Deletion of Vaccinia Virus A40R Gene Improves the Immunogenicity of the HIV-1 Vaccine Candidate MVA-B

Development of a safe and efficacious vaccine against the HIV/AIDS pandemic remains a major scientific goal. We previously described an HIV/AIDS vaccine based on the modified vaccinia virus Ankara (MVA) expressing HIV-1 gp120 and Gag-Pol-Nef (GPN) of clade B (termed MVA-B), which showed moderate immunogenicity in phase I prophylactic and therapeutic clinical trials. Here, to improve the immunogenicity of MVA-B, we generated a novel recombinant virus, MVA-B ΔA40R, by deleting in the MVA-B genome the vaccinia virus (VACV) A40R gene, which encodes a protein with unknown immune function. The innate immune responses triggered by MVA-B ΔA40R in infected human macrophages, in comparison to parental MVA-B, revealed an increase in the mRNA expression levels of interferon (IFN)-β, IFN-induced genes, and chemokines. Compared to priming with DNA-B (a mixture of DNA-gp120 plus DNA-GPN) and boosting with MVA-B, mice immunized with a DNA-B/MVA-B ΔA40R regimen induced higher magnitude of adaptive and memory HIV-1-specific CD4+ and CD8+ T-cell immune responses that were highly polyfunctional, mainly directed against Env. and of an effector memory phenotype, together with enhanced levels of antibodies against HIV-1 gp120. Reintroduction of the A40R gene into the MVA-B ΔA40R genome (virus termed MVA-B ΔA40R-rev) promoted in infected cells high mRNA and protein A40 levels, with A40 protein localized in the cell membrane. MVA-B ΔA40R-rev significantly reduced mRNA levels of IFN-β and of several other innate immune-related genes in infected human macrophages. In immunized mice, MVA-B ΔA40R-rev reduced the magnitude of the HIV-1-specific CD4+ and CD8+ T cell responses compared to MVA-B ΔA40R. These results revealed an immunosuppressive role of the A40 protein, findings relevant for the optimization of poxvirus vectors as vaccines.


Introduction
The acquired immune deficiency syndrome (AIDS) pandemic caused by the human immunodeficiency virus (HIV)-1 is spreading worldwide, with high impact and severity in human health. In spite of active antiretroviral therapy (ART), in 2017, an estimated 1.8 million individuals became newly infected with HIV-1 and 940,000 people died from AIDS-related illnesses worldwide, according to the Joint United Nations Programme on HIV/AIDS. Therefore, the discovery of an effective animal studies were approved by the Ethical Committee of Animal Experimentation (CEEA) of the CNB (Madrid, Spain) and by the Division of Animal Protection of the Comunidad de Madrid (PROEX 331/14). All animal procedures were conformed to international guidelines and to the Spanish law under the Royal Decree (RD 53/2013). THP-1 cells were differentiated into macrophages by treatment with 0.5 mM phorbol 12-myristate 13-acetate (Sigma-Aldrich, St. Louis, MO, USA) for 24 h before use. Cell cultures were maintained at 37 • C in a humidified incubator containing 5% CO 2 . Cell lines were infected with viruses and, after 1 h of adsorption, the virus inoculum was removed and DMEM-2% FCS or DMEM-2% NCS was added to the cell cultures.

Viruses
The viruses used in this study included the attenuated MVA wild-type (MVA-WT) strain (kindly provided by G. Sutter) obtained from the Chorioallantois vaccinia virus Ankara (CVA) strain after 586 serial passages in CEF cells [38], and the recombinant MVA-B expressing HIV-1 IIIB GPN as an intracellular polyprotein and the HIV-1 BX08 gp120 protein as a cell-released product from HIV-1 clade B isolates [16], which are inserted into the TK locus of the MVA-WT genome under the transcriptional control of a VACV sE/L promoter. MVA-B was used as the parental virus for the generation of the MVA-B ∆A40R deletion mutant, and MVA-B ∆A40R was used as the parental virus for the generation of the MVA-B ∆A40R-rev virus. All viruses were grown in DF-1 cells to obtain a master seed stock (P2 stock), and titrated in DF-1 cells by plaque immunostaining, using rabbit polyclonal antibody against VACV strain WR (CNB; diluted 1:1000), followed by an anti-rabbit horseradish peroxidase (HRP)-conjugated secondary antibody (Sigma-Aldrich, St. Louis, MO, USA; diluted 1:1000), as previously described [49]. Determinations of the titers of the different viruses were performed at least two times. Furthermore, viruses grown in primary CEF cells were purified by centrifugation through two 36% (wt./vol.) sucrose cushions in 10 mM Tris-HCl pH 9. All viral stocks were free of contamination with mycoplasma (checked by specific polymerase chain reaction (PCR) for mycoplasma), bacteria (checked by growth in LB plates without ampicillin), or fungi (checked by growth in Columbia blood agar plates; Oxoid, Hampshire, UK).

Construction of Plasmid Transfer Vector pGem-RG-∆A40R wm
This plasmid was used for the construction of the MVA-B ∆A40R deletion mutant from which the MVA A40R gene has been deleted (A40R in the Copenhagen strain of VACV is equivalent to MVA152R in MVA. For simplicity, throughout this work, we have used the open reading frame nomenclature of the Copenhagen strain to refer to the MVA genes). The plasmid transfer vector pGem-RG-∆A40R wm ("wm" stands for "without markers", indicating that the final recombinant virus lacks the markers used for its selection, dsRed2 and rsGFP) was obtained by sequential cloning of MVA A40R flanking sequences into plasmid pGem-RG wm (4540 bp), the generation of which has been previously described [21], and contains the genes for dsRed2 and red-shifted green fluorescent protein (rsGFP) fluorescent markers. The MVA-B genome was used as the template to amplify by PCR the left flank of the A40R gene (352 bp) with oligonucleotides LFA40R-AatII-F (5 -ACGTTTGACGTCATAGAAAAATATAA-3 ) (AatII site underlined) and LFA40R-XbaI-R (5 -TACACCGCACGACAATGAACAAACAT-3 ) (XbaI site underlined). The left flank was digested with AatII and XbaI and cloned into plasmid pGem-RG wm, which had previously been digested with the same restriction enzymes to generate pGem-RG-LFsA40R wm (4859 bp). The repeated left flank of the A40R gene (352 bp) was amplified by PCR from the MVA-B genome with oligonucleotides LF A40R-EcoRI-F (5 -ACGTTTGAATTCATAGAAAAATATAA-3 ) (EcoRI site underlined) and LF A40R-ClaI-R (5 -TACACCGCACGACAATGAACAAACAT-3 ) (ClaI site underlined), digested with EcoRI and ClaI, and inserted into EcoRI/ClaI-digested pGem-RG-LFsA40R wm to generate plasmid pGem-RG-LFdA40R wm (5170 bp). Finally, the right flank of the A40R gene (372 bp) was amplified by PCR from the MVA-B genome with oligonucleotides RFA40R-ClaI-F (5 -AGAAAAATCGATATATCGCCGTACCG-3 ) (ClaI site underlined) and RFA40R-BamHI-R (5 -CTGTTAATTTTACTAGATCGTCAT GG-3 ) (BamHI site underlined), digested with ClaI and BamHI, and inserted into ClaI/BamHI-digested pGem-RG-LFdA40R wm plasmid. The resulting plasmid transfer vector generated was termed pGem-RG-∆A40R wm (5512 bp), and directs the deletion of the A40R gene from the MVA-B genome. Its correct construction was confirmed by DNA sequence analysis.

Construction of Plasmid Transfer Vector pHA-A40R
This plasmid was used for the insertion of the MVA A40R gene into the MVA hemagglutinin (HA) locus of the MVA-B ∆A40R recombinant virus to generate the MVA-B ∆A40R-rev revertant virus. To construct the plasmid transfer vector pHA-A40R (7126 bp), the MVA A40R gene (526 bp) was amplified by PCR from the MVA-B genome with oligonucleotides LFA40R-XmaI-F (5 -TCCCCCCGGGATGAACAAACATAAGAC-3 ) (XmaI site underlined) and RFA40R-SacII-R (5 -AGGCCGCGGTTATTTTTTTCTAAAACACTC-3 ) (SacII site underlined), digested with XmaI and SacII restriction enzymes, and then inserted into the XmaI/SacII-digested pHA plasmid (6600 bp), the generation of which has been previously described [42]. Thus, pHA-A40R contained the MVA A40R gene under the control of the VACV sE/L promoter introduced in a multiple-cloning site between the MVA HA-L and HA-R flanking regions, and the selectable marker genes for ampicillin and β-glucuronidase (β-gus). The β-gus gene was inserted among two repetitions of the left HA flanking region, allowing their deletion from the final recombinant virus by homologous recombination after consecutive plaque purification steps. The correct construction of plasmid pHA-A40R was confirmed by DNA sequence analysis.

Generation of Recombinant MVA-B ∆40R and MVA-B ∆A40R-rev Viruses
The MVA-B ∆A40R deletion mutant contained a deletion of the MVA A40R gene in the MVA-B genome and the MVA-B ∆A40R-rev revertant virus contained an insertion of the MVA A40R gene in the MVA HA locus of MVA-B ∆A40R. MVA-B ∆A40R and MVA-B ∆A40R-rev were generated using MVA-B and MVA-B ∆A40R as parental viruses, respectively, and pGem-RG-∆A40R wm and pHA-A40R as plasmid transfer vectors, respectively, employing an infection/transfection protocol previously described [18,20,23,42,43,50]. In the case of MVA-B ∆A40R, after the infection/transfection in DF-1 cells, we selected plaques expressing both Red2/GFP fluorescent proteins, then plaques expressing only GFP, and lastly we selected viruses from plaques that did not express any marker due to the loss of the fluorescent marker, as previously described [18,21]. Thus, MVA-B ∆A40R was finally obtained after six consecutive rounds of plaque purification. In the case of MVA-B ∆A40R-rev, after the infection/transfection in DF-1 cells, we initially selected blue plaques stained with 5-bromo-4-chloro-3-indolyl-beta-d-glucuronic acid (β-Gus, Sigma-Aldrich, St. Louis, MO, USA). In the first three passages viruses from selected blue plaques expressing β-Gus were picked, and in the last three passages (six passages in total) viruses from selected plaques did not express any marker due to the loss of the β-Gus marker. In both cases, the isolated plaques were expanded in DF-1 cells until a cytopathic effect was observed, and then crude viral extracts obtained were used for the next plaque purification round. After these recombination events in cell culture, the final plaques of both recombinant viruses, MVA-B ∆A40R and MVA-B ∆A40R-rev, were selected and expanded in DF-1 cells to obtain a master seed stock (P2 stock).

PCR Analysis
To verify that the MVA A40R gene had been correctly deleted in MVA-B ∆A40R or correctly inserted in MVA-B ∆A40R-rev, viral DNA was extracted from DF-1 cells mock infected or infected at 5 PFU/cell with the different viruses, as previously described [43], and the correct deletion or insertion of the MVA A40R gene was confirmed by PCR analysis. Primers LFA40R-AatII-F and RFA40R-BamHI-R (described above), spanning the A40R flanking regions, were used for PCR analysis of the A40R locus to verify the correct deletion of the MVA gene A40R in MVA-B ∆A40R, while primers HA-2 (5'-GATCCGCATCATCGGTGG-3') and HA-MVA (5'-TGACACGATTACCAATAC-3'), annealing in the MVA HA gene-flanking regions, were used for PCR analysis of the MVA HA locus, to verify the correct insertion of the MVA gene A40R in MVA-B ∆A40R-rev. The A40R deletion and insertion were also confirmed by DNA sequence analysis (Secugen, Madrid, Spain). The amplification protocols were performed using PuReTaq™ Ready-To-Go™ PCR beads (GE Healthcare, Chicago, IL, USA), in accordance with the manufacturer's protocol. PCR products were run in 1% agarose gel and visualized by SYBR Safe staining (Invitrogen, Carlsbad, CA, USA).

Analysis of Virus Growth
To study the virus growth profile of MVA-B, MVA-B ∆A40R, and MVA-B ∆A40R-rev, monolayers of DF-1 cells grown in 12 well plates were infected in duplicate at 0.01 PFU/cell with the different viruses. Following virus adsorption for 60 min at 37 • C, the inoculum was removed. The infected cells were washed with DMEM and incubated with fresh DMEM containing 2% FCS at 37 • C in a 5% CO 2 atmosphere. At different times (0, 24, 48, and 72 h.p.i.), cells were collected by scraping, freeze-thawed three times, and briefly sonicated. Virus titers in cell lysates were determined by immunostaining plaque assay as previously described [49].

RNA Analysis by Reverse Transcription Real-Time Quantitative PCR (RT-qPCR)
Total RNA was isolated using the RNeasy Kit (Qiagen, Hilden, Germany), from THP-1 cells mock infected or infected at 5 PFU/cell with the different viruses, and harvested at 3 and/or 6 h.p.i. Reverse transcription of maximum 1000 ng of RNA was performed with the QuantiTect reverse transcription kit (Qiagen, Hilden, Germany), according to the manufacturer's recommendations. Quantitative PCR was performed with a 7500 Real-Time PCR system (Applied Biosystems, Foster City, CA, USA) using Power SYBR green PCR Master Mix (Applied Biosystems, Foster City, CA, USA), as previously described [51]. The mRNA expression levels of the genes for IFN-β, interferon-induced protein with tetratricopeptide repeats 1 (IFIT1), IFIT2, melanoma differentiation-associated protein 5 (MDA-5), retinoic acid-inducible gene I (RIG-I), tumour necrosis factor alpha (TNF-α), regulated on activation, normal T cell expressed and secreted (RANTES), and macrophage inflammatory protein 1 alpha (MIP-1α) were analyzed by real-time PCR with specific oligonucleotides (sequences are available upon request). Specific gene expression was expressed relative to the expression of the cellular hypoxanthine phosphoribosyltransferase (HPRT) gene and/or the VACV E3L gene in arbitrary units (A.U.) using the 2 −∆∆Ct method [52]. All samples were tested in triplicate, and two independent experiments were performed.

Mouse Immunization Schedule
Female BALB/c mice (6 to 8 weeks old) were purchased from Envigo Laboratories and stored in a pathogen-free barrier area of the CNB in accordance with the recommendations of the Federation of European Laboratory Animal Science Associations. A DNA prime/MVA boost immunization protocol was performed as previously described [16,18,20,21,23]

ICS Assay
The magnitude, breadth, polyfunctionality, and phenotypes of the HIV-1-specific T-cell adaptive and memory responses were analyzed by ICS as previously described [18,20,23,43,53], with some modifications. After spleen processing, 4 × 10 6 fresh splenocytes (depleted of red blood cells) were seeded onto M96 plates and stimulated for 6 h in complete RPMI 1640 medium supplemented with 10% FCS containing 1 µL/ml Golgiplug (BD Biosciences, Franklin Lakes, NJ, USA) to inhibit cytokine secretion; monensin 1X (eBioscience, Thermo Fisher Scientific, Waltham, MA, USA), anti-CD107a-FITC (BD Biosciences, Franklin Lakes, NJ, USA); and the different HIV-1 Env, Gag, and GPN pools of peptides (5 µg/mL). Cells were then washed, stained for surface markers, fixed, permeabilized (Cytofix/Cytoperm kit; BD Biosciences, Franklin Lakes, NJ, USA), and stained intracellularly with the appropriate fluorochromes. Dead cells were excluded using the violet LIVE/DEAD stain kit (Invitrogen, Carlsbad, CA, USA). The fluorochrome-conjugated antibodies used for functional analyses were CD3-phycoerythrin (PE)-CF594, CD4-allophycocyanin (APC)-Cy7, CD8-V500, IFN-γ-PE-Cy7, TNF-α-PE, and IL-2-APC. In addition, the antibodies used for phenotypic analyses were CD62L-Alexa 700 and CD127-peridinin chlorophyll protein (PerCP)-Cy5.5. All antibodies were from BD Biosciences. Cells were acquired with a Gallios flow cytometer (Beckman Coulter). Analyses of the data were performed with the FlowJo software version 8.5.3 (Tree Star, Ashland OR, USA). After gating, Boolean combinations of single functional gates were created using FlowJo software to determine the frequency of each response based on all possible combinations of cytokine expression or all possible combinations of differentiation marker expression. Background responses detected in negative control samples were subtracted from those detected in stimulated samples for every specific functional combination.

Statistical Procedures
For statistical analysis of cytokine/chemokine expression, one-way analysis of variance (ANOVA) with Tukey's honestly significant difference (HSD) post hoc tests were applied. Student's t-test was used for antibody analysis to establish the differences between two groups. Statistical analysis of the ICS assay results was realized as previously described [21,53], using an approach that corrects measurements for the medium response (RPMI), calculating confidence intervals and p values. Only antigen response values significantly larger than the corresponding RPMI are presented. Background values were subtracted from all of the values used to allow analysis of proportionate representation of responses. The statistical significances are indicated as follows: *, p < 0.05; **, p < 0.005; ***, p < 0.001.

Generation and In Vitro Characterization of MVA-B ∆A40R
To determine whether the MVA A40R gene might have an immunomodulatory role that, in turn, could influence the immunogenicity profile of antigens delivered from a poxvirus vector, we deleted the MVA A40R gene from the HIV/AIDS vaccine candidate MVA-B (expressing HIV-1 Env, Gag, Pol, and Nef antigens from clade B) [16], generating the MVA-B deletion mutant termed MVA-B ∆A40R (see Materials and Methods) ( Figure 1A). The presence of the A40R deletion was confirmed by PCR of the MVA A40R viral locus ( Figure 1B), and was also validated by DNA sequencing. Analysis by Western blotting demonstrated that MVA-B ∆A40R expressed HIV-1 BX08 gp120 and HIV-1 IIIB GPN antigens similarly as the parental MVA-B ( Figure 1C).

A40 was Non-Essential in Cell Culture
The mere isolation of MVA-B ΔA40R confirmed that the A40 protein is not essential for MVA replication. Next, to examine whether deletion of A40R altered virus multiplication, we compared the growth of MVA-B ΔA40R and MVA-B in cultured permissive DF-1 cells. The growth kinetics of parental MVA-B and deletion mutant MVA-B ΔA40R were similar ( Figure 1D), confirming that the MVA A40 protein is not required for MVA replication.

A40 Was Non-Essential in Cell Culture
The mere isolation of MVA-B ∆A40R confirmed that the A40 protein is not essential for MVA replication. Next, to examine whether deletion of A40R altered virus multiplication, we compared the growth of MVA-B ∆A40R and MVA-B in cultured permissive DF-1 cells. The growth kinetics of parental MVA-B and deletion mutant MVA-B ∆A40R were similar ( Figure 1D), confirming that the MVA A40 protein is not required for MVA replication.

Deletion of MVA A40R Gene Enhanced the MVA-B Innate Immune Responses in Human Macrophages
The production of type I IFN, pro-inflammatory cytokines, and chemokines is an important initial step in the induction of antiviral immunity [55,56]. Thus, to study whether MVA A40 protein impacts innate immune responses, human THP-1 macrophages were mock infected or infected for 3 and 6 h with MVA-WT, MVA-B, and MVA-B ∆A40R at 5 PFU/cell, and the mRNA expression levels of type I IFN (IFN-β), type I IFN-induced genes (IFIT1 and IFIT2), the viral dsRNA sensor MDA-5, the proinflammatory cytokine TNF-α, and the chemokine MIP-1α were analyzed by reverse transcription real-time quantitative PCR (RT-qPCR). The results showed that, compared to parental MVA-B, MVA-B ∆A40R significantly upregulated the mRNA levels of IFN-β, IFIT1, IFIT2, MDA-5, and MIP-1α, at 3 or 6 h, but did not affect the mRNA expression of TNF-α ( Figure 2), suggesting an immunosuppressive function of the MVA A40 protein.

Deletion of MVA A40R Gene Enhanced the MVA-B Innate Immune Responses in Human Macrophages
The production of type I IFN, pro-inflammatory cytokines, and chemokines is an important initial step in the induction of antiviral immunity [55,56]. Thus, to study whether MVA A40 protein impacts innate immune responses, human THP-1 macrophages were mock infected or infected for 3 and 6 h with MVA-WT, MVA-B, and MVA-B ΔA40R at 5 PFU/cell, and the mRNA expression levels of type I IFN (IFN-β), type I IFN-induced genes (IFIT1 and IFIT2), the viral dsRNA sensor MDA-5, the proinflammatory cytokine TNF-α, and the chemokine MIP-1α were analyzed by reverse transcription real-time quantitative PCR (RT-qPCR). The results showed that, compared to parental MVA-B, MVA-B ΔA40R significantly upregulated the mRNA levels of IFN-β, IFIT1, IFIT2, MDA-5, and MIP-1α, at 3 or 6 h, but did not affect the mRNA expression of TNF-α ( Figure 2), suggesting an immunosuppressive function of the MVA A40 protein.

MVA-B ΔA40R Increased the Magnitude of HIV-1-Specific T-Cell Adaptive Immune Responses
Given the apparent immunosuppressive role of the MVA A40 protein in impairing the innate immune responses in human macrophages, we next asked whether deletion of MVA A40R from MVA-B could have an impact on the immunogenicity of the vector. Therefore, to study in vivo the effect of the A40R deletion on the HIV-1-specific T-cellular immunogenicity elicited by the HIV/AIDS vaccine candidate MVA-B, we analyzed the HIV-1-specific CD4+ and CD8+ T-cell immune responses induced by MVA-B ΔA40R in Balb/c mice immunized with a DNA prime/MVA boost immunization regimen, as this protocol amplifies the levels of T-and B-cell responses, while the homologous MVA prime/MVA boost immunization triggers lower responses [4,16]. Mice  were measured 10 days post-boost by intracellular cytokine staining (ICS) assay, after the stimulation of splenocytes with pools of peptides (Env, Gag, and GPN peptide pools) that spanned the HIV-1 Env, Gag, Pol, and Nef antigens from an HIV-1 clade B consensus sequence.
The magnitude of the total HIV-1-specific CD4+ and CD8+ T-cell adaptive immune responses (determined as the sum of the individual responses producing IFN-γ, TNF-α, and/or IL-2 cytokines, as well as the expression of CD107a on the surface of activated T cells as an indirect marker of cytotoxicity; obtained for the Env, Gag, and GPN peptide pools) was significantly greater in the DNA-B/MVA-B ∆A40R immunization group than in DNA-B/MVA-B (2.3-and 1.4-fold times higher, respectively), with both vaccinated groups triggering an overall HIV-1-specific immune response mediated mainly by CD8+ T cells (91% and 95%, respectively) ( Figure 3A). post-boost by intracellular cytokine staining (ICS) assay, after the stimulation of splenocytes with pools of peptides (Env, Gag, and GPN peptide pools) that spanned the HIV-1 Env, Gag, Pol, and Nef antigens from an HIV-1 clade B consensus sequence.
The magnitude of the total HIV-1-specific CD4+ and CD8+ T-cell adaptive immune responses (determined as the sum of the individual responses producing IFN-γ, TNF-α, and/or IL-2 cytokines, as well as the expression of CD107a on the surface of activated T cells as an indirect marker of cytotoxicity; obtained for the Env, Gag, and GPN peptide pools) was significantly greater in the DNA-B/MVA-B ΔA40R immunization group than in DNA-B/MVA-B (2.3-and 1.4-fold times higher, respectively), with both vaccinated groups triggering an overall HIV-1-specific immune response mediated mainly by CD8+ T cells (91% and 95%, respectively) ( Figure 3A).  The pattern of HIV-1-specific T-cell adaptive immune responses showed that CD4+ and CD8+ T cell responses were directed mainly against the Env pool in both vaccinated groups, with CD8+ T cell responses broadly distributed among Env, Gag, and GPN ( Figure 3B). However, DNA-B/MVA-B ∆A40R significantly enhanced the magnitude of Env-specific CD4+ T cell responses and Env-and GPN-specific CD8+ T cell responses ( Figure 3B).
Furthermore, the quality of the HIV-1-specific T-cell adaptive immune response was characterized in part by the pattern of cytokine production and its cytotoxic potential. On the basis of the production of CD107a, IFN-γ, TNF-α, and IL-2 from HIV-1-specific CD4+ and CD8+ T cells, 15 different HIV-1-specific CD4+ and CD8+ T cell populations could be identified ( Figure 3C,D). As shown in Figure 3C

MVA-B ∆A40R Improved HIV-1-Specific T-Cell Memory Immune Responses
Memory T cell responses might be critical for protection against HIV-1 infection [57][58][59][60], and the durability of a vaccine-induced T-cell response is an important feature, since long-term protection is a requirement for prophylactic vaccination. Thus, we next analyzed the HIV-1-specific T-cell memory immune responses elicited by the different immunization groups 53 days after the boost, following the same ICS assay described in the adaptive phase.
Similar to the results obtained in the adaptive phase, the magnitude of the total HIV-1-specific CD4+ and CD8+ T-cell memory immune response was again significantly greater in the DNA-B/MVA-B ∆A40R immunization group than in DNA-B/MVA-B (2-and 2-fold higher, respectively), with both vaccinated groups triggering an overall HIV-1-specific immune response mediated mainly by CD8+ T cells ( Figure 4A).
The pattern of HIV-1-specific T-cell memory immune responses showed that CD4+ and CD8+ T cell responses were directed mainly against the Env pool in both vaccinated groups, with both CD4+ and CD8+ T-cell responses broadly distributed among Env, Gag, and GPN ( Figure 4B). However, DNA-B/MVA-B ∆A40R significantly enhanced the magnitude of Env-and GPN-specific CD4+ T-cell memory responses and Env-, Gag-and GPN-specific CD8+ T-cell memory responses ( Figure 4B).  on the x axis, while the percentages of T cells expressing CD107a and/or IFN-γ and/or TNF-α and/or IL-2 against Env, Gag, and GPN peptide pools are shown on the y axis. Responses are grouped and color-coded on the basis of the number of functions (4, 3, 2, or 1). The pie charts summarize the data. Each slice corresponds to the proportion of the total HIV-1-specific CD4+ and CD8+ T cells exhibiting 1, 2, 3, or 4 functions (CD107a and/or IFN-γ and/or TNF-α and/or IL-2) within the total HIV-1-specific CD4+ and CD8+ T cells.

MVA-B ∆A40R Enhanced HIV-1-Specific T Cells with an Effector Memory Phenotype in the Adaptive and Memory Phases
It has been described that HIV-1-specific T cells with a mature effector memory phenotype are more frequently detectable in HIV-1 controllers than in HIV-1 progressors [61][62][63]. Thus, we next determined the phenotype of the adaptive and memory HIV-1-specific CD4+ and CD8+ T cells by measuring the expression of the CD127 and CD62L surface markers, which allowed the definition of the different memory subpopulations: T central memory (TCM, CD127 + /CD62L + ), T effector memory (TEM, CD127 + /CD62L − ), and T effector (TE, CD127 − /CD62L − ) cells [64], and determined the sums of the individual responses expressing CD107a, IFN-γ, TNF-α, and/or IL-2 obtained for the Env, Gag, and GPN peptide pools ( Figure 5). The results showed that in both vaccinated groups, adaptive and memory HIV-1-specific CD4+ and CD8+ T cells were mainly of the TEM phenotype, followed by the TE phenotype. However, immunization with DNA-B/MVA-B ∆A40R induced a significant increase in the percentage of adaptive and memory HIV-1-specific CD4+ and CD8+ TEM and TE cells ( Figure 5A,B). Representative flow cytometry plots of memory HIV-1-specific CD8+ T cells against Env, Gag, and GPN peptide pools are shown in Figure 5C.

MVA-B ∆A40R Increased the Levels of Binding IgG Antibodies Against HIV-1 gp120
Since MVA-B is able to release monomeric gp120 from infected cells and induce humoral immune responses [16], which are thought to be necessary to control HIV-1 infection [65], we next analyzed the humoral immune responses elicited after immunization of mice with DNA-B/MVA-B and DNA-B/MVA-B ∆A40R. Thus, we quantified by ELISA the total IgG and subclass IgG1, IgG2a, and IgG3 levels of antibodies against HIV-1 gp120 (clade B, isolate BX08) in pooled sera obtained from mice 10 and 53 days post-boost ( Figure 6). The results showed that DNA-B/MVA-B ∆A40R elicited significantly higher levels of total IgG anti-gp120 antibodies than DNA-B/MVA-B in the adaptive and memory phases ( Figure 6A). Furthermore, analysis of the IgG subtypes showed that DNA-B/MVA-B ∆A40R induced significantly higher levels of IgG1, IgG2a, and IgG3 anti-gp120 antibodies than DNA-B/MVA-B ( Figure 6B), with IgG1 levels being higher than IgG3 and IgG2a levels, in both groups indicating a Th2 response. However, the analysis of the IgG2a/IgG1 ratio showed that DNA-B/MVA-B ∆A40R induced a higher ratio than DNA-B/MVA-B, suggesting a shift toward a Th1 response.  DNA-B/MVA-B ΔA40R induced significantly higher levels of IgG1, IgG2a, and IgG3 anti-gp120 antibodies than DNA-B/MVA-B ( Figure 6B), with IgG1 levels being higher than IgG3 and IgG2a levels, in both groups indicating a Th2 response. However, the analysis of the IgG2a/IgG1 ratio showed that DNA-B/MVA-B ΔA40R induced a higher ratio than DNA-B/MVA-B, suggesting a shift toward a Th1 response.   Moreover, expression analysis by RT-qPCR and Western blotting showed that MVA-B ∆A40R-rev expressed similar levels of HIV-1 BX08 gp120 mRNA ( Figure 7D) and HIV-1 BX08 gp120 and HIV-1 IIIB GPN proteins ( Figure 7E) to MVA-B and MVA-B ∆A40R, confirming that the expression of HIV-1 antigens was not impaired because of the reintroduction and overexpression of the MVA A40R gene.
Finally, the growth kinetics of MVA-B ∆A40R-rev in cultured permissive DF-1 cells was similar to that of MVA-B and MVA-B ∆A40R, confirming that the reintroduction of the MVA A40R gene did not affect MVA replication in vitro ( Figure 7F).

MVA A40 Protein Localized at the Cell Membrane
Previous studies have reported that VACV WR A40 protein is expressed at the cell surface [44]. Thus, the expression and intracellular localization of the MVA A40 protein expressed by MVA-B ∆A40R-rev was next studied by confocal immunofluorescence microscopy in non-permissive HeLa cells. Therefore, cells were infected with MVA-B, MVA-B ∆A40R, and MVA-B ∆A40R-rev for 18 h, and then non-permeabilized or permeabilized fixed cells were stained with a polyclonal antibody against VACV A40 protein and the specific wheat germ agglutinin (WGA) probe to label the cell surface and Golgi reticulum (Figure 8). The results showed that in non-permeabilized cells, A40 protein (in green) was expressed in cells infected with MVA-B ∆A40R-rev and, as expected, co-localized with the cell membrane, whereas it was not detected in cells infected with MVA-B or MVA-B ∆A40R ( Figure 8A).
On the other hand, in permeabilized cells, A40 protein (in green) was expressed in cells infected with MVA-B ∆A40R-rev with a diffused cytoplasmic pattern; again, no detection of A40 protein was observed in cells infected with MVA-B or MVA-B ∆A40R ( Figure 8B), probably due to poor reactivity of the anti-A40 antibody used.

Reintroduction of MVA A40R Gene in MVA-B ∆A40R Inhibited Innate Immune Responses In Vitro
Next, to confirm whether the reintroduction of MVA A40R gene in MVA-B ∆A40R restored the previous enhancement in type I IFN innate immune responses ( Figure 2) and to further demonstrate the immunosuppressive role of the MVA A40 protein, we infected human THP-1 macrophages for 3 and 6 h with MVA-WT, MVA-B, MVA-B ∆A40R, and MVA-B ∆A40R-rev at 5 PFU/cell, and analyzed by RT-qPCR the mRNA expression levels of several innate immune-related genes (Figure 9). Interestingly, the results showed that, compared to parental MVA-B ∆A40R and MVA-B, MVA-B ∆A40R-rev significantly downregulated the mRNA levels of IFN-β, IFIT1, IFIT2, MDA-5, RIG-I, and MIP-1α (Figure 9). Moreover, MVA-B ∆A40R-rev downregulated the mRNA levels of other genes not affected by the deletion of MVA A40R gene in MVA-B ∆A40R, such as TNF-α and RANTES, while it did not affect the mRNA expression of the endogenous cellular gene HPRT (Figure 9). These results confirm the immunosuppressive role of the MVA A40 protein.

Reintroduction of MVA A40R Gene in MVA-B ΔA40R Inhibited Innate Immune Responses in vitro
Next, to confirm whether the reintroduction of MVA A40R gene in MVA-B ΔA40R restored the previous enhancement in type I IFN innate immune responses ( Figure 2) and to further demonstrate the immunosuppressive role of the MVA A40 protein, we infected human THP-1 macrophages for 3 and 6 h with MVA-WT, MVA-B, MVA-B ΔA40R, and MVA-B ΔA40R-rev at 5 PFU/cell, and analyzed by RT-qPCR the mRNA expression levels of several innate immune-related genes (Figure 9). Interestingly, the results showed that, compared to parental MVA-B ΔA40R and MVA-B, MVA-B ΔA40R-rev significantly downregulated the mRNA levels of IFN-β, IFIT1, IFIT2, MDA-5, RIG-I, and MIP-1α (Figure 9). Moreover, MVA-B ΔA40R-rev downregulated the mRNA levels of other genes not affected by the deletion of MVA A40R gene in MVA-B ΔA40R, such as TNF-α and RANTES, while it did not affect the mRNA expression of the endogenous cellular gene HPRT (Figure 9). These results confirm the immunosuppressive role of the MVA A40 protein.

Reintroduction of MVA A40R Gene in MVA-B ΔA40R Impaired HIV-1-Specific T-Cell Immune Responses
Next, to ascertain whether the reintroduction of MVA A40R gene in MVA-B ΔA40R could restore HIV-1-specific T-cell immunogenicity to levels similar to those induced by parental MVA-B containing the A40R gene (see Figure 3 and 4), we analyzed at 10 days after the last immunization the HIV-1-specific CD4+ and CD8+ T-cell immune responses induced by MVA-B ΔA40R-rev in mice   Figure  10B, right panel). However, the heterologous DNA-B/MVA-B ΔA40R-rev immunization group did not decrease the magnitude of HIV-1-specific CD8+ T-cell immune responses compared to DNA-B/MVA-B ΔA40R, probably influenced by the DNA priming that frames the quality of the immune responses prior to a poxvirus protein boost, as recently described [66]. Collectively, these results confirmed that the reintroduction of MVA A40R gene in MVA-B ΔA40R restores the magnitude of the HIV-1-specific T-cell immune responses.   Figure 10B, right panel). However, the heterologous DNA-B/MVA-B ∆A40R-rev immunization group did not decrease the magnitude of HIV-1-specific CD8+ T-cell immune responses compared to DNA-B/MVA-B ∆A40R, probably influenced by the DNA priming that frames the quality of the immune responses prior to a poxvirus protein boost, as recently described [66]. Collectively, these results confirmed that the reintroduction of MVA A40R gene in MVA-B ∆A40R restores the magnitude of the HIV-1-specific T-cell immune responses.

Discussion
Many recombinant poxvirus vectors [such as MVA, New York vaccinia virus (NYVAC), canarypox, and fowlpox viruses] expressing different HIV-1 antigens have been widely used in several human clinical trials in the last few years, proving that they are safe and immunogenic, inducing HIV-1-specific humoral and cellular immune responses [12,67,68]. In fact, the canarypox ALVAC combined with HIV-1 gp120 protein is actually the only effective HIV/AIDS vaccine candidate, and it showed a 31.2% protective effect in a phase III clinical trial [9]. However, this efficacy is modest, and the immunogenicity against HIV-1 antigens induced by modified poxvirus vector-based vaccines tested in human clinical trials is limited. Thus, more efficient and optimized poxvirus vector-based HIV/AIDS vaccines able to enhance HIV-1-specific humoral and cellular immunogenicity are needed [35].
Among the different approaches developed to enhance the immune response induced by poxvirus vectors, one promising strategy is the removal of the VACV genes that antagonize the immune system, as the virus genome still contains several genes that interfere with host immune responses [39,40]. Several recombinant MVA vectors expressing HIV-1 antigens and containing deletions in different immunomodulatory VACV genes have been generated, and have been shown to be able to enhance immune responses to HIV-1 antigens in animal models [18,20,21,23,[69][70][71].
In this work we described the immunosuppressive function of MVA A40R gene and tested its role on antigen-specific immune responses in vivo after deleting the MVA A40R gene in the vector backbone of MVA-B, an HIV/AIDS vaccine candidate that expresses HIV-1 IIIB GPN as an intracellular polyprotein and HIV-1 BX08 gp120 as a cell-released product from HIV-1 clade B isolate [16]. Previously to this work, the role of the VACV gene A40R was controversial. On one hand, it has been proposed that the VACV gene A40R encodes a type II membrane glycoprotein that is expressed early during infection on the cell surface, but is not incorporated into IMV or EEV particles [44]. These authors also showed that the A40 protein shares amino acid similarity with the CDR domain of C-type lectins including Clr-b, natural killer cell receptors, the human IgE receptor, and CD69, an early activation marker on lymphocytes. These C-type lectins are key players in pathogen recognition and innate immunity [45] and, in this regard, the A40 protein might have a role in interfering with the host response to infection. Moreover, the localization of A40 protein at the cell surface suggests that it may modulate the immune response by interacting at the plasma membrane level with signaling pathways and/or with other cells, but there is no evidence for any of these interactions. Interestingly, deletion of the VACV A40R gene from the VACV strain WR resulted in a modest attenuation after intradermal inoculation of mice, which could perhaps reflect an immunomodulatory role for this protein [46]. On the other hand, other reports have affirmed that VACV A40 is an early protein that is partially SUMO-1-modified and associated with the viral "mini-nuclei" [48]. Although the small amount of non-sumoylated A40 protein has a role in the VACV life cycle, joining the cytosolic side of the ER and inducing the proper apposition of several ER cisternae before its fusion to generate the ER envelope that surrounds the viral replication sites, the role that sumoylated A40 protein could play in the VACV life cycle still remains unknown; possibly it could participate in the process of replication itself or in the late transcription that happens at VACV replication sites, making A40 essential for VACV life cycle [47]. However, nothing was previously known about the immune function of this VACV gene.
Consequently, to evaluate whether the VACV A40 protein has an immunomodulatory role, an MVA recombinant vector lacking the MVA A40R gene was generated from the HIV/AIDS vaccine candidate MVA-B (termed MVA-B ∆A40R). It is of note that the amino acid sequence of WR A40 protein differs from MVA A40 protein, in which the last five amino acids at the C-terminus are substituted with another 14 unrelated ones. The results showed that A40R deletion had no effect on virus growth, demonstrating that MVA A40R gene is not essential for VACV life cycle, in contrast to what has been suggested by others [47]. As the loss of immunomodulatory genes in the MVA backbone impairs the innate immune response to this vector [18,20,23,72], the first step to elucidating the supposed immunomodulatory role of the MVA A40R gene was to study the innate immune responses triggered in THP-1 human macrophages infected with MVA-WT, parental MVA-B, and MVA-B ∆A40R deletion  I IFN signaling pathway, such as IFN-β, IFIT1, and IFIT2, as well as the viral dsRNA sensor MDA-5, and the pro-inflammatory chemokine MIP-1α, suggesting that MVA A40 protein could have an immunomodulatory role, blocking innate immune responses during virus infection. The enhanced levels of these mRNAs in MVA-WT versus MVA-B were likely due to a suppressive effect mediated by expression of the HIV-1 antigens. Moreover, since it has been described that innate immune responses play a critical role in the control and resolution of HIV-1 infection, providing signals for the efficient priming of the adaptive branch of immune response [73], the enhanced IFN-signaling could be an advantage of this MVA-B ∆A40R recombinant vector as HIV/AIDS vaccine candidate.
To further define whether MVA A40 could impair the immune system in vivo, a DNA prime/MVA boost immunization protocol was performed in mice to compare adaptive and memory immune responses to HIV-1 antigens induced by parental MVA-B and the deletion mutant MVA-B ∆A40R. Results showed that the DNA-B/MVA-B ∆A40R immunization group presented significantly enhanced magnitude of the overall adaptive and memory HIV-1-specific CD4+ and CD8+ T cells expressing CD107a, IFN-γ, TNF-α, and/or IL-2 compared to DNA-B/MVA-B. These results further suggest the immunosuppressive role of MVA A40 protein, as deletions of well-known immunosuppressive VACV genes from MVA or NYVAC vectors expressing HIV-1 antigens have produced similar results [18,20,21,23,[69][70][71][74][75][76][77][78]. Furthermore, both immunization groups elicited an adaptive and memory HIV-1-specific T-cell immune response with a similar polyfunctional profile, and mainly TEM and to a lesser extent TE phenotypes. However, once again, MVA-B ∆A40R significantly enhanced the magnitude of these populations, an important and relevant feature because the presence of TEM has been correlated with protection in the macaque SIV model [79,80]. The fast acquisition of TEM and TE phenotypes in the adaptive phase could be important in the development of the T-cell memory responses and in the mounting of a more effective immunity during a primary pathogen encounter. Moreover, adaptive and memory CD4+ T cell immune responses were directed mainly against Env in both immunization groups. However, in contrast to other MVA-B deletion mutants previously characterized (with deletions in C6L, C6L-K7R, and A41L-B16R MVA genes), where a pattern of GPN-specific CD8+ T-cell immune responses was mainly induced [18,20,21], MVA-B ∆A40R triggered CD8+ T-cell immune responses preferentially directed against Env, similarly to the MVA-B deletion mutant lacking the N2L gene, which encodes for a nuclear inhibitor of IRF3 [23]. The biological relevance of this T cell immune shift is not known.
The analysis of the gp120-specific humoral immune responses at the adaptive and memory phases showed that DNA-B/MVA-B ∆A40R immunization induced higher levels of total IgG, IgG1, IgG2a, and IgG3 anti-gp120 antibodies than DNA-B/MVA-B. This enhancement in antibodies against gp120 may have been mediated by the increase in innate immune responses and HIV-1-specific CD4+ T helper cells triggered by the MVA-B ∆A40R deletion mutant, and it could be a positive immune parameter, as it has been described that anti-HIV-1 V2-specific IgG1 or IgG3 antibodies are able to drive ADCC correlated with efficacy in the RV144 phase III clinical trial [81].
Moreover, evaluation of the Th1 or Th2 response by analyzing the ratio of IgG2a/IgG1 or IgG3/IgG1 antibodies showed that all the viruses induced higher levels of IgG1 than IgG2a or IgG3 antibodies, indicating a Th2 response. However, MVA-B ∆A40R induced in the adaptive and memory phases a ratio of IgG2a/IgG1 significantly higher than MVA-B (i.e., in the memory phase: 0.16 vs. 0.73 for MVA-B and MVA-B ∆A40R, respectively, at a dilution of 1/200), indicating that MVA-B ∆A40R may induce a more pronounced shift towards a Th1 response than MVA-B. According to current thinking, any therapeutic vaccination approach against HIV-1 should stimulate the induction of cytotoxic T lymphocytes (CTLs) and Th1 cells. It has been demonstrated that Th1 cells are more resistant to HIV-1 replication than Th2 cells [82][83][84], and HIV-1 has evolved to subvert a "protective" Th1 response to a "permissible" Th2/Th0-type immune response [85][86][87]. For example, ALVAC-specific CD4+ T cells from RV144 vaccinees that show protection against HIV-1 infection display a polarized Th1-like phenotype shown to be less susceptible to HIV-1 infection [88], which could positively influence the efficacy of the RV144 trial.
To finally demonstrate that the in vivo effects triggered by MVA-B ∆A40R in immunized mice were due to the deletion of the MVA A40R gene, and to confirm our previously suggested immunosuppressive role exerted by the MVA A40R gene, a revertant virus termed MVA-B ∆A40R-rev was generated. Due to the lack of detection (using Western blot or immunofluorescence assays) of A40 protein when expressed from its natural locus in MVA or MVA-B viruses, we decided to reintroduce the MVA A40R gene into the HA locus of the MVA-B ∆A40R under the control of a stronger sE/L viral promoter. A similar strategy was used previously to generate NYVAC-based revertant viruses, which allowed us to demonstrate the important role of NFκB activation by VACV in enhancing neutrophil migration and HIV-1-specific T-cell responses [77,78]. Moreover, it has been reported that insertion of heterologous genes into the MVA HA locus does not affect the expression of neighboring MVA genes and the recombinant viruses generated are functional, indicating that the HA gene region is a suitable insertion site [89]. In fact, we have used the HA locus to insert in the MVA genome heterologous antigens from Leishmania [90] and Ebola virus [42], and this strategy was effective to trigger antigen-specific T cells, humoral immune responses, and protective efficacy.
The MVA-B ∆A40R-rev expressed the MVA A40R gene at higher mRNA levels (5-7-fold) than MVA-B or MVA-WT, allowing the amplification of the A40 protein signal during viral infection. In the MVA-B or MVA-WT viruses, the protein levels of A40 were very low since the rabbit polyclonal antibody anti-A40 used could not detect the A40 protein by Western blot or by immunofluorescence, while in MVA-B ∆A40R-rev-infected cells the A40 protein was readily detected. This was confirmed by Western blot analysis with a multiplicity of infection (MOI) of 5 PFU/cell. For these experiments, we used the same antibody that was described in the first report of the A40 protein from the WR strain [44]. The Western blot with this anti-A40 antibody detected three major bands at 18, 28, and 35 kDa. It has been suggested that the 18 kDa form corresponds with the unglycosylated A40 protein, whereas the 28 and 35 kDa forms (together with a 38 kDa form that was not detected here) correspond with the Nand O-linked glycosylated forms of A40 [44]. The expression of A40 by MVA-B ∆A40R-rev under the stronger sE/L virus promoter in place of its own promoter provided the means to follow the subcellular localization of the A40 protein. The immunofluorescence analysis of MVA-B ∆A40R-rev-infected HeLa cells detected the A40 protein at the cell surface in non-permeabilized cells, appearing as punctuate cytoplasmic structures in permeabilized cells that co-localized with the Golgi network and exocytic vesicles. These results were similar to the results obtained previously [44] and reinforced the membrane localization of A40 protein, in contrast to what was stated by another report [48]. Although it could be hypothesised that the amino acid sequence differences at the C-terminal region of VACV WR and MVA A40 proteins might lead to a less stable and functional MVA A40 protein, the inability to detect the MVA A40 protein (either by Western blot or immunofluorescence assays) using antibodies readily recognizing the VACV WR A40 protein [44] was mainly due to the low levels of A40 naturally expressed from the MVA genome. Interestingly, when the MVA A40R gene was inserted into the MVA HA locus (A56R gene) controlled by a stronger VACV promoter than its naturally occurring promoter, the MVA A40 protein was clearly detectable by Western blot and immunofluorescence assays using the anti-A40 antibody recognizing the VACV WR A40 protein, indicating that the MVA A40 protein was stable, correctly expressed, and detected in the cell membrane, as previously described with the WR A40 protein [44].
Importantly, the RT-qPCR experiments showed that mRNA levels of IFN-β, IFIT1, IFIT2, MDA-5, TNF-α, MIP-1α, RANTES, and RIG-I in MVA-B ∆A40R-rev-infected human THP-1 macrophages were significantly lower than those induced by MVA-B ∆A40R or MVA-B. However, mRNA levels of the constitutive cellular gene, HPRT, and the heterologous HIV-1 gp120 gene were not affected, suggesting that A40 blocks specifically innate immune sensing genes. The clear decrease in the levels of IFN-β, IFN-induced genes, proinflammatory cytokines, and chemokines induced by the revertant virus MVA-B ∆A40R-rev, expressing high mRNA levels of the MVA A40R gene and of A40 protein, strongly suggest the immunosuppressive function of A40 by blocking the type I IFN signaling pathway.
Interestingly, the enhancing effect on adaptive HIV-1-specific CD4+ and CD8+ T-cell immune responses observed with MVA-B ∆A40R was restored, in general, to levels similar or lower to those induced by MVA-B when we immunized mice with MVA-B ∆A40R-rev, either in DNA/MVA or in MVA/MVA regimens, further confirming in vivo the immunosuppressive function of MVA A40 protein.

Conclusions
Overall, our findings revealed that MVA A40R gene has an immunosuppressive role, with its deletion enhancing innate immune responses in cell cultures, and adaptive and memory HIV-1-specific CD4+ and CD8+ T-cell and humoral immune responses in vivo, while its overexpression inhibited innate responses in human macrophages and adaptive HIV-1-specific CD4+ and/or CD8+ T-cell immune responses in immunized mice. Thus, deletion of the A40R gene has an immunomodulatory role in VACV infection and ablation of this gene provides an important strategy for the optimization of MVA vectors as vaccines. Funding: This research was supported by Spanish grants SAF2013-45232-R and SAF2017-88089-R (AEI/MINECO/FEDER-EU), Red de SIDA RD16/0025/0014 and, in part, by H2020 EU-projects HIVACAR and EHVA to Mariano Esteban.