the presence of DOX, the two constructs replicated to comparable titer and formed plaques that were the same size as the wild-type VACV. The safety profile was confirmed in vivo using mouse models. CB6F1/J mice when infected intranasally with viP11A6L in the absence of DOX did not develop any clinical signs or lose weight but quickly succumbed to infection when exposed to 0.125 mg/mL or higher of DOX in drinking water, just like those given wild-type VACV. No plaque-forming virions could be recovered from ovaries of mice infected intraperitoneally with viP11A6L without DOX, while a significantly higher level of viral load was detected with DOX supplied in drinking water, to a level indistinguishable from WT VACV-infected mice. Another study adapting this strategy was published by the same group to confirm the feasibility of using a replication-inducible VACV as a vaccine vector for Zika virus (vIND-ZIKV) [50]. The vaccine did not lead to any weight loss or clinical signs in intranasally infected CB6F1 mice in the absence of DOX, indicating a consistent safety profile. C57BL/6 mice vaccinated intramuscularly with a single dose of vIND-ZIKV developed high levels of ZIKV E-specific IgG and cell-mediated responses as well as being fully protected from viremia. Figure 1. VACV with extra safety features cannot propagate in host cells. Wild-type VACVs without built-in safety mechanisms can establish an active infection. The outcomes include viral DNA replication, generation of progeny, formation of plaques, and potentially the development of diseases or death. VACVs with enhanced safety features can produce the same outcomes only if an inducing or complementing condition is provided but will not lead to an active infection in host cells. GOIs: genes of interest. Figure 1. VACV with extra safety features cannot propagate in host cells. Wild-type VACVs without built-in safety mechanisms can establish an active infection. The outcomes include viral DNA replication, generation of progeny, formation of plaques, and potentially the development of diseases or death. VACVs with enhanced safety features can produce the same outcomes only if an inducing or complementing condition is provided but will not lead to an active infection in host cells. GOIs: genes of interest.
Coupling a replication-deficient deletion mutant with a specialized cell line that supplements the product(s) of deleted gene(s) can also greatly improve the safety profile as the vector itself is unable to replicate in vaccinees. Wyatt and colleagues reported a vector system using an A23R-deletion-mutant VACV that can only replicate in a complementing cell line [51]. The intermediate transcription factor encoded by the A23R gene is indis- pensable for viral replication and can be provided constitutively by complementing cells to support propagation [52,53]. To ensure high-level expression of the heterologous gene even in non-complementing cells, a T7 promoter is placed immediately upstream, and the T7 polymerase needed is expressed from a VACV early promoter in the vector. This is because viral DNA replication can still be achieved via early gene expression to generate sufficient templates containing the gene of interest. The mutant VACV could only form plaques and produce progeny in complementing RK/A8A23 cells but not in regular HeLa or RK13 cells. However, the expression of heterologous genes was not affected in BALB/c mice, evidenced by both protein expression and specific antibody responses. Eldi et al. described a similar Sementis Copenhagen Vector (SCV) system that uses the Copenhagen strain of VACV in lack of D13L gene with an engineered CHO cell line that provides D13 and CP77 proteins to allow high-titer production only in vitro [54]. D13L is responsible for the formation of infectious virions [55,56], while CP77 ensures that the replication of viral DNA is not inhibited by a cellular protein HMG20A [57]. Therefore, the genome retains the capacity to replicate in any non-permissive cell line, and high-level expression of transgenes during the late phase of the infection remains intact, but no infectious progeny can be produced. The resultant recombinant virus could not form plaques, an indicator of viral replication and cell-to-cell spread, in HEK293, MRC-5, 143b, A431, or regular CHO cell lines, but replicated to a comparable level as the replication-competent VACV in the cell substrate cell line (SCS) even after thirty passages. Electron microscopy confirmed the presence of viroplasma without immature or mature virions, suggesting arrested viral multiplication in non-permissive cell lines. An excellent safety feature was demonstrated by virulence studies in mice. Compared to severe combined immunodeficiency (SCID) mice injected with replication-competent VACV, SCV-infected ones displayed no weight loss, 100% survival rate up to 100 dpi, absence of infectious virions in multiple organs, and significantly lower clinical scores. Multiplication competence was not regained within twenty passages, as claimed by the authors.

Rapid Generation and Purification of Recombinant VACV
Homologous recombination has been the mainstay to introduce transgenes or deletions to VACV genome [58]. This is carried out by infecting cells with a parental virus to be genetically edited and transfecting with a shuttle plasmid that contains the gene of interest flanked by homologous sequences targeting a gene in the parental virus. However, DNA recombination is a rare event where thousands of progenies may give rise to only one desired recombinant. Therefore, a selection system is essential to distinguish between parental and recombinant viruses ( Figure 2). Any system that is labor intensive or time consuming, such as transient dominant selection [59], will slow down the purification progress and reduce the flexibility of testing various genes of interest for vaccine or therapeutic purpose. On the other hand, the use of some highly efficient systems can be limited by the requirement for special equipment or reagents and cannot be widely transplanted to most laboratories. A balance between convenience and accessibility needs to be reached as innovative ideas and technologies aiming at generating recombinant VACVs advance in the future.
The difference in plaque size between parental and recombinant viruses can be a simple way for selection as reported by Siciliano et al. [60]. The parental strain vRB12 forms small plaques due to the deletion of F13L gene but will form larger plaques upon its restoration by a plasmid that also encodes the gene of interest [60]. However, distinguishing between large and small plaques is not a clear binary system, and parental viruses can be easily carried over, leading to the need for repetitive and additional purification. Cao and co-workers constructed a double-deletion mutant that can only propagate in permissive cell lines to simplify the selection of recombinants [61]. The E3L and K3L genes in the parental virus are, respectively, interrupted by an EGFP and a mCherry gene. In the shuttle vector, homology arms targeting the K3L gene in the parental virus flank a taterapox virus K3L ortholog and a heterologous gene of interest. Homologous recombination occurs in permissive HeLa/PKR knockout or A549/PKR+Rnase L knockout cells and subsequent purification happens in non-permissive BHK21 cells. Nearly all green plaques are successful recombinants. Despite accelerated purification, this system still relies on a fluorescent protein marker. An EPPIC (Efficient Purification by Parental Inducer Constraint) system independent of a selection marker was developed by Jasperse et al. to achieve both generation and purification of recombinant VACVs within one week [62], using minimal equipment available in any lab setting. Taking advantage of their previously reported replication-inducible system [36], a lactose-inducible (replicates only in the presence of isopropyl β-D-1-thiogalactopyranoside, or commonly known as IPTG) parental virus undergoes homologous recombination with a shuttle vector containing gene(s) of interest to give rise to a tet-inducible (replicates only in the presence of DOX) recombinant virus. Since the inducing conditions for parental and recombinant viruses are mutually exclusive, removing IPTG and adding DOX to the culture media will make it solely conducive to the growth of recombinant viruses. Cells infected by parental viruses will not show any cytopathic effect, enabling easy isolation and sequential purification. This is particularly important for making marker-free vectors as fluorescent protein genes, despite convenience in selecting for recombinant virus, cannot be present in a vaccine candidate. Similar to this idea, White and co-workers added a negative selection marker to inhibit viral replication in parental virus to greatly increase the percent of recombinants [63]. The inhibition is achieved by blocking protein synthesis in infected cells via coumermycin-activated phosphorylation of translation initiation factor eIF2α. The coumermycin-binding domain is derived from Escherichia coli gyrase B (GyrB) and is fused to a mammalian double-stranded RNA-dependent protein kinase PKR [64]. In the presence of the antibiotic coumermycin which binds to two molecules of GyrB, dimerization of GyrB will occur and activate PKR catalytic domain to phosphorylate eIF2α. This system significantly decreased the ratio of parental to recombinant viruses from 4000:1 (without coumermycin) to 1:80 (with coumermycin). Although distinguishing between parental and recombinant viruses is more straightforward, the necessity for multiple rounds of plaque purification remains.
Viruses 2023, 15, x FOR PEER REVIEW 6 of 20 populated by colorless plaques of parental virus. An elaborated version of this approach by Gowripalan and co-workers [70] provided an explanation and evidence for the limited improvement in homologous recombination rate by the CRISPR/Cas9 system as, in spite of the efficient DNA cleavage by Cas9, neither of the two repairing pathways (non-homologous end joining or homology-directed repair) was highly functional. Instead, they took advantage of the inefficient DNA repair in parental VACV to render them replication deficient, which in turn significantly increased the chance of isolating recombinant virus. After introducing Cas9 and single-guide RNA that targets a fluorescence gene (only present in parental viruses) as ribonucleoprotein complexes to a mixture that contained just 5% of successfully recombined viruses, this percentage surged to 80%. With a second round of editing, nearly all residual parental viruses were eliminated. A parallel experiment in this study tested the feasibility of abandoning the fluorescent marker and yielded a similar increase in the purity of recombinant virus from 4% to 71% with just two rounds of Cas9 selection. A further extension of combining CRISPR/Cas9 with trimolecular recombination was reported by Laudermilch and Chandran, termed MAVERICC (markerfree vaccinia virus engineering of recombinants through in vitro CRISPR/Cas9 cleavage) [72]. This strategy utilized cell-free viral DNA cleavage by Cas9/sgRNA to first generate cleaved VACV DNA before joining it with helper virus (fowlpox virus) infected cells cotransfected with a rescue amplicon that contains the transgene flanked by homologous regions specific to cleaved parental VACV DNA. In a single step, without need for sequential purification, both wild type and recombinant virus could be rescued after trimolecular recombination [81]. Among the randomly selected plaques, recombinant viruses accounted for a stunning 91% (10 out of 11 plaques for a truncation mutant), 100% (12 out of 12 plaques for a K151E mutant), or 92% (11 out 12 plaques for an insertion mutant). Simultaneous modifications in two genes were also tested in this study and showed a similarly high percentage (70%, 7 out of 10) of desired recombinants in the rescued viruses.

Figure 2.
Rapid purification of recombinant VACVs is achieved by selectively suppressing the parental population. The purification process begins with homologous recombination in cells, followed by the selection of desired recombinants. Parental vaccinia viruses (pVACV) will account for a predominantly high percentage after the recombination event and need to be inhibited to accelerate the selection of recombinant vaccinia viruses (rVACV). Several major approaches to inhibition and selection are listed.

Modification in Viral Immunomodulatory Genes
Vaccinia virus has a large genome that includes a plethora of immunomodulatory genes interfering with host antiviral immune responses [82,83]. For example, the complement system is targeted by a viral protein VCP that functions to cleave cellular C3b and Rapid purification of recombinant VACVs is achieved by selectively suppressing the parental population. The purification process begins with homologous recombination in cells, followed by the selection of desired recombinants. Parental vaccinia viruses (pVACV) will account for a predominantly high percentage after the recombination event and need to be inhibited to accelerate the selection of recombinant vaccinia viruses (rVACV). Several major approaches to inhibition and selection are listed.
Several other approaches to generating marker-free recombinant VACV have also been explored. Introducing a fluorescent marker during recombination then cleaving it using Cre/loxP system [65,66] once successful recombinants are purified is a strategy used in Rintoul and co-workers' work [67]. They placed loxP sites on both sides of the fluorescence gene which is flanked by homology arms specific to the VACV genome. Isolation of recombinant plaques could be achieved by identifying fluorescence and facilitated using fluorescence-activated cell sorting (FACS). Purified recombinant viruses were then grown in a cell line that expresses Cre recombinase to excise the marker gene. The same strategy was also applied in the work of Guo et al. to generate multiple versions of VACV vectors expressing a human C-C motif chemokine ligand 5 (CCL5) or cytokines [68]. A next-level technique to accelerate the generation and purification of recombinant VACV is CRISPR/Cas9 system [69][70][71][72]. It is an undeniably popular gene editing tool in multiple living organisms [73][74][75][76][77][78][79][80]. Yuan et al., in their two studies [69,71], increased the efficiency of homologous recombination by expressing the Cas9 endonuclease gene from Streptococcus pyogenes and single-guide RNA in VACV infected cells co-transfected with a repair plasmid that contains a cassette of red fluorescent protein (RFP) marker gene and a tumor antigen TRP2 flanked by homology arms targeting a VACV gene. Modification of two VACV genes simultaneously was also confirmed using this method. A follow-up study of theirs sought to remove the RFP marker gene by adding loxP sites on both sides, so that infected cells expressing Cre recombinase would excise the RFP gene [71]. However, this strategy did not address the challenge of isolating red plaques from a background overwhelmingly populated by colorless plaques of parental viruses. An elaborated version of this approach by Gowripalan and co-workers [70] provided an explanation and evidence for the limited improvement in homologous recombination rate by the CRISPR/Cas9 system as, in spite of the efficient DNA cleavage by Cas9, neither of the two repairing pathways (non-homologous end joining or homology-directed repair) was highly functional. Instead, they took advantage of the inefficient DNA repair in parental VACV to render them replication deficient, which in turn significantly increased the chance of isolating recombinant virus. After introducing Cas9 and single-guide RNA that targets a fluorescence gene (only present in parental viruses) as ribonucleoprotein complexes to a mixture that contained just 5% of successfully recombined viruses, this percentage surged to 80%. With a second round of editing, nearly all residual parental viruses were eliminated. A parallel experiment in this study tested the feasibility of abandoning the fluorescent marker and yielded a similar increase in the purity of recombinant virus from 4% to 71% with just two rounds of Cas9 selection. A further extension of combining CRISPR/Cas9 with trimolecular recombination was reported by Laudermilch and Chandran, termed MAVERICC (marker-free vaccinia virus engineering of recombinants through in vitro CRISPR/Cas9 cleavage) [72]. This strategy utilized cell-free viral DNA cleavage by Cas9/sgRNA to first generate cleaved VACV DNA before joining it with helper virus (fowlpox virus) infected cells co-transfected with a rescue amplicon that contains the transgene flanked by homologous regions specific to cleaved parental VACV DNA. In a single step, without the need for sequential purification, both wild type and recombinant viruses could be rescued after trimolecular recombination [81]. Among the randomly selected plaques, recombinant viruses accounted for a stunning 91% (10 out of 11 plaques for a truncation mutant), 100% (12 out of 12 plaques for a K151E mutant), or 92% (11 out of 12 plaques for an insertion mutant). Simultaneous modifications in two genes were also tested in this study and showed a similarly high percentage (70%, 7 out of 10) of desired recombinants in the rescued viruses.

Modification in Viral Immunomodulatory Genes
Vaccinia virus has a large genome that includes a plethora of immunomodulatory genes interfering with host antiviral immune responses [82,83]. For example, the complement system is targeted by a viral protein VCP that functions to cleave cellular C3b and C4b [84,85]. VACV is also well known to inhibit the production and binding of several interferons (IFN), chemokines, and cytokines. As a vector bearing genes of interest, one would hope that the characteristics of the vector would enhance, rather than diminish, host immunity specific to the target antigen. Therefore, immunomodulatory genes in VACV may exert negative effects on desired immune responses, and deleting or modifying these genes has become a promising strategy to improve the immunogenicity of VACV vectors.
To quantitatively evaluate the effects of such a gene on antigen-specific immune responses after immunization, it is common to first generate a deletion mutant identical to the parental virus in every way but the presence of the immunomodulatory gene of interest. Analysis of growth kinetics in several permissive cell lines, such as chick embryo fibroblasts in vitro, will follow to confirm if the deleted gene is essential for viral replication, which can be reflected by plaque formation and titer. An infection experiment will also be conducted in human macrophages and dendritic cells to compare the mRNA and protein levels of gene expression involved in innate immunity, including interleukins (IL), IFN-β, interferoninduced proteins with tetratricopeptide repeats (IFIT), melanoma differentiation-associated protein 5 (MDA-5), tumor necrosis factor alpha (TNF-α), and macrophage inflammatory protein-1 alpha (MIP-1α). Subsequent in vivo studies will focus on evaluating antigenspecific adaptive immune responses during both the primary and memory phases. Taking the HIV-1 vaccine with DNA priming and VACV boosting as an example, CD4+ and CD8+ T-cell as well as antibody responses have been extensively investigated, to address the potential improvements resulting from the deletion of immune modulators in the vector. The secretion of IFN-γ, TNF-α, IL-2, and surface expression of CD107a by T cells upon peptide stimulation, targeting HIV Gag, Env, and Gag-Pol-Nef fusion proteins, can well represent the magnitude, breadth, and polyfunctionality of antigen-specific immune responses. Total IgG responses against HIV-1 gp120 protein and the VACV vector serve as another arm for quantification, revealing the quality of immune responses in the absence of the vaccinia immunomodulatory gene.
Multiple VACV genes have gained particular interest. The A40R gene is involved in the early phase of infection and is reported to play a role in virulence [86] or is an essential gene [87]. It encodes a type II membrane glycoprotein found on the surface of infected cells [88], but it has also been shown to form a complex with a small SUMO-1 peptide, which is believed to aid in viral replication [87]. The A46R gene encodes a viral protein containing a Toll-interleukin receptor (TIR) domain which binds to adaptor molecules involved in Toll-like receptor (TLR) signaling pathways [88], functioning essentially as a TLR signaling inhibitor. A41L is known to counteract the function of CC-chemokines by reducing their concentration instead of antagonizing the receptor binding process [89]. A52R is involved in disrupting several TLR signaling pathways by blocking the activation of nuclear factor kappa B (NF-κB) [90]. The expression product of B2R has nuclease activity and destabilizes the cellular cyclic GMP-AMP (2 3 cGAMP) messenger that is involved in stimulator of interferon genes (STING)-dependent antiviral responses [91]. B7R encodes a soluble decoy for TNF and binds specifically to several chemokines [92]. B8 is a soluble glycoprotein binding and inhibiting a wide spectrum of IFN-γ [93]. B9R produces a small intracellular protein responsible for viral replication in vivo [94]. B13R is an early gene, and its product is known to prevent apoptosis in infected cells [95]. B15R (equivalent to B14 in the Western Reserve strain) encodes an intracellular virulence factor that disrupts TLR signaling by inhibiting I kappa B (IκB) kinase [96]. A soluble IL-1β receptor is expressed by B16R and interferes with the acute phase of innate immunity [97]. B19 is a soluble glycoprotein that broadly binds to type I IFNs from a wide host range [98]. C6L is an early gene that serves as an IFN-β inhibitor to block interferon regulatory factor (IRF) signaling [99]. C7L is a viral host range gene, and its expression product is anti-apoptotic, inhibiting type I interferons [100][101][102]. C12L codes for an IL-18-binding protein that blocks IFN-γ expression [103]. E5R encodes a virulence factor that also directly inhibits cellular cyclic GMP-AMP synthase (cGAS) to evade the host cGAS/STING pathway [104]. K7R is a member of the VACV Bcl-2 family and upregulates histone methylation as well as blocks TLR signaling by inhibiting NF-κB and IRF3 activation [105][106][107]. N2L also belongs to the Bcl-2 family, functioning as an inhibitor of TLR signaling pathways [107,108].
A list of reported attempts to delete immunomodulatory genes in VACV vectors over the past ten years can be seen in Table 1 (not exhaustive). However, enhanced immune responses do not always result from the generation of deletion mutants, largely due to inadequate knowledge of gene functions and the complexity of systematic interactions, especially when multiple genes are deleted. An example of improved immunogenicity was reported in a study by Pérez and co-workers on an HIV-1 vaccine candidate based on an A40R-deleted MVA [109]. They showed that A40 protein is not required for viral growth in vitro, as the deletion mutant grew to similar titers as the parental virus in DF-1 cells. In the absence of A40R gene, the mRNA levels of IFN-β, IFIT1&2, MDA-5, and MIP-1α significantly increased in human THP-1 cells, except for IFN-β, which showed an increase at 3 h post-infection (hpi), while the others showed an increase at 6 hpi. The magnitude of HIV-1-specific primary CD4+ and CD8+ T-cell responses were, respectively, 2.3-and 1.4-fold higher in mice boosted with the deletion mutant. Additionally, the polyfunctionality of antigen-specific T cells was improved, as 90% of stimulated CD4+ T cells expressed at least two types of cytokines or surface markers with CD107a-IFNγ-TNF-α-IL-2and IFN-γ-TNF-α-IL-2-populations being the dominant ones. CD8+ T cells were also polyfunctional upon stimulation with 80% of them displaying at least two hallmarks. The most abundant populations were CD107a-IFN-γ-TNF-α and CD107a-IFN-γ. Elevated memory T-cell responses were detected almost 8 weeks post boost vaccination, with twice as high percentages of antigen-specific CD4+ and CD8+ T cells in mice that received the deletion mutant. These memory T cells exhibited polyfunctionality similar to those in primary responses. Particularly, during both the primary and memory phases, the deletion of the A40R gene elicited a remarkably higher percentage of T effector memory and T effector cells. Quantitation of total anti-gp120 IgG (IgG1, IgG2a, and IgG3) revealed a significant increase in humoral responses in the absence of A40R gene both 10 and 53 days post-boost. Contrary to the phenotypes reported above, Marín et al. developed an MVA based hepatitis C virus vaccine candidate with C6L gene deletion but did not obtain comparably improved immune responses [110]. Despite similar growth kinetics to the parental virus without C6L deletion in DF-1 cells, the mutant still indistinguishably downregulated IFN-β, IFIT1&2, and TNF-α in infected human macrophages and monocytederived dendritic cells. The levels of HCV-specific CD8+ T-cell responses were essentially unchanged upon deletion during both primary and memory phases except that the deletion mutant was more NS3 specific while the parental virus was more p7 and NS2 specific. Anti-E2 and anti-vector total IgG responses were also similar.

Other Aspects of Innovation
Information can become overwhelming if disorganized, and it is particularly true for research on VACV vectors. Even within the scope of this article, the extensive discussion on safety mechanisms, rapid purification of recombinant VACVs, and modifications in viral immunomodulatory genes will need to be updated in the next few years. Identifying and locating new reports on construction and application of VACV vectors can be time consuming. The search results may be inexhaustive and new experiments based on the incomplete list may lead to repeatedly produced data. A user-friendly database that can be conveniently updated should unquestionably receive more attention. An example of a web-based database, named Vaxvec, for recombinant vaccine vectors was reported by Deng and co-workers [126]. It falls under the umbrella of VIOLIN (Vaccine Investigation and Online Information Network), the first platform that includes thousands of vaccines and hundreds of diseases [127,128] but provides additional information specifically for vaccine vectors and antigens of interest in these recombinant vectors. For instance, a page displaying recombinant vectors will include links to all vaccines that use the vector as well as the protective antigens. The authors mentioned that the addition of immune responses to this database will tailor this platform to the booming research on vectored vaccine development.
We still have a long way to go to understand the VACV genome. The very fact that viral replication occurs intracellularly and is hard to track necessitates an advancement in labeling methods. Luminescence imaging can serve as a powerful tool to visualize infection in cells or live animals, e.g., the expression product of a luciferase reporter gene inserted into the VACV genome can react with substrates available in cell culture or circulation in an animal and emit light that can be captured [129]. Similarly, Kieser and co-workers combined live time-lapse fluorescence microscopy with DNA labeling to track the replication and recombination events of VACV [130]. The bacteriophage lambda DNA-binding protein, cro, when fused with a fluorescent protein such as GFP, can nonspecifically bind to DNA and therefore locate any cytoplasmic DNA during infection. Another DNA labeling technology, ANCHOR, was adopted by Gallardo et al. to evaluate VACV infection kinetics by inserting to its genome an ANCH sequence and an OR3-Santaka sequence [131]. The ANCH sequence in viral DNA copies provides a docking site for OR-Santaka fluorescent protein which oligomerizes upon binding to this sequence. Livecell imaging can then detect the behaviors of labeled viral DNA particles in the cytoplasm and reveal dynamic viral replication. This fluorescent tagging signal remained strong even after cell fixation and other processing procedures used in common staining methods. However, early replication steps with minimal protein involvement can pose a challenge to this approach. To address this limit, Mok et al. demonstrated a tracking system using a nucleoside analog, 5-ethynyl-2 -deoxyuridine (EdU), that can locate single viral genomes in replication factories or even when they are still packed in the viral core [132]. An azide (N3 − )-containing fluorescent dye can then covalently form a heteroatom ring with the nucleoside analog for easy visualization, and this process is not sensitive to crowded cytoplasmic compartments due to the involvement of only small molecules.
The research and development of novel VACV vectors admittedly draw the most attention, but the manufacturing and delivery of vaccines also deserve recognition. To replace chick embryo fibroblasts (CEF), which are primary cells with a limited lifespan, new permanent cell lines have been generated for MVA production in bioreactors, including cell lines derived from ducks, quails, and monkeys [133][134][135][136]. Jordan and colleagues reported a biphasic suspension culture system, scalable to 200-L bioreactors, aiming for high cell density during phase 1, while promoting aggregation when a chemical inducer is added in phase 2 to facilitate viral spread and growth [137]. A titer higher than 10 8 PFU/mL of viruses was achieved using just a few million cells [138]. Chen et al. reported two new freeze-drying formulations to significantly increase the storage time of a VACV vectored vaccine without compromising antigen stability [139]. The vaccine was mixed with a formula consisting of polyethylene glycol, dextran, bovine serum albumin, and L-glutamic acid at different ratios, lyophilized, and kept at either 4 • C or 25 • C for up to four weeks. Notwithstanding that all formulae similarly preserved titer at 4 • C for four weeks, with the maximal loss being less than 30-fold even for PBS-formulated vaccine, only two of the five formulae managed to limit the loss of titer at 25 • C to less than 100-fold, while the other groups had a nearly million-fold decrease. The antigen-specific T-cell and antibody responses induced by the formulated vaccines were indistinguishable from those of the non-freeze-drying control.

Future Perspectives
Vaccinia virus has been long known to us since its use as a vaccine to eradicate smallpox. The advancement of genetic and immunologic research not only improved the safety profile of VACV as a smallpox vaccine but also as a recombinant vector [204]. A multitude of approaches to finely controlling viral replication by introducing inducible systems or complementing cell lines yielded exciting results in animal studies. Innovative methods to accelerate the generation of recombinant VACVs considerably reduced the turnover time to obtain and test a desired phenotype. Deepening the understanding of viral gene functions offered the opportunity to bring vector modification to the next level for a more immunogenic and protective vaccine or therapeutic vector. The participation of multidisciplinary efforts enabled the rational design of new VACV vectors with easily accessible data support, the exposure and monitoring of previously hidden replication events using novel labeling and imaging techniques, and the cost-efficient manufacture and distribution of vaccines beyond the R&D stage.
We are at the dawn of explosive technological advancement and should welcome changes to traditional research practices as well. Artificial intelligence (AI) integrated platforms may well mark the beginning of a new era of literature review and data management. For example, ChatGPT is a chatbot developed by OpenAI. Since its release in November 2022, billions of users have been amazed by its outstanding performance. It can handle computer coding work, consolidate information, write poems and songs, and take tests or exams. Limitations do exist, such as providing ostensible answers that are factually nonsensical. However, we cannot neglect the potential of incorporating AI into scientific research, including in the development of VACV-vectored vaccines. Imagine an AI-assisted or even AI-operated platform that collects all publications on VACV by exhaustion and categorizes them by topic or field. Critical findings and conclusions are grouped and cross-compared to return a pattern of research interests shared among researchers. Input from new studies, such as identification of gene functions, mapping of epitopes, elucidation of immune responses, inter alia, can be conveniently added in a user-friendly manner. Redundancy in similar study designs will be minimized, if not eliminated altogether, and the consolidation of information in one comprehensive view should help us shift from reductionist biology to systems biology in the next step. If the process of literature review could transition from a brick-and-mortar library to the Internet around two decades ago, the same transition should be embraced now with the advent of AI.
High-throughput automation is no longer a dream, even in a small laboratory, as the exorbitant prices of robots for PCR, ELISA, imaging, etc., are no longer an obstacle. One can anticipate the streamlined full spectrum of VACV research in most lab settings. The generation and purification of recombinant viruses will be handled automatically in 96-well or 384-well microplates in addition to the built-in genetic modifications that ensure predominant populations of desired recombinants. All possible combinations of gene deletion or modification can be tested in vitro before modeling in animals, after which the automatic characterization of T-cell and antibody responses would also be achieved. The viral infection and replication will be visualized and imaged live with innovative DNA labeling to allow for high-resolution monitoring at every step involved in these events. "A picture is worth a thousand words, then a video has to be worth at least 1.8 million words", says McQuivey in "How Video Will Take Over the World". While live images of biological events provide unprecedented insights into the interactions between biomacromolecules, prolonged dynamic imaging, enabled by automation, can transform a picture into a video, capturing the trajectory of molecules and adding a temporal axis.
Incessant investigations into the biology and immunology of VACV, coupled with technological innovations in automation, centralized information management, highthroughput purification, cost-efficient manufacturing, and optimized formulation, are now rejuvenating this 'ancient' virus that once made it possible to claim victory against smallpox. Its path is now converging with ours yet again and will surely bear more fruit in the future.

Conflicts of Interest:
The author declares no conflict of interest.