A Novel Effective and Safe Vaccine for Prevention of Marek’s Disease Caused by Infection with a Very Virulent Plus (vv+) Marek’s Disease Virus

Marek’s disease virus (MDV) is a highly contagious alphaherpesvirus that causes rapid onset lymphoma in chickens. Marek’s disease (MD) is effectively controlled using vaccination; however, MDV continues to break through vaccinal immunity, due to the emergence of highly virulent field strains. Earlier studies revealed that deletion of the meq gene from MDV resulted in an attenuated virus that protects against MD in chickens challenged with highly virulent field strains. However, the meq deleted virus retains the ability to induce significant lymphoid organ atrophy. In a different study, we found that the deletion of the vIL8 gene resulted in the loss of lymphoid organ atrophy in inoculated chickens. Here, we describe the generation of a recombinant MDV from which both meq and vIL8 genes were deleted. In vitro studies revealed that the meq and vIL8 double deletion virus replicated at levels similar to the parental very virulent plus (vv+) virus. In addition, in vivo studies showed that the double deletion mutant virus (686BAC-ΔMeqΔvIL8) conferred protection comparable to CVI988, a commercial vaccine strain, when challenged with a vv+ MDV virus, and significantly reduced lymphoid organ atrophy, when compared to meq null virus, in chickens. In conclusion, our study describes the development of a safe and effective vaccine candidate for prevention of MD in chickens.

MD is the first tumor disease that has been successfully controlled using vaccination [6], and strains from MDV-1 (refers to MDV in following text), MDV-2, and HVT have been used as vaccines to protect against MD [3,7,8]. An MDV cell culture passage attenuated viral strain, HPRS-16/att, was first used to protect chickens against challenge with v MDV strains [9]. HVT started being used in the 1970s and was followed by the

Construction of meq and/or vIL8 Single and Double Deletion Viruses
To delete meq and vIL8 genes, individually or together, from the 686BAC, a two-step Red-mediated recombination procedure was performed, as previously described [36,37]. Briefly, the Kan R -I-SecI cassette was amplified from pEPkan-S and electroporated into the 686BAC-containing E. coli, where meq or vIL8 genes were replaced with the Kan R -I-SecI cassette. The Kan R sequence was then deleted by the addition of arabinose to generate 686BAC with a single copy of meq or vIL8 genes. Due to the presence of two copies of meq and vIL8 in the MDV genome, this procedure was repeated to generate 686BAC-∆Meq and 686BAC-∆vIL8 mutants where both copies of meq or vIL8 were deleted. Then, 686BAC-∆Meq was used as the backbone to generate the meq and vIL8 double deletion mutant (686BAC-∆Meq∆vIL8) using the same procedure. All mutant BAC clones were screened by PCR, followed by DNA sequencing and restriction fragment length polymorphism (RFLP) to confirm the deletion of meq and vIL8 and the absence of unexpected mutations. Primers used to construct all mutant BAC clones are shown in Table 1. All BAC DNAs were transfected into CEF to produce meq or vIL8 single and double deletion mutant viruses.

Immunofluorescence (IFA) and Immunohistochemistry (IHC) Assays
IFA was carried out as previously described with modifications [38]. Briefly, infected or transfected CEF were washed with phosphate-buffered saline (PBS) and fixed with ice-cold acetone:methanol (6:4) for 10 min. After blocking with non-fat dry milk, the cells were incubated with MDV pp38 monoclonal antibody (1:500) for 1 hour, followed by another hour of incubation with goat anti-mouse fluorescein isothiocyanate (FITC)-labeled secondary antibodies (KPL, Gaithersburg, MD, USA). Cells were then washed three times with PBS and examined under a fluorescence microscope.
For IHC, lymphoid organs (thymus, spleen, and bursa of Fabricius) and skin with feather follicles were embedded in optimal cutting temperature compound (Tissue-Tek, OCT, Sakura Finetek, Torrance, CA, USA), frozen immediately in liquid nitrogen, and stored at −80 • C until use. Six to eight µm-thick cryostat sections were generated and subjected to immunostaining with MDV pp38 monoclonal antibody (1:3200) and the Vectastain ABC kit (Vector Laboratories, Burlingame, CA, USA) according to the manufacturer's instructions.

MDV Genome Copy Numbers
To compare the replication of 686BAC, 686BAC-∆Meq, 686BAC-∆vIL8, and 686BAC-∆Meq∆vIL8 viruses in vivo, one-day-old specific pathogen free (SPF) and MDV maternal antibodies free (Ab-) chickens (Charles Rivers Laboratories, Wilmington, MA, USA) were inoculated subcutaneously with 2000 PFU of each virus; one group was not inoculated and served as the negative control. Three chickens from each group were euthanatized at 5, 14, or 56 days post-inoculation (dpi) and spleen samples were collected. Genomic DNA was extracted from chicken splenocytes using the phenol-chloroform method, and MDV genome copy number was measured by quantitative polymerase chain reaction (qPCR) using primers specific for MDV-infected cell protein 4 (ICP4) and chicken GAPDH genes (Table 1) [39,40]. All qPCR assays were carried out in the Bio-Rad iCycler iQ Multicolor Real-Time Detection System (Bio-Rad, USA), using iTag SYBR supermix buffer (Bio-Rad, USA). Results are presented as the ratio of MDV DNA copy numbers divided by the copy number of GAPDH, with error bars representing standard error of the mean (SEM).

Virus Reactivation Assay
To examine the reactivation efficacy of mutant viruses, 3 chickens from each experimental group were randomly selected and bled at 14 dpi for virus reactivation assay. Briefly. CEF monolayers (seeded on 35 mm plates, in duplicate) were co-seeded with 10 6 peripheral blood lymphocytes (PBL) isolated from 686BAC, 686BAC-∆Meq, 686BAC-∆vIL8, or 686BAC-∆Meq∆vIL8 viruses inoculated chickens (SPF, Ab-) or negative control chickens at 14 dpi. Plaques were counted 7 days after infection, and data were presented as average plaque numbers ± SEM.

Pathogenesis of meq or vIL8 Single and Double Deletion Mutant Viruses in SPF Ab-Chickens
One-day-old SPF Ab-chickens were wing-banded and randomly sorted into 5 experimental groups. Chickens were inoculated subcutaneously with 2000 PFU of parental 686BAC, 686BAC-∆Meq, 686BAC-∆vIL8, and 686BAC-∆Meq∆vIL8 viruses, or were not inoculated and kept as negative control. All animal experiments were carried out in accordance with Texas A&M University Institutional Animal Care and Use Committee (IACUC) approved protocol. Lymphoid organ atrophy: To evaluate the effect of virus replication in lymphoid organ atrophy, 5 randomly selected chickens from each experimental group were euthanatized at 14 dpi, and thymus and bursa were weighed. Results were presented as the average ratio of lymphoid organs weight to body weight of 5 chickens multiplied by one hundred. b Pathogenesis: To compare the pathogenic properties of parental 686BAC, 686BAC-∆Meq, 686BAC-∆vIL8, and 686BAC-∆Meq∆vIL8 viruses, the mortality of each experimental group was recorded daily for 65 days. All chickens that died during the experiment or were euthanized at the end of the experiment were necropsied and examined for MD associated gross tumors in visceral organs and nerves.

Vaccine Protection Experiments
To study the protection efficacy of 686BAC-∆Meq∆vIL8, one-day-old SPF Ab-and MDV maternal-antibody-positive (Ab+) commercial White Leghorn male chickens (Hyline International, Bryan, TX, USA) were not vaccinated or vaccinated with 2000 PFU of 686BAC-∆Meq, 686BAC-∆Meq∆vIL8, or CVI988 by the subcutaneous route. Five days later, vaccinated and unvaccinated chickens were challenged subcutaneously with 500 PFU of 686. Chickens that died or survived to the end of the experiment (56 days post challenge) were necropsied and examined for MD associated gross tumors. Vaccine protection efficacy was expressed as protective index (PI), as previously described [29,32].

Data and Statistical Analysis
For in vitro growth kinetics, data represent an average of duplicates and were analyzed by ANOVA for each individual time point. The relative lymphoid organs to body weight ratios were averaged over 5 independently collected samples in each group and analyzed by Student t test. The trends of the chicken survival curve were examined with LogRank and Wilcoxon tests. The protective index (PI) among the different groups was analyzed by Chi Square test. All statistical analyses were performed with GraphPad Prism software (GraphPad Software, Inc. La Jolla, CA, USA). A value of p < 0.05 was considered statistically significant.

Construction and In Vitro Characterization of meq or vIL8 Single and Double Deletion Viruses
MDV meq and vIL8 genes ( Figure 1A) alone or in combination were deleted from 686BAC DNA via a two-step Red-mediated recombination method to generate 686BAC-∆Meq, 686BAC-∆vIL8, and 686BAC-∆Meq∆vIL8, respectively. To confirm the absence of an unexpected recombination associated with the deletion processes, BAC DNAs were subjected to RFLP analysis using BamHI and EcoRI enzymes. RFLP results showed that the patterns obtained correspond exactly to those predicted by in silico analysis without unexpected rearrangements in the mutant MDV genomes ( Figure 1B). The BamHI fragment corresponding to the meq gene in 686BAC is 5133 bp and 1288 bp, whereas in 686BAC-∆Meq, we expect a single 5404 bp fragment (unable to resolve from 5133) with the loss of the 1288 bp fragment (*). The BamHI fragment corresponding to the vIL8 gene in 686BAC is 3092 bp (*), whereas in 686BAC-∆vIL8 it is 2414 bp (*). Digestion of 686BAC-∆Meq∆vIL8 with BamHI resulted in 5404 bp and 2414 bp fragments, with the loss of 1288 bp (*) and 3092 bp fragments. On the other hand, the EcoRI fragment corresponding to the meq gene in 686BAC is 2456 bp (*), whereas in 686BAC-∆Meq it is 1439 bp (*). The EcoRI fragment corresponding to the vIL8 gene in 686BAC is 9476 bp, which is reduced to 8798 bp in 686BAC-∆vIL8 (unable to resolve from 9476 bp fragment). Digestion of 686BAC-∆Meq∆vIL8 with EcoRI resulted in 8798 bp (unable to resolve from 9476 bp fragment) and 1439 bp fragments (*).
were detected.
Immunofluorescence assay (IFA) of BAC transfected cells showed that expression of viral protein pp38 and plaque size were similar in parental and deletion mutant viruses infected CEF ( Figure 1D). In addition, parental 686BAC, 686BAC-∆Meq, 686BAC-∆vIL8, and 686BAC-∆Meq∆vIL8 showed similar in vitro growth kinetics ( Figure 1E), suggesting that deletion of meq and/or vIL8 did not affect the growth of any of the viruses in cell culture.  . 686BAC, 686BAC-∆Meq, 686BAC-∆vIL8 or 686BAC-∆Meq∆vIL8 BAC transfected CEF were subjected to IFA using MDV pp38 specific monoclonal antibody and FITC conjugated secondary antibody. Scale bar = 100 µm. (E) In vitro growth kinetics. CEF were infected with 100 plaque-forming units (PFU) of 686BAC, 686BAC-∆Meq, 686BAC-∆vIL8 and 686BAC-∆Meq∆vIL8 viruses. Infected cells were trypsinized, diluted, and co-seeded with fresh CEF at the indicated time points, and plaques were counted 6 days post infection. Each point represents two independent experiments and data present average plaque numbers ± standard error of the mean (SEM).
In addition, PCR amplification was performed to confirm the deletions of meq and/or vIL8, and the MDV ribonucleotide reductase (RR) gene served as an internal control, as it should not be affected by the mutation process. As shown in Figure 1C, the RR gene was amplified from genomic DNA isolated from chicken embryonic fibroblasts (CEF) infected with all viruses ( Figure 1C, RR: lanes 1-5), while meq was amplified only from 686, 686BAC, and 686BAC-∆vIL8 viruses ( Figure 1C, meq: lanes 1, 2, and 4), and vIL8 was amplified only from 686, 686BAC, and 686BAC-∆Meq ( Figure 1C, vIL8: lanes 1, 2, and 3). Neither meq nor vIL8 were amplified from genomic DNA isolated from 686BAC-∆Meq∆vIL8 virus infected CEF ( Figure 1C, meq and vIL8, lanes 5). The manipulated regions of meq and vIL8 were sequenced using PCR products and no unexpected mutations were detected.
Immunofluorescence assay (IFA) of BAC transfected cells showed that expression of viral protein pp38 and plaque size were similar in parental and deletion mutant viruses infected CEF ( Figure 1D). In addition, parental 686BAC, 686BAC-∆Meq, 686BAC-∆vIL8, and 686BAC-∆Meq∆vIL8 showed similar in vitro growth kinetics ( Figure 1E), suggesting that deletion of meq and/or vIL8 did not affect the growth of any of the viruses in cell culture.

In Vivo Replication and Reactivation of meq or vIL8 Single and Double Deletion Viruses
Early cytolytic infection of MDV occurs around 2-6 days post-infection followed by the latent phase, starting at 7-8 days post-infection, and transformed cells appear as early as 2 weeks post-infection. To examine the replication of deletion mutant viruses in vivo, one-day-old MDV Ab-chickens were inoculated with 2000 plaque-forming units (PFU) of parental 686BAC, 686BAC-∆Meq, 686BAC-∆vIL8 or 686BAC-∆Meq∆vIL8 viruses and their replication rates (virus genome copy number) were measured in spleen samples taken at 5, 14, and 56 days post-inoculation (dpi). At 5 dpi, parental 686BAC and 686BAC-∆Meq inoculated chickens had similar virus genome copy numbers and these were significantly higher than for 686BAC-∆vIL8 and 686BAC-∆Meq∆vIL8 inoculated chickens (Figure 2A). At 14 dpi, the genome copy number in 686BA-∆Meq and 686BAC-∆Meq∆vIL8 inoculated chickens was significantly lower than in the parental 686BAC group. While the genome copy number in the 686BAC-∆vIL8 group increased by~100× compared to day 5, at 14 dpi it was still lower than in the parental 686BAC group (Figure 2A). These results are consistent with the virus reactivation assay results at 14 dpi ( Figure 2B). At 56 dpi, the genome copy number in all groups was low or not tested (NT), due to the death of inoculated chickens (Figure 2A). lane 4: 686BAC-ΔvIL8 virus; lane 5: 686BAC-ΔMeqΔvIL8 virus. (D) Immunofluorescence assay (IFA). 686BAC, 686BAC-ΔMeq, 686BAC-ΔvIL8 or 686BAC-∆Meq∆vIL8 BAC transfected CEF were subjected to IFA using MDV pp38 specific monoclonal antibody and FITC conjugated secondary antibody. Scale bar = 100 µm. (E) In vitro growth kinetics. CEF were infected with 100 plaque-forming units (PFU) of 686BAC, 686BAC-ΔMeq, 686BAC-ΔvIL8 and 686BAC-ΔMeqΔvIL8 viruses. Infected cells were trypsinized, diluted, and co-seeded with fresh CEF at the indicated time points, and plaques were counted 6 days post infection. Each point represents two independent experiments and data present average plaque numbers ± standard error of the mean (SEM).

In Vivo Replication and Reactivation of meq or vIL8 Single and Double Deletion Viruses
Early cytolytic infection of MDV occurs around 2-6 days post-infection followed by the latent phase, starting at 7-8 days post-infection, and transformed cells appear as early as 2 weeks post-infection. To examine the replication of deletion mutant viruses in vivo, one-day-old MDV Ab-chickens were inoculated with 2,000 plaque-forming units (PFU) of parental 686BAC, 686BAC-∆Meq, 686BAC-∆vIL8 or 686BAC-∆Meq∆vIL8 viruses and their replication rates (virus genome copy number) were measured in spleen samples taken at 5, 14, and 56 days post-inoculation (dpi). At 5 dpi, parental 686BAC and 686BAC-∆Meq inoculated chickens had similar virus genome copy numbers and these were significantly higher than for 686BAC-∆vIL8 and 686BAC-∆Meq∆vIL8 inoculated chickens (Figure 2A). At 14 dpi, the genome copy number in 686BA-∆Meq and 686BAC-∆Meq∆vIL8 inoculated chickens was significantly lower than in the parental 686BAC group. While the genome copy number in the 686BAC-∆vIL8 group increased by ~100x compared to day 5, at 14 dpi it was still lower than in the parental 686BAC group (Figure 2A). These results are consistent with the virus reactivation assay results at 14 dpi ( Figure 2B). At 56 dpi, the genome copy number in all groups was low or not tested (NT), due to the death of inoculated chickens (Figure 2A).  To further examine virus replication in chickens, MDV pp38 expression was evaluated by immunohistochemistry (IHC) assay. Three MDV Ab-chickens inoculated with parental 686BAC or 686BAC-∆Meq∆vIL8 viruses were euthanized at 5 dpi and their lymphoid organs (thymus, bursa of Fabricius, and spleen) examined for pp38 expression. Our results show that there were high levels of pp38 expression in the lymphoid organs of parental 686BAC inoculated chickens, while almost no pp38 expression was detected in 686BAC-∆Meq∆vIL8 inoculated chickens (Figure 3). Taken together, the in vivo MDV genome copy number data and pp38 expression results suggest that double deletion of meq and vIL8 significantly reduces virus replication in chickens. We also examined the replication of 686BAC-∆Meq∆vIL8 virus in feather follicle epithelium (FFE), where fully infectious viral particles are produced. IHC results showed that chickens inoculated with 686BAC-∆Meq∆vIL8 virus had levels of pp38 expression similar to parental 686BAC virus at 14 dpi (Figure 3, FFE), suggesting that double deletion of meq and vIL8 did not abrogate virus replication in FFE.
parental 686BAC or 686BAC-∆Meq∆vIL8 viruses were euthanized at 5 dpi and their lymphoid organs (thymus, bursa of Fabricius, and spleen) examined for pp38 expression. Our results show that there were high levels of pp38 expression in the lymphoid organs of parental 686BAC inoculated chickens, while almost no pp38 expression was detected in 686BAC-∆Meq∆vIL8 inoculated chickens (Figure 3). Taken together, the in vivo MDV genome copy number data and pp38 expression results suggest that double deletion of meq and vIL8 significantly reduces virus replication in chickens.
We also examined the replication of 686BAC-∆Meq∆vIL8 virus in feather follicle epithelium (FFE), where fully infectious viral particles are produced. IHC results showed that chickens inoculated with 686BAC-∆Meq∆vIL8 virus had levels of pp38 expression similar to parental 686BAC virus at 14 dpi (Figure 3, FFE), suggesting that double deletion of meq and vIL8 did not abrogate virus replication in FFE. , lymphoid organs (spleen, thymus, and bursa) and feather follicle epithelium (FFE) were collected, respectively, from chickens inoculated with 686BAC or 686BAC-∆Meq∆vIL8 viruses, or negative control chickens. Tissue samples were frozen immediately using the O.C.T. (optimal cutting temperature) compound and 6 to 8 µm-thick cryostat sections were stained with MDV pp38 specific monoclonal antibody. Scale bar = 50 µm.

Evaluation of Lymphoid Organ Atrophy and Pathogenesis Induced by Inoculation of Meq or Vil8 Single and Double Deletion Viruses
We have reported earlier that the deletion of meq results in the complete loss of oncogenicity but does not eliminate lymphoid organ atrophy; while the deletion of vIL8 significantly reduces MDV transformation, due to reduced virus replication in lymphoid organs [21,33]. Therefore, we hypothesized that the deletion of both genes would result in a non-oncogenic virus that does not cause lymphoid organ atrophy.
To evaluate the lymphoid organ atrophy and pathogenic properties of 686BAC-∆Meq∆vIL8 virus, one-day-old MDV Ab-chickens were inoculated with parental and Figure 3. Replication of meq and vIL8 double deletion virus in lymphoid organs and feather follicle epithelium (FFE) of inoculated chickens. At 5 or 14 days post-inoculation (dpi), lymphoid organs (spleen, thymus, and bursa) and feather follicle epithelium (FFE) were collected, respectively, from chickens inoculated with 686BAC or 686BAC-∆Meq∆vIL8 viruses, or negative control chickens. Tissue samples were frozen immediately using the O.C.T. (optimal cutting temperature) compound and 6 to 8 µm-thick cryostat sections were stained with MDV pp38 specific monoclonal antibody. Scale bar = 50 µm.

Evaluation of Lymphoid Organ Atrophy and Pathogenesis Induced by Inoculation of meq or vIL Single and Double Deletion Viruses
We have reported earlier that the deletion of meq results in the complete loss of oncogenicity but does not eliminate lymphoid organ atrophy; while the deletion of vIL8 significantly reduces MDV transformation, due to reduced virus replication in lymphoid organs [21,33]. Therefore, we hypothesized that the deletion of both genes would result in a non-oncogenic virus that does not cause lymphoid organ atrophy.
To evaluate the lymphoid organ atrophy and pathogenic properties of 686BAC-∆Meq∆vIL8 virus, one-day-old MDV Ab-chickens were inoculated with parental and mutant viruses or remained uninoculated and served as negative control. At 14 dpi, lymphoid organs (bursa and thymus) collected from five chickens from each group were weighted. As shown in Figure 4A, compared to negative control chickens, parental 686BAC and 686BAC-∆Meq viruses, but not 686BAC-∆vIL8 and 686BAC-∆Meq∆vIL8 viruses, induced severe bursa and thymus atrophy in inoculated chickens. before the end of the experiment, while MD associated mortality began at 37 dpi in 686BAC-∆vIL8 group, and 6 out of 15 (40%) chickens survived for the duration of the experiment ( Figure 4B). All the chickens that died in the 686BAC and 686BAC-∆vIL8 groups had MD specific gross tumors. However, no MD specific mortality and tumors were observed in the uninoculated control, 686BAC-∆Meq, and 686BAC-∆Meq∆vIL8 groups (Figure 4B). Taken together, these results indicated that 686BAC-∆Meq∆vIL8 was fully attenuated.

686BAC-∆Meq∆vIL8 Provides Protection Comparable to CVI988 Against Challenge with a vv+ MDV
The protection efficacy of 686BAC-∆Meq and 686BAC-∆Meq∆vIL8 was compared to that of CVI988, the gold standard MD vaccine, in both Ab-and Ab+ chickens challenged with a vv+ MDV (686) strain. In the Ab-group, MD specific mortality and gross tumors were observed in 100% of the unvaccinated chickens ( Figure 5A and Table 2). In the CVI988 vaccinated group, MD specific mortality was observed in 13% of the chickens, while no mortality was observed in the 686BAC-∆Meq or 686BAC-∆Meq∆vIL8 vaccinated To further evaluate the pathogenic properties of 686BAC-∆Meq∆vIL8, all inoculated and uninoculated chickens were closely monitored for 65 days. Our results show that MD associated mortality began at 16 dpi in the parental 686BAC group, and all chickens died before the end of the experiment, while MD associated mortality began at 37 dpi in 686BAC-∆vIL8 group, and 6 out of 15 (40%) chickens survived for the duration of the experiment ( Figure 4B). All the chickens that died in the 686BAC and 686BAC-∆vIL8 groups had MD specific gross tumors. However, no MD specific mortality and tumors were observed in the uninoculated control, 686BAC-∆Meq, and 686BAC-∆Meq∆vIL8 groups ( Figure 4B). Taken together, these results indicated that 686BAC-∆Meq∆vIL8 was fully attenuated.

Discussion
Since the early 1970s, the poultry industry has relied on the use of vaccines to control losses due to MD. These vaccines, although effective at preventing mortality and tumor formation, do not prevent infection by field strains, and are thought to have contributed to the emergence of more virulent field strains [5,[41][42][43]. The lack of more effective MD vaccines has led the poultry industry to rely on the use of multiple vaccine doses and/or multivalent vaccines [44], significantly increasing vaccination cost. Although several MDV-derived vaccine candidates have been generated, only few have been successful. MDV CVI988/Rispens is the most efficacious vaccine in the market and is considered the 'gold-standard' [45][46][47].
The mechanisms of MD vaccine protection are still not well understood, but it has been suggested that in order to be effective, an MD vaccine has to be able to replicate to sufficient levels to stimulate an immune response. Attempts have been made to attenuate highly virulent MDV strains, by serial passage in cell culture, to develop more efficacious MD vaccines. It has been shown that partially attenuated strains, which were still mildly oncogenic, induced higher protection than CVI988/Rispens, while fully attenuated passages conferred only limited protection [48,49]. It was also been shown that highly protective vaccines replicate to higher levels than low protective vaccines in lymphoid organs of vaccinated chickens [50]. These data suggest that the protection efficacy of MD vaccines is highly associated with its replication.
A number of laboratories have used molecular techniques to identify and study MDV genes involved in MDV pathogenesis, viral replication, or tumor formation. These techniques have facilitated, through the deletion or mutation of genes associated with pathogenesis, the development of attenuated recombinant viruses which could be used as potential vaccines. One of the most promising candidates was an MDV mutant virus lacking both copies of the meq oncogene (MDV-∆Meq) [21]. Deletion of meq rendered the virus non-oncogenic, while maintaining a normal early cytolytic infection, and proved to be an effective vaccine candidate in laboratory and/or field conditions [29,30]. However, like the parental vv MDV, MDV-∆Meq causes lymphoid organ (bursa and thymus) atrophy in highly susceptible MDV Ab-chickens [31,32], an important safety concern that has interfered with its commercialization. In order to improve the safety of MDV-∆Meq virus, Lee et al. attempted to further attenuate it by serial passage in cell culture. Although at passage 40 MDV-∆Meq virus did not cause lymphoid organ atrophy in MDV susceptible chickens, the protection efficacy in MDV Ab+ chickens was reduced [32], potentially due to the introduction of random mutations in genes essential for virus replication. Thus, this approach was not satisfactory for improving the safety of MDV-∆Meq virus without negatively affecting its efficacy as vaccine.
An alternative strategy, used in this study, was to introduce known secondary mutations in the MDV genome, by deleting or mutating a second gene involved in virus replication. It has been shown that the introduction of a point mutation in UL5, a helicase primase subunit involved in virus replication in vivo, provided partially protection against v and vv MDV [51]. A subsequent study showed that the introduction of the UL5 point mutation in MDV-∆Meq virus eliminated the lymphoid organ atrophy associated with MDV-∆Meq virus; however, the protection efficacy of the resulting virus (PI = 51%) was significantly lower than MDV-∆Meq virus (PI = 94%) and CVI988/Rispens (PI = 94%) [52]. Thus, the introduction of a random or site-specific mutation in certain gene/s involved in virus replication cannot improve the safety of the MDV-∆Meq virus without affecting its protection efficacy.
It was previously shown that the deletion of vIL8 from the MDV genome resulted in reduced oncogenicity [20,33], primarily due to severely restricted early cytolytic replication in lymphoid organs [33]. Interestingly, MDV-∆vIL8 still conferred good protection against challenge with vv+ MDV strains [34]. On the basis of these characteristics, we hypothesized that the double deletion of meq and vIL8 may reduce lymphoid organ atrophy but still retain the protective efficacy of MDV-∆Meq. In the present study, we generated a double gene deletion mutant virus in which both meq and vIL8 were deleted from the vv+ MDV strain 686 (686BAC-∆Meq∆vIL8). Like the single deletion mutant viruses, 686BAC-∆Meq∆vIL8 replicated well in vitro (Figure 1). Further characterization of this virus in vivo showed that it replicates to a significantly lower level than parental 686BAC and 686BAC-∆Meq viruses in spleen, as measured by qPCR at 5, 14, and 56 dpi (Figure 2A). Similarly, during early cytolytic infection (day 5), 686BAC-∆Meq∆vIL8 showed very limited to no viral antigen expression in lymphoid organs (spleen, thymus, and bursa) (Figure 3). These results suggest that the double deletion of meq and vIL8 did not affect the virus replication in vitro or FFE, but significantly reduced its replication in lymphoid organs (spleen, thymus, and bursa). Furthermore, 686BAC-∆Meq∆vIL8 s limited replication during the early cytolytic phase was comparable to that of 686BAC-∆vIL8 (Figure 2A). This limited virus replication in the lymphoid organs probably contributed to the reduction of lymphoid organ (bursa and thymus) atrophy ( Figure 4A) and mortality rate ( Figure 4B), in Ab-chickens inoculated with 686BAC-∆Meq∆vIL8.
Additionally, in agreement with previous results, 686BAC-∆vIL8 replicated poorly during early cytolytic infection (day 5) resulting in a low viral load compared to parental 686BAC and 686BAC-∆Meq viruses (Figure 2A). However, since the 686BAC-∆vIL8 virus causes lymphocyte transformation, by 14 dpi there was a~100 fold increase in virus genome copy number. On the other hand, at the same time point, the genome copy number in the 686BAC-∆Meq group decreased by~100 fold in agreement with the lack of transformation. As a result of the role of Meq and vIL8 in virus replication and transformation, the virus genome copy number of 686BAC-∆Meq∆vIL8 virus was low during early cytolytic infection (due to the absence of vIL8) and absent during the transformation phase (due to the absence of Meq). Interestingly, we and others have earlier shown that MDV-∆Meq is able to confer long term protection even though the virus could not be efficiently recovered from infected chickens after the initial cytolytic phase [29,53]. Similarly, in this study, we showed that 686BAC-∆Meq∆vIL8 replicated to low levels during the early cytolytic infection (due to vIL8 deletion) and was defective in latency and/or reactivation and transformation (due to meq deletion) (Figure 2), while it was still capable of inducing a vigorous immune response able to provide excellent protection against challenge with a vv+ MDV ( Figure 5 and Table 2). These results suggest that the 686BAC-∆Meq∆vIL8 may use an alternative mechanism of protection, which needs further investigation. We speculate that even though the replication of 686BAC-∆Meq∆vIL8 is severely impaired in lymphocytes, its replication in other cells, yet to be identified, may be important for MD protection. In addition, a single trial of the animal experiment was performed in both MDV maternal Ab-and Ab+ chickens, future studies will include a larger number of chickens to verify the data and study the potential mechanisms behind the protection conferred by the double deletion mutant virus.
Another drawback of currently available MD vaccines, including the "gold standard", CVI988/Rispens, is that they cannot completely overcome field virus infection-induced immunosuppression (MDV-IS). It has been recently shown that infection of highly virulent MDV could induce MDV-IS in chickens, which is classified into three phases, including early-MDV-IS (early IS associated with early cytoytic infection of MDV in lymphoid organs), late-MDV-IS-R (late IS associated with reactivation of MDV), and late-MDV-IS-T (late IS associated with development of tumors) [54]. Interestingly, early-MDV-IS and late-MDV-IS-T are well controlled by current MD vaccines; however, only MDV-∆Meq has been shown to protect against late-MDV-IS-R [55]. Future studies will examine if 686BAC-∆Meq∆vIL8 is capable of overcoming MDV-IS, especially late-MDV-IS-R.
MDV undergoes secondary replication in the FFE and sheds infectious virus into the environment through dander, infecting susceptible chickens [3]. We showed earlier that deletion of either vIL8 [33] or meq [21] genes did not interfere with the ability of the virus to replicate in the FFE, suggesting that these genes are not essential for horizontal transmission. Here, we showed that the deletion of both vIL8 and meq did not affect virus replication in the FFE (Figure 3). Thus, it is our expectation that the 686BAC-∆Meq∆vIL8 mutant virus may spread efficiently among chickens even though it was unable to replicate effectively and transform lymphocytes in chickens. Since the double deletion virus reported here replicated efficiently in the FFE (Figure 3), it would be interesting to study if it could interfere with the transmission of field viruses, thus slowing the progression of evolution of field viruses to greater virulence.

Conclusions
In summary, in this report, we show that the double deletion of meq and vIL8 did not affect virus growth in cell culture and FFE but significantly impaired virus replication in lymphoid organs. In addition, we demonstrate that the 686BAC-∆Meq∆vIL8 virus overcame the disadvantage of 686BAC-∆Meq virus-induced lymphoid organ atrophy while providing good protection against vv+ MDV challenge. Our study supports double gene deletion/mutation as a new strategy that can be exploited to generate the next generation of MD vaccines.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.