Type I Interferon Promotes Humoral Immunity in Viral Vector Vaccination

ABSTRACT

Recombinant viral vectors represent an important platform for vaccine delivery. Our recent studies have demonstrated distinct innate immune profiles in responding to viral vectors of different families (e.g., adenovirus versus poxvirus): while human Ad5 vector is minimally innate immune stimulatory, the poxviral vector ALVAC induces strong innate response and stimulates type I interferon (IFN) and inflammasome activation. However, the impact of the innate immune signaling on vaccine-induced adaptive immunity in viral vector vaccination is less clear. Here, we show that Modified Vaccinia Ankara (MVA), another poxviral vector, stimulated a type I IFN response in innate immune cells through cGAS-STING. Using MVA-HIV vaccine as a model, we found that type I IFN signaling promoted the generation of humoral immunity in MVA-HIV vaccination in vivo. Following vaccination, type I IFN receptor-knockout (IFNAR1^–/–) mice produced significantly lower levels of total and HIV gp120-specific antibodies compared to wild-type (WT) mice. Consistent with the antibody response, a type I IFN signaling deficiency also led to reduced levels of plasma cells and memory-like B cells compared to WT mice. Furthermore, analysis of vaccine-induced CD4 T cells showed that type I IFN signaling also promoted the generation of a vaccine-specific CD4 T-cell response and a T follicular helper (Tfh) response in mice. Together, our data indicate a role for type I IFN signaling in promoting humoral immunity in poxviral vector vaccination. The study suggests that modulating type I IFN and its associated innate immune pathways will likely affect vaccine efficacy.

IMPORTANCE Viral vectors, including MVA, are an important antigen delivery platform and have been commonly used in vaccine development. Understanding the innate host-viral vector interactions and their impact on vaccine-induced immunity is critical but understudied. Using MVA-HIV vaccination of WT and IFNAR1^–/– mice as a model, we report that type I IFN signaling promotes humoral immunity in MVA vaccination, including vaccine-induced antibody, B-cell, and Tfh responses. Our findings provide insights that not only add to our basic understanding of host-viral vector interactions but also will aid in improving vaccine design by potentially modulating type I IFN and its associated innate immune pathways in viral vector vaccination.

INTRODUCTION

Viral vectors are an important platform for antigen delivery and have been widely used in vaccine development (1). Viral vectors as a vaccine platform provide some advantages, including their ability to infect a broad range of host cells, induction of high-level transgene expression, and stimulation of strong adaptive immunity (2, 3). To date, a number of viral vectors derived from different families, such as adenovirus (4, 5) and poxvirus (6, 7), have been generated for the development of vaccines against a variety of infections and other human diseases. Similar to their parental viruses, viral vectors contain pathogen-associated molecular patterns and therefore can stimulate innate immune response (8, 9). Our recent study compared the innate immune stimulatory properties of the human Ad5 vector and the canarypox viral vector (ALVAC), showing that these two vectors induce largely distinct innate immune profiles: while the Ad5 vector is minimally innate immune stimulatory, the poxviral vector ALVAC induces strong type I interferon (IFN) and inflammasome activation (10).

Type I IFNs serve as a first line of host defense against pathogenic infections; however, divergent findings are reported regarding the roles of type I IFN signaling in viral infections and/or vaccination. For example, deficiency in type I IFN in acute viral infection can result in impaired immunity and dissemination of viral infections (11–17). In persistent viral infections, however, prolonged type I IFN signaling seems to be associated with hyperimmune activation and disease progression, since blockade of type I IFN signaling can hinder chronic immune activation and increase immune parameters associated with viral control (18–20). Similarly, opposing roles for type I IFN in modulating vaccine-mediated responses, from stimulatory to inhibitory, have both been reported (21–23). A number of earlier studies suggest that robust type I IFN signaling is considered a hallmark of effective vaccine response (24–26). More recently, it was reported that transient blockade of type I IFN signaling could improve immune memory and viral vaccine efficacy (27), which is thought to attribute to the enhanced antigen expression in the absence of a type I IFN response. Thus, the potential impact of type I IFN on vaccine-induced immunity, especially in viral vector vaccination, remains less clear.

Modified Vaccinia Ankara (MVA) is an attenuated poxviral vector and has been used in vaccine development for a variety of pathogens (28), including HIV (6, 29). Multiple MVA-HIV candidate vaccines have been tested in animal models and are at various stages of clinical development as prophylactic and/or therapeutic vaccines (30–33). Here, we explored the innate immune stimulatory property of MVA vector and show that, like the canary poxviral vector (10), MVA vector induced strong type I IFN expression and production in human innate immune cells through the cGAS-STING pathway. Using MVA-HIV vaccine as a model, we examined the impact of type I IFN signaling on vaccine-induced immune response in viral vector vaccination in wild-type (WT) and IFNAR-deficient mice. We report that compared to WT mice, a deficiency in type I IFN signaling impaired the generation of antibody, B-cell, and T follicular helper (Tfh) responses following MVA vaccination, suggesting that vector-stimulated type I IFN signaling is important for the generation of humoral immunity in poxviral vector vaccination.

RESULTS

MVA vector induces type I IFN production in human and mouse monocytic cells.

THP-1 is a human monocytic cell line that is widely used to study innate host-pathogen interactions. To understand molecular interactions between viral vectors and host innate immune cells, our previous study generated multiple THP-1 cell lines that are deficient in molecules involved in DNA sensing or type I IFN induction, including cGAS-KO THP-1 and STING-KO THP-1, by using CRISPR/Cas9 (10). Here, we first examined the ability of MVA vector to induce type I IFN gene expression and protein production in WT THP-1 cells (transduced with CRISPR/Cas9 empty vector as a control). Cells were infected with MVA vector (MOI = 2) or mock-treated control. At 24 h after infection, type I IFN (IFN-α and IFN-β) RNA expression in cells was quantified by qPCR (Fig. 1A); at 48 h after infection, type I IFN production in culture supernatants was quantified by Bio-Plex Immunoassay (Fig. 1B). As expected, compared to the mock-treated control, MVA vector induced significant IFN-β gene expression (Fig. 1A) and protein production (Fig. 1B) in WT THP-1 cells. Unlike IFN-β, little IFN-α protein was induced by MVA vector in THP-1 cells (data not shown), consistent with our previous finding on ALVAC vector (10). As a control, human Ad5 vector was used to infect the same THP-1 cells in our previous study and was found to be unable to stimulate type I IFN production (10). Lack of type I IFN stimulation by Ad5 vector was further confirmed in human monocyte-derived dendritic cells and mouse bone-marrow derived dendritic cells in the previous study (10).

FIG 1.

MVA vector induces IFN-β response in human THP-1 cells and mouse J774A.1 cells. (A) WT (CRISPR/Cas9 empty vector control), cGAS-KO, or STING-KO THP-1 cells were infected with MVA-HIV vector (MOI = 2). Mock-infected THP-1 cells were used as controls. At 24 h after infection, cellular RNA was extracted for PCR quantification of human IFN-β and GAPDH expression. PCR data were normalized to GAPDH and are shown as the fold change in RNA copies versus the WT THP-1 mock infection control. (B) At 48 h after infection, culture supernatants were collected, and human IFN-β production was quantified by Bio-Plex. The levels of human IFN-β under different conditions are shown as pg/ml. (C) Mouse J774A.1 cells were infected with either MVA-HIV vector (MOI = 2) or mock infected as a control. At 24 and 48 h after infection, cellular RNA was extracted for PCR quantification of mouse IFN-β and GAPDH expression. PCR data were normalized to GAPDH, and results are shown as the fold change versus the mock infection control. (D) At 24 and 48 h after infection, culture supernatants were collected, and mouse IFN-β production was quantified by ELISA. The levels of mouse IFN-β are shown as pg/ml. Experiments were independently repeated three times. Two-tailed P values were determined using one-way ANOVA. n.s., not significant.

To explore pathways involved in MVA-stimulated type I IFN production, we tested MVA infection in THP-1 cells deficient in cGAS or STING. We observed that cGAS-KO or STING-KO remarkably reduced MVA-activated type I IFN expression and production in THP-1 (Fig. 1A and B). Notably, STING-KO completely abrogated IFN-β production in THP-1 cells (Fig. 1B). The data suggest that MVA stimulates type I IFN production in THP-1 cells via the c-GAS/STING pathway.

In addition to human THP-1, we also examined the activity of MVA to stimulate type I IFN in a mouse monocytic cell line (J774A.1). Consistent with the findings in THP-1 cells, MVA vector also induced significant mouse IFN-β gene expression (Fig. 1C) and mouse IFN-β protein production (Fig. 1D) in J774A.1 cells. Together, these data indicate that MVA as a poxviral vector induces type I IFN responses in both human and mouse innate immune cells.

Type I IFN signaling promotes vaccine-induced antibody response in MVA vaccination.

Next, we explored the impact of type I IFN signaling on vaccine-induced adaptive immune response in viral vector vaccination by using recombinant MVA-HIV vaccine as a model (30). WT or type I IFN receptor-deficient (IFNAR1^–/–) mice (34) were intramuscularly (i.m.) vaccinated with MVA-HIV vaccine at week 0 (prime) and week 2 (boost), followed by analysis of vaccine-induced immune response 2 weeks after the second vaccination (week 4) (Fig. 2A). We first examined vaccine-induced antibody responses in the two vaccinated groups. Total and vaccine-induced, HIV Env (gp120)-specific binding IgG in sera were quantified by enzyme-linked immunosorbent assay (ELISA). The data showed that there was a modest but significant reduction in total IgG in IFNAR1^–/– mice compared to WT mice (P = 0.028) (Fig. 2B). Analysis of gp120-specific antibodies revealed a more profound reduction in the HIV gp120-specific binding IgG in IFNAR1^–/– mice compared to WT mice (P = 0.001) (Fig. 2C). We also analyzed gp120-specific IgG subclasses, including IgG1, IgG2a, and IgG3, in both groups of mice after vaccination and found that compared to IgG3, IgG1, and IgG2a were the predominant subclasses induced by the vaccine, both of which were significantly reduced in IFNAR^–/– mice compared to WT mice (Fig. 2D and E). Unlike IgG, no gp120-specific binding IgA was detected in the sera of WT and IFNAR^–/– mice (Fig. 2G). Together, these data indicate that type I IFN is important for the generation of antibody response (binding IgG) in mice following MVA-HIV vaccination.

FIG 2.

Vaccination schedule and antibody response after MVA-HIV vaccination. (A) Study design and vaccination schedule. WT mice (C57BL/6J) and matched IFNAR1^–/– mice (6/group) were vaccinated i.m. with MVA-HIV vaccine (10⁷ PFU) at weeks 0 and 2, followed by the collection of blood and tissue samples at week 4 (2 weeks after boost vaccination) for immune analyses. (B to G) Levels of total IgG (B), gp120-specific binding IgG (C), gp120-specific binding IgG1 (D), gp120-specific binding IgG2a (E), gp120-specific binding IgG3 (F), or gp120-specific binding IgA (G) in the sera of WT and IFNAR1^–/– mice. Serum antibodies were measured by ELISA (in duplicate for each sample) and are shown as mean OD₄₅₀ values (serum dilution, 1:100). Two-tailed P values were determined using a nonpaired Student t test.

Type I IFN is important for the generation of plasma cells and plasmablasts in bone marrow.

Antibodies are produced by plasma cells (PCs) that can be grouped into short-lived and long-lived PCs based on their longevity (35). Compared to short-lived PCs, long-lived PCs can sustain long-term antibody generation to provide immune protection for the host (35–37). After infection or vaccination, newly generated plasma cells are typically short-lived, except those homing to the bone marrow where they receive microenvironmental signals for long-term survival (35, 38). Plasmablasts (PBs), which are precursor PCs, can also migrate to the bone marrow (BM) and may develop into mature PCs that can survive for years (39). Thus, we examined the impact of type I IFN signaling on PCs and PBs in mouse BM after MVA-HIV vaccination (Fig. 2). Mononuclear cells in the BM were subjected to immune staining and flow cytometric analysis. According to previous studies (40), PCs and PBs in BM were identified based on the phenotype of B220 negative/CD138 positive (B220^– CD138⁺) and B220 intermediate/CD138 positive (B220^Int CD138⁺), respectively. The gating strategy for identifying PCs or PBs in BM mononuclear cells is shown in Fig. 3A. Our data showed that the percentage of PCs in BM mononuclear cells was slightly higher in WT mice than in IFNAR1^–/– mice, with a weak statistically significant difference detected (P = 0.048) (Fig. 3B, left). When comparing the absolute number of PCs between the two groups, we observed that IFNAR1^–/– mice had significantly lower numbers of PCs in BM than WT mice did (P = 0.0023; Fig. 3B, right). Similar to PCs, IFNAR1^–/– mice also had significantly lower numbers of PBs in BM compared to WT mice (P = 0.04) (Fig. 3C and D). These data are consistent with the results of vaccine-induced antibody responses (Fig. 2) and indicate that type I IFN signaling is probably important for the generation of PCs and PBs in bone marrow after MVA vaccination.

FIG 3.

Levels of plasma cells and plasmablasts in bone marrow of WT and IFNAR1^–/– cells after vaccination. (A) Gating strategy to identify plasma cells (PCs) or plasmablasts (PBs) in bone barrow by flow cytometry. First, live singlets were gated, followed by gating on CD3^– and CD11b^– cells. Based on the B220 expression levels, two subsets were further gated: B220^– and B220-Intermediate (B220-Int). PCs and PBs in bone marrow are identified as B220^– CD138⁺ and B220^Int CD138⁺, respectively. The final percentages of PCs or PBs in total bone marrow mononuclear cells were calculated. (B) Comparison of the percentages of PCs in total bone marrow cells (left panel) or the absolute numbers of PCs in bone marrow (right panel) between WT and IFNAR1^–/– mice. (C) Comparison of the percentages of PBs in total bone marrow cells (left panel) or the absolute numbers of PBs in bone marrow (right panel) between WT and IFNAR1^–/– mice. For each sample, immune staining and flow cytometry were conducted in duplicate. Two-tailed P values were determined using a nonpaired Student t test.

Type I IFN promotes the induction of memory-like B cells in spleen and mucosal lymphoid tissues.

In contrast to long-lived PCs that constitutively produce antibodies, memory B cells can be reactivated by antigen reexposure to differentiate into PCs for antibody production and thus are important for vaccine-induced immunity (37). Next, we examined frequency of memory-like B cells in WT and IFNAR1^–/– mice after vaccination. Multiple types of memory B cells with distinct phenotypes exist (37), among which IgM-expressing B cells are more likely to proliferate and reenter the germinal center (GC) during a secondary immune response (37). In addition, memory B cells rarely express IgD (41). Thus, we examined a memory-like B-cell subset with the B220⁺ IgM⁺ IgD^– phenotype. The gating strategy used to identify this population is shown in Fig. 4A. We observed that compared to WT mice, IFNAR1^–/– mice had significantly lower percentage of B220⁺ IgM⁺ IgD^– B cells in the spleen compared to WT mice (P = 0.0007; Fig. 4B). Comparison of the absolute numbers of B220⁺ IgM⁺ IgD^– B cells in the spleen revealed a more profound difference between WT and IFNAR1^–/– mice (P < 0.0001; Fig. 4C). The mean number of B220⁺ IgM⁺ IgD^– B cells in WT mice was >1-fold higher than in IFNAR1^–/– mice (Fig. 4C). Other than B220⁺ IgM⁺ IgD^– memory-like B cells, we observed that total B-cell levels in IFNAR1^–/– mice were also lower than in WT mice, either shown as the percentage in total splenocytes (P = 0.029, Fig. 4D) or in absolute numbers (P = 0.00028, Fig. 4E).

FIG 4.

Levels of memory B cells and total B cells in the spleens, MLN, and PP of WT and IFNAR1^–/– mice after vaccination. (A) Gating strategy used to identify memory-like B cells and total B cells by flow cytometry: singlets (FSC-A and FSC-H), live cells (Aqua Blue-), B cells (CD3⁻ B220⁺), and memory B (IgM⁺ IgD^–). (B and C) Comparison of levels of CD3⁻ B220⁺ IgM+ IgD⁻ memory-like B cells, shown as percentages in total B cells (B) or as absolute numbers (C), in the spleens of WT and IFNAR1^–/– mice. (D and E) Comparison of the levels of total B cells (D, percentage; E, absolute numbers) in the spleens of WT and IFNAR1^–/– mice. (F and G) Comparison of levels of CD3⁻ B220⁺ IgM⁺ IgD^– memory-like B cells, shown as percentages in total B cells (F) or as absolute numbers (G), in the MLN of WT and IFNAR1^–/– mice. (H and I) Comparison of levels of total B cells (H, percentage; I, absolute numbers) in the MLN of WT and IFNAR1^–/–mice. (J and K) Comparison of levels of CD3⁻ B220⁺ IgM⁺ IgD^– memory-like B cells, shown as percentages in total B cells (J) or as absolute numbers (K), in PP of WT and IFNAR1^–/– mice. (L and M) Comparison of levels of total B cells (L, percentage; M, absolute numbers) in PP of WT and IFNAR1^–/– mice. For each sample, immune staining and flow cytometric analysis were conducted in duplicate. Two-tailed P values were determined using a nonpaired Student t test.

Gut-associated lymphoid tissues (GALT), including Peyer’s patches (PP) and mesenteric lymph nodes (MLN) (42), are critical for the generation of mucosal immunity. The induction of a strong B-cell response in GALT is likely important for vaccine-induced protective immunity against mucosal pathogens, including HIV. Therefore, in addition to examining B cells in the spleen, as described above, we also analyzed total and memory-like B-cell populations in the MLN and PP of vaccinated WT and IFNAR1^–/– mice (Fig. 4F to M). We observed that the levels of B220⁺ IgM⁺ IgD^– memory-like B cells in the MLN, shown either as percentage of total B cells (Fig. 4F) or as absolute numbers (Fig. 4G) of IFNAR1^–/– mice, were significantly lower than those in WT mice, which is consistent with the results observed in the spleen. However, unlike the spleen, no significant differences were detected in the levels of total B cells in MLN (Fig. 4H and I) between WT and IFNAR1^–/– mice. A similar pattern was also observed for memory-like B cells (Fig. 4J and K) or total B cells (Fig. 4L and M) in PP. Together, these data indicate that type I IFN signaling is also important for the generation of memory-like B cells in the spleen, as well as in mucosal lymphoid tissues, of mice following MVA-HIV vaccination.

Comparable levels of GL-7⁺ germinal center B cells between WT and IFNAR1^–/– mice.

Germinal centers (GCs) are the site for antibody diversification and affinity maturation and thus are important for the generation of humoral immunity (43). During the induction of immune response, a panel of mutated B cells was selected based on their affinity in GCs to proliferate and then differentiate into antibody-producing plasma cells and memory B cells (43). GL7 is a marker for GCs in the immunized spleen (44, 45) and has been commonly used to identify GC B cells. Therefore, we examined the levels of GC B cells in the spleens and GALT of WT and IFNAR1^–/– mice after MVA-HIV vaccination. GC B cells were identified as CD3^– B220⁺ GL7^Hi by flow cytometry. Of interest, the levels of GC B cells, shown either as percentage of total B cells or as absolute numbers, were comparable in the spleen (Fig. 5A and B), MLN (Fig. 5C and D), or PP (Fig. 5E and F) between WT and IFNAR1^–/– mice. These data are different from the results of BM plasma cells (Fig. 3) and memory-like B cells (Fig. 4), indicating that type I IFN signaling does not appear to affect GC B-cell expansion but likely affects their subsequent differentiation into plasma cells or memory B cells.

FIG 5.

Levels of GC B cells in the spleens, MLN, and PP of WT and IFNAR1^–/– mice after vaccination. Single cell suspensions harvested from spleens, MLN, and PP of mice were stained for viability (Live/Dead Aqua Blue), as well as with a panel of antibodies that characterize mouse B-cell phenotype. GC B cells are identified as CD3⁻ B220⁺ GL7^hi B cells. The levels of GC B cells, shown either as percentages in total B cells (A, C, and E) or as absolute numbers (B, D, and F), were compared between WT and IFNAR1^–/– mice in the spleen (A and B), MLN (C and D), or PP (E and F). For each sample, immune staining and flow cytometric analysis were conducted in duplicate. Significance values were determined using a nonpaired Student t test.

Type I IFN promotes vaccine-induced CD4 T cell and Tfh response after vaccination.

After demonstrating the impact of type I IFN on antibody and B-cell responses after MVA-HIV vaccination, we next investigated its effect on vaccine-induced follicular help T cells (Tfh). Tfh cells are able to select proliferated GC B cells and promote their differentiation into plasma cells and memory B cells (43). Accumulating evidence has indicated that the interaction of cognate Tfh cells with GC B cells is critical for determining the fate of GC B cells (46–48). Thus, we examined vaccine-induced Tfh cells in the mouse spleen following MVA-HIV vaccination. To identify vaccine Env-specific total CD4 T cells and Tfh cells, splenocytes of vaccinated WT and IFNAR1^–/– mice were restimulated with a vaccine-matched HIV Env peptide pool for 6 h, followed by immune staining and flow cytometric analysis. Vaccine-specific CD4 T cells were identified based on a CD3⁺ CD4⁺ IFN-γ⁺ phenotype, and vaccine-specific Tfh cells were identified based on a CD3⁺ CD4⁺ IFN-γ⁺ PD1⁺ CXCR5⁺ phenotype and also displayed a PD1⁺ CXCR5⁺ phenotype. The numbers of Env-specific CD4 T cells and Tfh cells were calculated based on the frequency (%) and total number of splenocytes isolated from each mouse. Interestingly, compared to WT mice, IFNAR1^–/– mice had markedly lower levels of vaccine Env (gp120)-specific CD4 T cells (IFN-γ⁺) in the spleen after vaccination (mean % IFN-γ⁺, Env-specific CD4 T cells for WT versus IFNAR1^–/–: 0.26% versus 0.15%, P = 0.002) (Fig. 6A, left). Comparison of the absolute numbers of Env-specific CD4 T cells in the spleen between WT and IFNAR1^–/– mice revealed a more significant difference between them (P = 0.0007) (Fig. 6A, right). Next, we examined the levels of Env-specific Tfh cells and found that, compared to WT mice, IFNAR1^–/– mice also had significantly lower levels of Env-specific Tfh cells in the spleen, shown either as the percentage of Env-specific Tfh cells in the total CD4 T cells (P = 0.009) (Fig. 6B, left) or as absolute numbers (P = 0.025) (Fig. 6B, right).

FIG 6.

Analysis of Env-specific CD4 T cells and follicular T helper (Tfh) cells in WT and IFNAR1^–/– mice. (A and B) Comparison of levels of vaccine Env-specific CD4 T cells (A) or Env-specific Tfh cells (B) in the spleens of WT and IFNAR^–/– mice. Mouse splenocytes were ex vivo stimulated with an HIV Env peptide pool for 6 h, followed by surface and intracellular cytokine staining and flow cytometric analysis. Mock (DMSO)- or PMA-stimulated cells were included as negative or positive controls, respectively. For each stimulation condition, experimental duplicates were included. Env-specific CD4 T cells were identified as CD3⁺ CD4⁺ IFN-γ⁺. Env-specific Tfh cells were identified as CD3⁺ CD4⁺ IFN-γ⁺ CXCR5⁺ PD1⁺. (A) Percentages of Env-specific CD4 T cells in total CD4 T cells (left panel) or absolute numbers of Env-specific CD4 T cells (right panel) were compared between WT and IFNAR1^–/– mice. (B) Percentages of Env-specific Tfh T cells in total CD4 T cells (left panel) or absolute numbers of Env-specific Tfh cells (right panel) were compared between WT and IFNAR1^–/– mice. (C to E) Phenotypic analysis of total Tfh cells in spleen. Splenocytes were directly stained for CD3, CD4, CXCR5, PD1, CD25, Ki-67, CD44, and CD62L. Total Tfh cells were identified as CD3⁺ CD4⁺ CXCR5⁺ PD1⁺. (C) CD25 expression, shown either as % CD25⁺ in total Tfh cells (left panel) or as the CD25 MFI (right panel), was compared between WT and IFNAR1^–/– mice. (D) Ki-67 expression, shown either as % Ki-67⁺ in total Tfh cells (left panel) or as the Ki-67 MFI (right panel), was compared between WT and IFNAR1^–/– mice. (E) Memory versus effector phenotype of total Tfh cells. The % memory cells (CD44⁺ CD62L^–) in total Tfh cells (left panel) or the % effector cells (CD44⁻ CD62L^–) in total Tfh cells (right panel) was compared between WT and IFNAR1^–/– mice. Two-tailed P values were determined using a nonpaired Student t test.

To further characterize Tfh cells in WT and IFNAR^–/– mice, we examined their activation status, proliferation, and survival capability. Since the numbers of Env-specific Tfh cells (CD3⁺ CD4⁺ IFN-γ⁺ CXCR5⁺ PD1⁺) in mouse spleen were very low, further analysis of additional markers on this cell subset was challenging and less reliable. Thus, we analyzed these markers on total Tfh cells in the spleen. Splenocytes were directly stained for Tfh markers (CD3, CD4, CXCR5, and PD1), as well as for CD25 (activation), Ki-67 (proliferation), and CD44 and CD62L (memory phenotype as an indication of survival potential), followed by flow cytometric analysis. We found that most of the Tfh cells in the mouse spleen were CD25⁺ (∼70%) (Fig. 6C); however, no significant difference in CD25 expression (based on either the % CD25⁺ or CD25 mean fluorescence intensity [MFI]) was observed between the two groups (Fig. 6C), indicating that type I IFN signaling may not affect Tfh activation. Of note, compared to WT mice, Tfh cells in spleen of IFNAR1^–/– mice expressed significant lower levels of Ki-67 (Fig. 6D), suggesting that type I IFN may be important for the proliferation of Tfh cells. Lastly, to explore the potential longevity of Tfh cells in mice, we examined memory versus effector phenotypes of Tfh cells based on CD44 and CD62L. The data showed that majority of the splenic Tfh cells displayed a memory phenotype (CD44⁺ CD62L^–; 65 to 75%); however, no significant difference in memory versus effector phenotypes for Tfh cells was observed between WT and IFNAR1^–/– mice (Fig. 6E), indicating that the survival potential of Tfh cells might be comparable between these two groups of mice. Collectively, our data support that type I IFN signaling promotes the induction of vaccine-specific CD4 T cells and Tfh cells in mice after MVA-HIV vaccination.

DISCUSSION

In the present study, we compared antibody, B-cell, and CD4 T-cell responses in WT and IFNAR1^–/– mice following viral vector vaccination by using MVA-HIV vaccine as a model. Our in vitro data demonstrated that MVA vector stimulated type I IFN production in human innate immune cells through activating cGAS-STING, similar to our recent report on another poxviral vector, ALVAC (10). Importantly, our in vivo studies showed that compared to WT mice, IFNAR1^–/– mice manifested reduced level of humoral immunity, including antibody, plasma cells, memory-like B cells, and vaccine-specific CD4 and Tfh cells. Therefore, our data provided in vivo evidence that type I IFN signaling is important for the generation of humoral immunity in viral vector (MVA) vaccination.

The role of type I IFN in regulating adaptive immunity, especially the T-cell response, has been explored in several previous studies. Conventionally, type I IFN is considered important for priming the adaptive immune response (11–17), and is thus thought to be a hallmark of effective vaccines (24–26). However, recent studies indicate that the roles of type I IFN in regulating vaccine-induced immunity are likely divergent and context dependent. Opposing roles for type I IFN in modulating T-cell immunity induced by mRNA vaccines, from profoundly stimulatory to strongly inhibitory, have been reported (21–23). It was more recently shown that transient blockade of type I IFN signaling by blocking antibody improves protective immunity to viral infections and enhances vaccine efficacy (27). This effect was attributed to the enhanced viral antigen expression and presentation by type I IFN blockade. Compared to these earlier studies, our data showed that deficiency in type I IFN signaling impaired humoral immunity and Tfh responses following MVA-HIV vaccination, supporting the notion that type I IFN likely plays a beneficial role in viral vector vaccination. It is now believed that the activity of type I IFN in vaccination is complex and could be determined by the timing and intensity of type I IFN induction (21). On one hand, early and robust type I IFN could limit efficient viral or vaccine antigen expression and may thus impair antigen presentation and priming of adaptive immune response (27). On the other hand, type I IFN could promote an adaptive immune response by improving T- and B-cell survival and differentiation (49, 50). In our study, we speculate that in MVA vaccination, the promoting effect of type I IFN on B- and/or T-cell responses might outcompete its potential inhibitory effect on early vaccine antigen expression, thereby leading to improved humoral immunity compared to the mice deficient in type I IFN signaling.

The molecular mechanisms of type I IFN in promoting B-cell and antibody responses in the present study remain less clear. An early study conducted by Braun et al. indicated that type I IFN induces the expression of activation antigens involved in lymphocyte migration (CD69) or costimulation (CD86) (51), which might explain how type I IFN promotes humoral immunity, as observed in our study. In addition, our previous study dissecting innate immune pathways activated by different viral vectors (poxviral versus adenoviral vector) showed that canarypox viral vector selectively primes and activates AIM2 inflammasome through type I IFN signaling, leading to the production of the inflammatory cytokines IL-1β and IL-18 (10). Our ongoing studies employing an in vitro cell culture model, wherein MVA-infected antigen-presenting cells (APCs) were cocultured with autologous human lymphocytes, suggested that MVA triggers IL-1β production in APCs, which in turn promotes the proliferation and differentiation of human B cells in vitro (data not shown). Thus, a potential mechanism for type I IFN to promote humoral immunity in MVA vaccination is likely mediated by the inflammasome/IL-1β. This warrants further investigation in future studies.

An interesting observation of our study is that a deficiency in type I IFN signaling reduces memory-like B cells and plasma cells/plasmablasts in BM (Fig. 3 and 4) but does not appear to affect GC B-cell numbers (Fig. 5). The underlying mechanisms for this are not yet clear. During the induction of the immune response, antigen-activated B cells undergo clonal expansion and affinity selection in GCs, followed by differentiation into memory B cells and plasma cells (PCs). In our model, we speculate that type I IFN signaling may not affect GC B-cell expansion but promote its subsequent differentiation into memory B and/or PCs, which typically requires help signals from cognate Tfh cells. Indeed, our data showed that type I IFN promotes vaccine-specific Tfh response (Fig. 6). These data might help explain why deficiency in type I IFN signaling does not affect GC B cells but reduces memory B cells and PCs. Nevertheless, the molecular mechanisms for how type I IFN signaling regulates humoral immunity in vaccination, including the generation of GC B cells, memory B cells, and PCs, require further investigation.

In our study, we also noted that deficiency in type I IFN signaling does not alter total B-cell (Fig. 4H and I and Fig. 4L and M) and GC B-cell (Fig. 5C to F) numbers in the MLN and PP, while reducing total B-cell numbers in the spleen (Fig. 4D and E). These data indicate that the effect of type I IFN signaling on B cells may be specific to different immune compartments. We speculate that this observation might be related to the route of immunization. In our model, the vaccine was administered i.m., which is thought to mainly induce systemic immune response in the spleen. Thus, the impact of type I IFN on total B cells in gut-associated lymphoid tissues, such as MLN and PP, is likely limited compared to the spleen. In future research, it would be interesting to examine how B cells in different immune compartments are regulated by type I IFN signaling following systemic versus mucosal vaccination.

This study has some implications for viral vector vaccine development. Since type I IFN promotes humoral immunity and Tfh response, as indicated in our model, we speculate that an adjuvant that potentiates type I IFN signaling is likely beneficial in the context of viral vector vaccination. While viral vectors themselves are considered to confer adjuvant effects due to their innate immune stimulatory property, combining a viral vector vaccine with an adjuvant to induce stronger immunity has been reported and has demonstrated promise (52–54). Adjuvants that potentiate type I IFN responses, such as the STING agonist (55, 56), the TLR4 agonist (57), or water-in-oil adjuvants (58), might be interesting candidates for investigation. However, given the complexity of type I IFN in the generation of protective immunity to viral infections, the timing of type I IFN response is likely important in vaccination. These potential adjuvant approaches could be experimentally tested in additional vaccination models.

Our present study has some limitations. First, the study focused on humoral immunity and did not examine the impact of type I IFN on vaccine-induced CD8 T cells, which was mainly because the MVA-HIV vaccine used in our study does not induce a strong Env-specific CD8 T-cell response (33). The impact of type I IFN on vaccine-induced CD8 T cells should be investigated in future studies, which would provide an opportunity for direct comparison with several recent studies showing the divergent activities of type I IFN in mRNA vaccination (21–23) or viral infections (27). Second, other than magnitudes, the quality of antibody response, including the affinity and avidity, is important for assessing vaccine-induced protective immunity. Although we have measured vaccine-induced total binding IgG and its subclasses (Fig. 2), we were unable, due to limited serum samples, to address whether type I IFN signaling affects the affinity and avidity of these antibodies; this should be explored in future studies. Third, type I IFN plays a role in inducing autoantibodies (auto-Abs) as well (59). In our study, we did not measure auto-Abs, but our data on total IgG (Fig. 2B) suggest the possibility that WT mice may develop more auto-Abs than IFNAR^–/– mice after vaccination. Questions regarding whether type I IFN regulates auto-Ab production in vaccination and how this affects vaccine efficacy are interesting and should be further investigated. Lastly, while the present study showed that type I IFN signaling promoted humoral immunity, it remains unclear whether this vector-stimulated type IFN response is protective in MVA-HIV vaccination. A previous study reported that type I IFN prevents SIV infection and slows disease progression in SIV-infected nonhuman primates (NHPs) (17). Thus, future studies should test whether a type I IFN blockade, as well as the timing of the blockade, could modulate MVA-HIV vaccine efficacy in SIV- or SHIV-challenged NHP models.

In summary, our study examined the impact of type I IFN on the induction of an adaptive immune response in viral vector vaccination and provided in vivo evidence that type IFN signaling promotes the generation of humoral immunity and Tfh response after MVA-HIV vaccination. Further studies are needed to elucidate the molecular mechanisms for type I IFN modulation of B-cell and T-cell response in vaccination and define the role of type I IFN in regulating vaccine efficacy in virus challenge models. These studies are needed and important, given that MVA and its related poxviral vectors are being commonly used for vaccine development.

MATERIALS AND METHODS

Cells, vaccines, and mice.

Wild-type (WT) THP-1 cells, as well as cGAS-KO and STING-KO THP-1 cells generated using CRISPR/Cas9 in our previous study (10), were maintained in RPMI 1640 medium (Gibco) containing 10% FBS. The mouse monocytic cell line J774A.1 was purchased from Sigma (catalog no. 91051511) and cultured in Dulbecco modified Eagle medium (Gibco) containing 10% fetal bovine serum. THP-1 and J774A.1 cell lines were used for MVA vector infection and analysis of type I expression and production. The MVA-HIV vaccine (MVA-CMDR) was kindly provided by Merlin Robb and Nelson Michael at the U.S. Military HIV Research Program (MHRP). The vaccine encodes env/gag/pol (60) and has been shown to be safe and immunogenic in clinical trials (33). Wild-type (C57BL/6J) and IFNAR1^–/– (B6.129S2-Ifnar1^tm1Agt/Mmjax) mice were purchased from the Jackson Laboratory. Mice were maintained under specific-pathogen-free conditions in the accredited animal facility at the University of Texas Medical Branch (UTMB). All animal experiments were performed according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and according to the animal protocol approved by the Institutional Animal Care and Use Committee (IACUC) at UTMB.

MVA infection of THP-1 and J774A.1 cells.

THP-1 cells, including WT (CRISPR/Cas9 empty vector control), cGAS-KO, and STING-KO, or J774A.1 (5 × 10⁴) cells, were infected with MVA vector (MOI = 2) or mock as a control. At 24 h after infection, RNA was extracted from cells for quantification of type I IFN gene expression by real-time PCR. At 48 h after infection, supernatants were harvested from the cell culture for quantification of the type I IFN production by Bio-Plex.

PCR quantification of type I IFN gene expression.

Total RNA was extracted from cells using the RNA lysis buffer. RNA concentration and purity were determined using the multimode reader (BioTek). To quantify RNA expression, cDNA was synthesized from RNA using the iScript reverse transcription supermix for RT-qPCR (Bio-Rad). Human and mouse IFN-β were quantified by qPCR using iTaq Universal SYBR green Supermix (Bio-Rad) and the CFX Connect real-time PCR detection system (Bio-Rad). Primers for all genes are listed in Table 1. PCRs (20 μl) contained 10 μM primers, 90 ng of cDNA, 10 μl of iTaq universal SYBR green supermix (2×) (Bio-Rad), and molecular-grade water. The PCR cycling conditions were as follows: 95°C for 3 min and then 45 cycles of 95°C for 5 s and 60°C for 30 s. For each PCR, GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was also quantified for normalization.

TABLE 1.

Primers used for real-time PCR

Target	Orientationa	Primer sequence (5′–3′)
hIFN-α	F	GACTCCATCTTGGCTGTGA
	R	TGATTTCTGCTCTGACAACCT
hIFN-β	F	TGCTCTGGCACAACAGGTAG
	R	CAGGAGAGCAATTTGGAGGA
hGAPDH	F	ACAACTTTGGTATCGTGGAAGG
	R	GCCATCACGCCACAGTTTC
mIFN-β	F	CCCTATGGAGATGACGGAGA
	R	ACCCAGTGCTGGAGAAATTG
mGAPDH	F	GGGTGTGAACCACGAGAAAT
	R	CCTTCCACAATGCCAAAGTT

^aF, forward; R, reverse.

Quantification of type I IFN production in cell culture supernatants.

Human IFN-β in THP-1 cell culture supernatants was quantified by using Bio-Plex (Bio-Rad) according to the manufacturer’s instructions. Briefly, supernatants were diluted 1:4 in assay buffer. The plate was then rotated (850 rpm) in the dark at room temperature for 60 min. The plate was washed three times, 25 μl of the detection antibody (in the assay buffer) was added to the wells, and the plate was incubated again for 30 min. The plate was washed three times, and then 50 μl of 1× streptavidin-phycoerythrin (SA-PE) was added, followed by plate incubation for 30 min. The plate was washed finally another three times, followed by the addition of 125 μl of assay buffer. The plate was incubated with rotation for 5 min, and the signal was measured using a Bio-Plex 200 instrument (Bio-Rad) and analyzed using Bio-Plex Manager software version 6.1. Mouse IFN-β in J774A.1 cell culture supernatants was quantified using a Legend Max mouse IFN-β ELISA kit (BioLegend) according to the manufacturer’s instructions.

Mouse vaccination and sample collection.

Six-week-old female WT or IFNAR1^–/– mice (6/group) were i.m. immunized with MVA-HIV vaccines via gastrocnemius muscle injection of the left hind limb, using a prime-boost approach, at weeks 0 and 2 (Fig. 2A). Based on previous studies (61), 1 × 10⁷ PFU of MVA vaccine was used for each mouse per vaccination. All vaccines were diluted in phosphate-buffered saline (PBS) up to a total volume of 50 μl. Mice were euthanized 2 weeks after the second vaccination. BM, spleen, PP, and MLN cells, as well as blood sera, were collected from all vaccinated mice to measure the immune responses.

B-cell analysis by flow cytometry.

Single-cell suspensions prepared from the bone marrow, spleen, PP, and MLN were subjected to immune staining and flow cytometric analysis to identify plasma cells and plasmablasts in BM and various B-cell populations (total B cells, memory B cells, and GC B cells) in the spleen, MLN, and PP. The cells were first stained with Live/Dead Fixable Aqua (Life Technologies) to determine cell viability, followed by staining with a panel of surface antibodies (all from BioLegend), including CD3-APC (catalog no. 100236), B220-PE-Cy7 (catalog no. 103221), CD11b-BV421 (catalog no. 101235), CD138-APC-Cy7 (catalog no. 142529), IgD-Alexa Fluor 700 (catalog no. 405729), IgM-PerCP/Cy5.5 (catalog no. 406511), and GL7-FITC (catalog no. 144604). After antibody staining, the cells were washed and acquired on an LSRII Fortessa flow cytometer (BD). The data were analyzed using FlowJo version 10 (TreeStar).

CD4 T-cell and Tfh analysis by flow cytometry.

HIV Env-specific CD4 T cells and Env-specific Tfh cells were measured by intracellular cytokine staining (ICS) and flow cytometry as described previously (62). In brief, 1 × 10⁶ splenocytes were stimulated with HIV envelope peptide pool (PepMix HIV ENV Ultra; JPT) for 6 h. Mock (dimethyl sulfoxide [DMSO])- or phorbol myristate acetate (PMA)/ionomycin-stimulated splenocytes (1 × 10⁶) were included as negative and positive controls, respectively. In the last 4 h of stimulation, the cells were treated with the protein transport inhibitors GolgiStop and GolgiPlug (BD Bioscience), followed by viability and immune staining. Cells were first stained with Live/Dead Fixable Aqua (Life Technologies) for cell viability and then stained for surface markers (all from BioLegend): CD3-APC (catalog no. 100236), CD4-FITC (catalog no. 100406), CD8 PerCP (catalog no. 100732), CXCR5-BV605 (catalog no. 145513), and PD1-BV605 (catalog no. 135225). The cells were then fixed, permeabilized (BD Bioscience), and intracellularly stained with IFN-γ-APC-Cy7 (catalog no. 505850). Cells were acquired by using an LSRII Fortessa flow cytometer, and the data were analyzed using FlowJo version 10 (TreeStar). The frequency of Env-specific total CD4 T cells or of Env-specific Tfh cells was calculated by subtracting that of the unstimulated (mock) control. Absolute numbers of Env-specific total CD4 T cells or Tfh cells were quantified based on the frequency (%) and the total number of splenocytes isolated from each mouse.

To identify total Tfh cells, 0.5 × 10⁶ splenocytes of each mouse were directly stained without stimulation. Cells were first stained with Live/Dead Fixable Aqua (Life Technologies) to assess cell viability and then stained for surface markers (all from BioLegend): CD3-APC (catalog no. 100236), CD4-FITC (catalog no. 100406), CXCR5-BV605 (catalog no. 145513), PD1-BV605 (catalog no. 135225), CD25-PerCP (catalog no. 102028), CD44-APC/Cy7 (catalog no. 103027), and CD62L-BV421 (catalog no. 104435). The cells were then fixed and permeabilized (BioLegend, nuclear staining set, catalog no. 424401), followed by intracellular staining with Ki67-PE (catalog no. 652403). Cells were acquired by using an LSRII Fortessa flow cytometer, and the data were analyzed by using FlowJo version 10 (TreeStar).

ELISA measurement of serum antibodies.

ELISA was performed to examine serum binding antibody, as previously described, but with modifications (49). ELISA plates (Greiner Bio-One) were coated with mouse IgG capture antibody (for total IgG quantification) or 50 ng of recombinant gp120 protein (for gp120-specific IgG) overnight at 4°C. The plates were washed three times with wash buffer (Dulbecco-PBS with 0.05% Tween 20 [PBST]), for 5 min for each time, and then blocked with PBST containing 5% nonfat milk for 1.5 h at 37°C. The plates were then washed and incubated with diluted serum (1:100) for 1 h at 37°C. The plates were washed again and incubated with horseradish peroxidase-conjugated anti-mouse IgG (Cell Signaling, catalog no. 7076; 1:4,000), anti-mouse IgG1 (Thermo Fisher, catalog no. A10551; 1:2,000), anti-mouse IgG2a (Thermo Fisher, catalog no. A10684; 1:2,000), anti-mouse IgG3 (Thermo Fisher, catalog no. M32707; 1:2,000), or anti-mouse IgA (Thermo Fisher, catalog no. 62-6720; 1:2,000) for 1 h at 37°C. After a final wash, the plates were developed using TMB 1-component peroxidase substrate (Thermo Fisher), and the reaction was stopped with TMB stop solution (Thermo Fisher) after 15 min. The plates were read at a 450-nm wavelength within 30 min using a Synergy HTX Multi-Mode reader (Bio-Tek).

Statistical analysis.

Statistical analysis was performed using Prism 6.0 (GraphPad). Statistical comparison between groups was performed using a nonpaired Student t test or one-way analysis of variance (ANOVA) as appropriate. Two-tailed P values were determined, and P values of <0.05 were considered significant. Quantitative data displayed in the figures indicate means ± the standard deviations (SD; represented as error bars).