Immune Modulation of NYVAC-Based HIV Vaccines by Combined Deletion of Viral Genes that Act on Several Signalling Pathways

An HIV-1 vaccine continues to be a major target to halt the AIDS pandemic. The limited efficacy of the RV144 phase III clinical trial with the canarypox virus-based vector ALVAC and a gp120 protein component led to the conclusion that improved immune responses to HIV antigens are needed for a more effective vaccine. In non-human primates, the New York vaccinia virus (NYVAC) poxvirus vector has a broader immunogenicity profile than ALVAC and has been tested in clinical trials. We therefore analysed the HIV immune advantage of NYVAC after removing viral genes that act on several signalling pathways (Toll-like receptors—TLR—interferon, cytokines/chemokines), as well as genes of unknown immune function. We generated a series of NYVAC deletion mutants and studied immune behaviour (T and B cell) to HIV antigens and to the NYVAC vector in mice. Our results showed that combined deletion of selected vaccinia virus (VACV) genes is a valuable strategy for improving the immunogenicity of NYVAC-based vaccine candidates. These immune responses were differentially modulated, positive or negative, depending on the combination of gene deletions. The deletions also led to enhanced antigen- or vector-specific cellular and humoral responses. These findings will facilitate the development of optimal NYVAC-based vaccines for HIV and other diseases.


Introduction
UNAIDS (Joint United Nations Programme on HIV/AIDS) estimates that at the end of 2016, 36.7 million people worldwide lived with HIV (www.unaids.org); the number of newly infected individuals nonetheless continues to fall as the result of global implementation of preventive and therapeutic strategies. Vaccine development remains among the best hopes for controlling the HIV/AIDS pandemic.
To date, RV144 is the only HIV-1 vaccine efficacy trial that demonstrated a modest protection level (31%) [1]. This study combined a recombinant canarypox vector (ALVAC-HIV, vCP1521) as a prime component, with a recombinant HIV-1 envelope gp120 protein (AIDSVAX B/E) as a boost. Further studies seeking immune correlates of protection showed that non-neutralizing antibodies to HIV-1 Env V1/V2 regions were associated with reduced risk of HIV-1 acquisition, whereas IgA antibodies to the envelope correlated with decreased vaccine efficacy [2,3]. High levels of antibody-dependent cell-mediated cytotoxicity (ADCC) also correlated with reduced risk of Table 1. New York vaccinia virus (NYVAC) deletion mutants used in this study. The genes are named according to Copenhagen strain nomenclature.

Construction of NYVAC-Based Deletion Mutants
The different NYVAC-based deletion mutants generated and the corresponding parental viruses and plasmid transfer vectors used in the infection/transfection protocol are listed in Table 1. NYVAC-based deletion mutants were constructed using dsRed2 and rsGFP markers. BSC-40 cells (3 × 10 6 ) were infected with 0.005 pfu (plaque-forming units)/cell of parental virus and transfected 1 h later with 6 µg DNA of specific plasmid transfer vector using Lipofectamine (Invitrogen; Thermo Scientific Inc., USA). At 72 h post-infection, cells were harvested, lysed by freeze-thaw cycling, sonicated and used for recombinant virus screening. Deletion mutants were selected from progeny virus by consecutive rounds of plaque purification in BSC-40 cells, during which plaques were screened for Red2/GFP fluorescence. In the first three passages, viruses from selected plaques expressed both fluorescent proteins; in the next two passages, viral progeny from selected plaques expressed only one fluorescent marker. In the last two passages (seven passages total), viruses from selected plaques did not express a fluorescent marker due to marker loss by homologous recombination within the repeated flanking DNA sequences.

PCR Analysis of Deletion Mutants
To test for correct generation and purity of the deletion mutants, viral DNA was extracted from BSC-40 cells infected at 5 pfu/cell with NYVAC-WT, NYVAC-C, or the different NYVAC-C deletion mutants. Cell membranes were disrupted by proteinase K treatment (0.2 mg/mL proteinase K in 50 mM Tris-HCl pH 8, 100 mM EDTA (ethylenediaminetetraacetic acid) pH 8, 100 mM NaCl, 1% SDS; 1 h, 55 • C), followed by incubation with RNase A (80 µg/mL). Viral DNA was precipitated using 2-propanol. Different sets of primers annealing in the gene-flanking regions to be deleted were used for PCR analysis of the loci. The amplification reactions were carried out with Phusion High-Fidelity DNA polymerase (BioLabs, Ipswich, MA, USA). Primers used and size of the expected PCR products are shown in Table 2.

Analysis of Virus Growth
To determine virus growth profiles, CEF monolayers grown in 12-well tissue culture plates were infected at 0.01 pfu/cell with NYVAC-WT, NYVAC-C, or the different NYVAC-C deletion mutants. Following virus adsorption (60 min, 37 • C), the inoculum was removed. Infected cells were washed once with serum-free DMEM and incubated with fresh DMEM containing 2% FCS (37 • C, 5% CO 2 ). At different times post-infection (0, 24, 48 and 72 h), cells were removed by scraping and freeze-thawed three times; lysates were prepared from 5 × 10 5 cells/mL and briefly sonicated. Virus titres in cell lysates were determined by immunostaining in a plaque assay in BSC-40 cells, as described [19].

Mouse Immunization Schedule
BALB/c mice were purchased from Harlan. The DNA prime/poxvirus boost immunization protocol was used to assay the immunogenicity of the deletion mutants. Groups of 6-to 8-week-old female mice (n = 4) received 100 µg DNA-C (50 µg pcDNA-CN54 gp120 + 50 µg pcDNA-CN54 GPN) by the intramuscular route (i.m.); two weeks later, they received an intraperitoneal (i.p.) inoculation of 1 × 10 7 pfu of the corresponding virus. The control group was primed with sham DNA (DNA-ϕ) and boosted with non-recombinant NYVAC-WT. At 53 days after the last immunization (memory phase), mice were sacrificed and spleens and sera were processed for intracellular cytokine staining (ICS) and Enzyme-Linked ImmunoSorbent (ELISA) assays to analyse cellular and humoral immune responses to HIV-1 and VACV antigens, respectively. Representative data are shown for 2-3 experiments.

Intracellular Cytokine Staining Assay (ICS)
The magnitude and phenotype of HIV-or VACV-specific T cell responses were analysed by ICS. After an overnight rest, 4 × 10 6 splenocytes (erythrocyte-depleted) were seeded on 96-well plates and stimulated (6 h) in complete Roswell Park Memorial Institute (RPMI) 1640 medium with 10% FCS, 1 µL/mL Golgiplug (BD Biosciences, San Jose, CA, USA) and 1 µg/mL of the different HIV-1 peptide pools or 10 µg/mL of E3 peptide. After stimulation, cells were washed, stained for surface markers, permeabilized (Cytofix/Cytoperm kit; BD Biosciences) and stained intracellularly using appropriate fluorochromes. Fluorochrome-conjugated antibodies were used for functional analyses (CD4-Alexa 700, CD8-FITC or -V500, IL-2-APC, IFN-γ-PeCy7, TNF-α-PE) and for phenotypic analyses (CD62L-FITC, CD44-SPRD). Dead cells were excluded using the violet LIVE/DEAD stain kit (Invitrogen). All antibodies were from BD Biosciences. Cells were acquired on an LSRII flow cytometer (BD Immunocytometry Systems, Franklin Lakes, NJ, USA). Data analyses were performed using FlowJo software v.8.5.3 (TreeStar, Ashland, OR, USA). The number of lymphocyte-gated events ranged between 10 5 and 10 6 . After gating, Boolean combinations of single functional gates were generated using FlowJo to determine the frequency of each response, based on all possible combinations of cytokine expression or of differentiation marker expression. Background responses in negative control samples were subtracted from those in stimulated samples for each functional combination. Graphs were generated using GraphPad software (Version 6.01, San Diego, CA, USA).

Antibody Measurement by ELISA
Antibody binding to Env and vaccinia virus proteins in serum was assessed by ELISA as described [8]. Sera from naïve and immunized mice were diluted serially (3-fold) in duplicate and incubated with 2 µg/mL recombinant CN54gp120 purified protein (ARP683, HIV-1 CN54gp120 clade C; EU Programme EVA, NIBSC Centralised Facility for AIDS Reagents) or 10 µg/mL extract of BSC-40 cells infected (5 pfu/cell) for 24 h with VACV WR (Western Reserve) strain. Antibody titres of Env-or VACV-specific IgG were defined as the last serum dilution that gave three times the mean OD 450 value of the naïve control.

Plaque Neutralization Assay
To measure the neutralizing titre of anti-VACV antibodies, vaccinated mice were exsanguinated and sera were prepared and heated at 56 • C for 30 min to inactivate complement. Two-fold dilutions of serum in DMEM (Gibco, Waltham, MA, USA) supplemented with 2% NCS were prepared and incubated with about 400 pfu of purified VACV WR for 1 h at 37 • C before plaque assay on BSC-40 cells grown in 6-well plates. After 48 h of incubation at 37 • C, plaques were visualized by staining with 1% crystal violet in 2% ethanol. ND 50 (Neutralization dose 50) values represent the reciprocal of the serum dilution giving 50% reduction in plaque number compared with virus incubated without serum.

Data Analysis and Statistics
For statistical analysis of ICS data, we corrected measurements for the unstimulated control sample response (RPMI) and calculated confidence intervals and p values of hypothesis tests [30,31]. Only antigen response values significantly higher than the RPMI value are represented; background for the distinct cytokines in unstimulated controls were never >0.05%. For statistical analysis of the humoral response measured by ELISA, two-way ANOVA multiple comparison was used.

Generation and In Vitro Characterization of NYVAC-C Deletion Mutants
All NYVAC-C-based deletion mutants used and the site of action of the deleted genes are detailed in Table 1. They were generated as described (see Section 2) using as parental virus the NYVAC-C recombinant or NYVAC-C-based deletion mutants that express HIV-1 Env and GPN antigens from clade C [8]. Correct gene deletion was confirmed by PCR using primers that annealed in gene-flanking sequences. The distinct viral genes were deleted correctly, with no wild-type contamination in NYVAC preparations ( Figure 1A, deletion of VACV-TLR inhibitors A46R, A52R, K7R and B15R; Figure 2A, deletion of VACV-cytokine/chemokine inhibitors B8R and/or B19R and/or B6R-B10R cassette). Western blot analysis indicated that NYVAC-C-based deletion mutants expressed HIV-1 gp120 (120 kDa) and GPN (150 KDa) proteins at levels similar to the parental NYVAC-C virus ( Figures 1B and 2B). To determine whether deletion of specific viral genes affected virus replication, we compared virus growth kinetics of NYVAC-C-based deletion mutants with parental virus in CEF. Growth kinetics of parental and deletion mutants were similar ( Figures 1C and 2C), which indicated that the deleted genes are not necessary for virus replication in cultured CEF.

Generation and In Vitro Characterization of NYVAC-C Deletion Mutants
All NYVAC-C-based deletion mutants used and the site of action of the deleted genes are detailed in Table 1. They were generated as described (see Section 2) using as parental virus the NYVAC-C recombinant or NYVAC-C-based deletion mutants that express HIV-1 Env and GPN antigens from clade C [8]. Correct gene deletion was confirmed by PCR using primers that annealed in gene-flanking sequences. The distinct viral genes were deleted correctly, with no wild-type contamination in NYVAC preparations ( Figure 1A, deletion of VACV-TLR inhibitors A46R, A52R, K7R and B15R; Figure 2A, deletion of VACV-cytokine/chemokine inhibitors B8R and/or B19R and/or B6R-B10R cassette). Western blot analysis indicated that NYVAC-C-based deletion mutants expressed HIV-1 gp120 (120 kDa) and GPN (150 KDa) proteins at levels similar to the parental NYVAC-C virus ( Figures 1B and 2B). To determine whether deletion of specific viral genes affected virus replication, we compared virus growth kinetics of NYVAC-C-based deletion mutants with parental virus in CEF. Growth kinetics of parental and deletion mutants were similar ( Figures 1C and  2C), which indicated that the deleted genes are not necessary for virus replication in cultured CEF.  in CEF cells. CEF cell monolayers were infected with NYVAC-WT, NYVAC-C or NYVAC-C-ΔTLR1-4 (0.01 pfu/cell). At various times post-infection (0, 24, 48 and 72 h), cells were collected and infectious viruses quantified by immunostaining plaque assay in BSC-40 cells.

Cellular Immune Profile Induced by NYVAC-C Recombinants after Combined Deletion of VACV Immunomodulatory Genes
To define the effect of the distinct viral gene deletions in NYVAC-C vector immunogenicity, we used a mouse immunization protocol based on a DNA prime/poxvirus boost regime. This protocol is more immunogenic in activating T cell responses to HIV-1 antigens than the homologous combination of viral vectors and is being tested in clinical trials [8,11]. BALB/c mice (4 per group) were immunized as described (see Section 2). Memory T cell responses were analysed by polychromatic ICS assay 53 days after the last immunization. HIV-1-specific responses were measured after splenocyte stimulation with a panel of 464 peptides (15-mers overlapping by 11 amino acids) grouped in three pools (Env, 112 peptides; Gag, 121 peptides; and GPN, 231 peptides); vector-

Cellular Immune Profile Induced by NYVAC-C Recombinants after Combined Deletion of VACV Immunomodulatory Genes
To define the effect of the distinct viral gene deletions in NYVAC-C vector immunogenicity, we used a mouse immunization protocol based on a DNA prime/poxvirus boost regime. This protocol is more immunogenic in activating T cell responses to HIV-1 antigens than the homologous combination of viral vectors and is being tested in clinical trials [8,11]. BALB/c mice (4 per group) were immunized as described (see Section 2). Memory T cell responses were analysed by polychromatic ICS assay 53 days after the last immunization. HIV-1-specific responses were measured after splenocyte stimulation with a panel of 464 peptides (15-mers overlapping by 11 amino acids) grouped in three pools (Env, 112 peptides; Gag, 121 peptides; and GPN, 231 peptides); vector-specific responses were detected using the VACV E3 140-148 peptide. The percentages of T cells that produced IFN-γ and/or IL-2 and/or TNF-α indicated overall CD4 + and CD8 + T cell responses. The phenotype of the vaccine-induced memory responses was determined according to CD44 and CD62L surface marker expression on activated T cells (naïve, CD44 − CD62L + ; T central memory, TCM, CD44 + CD62L + ; T effector memory, EM, CD44 + CD62L − ; T terminal effector memory, TEM, CD44 − CD62L − ). Mice primed with sham DNA (DNA-ϕ) and boosted with non-recombinant NYVAC-WT were used as controls. A head-to-head comparison of immunogenicity among different NYVAC vectors is described below.
For CD8 + T cells, the NYVAC-C-∆TLR1, -∆TLR2 and -∆TLR4 deletion mutants induced significantly larger HIV-1-specific responses than parental NYVAC-C, with respective increases of 3.7-, 2.7-and 2.9-fold. In this cell subset, single deletion of the A46R gene yielded the highest HIV-1-specific CD8 + T cell response; additional deletions did not increase these values. Deletion of the K7R gene in NYVAC-C-∆TLR2 (NYVAC-C-∆TLR3) reduced CD8 + T cell responses, which recovered after B15R deletion (NYVAC-C-∆TLR4). In all groups, the HIV-1-specific memory CD8 T cells elicited were distributed between the EM and TEM phenotypes. For parental NYVAC-C, distribution of the response was 60% EM /40% TEM . This ratio was maintained in the NYVAC-C-∆TLR1 group but was reversed to 40% EM /60% TEM in NYVAC-C-∆TLR2 and NYVAC-C-∆TLR3 mutants. For NYVAC-C-∆TLR4, the response was distributed almost equally between the two phenotypes.
We also assessed the effect of these deletions on generation of VACV vector-specific responses, using the immunogenic VACV E3 protein as a marker. Compared to wild-type NYVAC virus, only NYVAC-C-∆TLR3 and NYVAC-C-∆TLR4 mutants showed increased anti-E3 CD8 + T cell responses, by about 1.5-fold. All deletion mutants nonetheless induced higher anti-VACV responses than parental NYVAC-C ( Figure 3B). As for the HIV-1-specific response, the phenotypic profile of the E3-specific memory CD8 + T cells elicited by all viruses was distributed between EM and TEM phenotypes. For the NYVAC-C-∆TLR2 and -∆TLR3 groups, distribution of the response was 25% EM /75% TEM . For wild-type NYVAC virus, NYVAC-C, NYVAC-C-∆TLR1 and -∆TLR4, the response was distributed almost equally between the two phenotypes.
Except for NYVAC-C-∆TLR3, the remainder of the gene deletion mutants thus improved the HIV immunogenicity of parental NYVAC-C, with enhanced memory cell HIV-1-specific responses by the single (NYVAC-C-∆TLR1) and double (NYVAC-C-∆TLR2) mutants. For anti-vector immunity, the sequential removal of three or four VACV-TLR inhibitors was necessary to enhance responses. by the single (NYVAC-C-∆TLR1) and double (NYVAC-C-∆TLR2) mutants. For anti-vector immunity, the sequential removal of three or four VACV-TLR inhibitors was necessary to enhance responses. Magnitude and phenotypic profiles of (A) memory HIV-1-specific CD4 + (top) and CD8 + (bottom) T cells and (B) memory VACV-specific CD8 + T cells. The vaccine-induced cellular immune response was characterized by multi-parameter flow cytometry at 53 days after the last immunization. Values indicate the sum of the percentages of T cells that secrete IFN-γ and/or TNF-α and/or IL-2 in response to Env plus Gag plus GPN peptide pools (for HIV-1specific responses) or to E3 peptide (for VACV-specific responses). Background percentages were subtracted from all data. The phenotype of the vaccine-induced memory responses was determined based on expression of CD44 and CD62L surface markers on activated T cells as follows: Naïve (CD44 − CD62L + ), T central memory (TCM; CD44 + CD62L + ), T effector memory (EM; CD44 + CD62L − ) or T terminal effector memory (TEM; CD44 − CD62L − ). * p < 0.05; ** p < 0.005, *** p < 0.001. Significant differences compared to the NYVAC-C group (for HIV-1 response) are indicated above each column.

top) and CD8 + (bottom) T cells and (B) memory VACV-specific CD8 + T cells.
The vaccine-induced cellular immune response was characterized by multi-parameter flow cytometry at 53 days after the last immunization. Values indicate the sum of the percentages of T cells that secrete IFN-γ and/or TNF-α and/or IL-2 in response to Env plus Gag plus GPN peptide pools (for HIV-1-specific responses) or to E3 peptide (for VACV-specific responses). Background percentages were subtracted from all data. The phenotype of the vaccine-induced memory responses was determined based on expression of CD44 and CD62L surface markers on activated T cells as follows: Naïve (CD44 − CD62L + ), T central memory (TCM; CD44 + CD62L + ), T effector memory (EM; CD44 + CD62L − ) or T terminal effector memory (TEM; CD44 − CD62L − ). * p < 0.05; ** p < 0.005, *** p < 0.001. Significant differences compared to the NYVAC-C group (for HIV-1 response) are indicated above each column.

Effect of Combined Deletion of Several Unknown Non-Essential Genes and Cytokine Inhibitors on NYVAC-C Recombinant Immunogenicity
Since deletion of VACV genes that block IFN type I and II pathways improves the immunogenicity of NYVAC-C recombinants in mice [14], we explored whether combined deletion of several unknown, non-essential genes (B6R, B9R and B10R) with IFN-(B8R and B19R) and cytokine-(B7R) VACV inhibitors increased NYVAC-C immunogenicity. We deleted the B6R-B10R ORF cassette from a recombinant virus that lacked B19R (NYVAC-C-∆B19R), to generate NYVAC-C-∆B19R/∆B6R-B10R. Using the DNA prime/poxvirus boost approach, we evaluated the immune response induced by the resulting virus compared with parental NYVAC-C and with the single (NYVAC-C-∆B19R) and double (NYVAC-C-∆B8R/∆B19R) mutants.

Effect of Combined Deletion of Several Unknown Non-Essential Genes and Cytokine Inhibitors on NYVAC-C Recombinant Immunogenicity
Since deletion of VACV genes that block IFN type I and II pathways improves the immunogenicity of NYVAC-C recombinants in mice [14], we explored whether combined deletion of several unknown, non-essential genes (B6R, B9R and B10R) with IFN-(B8R and B19R) and cytokine-(B7R) VACV inhibitors increased NYVAC-C immunogenicity. We deleted the B6R-B10R ORF cassette from a recombinant virus that lacked B19R (NYVAC-C-∆B19R), to generate NYVAC-C-∆B19R/∆B6R-B10R. Using the DNA prime/poxvirus boost approach, we evaluated the immune response induced by the resulting virus compared with parental NYVAC-C and with the single (NYVAC-C-∆B19R) and double (NYVAC-C-∆B8R/∆B19R) mutants.
Combined deletion of B6R-B10R genes with the IFN inhibitor B19R, thus increased immunogenicity of parental NYVAC-C but did not modify the HIV-1-specific cellular responses elicited by the single and double gene deletion mutants that lacked the IFN inhibitors B8R/B19R; the deletions nonetheless significantly improved VACV-specific memory responses.

Combined Deletion of Immunomodulatory Genes Alters Humoral Responses to Env and VACV Vector
Since cells infected with NYVAC-C recombinants release monomeric gp120 [8], we evaluated how combined deletion of VACV immunomodulatory genes affected humoral responses at the memory phase. In ELISA, we quantified the Env-and VACV-specific IgG using recombinant CN54gp120 purified protein and cell extract from BSC-40 cells infected with VACV-WR strain, respectively ( Figure 5).
In the group of viruses based on NYVAC-C after combined deletion of B6R-B10R with B19R, or the IFN inhibitors B19R/B8R, only the NYVAC-C-∆B19R/∆B6R-B10R mutant significantly enhanced Background percentages were subtracted from all data. The phenotype of vaccine-induced memory responses was determined based on expression of CD44 and CD62L surface markers on activated T cells as in Figure 3. * p < 0.05; ** p < 0.005, *** p < 0.001. Significant differences compared to the NYVAC-C group (for HIV-1 response) are indicated above each column.
Combined deletion of B6R-B10R genes with the IFN inhibitor B19R, thus increased immunogenicity of parental NYVAC-C but did not modify the HIV-1-specific cellular responses elicited by the single and double gene deletion mutants that lacked the IFN inhibitors B8R/B19R; the deletions nonetheless significantly improved VACV-specific memory responses.

Combined Deletion of Immunomodulatory Genes Alters Humoral Responses to Env and VACV Vector
Since cells infected with NYVAC-C recombinants release monomeric gp120 [8], we evaluated how combined deletion of VACV immunomodulatory genes affected humoral responses at the memory phase. In ELISA, we quantified the Env-and VACV-specific IgG using recombinant CN54gp120 purified protein and cell extract from BSC-40 cells infected with VACV-WR strain, respectively ( Figure 5).
In the group of viruses based on NYVAC-C after combined deletion of B6R-B10R with B19R, or the IFN inhibitors B19R/B8R, only the NYVAC-C-∆B19R/∆B6R-B10R mutant significantly enhanced the humoral Env-specific response compared with parental NYVAC-C. Single or double deletion of VACV-IFN inhibitors had no effect on the NYVAC-C-induced HIV-1 humoral immune response ( Figure 5B, left). Although all the deletion mutants induced higher anti-VACV antibody levels than parental NYVAC-C, they did not improve the humoral anti-vector response induced by wild-type NYVAC ( Figure 5B, right).
Viruses 2018, 10,7 14 of 20 the humoral Env-specific response compared with parental NYVAC-C. Single or double deletion of VACV-IFN inhibitors had no effect on the NYVAC-C-induced HIV-1 humoral immune response ( Figure 5B, left). Although all the deletion mutants induced higher anti-VACV antibody levels than parental NYVAC-C, they did not improve the humoral anti-vector response induced by wild-type NYVAC ( Figure 5B, right). ELISA assessment of levels of IgG antibodies to Env and vaccinia virus proteins in serum from naïve and immunized mice. Data shown as mean OD450 ± SD for each group at the dilutions assayed. * p < 0.05, ** p < 0.005, *** p < 0.001. Significant differences compared to the NYVAC-C group (anti-Env response, left) or the NYVAC-WT group (anti-VACV, right).

Single and Combined Deletion of VACV-TLR Inhibitors A46R, A52R, K7R and B15R Induced Similar "In Vivo" Protection against an Intranasal Challenge with a Lethal Dose of Wild-Type VACV WR
To determine to what extent single and sequential gene deletion of VACV-TLR inhibitors impact on the vaccine efficacy compared with the NYVAC-C parental virus, mice were vaccinated i.p. and challenged i.n. one month later with a lethal dose of wild-type VACV WR. As shown in Figure 6A, all the viruses induced protection against challenge, indicated by similar weight loss over a period of 15 days compared with the PBS control group. However, before the intranasal challenge the levels of both the binding VACV-specific antibodies ( Figure 6B) and the titre of neutralizing serum antibodies ( Figure 6C) were different between the groups of vaccinated mice. Single deletion of the A46R gene (NYVAC-C-∆TLR1) and double deletion of A46R and A52R genes (NYVAC-C-∆TLR2) induced the lowest anti-vector humoral and neutralizing responses after a single i.p. immunization; however, animals were protected. ELISA assessment of levels of IgG antibodies to Env and vaccinia virus proteins in serum from naïve and immunized mice. Data shown as mean OD 450 ± SD for each group at the dilutions assayed. * p < 0.05, ** p < 0.005, *** p < 0.001. Significant differences compared to the NYVAC-C group (anti-Env response, left) or the NYVAC-WT group (anti-VACV, right).

Single and Combined Deletion of VACV-TLR Inhibitors A46R, A52R, K7R and B15R Induced Similar "In Vivo" Protection against an Intranasal Challenge with a Lethal Dose of Wild-Type VACV WR
To determine to what extent single and sequential gene deletion of VACV-TLR inhibitors impact on the vaccine efficacy compared with the NYVAC-C parental virus, mice were vaccinated i.p. and challenged i.n. one month later with a lethal dose of wild-type VACV WR. As shown in Figure 6A, all the viruses induced protection against challenge, indicated by similar weight loss over a period of 15 days compared with the PBS control group. However, before the intranasal challenge the levels of both the binding VACV-specific antibodies ( Figure 6B) and the titre of neutralizing serum antibodies ( Figure 6C) were different between the groups of vaccinated mice. Single deletion of the A46R gene (NYVAC-C-∆TLR1) and double deletion of A46R and A52R genes (NYVAC-C-∆TLR2) induced the lowest anti-vector humoral and neutralizing responses after a single i.p. immunization; however, animals were protected.

Discussion
Given the modest efficacy observed in the RV144 clinical trial, the scientific community has focused on generating and optimizing vaccine candidates with improved immunogenicity, able to confer higher protection.
Poxviruses and particularly the highly attenuated VACV strains such as MVA and NYVAC, are being widely tested for potential HIV vaccines and are components in some of the clinical trials planned in the next few years [32][33][34] (https://clinicaltrials.gov/). Despite the safety and immunogenicity profiles of these attenuated VACV strains, it would be desirable to develop more efficient vectors that enhance the magnitude, breadth, polyfunctionality and durability of the T and B cell immune responses to exogenously expressed antigens. Various strategies are being used to achieve this purpose. One is deletion of viral immunomodulatory genes still present in the vector genome, whose products are predicted to interfere with optimal induction of cellular and humoral immune responses to vector-expressed antigens. We previously reported enhanced immunogenicity of MVA-and NYVAC-based recombinants with single, double or multiple deletions of VACV immunomodulatory genes such as C12L [35], C6L and/or K7R [36,37], A41L and/or B16R [30], F1L [38], B8R and/or B19R [14], A46R [15] or A52R, K7R and B15R in combination [16,17].
Here we extended these findings and explored the extent possibly improving NYVAC-C recombinant immunogenicity after combined deletion of genes involved in inhibition of TLR, IFN and cytokine/chemokine host-cell antiviral pathways or of genes with unknown immune function.

Discussion
Given the modest efficacy observed in the RV144 clinical trial, the scientific community has focused on generating and optimizing vaccine candidates with improved immunogenicity, able to confer higher protection.
Poxviruses and particularly the highly attenuated VACV strains such as MVA and NYVAC, are being widely tested for potential HIV vaccines and are components in some of the clinical trials planned in the next few years [32][33][34] (https://clinicaltrials.gov/). Despite the safety and immunogenicity profiles of these attenuated VACV strains, it would be desirable to develop more efficient vectors that enhance the magnitude, breadth, polyfunctionality and durability of the T and B cell immune responses to exogenously expressed antigens. Various strategies are being used to achieve this purpose. One is deletion of viral immunomodulatory genes still present in the vector genome, whose products are predicted to interfere with optimal induction of cellular and humoral immune responses to vector-expressed antigens. We previously reported enhanced immunogenicity of MVA-and NYVAC-based recombinants with single, double or multiple deletions of VACV immunomodulatory genes such as C12L [35], C6L and/or K7R [36,37], A41L and/or B16R [30], F1L [38], B8R and/or B19R [14], A46R [15] or A52R, K7R and B15R in combination [16,17].
Here we extended these findings and explored the extent possibly improving NYVAC-C recombinant immunogenicity after combined deletion of genes involved in inhibition of TLR, IFN and cytokine/chemokine host-cell antiviral pathways or of genes with unknown immune function. We generated a collection of NYVAC-C deletion mutants, all of which expressed Env (gp120) and the HIV-1 clade C polyprotein Gag-Pol-Nef and tested them in mice for their immunogenic characteristics (CD4 + /CD8 + T cells and antibodies) in response to HIV antigens and to the VACV vector (Table 1).
We analysed NYVAC-C recombinant immunogenicity after sequential deletion of the VACV-TLR inhibitors A46R, A52R, K7R and B15R. Double deletion of A46R and A52R was the best combination for enhancing cellular and humoral HIV-1-specific responses and increasing the percentage of CD8 + T cells with the TEM phenotype. Additional deletion of K7R and B15R did not further enhance the magnitude or quality of HIV-1-specific responses but significantly improved anti-vector immune responses.
We also examined whether deletion of various unknown non-essential genes in combination with VACV-IFN and -cytokine inhibitors enhanced NYVAC-C immunogenicity.
The NYVAC-C-∆B19R/∆B6R-B10R deletion mutant significantly improved the HIV-1-specific humoral response but did not augment the HIV-1-specific cellular responses elicited by single or double gene deletion mutants lacking VACV-IFN inhibitors. This mutant significantly increased cellular and humoral VACV-specific memory responses. When we analysed in immunized mice the protection induced by the vectors lacking TLR inhibitors following an intranasal challenge with a lethal dose of wild-type VACV WR, we observed similar levels of protection between the vectors, with differences in total binding antibodies and neutralizing VACV titres. These findings highlight that deletions do not reduce the protective efficacy of the VACV-TLR inhibitors but influenced humoral responses.
Our overall analysis of the immunogenicity profiles induced by the NYVAC-C recombinants with selected deletions is summarized in Table 3, in which each NYVAC vector is ranked by levels of immune T and B cell activation compared with parental NYVAC-C. These results indicate that combined deletion of VACV immunomodulatory genes is a valuable strategy for improving immunogenicity of NYVAC-based vaccine candidates. The specific combination of gene deletions allows differential control of an immune response towards antigen-or vector-specific cellular and humoral responses. Deletion of more immunomodulatory genes from the NYVAC-C genome did not always guarantee more a robust anti-HIV-1 immune response, although some of these genes were necessary to improve VACV-specific responses.
These results reflect the complexity and unpredictability of virus-host interactions in the context of an attenuated strain such as NYVAC.
Other groups have analysed the effect of multiple deletions in immunomodulatory genes on immune responses in the context of another attenuated poxvirus strain, MVA. Garber et al. assayed the effect on antigen-specific immune responses of simultaneous deletion of C12L, B15R, A41L and A46R genes, alone (MVA∆4) or combined with deletion of an essential viral replication gene (udg) (MVA∆5) in the genome of a recombinant MVA vector that expressed HIV gag and env genes [39]. Following a homologous prime-boost combination in rhesus macaques, they observed that both modified vectors significantly increase cellular and humoral HIV-specific immune responses compared to the control virus. Removal of the udg gene did not further improve HIV responses, however and offset the enhancement of vector-specific antibody titres due to immunization with the parental virus (∆4). Holgado et al. similarly reported that simultaneous deletion of A44L, A46R and C12L genes improved both innate and adaptive VACV-specific T cell immune responses in immunized mice, although the effect of these deletions in immunogenicity for a heterologous antigen was not tested [40]. Using MVA-BAC technology to examine the effect of deleting a gene cluster on immunogenicity of MVA deletion mutants, Alharbi et al. reported that none of the derived MVA vectors improved immunogenicity to MVA antigens or to the encoded heterologous antigen; this suggests that this approach should be assessed carefully for each recombinant antigen and epitope, rather than being used generically [41].
Although the results for MVA could not be extrapolated to NYVAC because of differences in vector genomes, it seems clear that combined deletion of selected immunomodulatory genes in the NYVAC genome should be considered for the design of viral recombinants as vaccine candidates. Terminal effector memory. The symbols "+", "−" or "=" indicate an increase, decrease or no effect in the immune response elicited compared to the parental NYVAC-C.