Roles of Aluminum Hydroxide and Monophosphoryl Lipid A Adjuvants in Overcoming CD4⁺ T Cell Deficiency To Induce Isotype-Switched IgG Antibody Responses and Protection by T-Dependent Influenza Vaccine

Successful vaccination is the most effective measure to prevent disease against future pathogens. In general, nonreplicating subunit vaccines are safer than live attenuated vaccines but typically require adjuvants for successful Ag-specific immune responses. Aluminum hydroxide (Alum) has a long history of usage as an adjuvant in human vaccines, and a gel suspension of Alum (Alhydrogel) is the common clinical formulation. Alum adjuvant effects have a bias toward promoting Th2 Ab responses and would not be highly effective against intracellular viral pathogens. Inactivated H5 influenza split vaccine (Emerflu) and formalin-inactivated human respiratory syncytial virus vaccine with the Alum adjuvant formulations are symbolic failures of vaccination in humans (1, 2). Adjuvant system 04 (AS04) contains Alum and monophosphoryl lipid A (MPL), a TLR4 agonist, was tested by GlaxoSmithKline, and AS04-adjuvanted vaccines were licensed for humans (3–5). AS04 is a component of the hepatitis B virus vaccine Fendrix and the human papillomavirus vaccine Cervarix. Fendrix is licensed in the European Union, and Cervarix is licensed in the United States, European Union, Australia, and other countries.

Most adjuvants are involved in activating components of the innate immune system, which result in the outcomes of increasing adaptive immune responses (6, 7). In general, B cells require CD4⁺ T cell help to produce isotype-switched IgG Abs against protein Ags (8, 9). The production of influenza-specific Abs was highly impaired in thymectomized mice (10) and in T cell–deficient mice (11). These studies demonstrate that induction of isotype-switched IgG Abs to viral protein Ags is dependent on the help from CD4⁺ T cells. Professional APCs such as dendritic cells (DCs) of the innate immune system are required to activate a certain type of T cells. The activation status of DCs and cytokine milieu are known to determine whether T cells differentiate into Th1, Th2, Th17, follicular Th, or regulatory T cells (12). Bacterial LPS is a TLR4 agonist and a natural endotoxin adjuvant and has significantly contributed to our understanding of how vaccine adjuvant works (7). TLR4 signaling activates the NF-κB pathway, eventually producing inflammatory cytokines (IL-6, TNF-α) (13). Also, induction of type 1 IFNs by LPS stimulates DCs to express costimulatory molecules (CD40, CD80, CD86). TLR-mediated activation of DCs is thought to be a major mechanism dictating the type of T cells recruited by presenting an Ag to specific T cells in lymphoid tissues, and leading to specific CD4⁺ T cell clonal expansion and differentiation (14). Thus, it is thought that vaccine adjuvants educate certain CD4⁺ T cells, subsequently orchestrating the quantity and quality of B cell responses via Ab isotype switching, as well as long-lived plasma and memory B cells.

In this study, we sought to determine whether the effects of combined MPL and Alum (MPL+Alum), MPL, and Alum adjuvants would require CD4⁺ T cells in inducing IgG isotype-switched Abs and protection in the context of T-dependent influenza virus vaccine. MPL is an attenuated version of LPS. IgG isotype Abs and protective efficacy were determined in wild-type (WT) and CD4 knockout (CD4KO) mice after immunization with adjuvanted T-dependent influenza virus vaccine. In vitro and in vivo mechanistic studies were carried out to gain further insight into the action mechanisms of MPL+Alum adjuvant. This study reveals a new paradigm of CD4-independent MPL+Alum combination adjuvant mechanism as well as possible roles of Alum and MPL in combination MPL+Alum adjuvant effects.

Female and male 6- to 8-wk-old C57BL/6, CD4KO (B6.129S6-Cd4^tm1Mak/J), and MHC class II (MHCII) KO (I-Aβ^−/−) mice were purchased from The Jackson Laboratory and maintained in the animal facility at Georgia State University. All mouse experiments followed the approved Georgia State University Institutional Animal Care and Use Committee protocol (A14025). Commercial human monomeric influenza vaccine (inactivated and detergent split virus [Vac]), derived from the 2009 pandemic strain of A/California/07/2009 H1N1 virus, was provided by Green Cross (South Korea), a World Health Organization–approved vaccine manufacturing company. MPL and Alum were purchased from Sigma-Aldrich. MTT was obtained from Sigma-Aldrich for cell death analysis.

Mice were immunized i.m. with influenza vaccine alone or adjuvanted with 5 μg of MPL, 50 μg of Alum, or 5 μg of MPL plus 50 μg of Alum (MPL+Alum). For Ag adsorption of Alum or MPL+Alum adjuvant, influenza vaccine was incubated with adjuvant at 37°C for 30 min before immunization. The immunizations were performed twice (prime and boost) in WT, CD4KO, and MHCII KO mice with an interval of 4 wk, and immune sera were collected at 3 wk after each immunization (see Fig. 1A). To determine MPL+Alum adjuvant effects in primed mice, WT mice were primed with split influenza vaccine only, and then boosted with vaccine only or vaccine plus (Vac+)MPL+Alum. For CD4⁺ T cell acute depletion in WT mice, naive WT mice were injected with anti-mouse CD4 mAb (200 μg/mouse, clone GK1.5) i.p. 2 d before each prime and boost immunization. To maintain CD4 depletion status, the mice were injected with CD4-depleting Abs every 7 d. At 20 wk after boost, naive and immunized mice were challenged with 17-fold the LD₅₀ of A/California/04/2009 (H1N1) virus. The challenged mice were monitored to determine survival rates and body weight changes for 14 d. Kaplan–Meier analysis and log rank were applied for the survival graphs. Additional sets of challenged mice were sacrificed at day 5 after infection to determine protective efficacy. After sacrifice, bronchoalveolar lavages fluid (BALF), lung, bone marrow (BM), and spleens were harvested for further analysis.

For ELISA of serum IgG and IgG isotypes, the serially diluted sera were applied to the ELISA plates (Costar) coated with inactivated A/California/04/2009 H1N1 virus as previously described (15). Levels of IgG and IgG isotypes were determined by using anti-mouse IgG, IgG1, or IgG2c HRP conjugates as secondary Abs and tetramethylbenzidine as a substrate (16). Cytokines and chemokines in BALF, lung extract, peritoneal exudates, sera, and cell culture supernatants were measured by using ELISA kits from eBioscience and R&D Systems.

Sera and receptor destroying enzyme were mixed at a 3:1 ratio and incubated in 37°C for 16 h, and then followed by heat treatment at 56°C for 30 min. The sera were serially diluted (final volume of 25 μl) and incubated with 8 hemagglutination units of A/California/04/2009 H1N1 virus (final volume of 25 μl) in V-bottom plates for 30 min. Chicken RBCs (50 μl, 0.5%) were added to the plates and hemagglutination inhibition (HAI) titers were determined after 40 min.

Naive and immune sera were incubated at 56°C for 30 min for complement inactivation and mixed with a lethal dose of A/California/04/2009 (H1N1) virus. After 30 min, the mixture of sera (25 μl of 2-fold serial diluted sera) and virus was used to infect naive WT mice (n = 5) intranasally. Body weight changes and survival rates of the infected mice were daily monitored for 14 d.

Lung samples were harvested from the immunized mice at day 5 postinfection. The lung extracts were diluted and injected into 9- to 10-d-old embryonated chicken eggs. The allantoic fluids were collected after 3 d of incubation and used to perform hemagglutination activity assays to determine 50% egg infective dose.

Mouse lung tissue samples were harvested at 5 d after virus infection and treated with 10% of neutral formalin for fixation. The fixed lung tissues were sectioned and stained with H&E as described (17, 18). Images were acquired by a microscope (Zeiss Axiovert 100) at ×100 magnification and an attached camera (Canon 30D).

WT and CD4KO mice were injected with 200 μl of PBS, MPL (5 μg) + Alum (50 μg), MPL (5 μg), or Alum (50 μg) i.p. Sera were collected at 1.5, 6, and 24 h postinjection for detection of cytokines and chemokines. Peritoneal exudates were harvested at 24 h postinjection by PBS flushing. Peritoneal cells were used to determine phenotypes and cellularity of infiltrating cells by flow cytometry, and peritoneal exudates were used for detection of cytokines and chemokines.

BM-derived DCs (BMDCs) were generated from BM cells with mouse GM-CSF as described (19). The cells were treated with MPL (5 μg) + Alum (50 μg), MPL (5 μg), or Alum (50 μg) to determine proinflammatory cytokine production, DC activation marker expression, and cell death by adjuvants. For measuring in vitro IgG production by adjuvant-stimulated BMDCs, Vac (3 μg) + MPL (5 μg) + Alum (50 μg) combination was used to pretreat BMDCs for 2 d for preparation of mature DCs (mDCs). Spleen cells from naive CD4KO mice were cultured with MPL+Alum, mDCs, or mDCs plus mDC culture supernatant for 7 d. IgG levels in culture supernatants were determined by ELISA. To determine proliferated double-negative (DN) T cells, spleen cells were harvested from Vac+MPL+Alum–immunized CD4KO mice and stained with CFSE. The CFSE-labeled cells were cocultured with immature DCs, mDCs, or mDC plus anti-mouse MHCII mAbs (1 μg/ml, clone M5/114.15.2) for 5 d. For Ag-specific DN T cell proliferation, OVA (1 μg/ml)-, inactivated A/Philippine H3N2 (3 μg/ml)–, or inactivated A/California H1N1 vaccine strain (3 μg/ml)–treated BMDCs were cultured with CFSE-labeled cells for 5 d. The percentages of DN T cell proliferation were measured by flow cytometry.

Cells harvested after i.p. injection were stained with fluorescence-labeled anti-mouse F4/80, Ly6c, CD11b, CD11c, Siglec-F, CD103, B220, CD3, CD4, CD8, and CD49b Abs after blocking Fc receptor (anti-mouse CD16/32) to determine cellular phenotypes recruited by adjuvant injection. Adjuvant-treated BMDCs were harvested and stained with fluorescence-labeled anti-mouse CD40, CD80, CD86, and MHCII Abs after blocking Fc receptors (anti-mouse CD16/32). The stained cells were acquired by BD LSRFortessa (BD Biosciences, Mountain View, CA) and analyzed by FlowJo (Tree Star, Ashland, OR).

Experimental data were presented as means ± SEM. The statistical significance was determined by one-way ANOVA followed by a Tukey multiple comparison test or by two-way ANOVA followed by a Bonferroni posttest. A p value <0.05 was considered significant. We analyzed all data with Prism software (GraphPad Software, San Diego, CA).

To determine whether adjuvanted vaccination would overcome a defect in CD4⁺ T cells for inducing isotype-switched IgG Abs, WT and CD4KO mice were i.m. immunized with influenza vaccine or in the presence of MPL+Alum, MPL, or Alum adjuvant (n = 10). At 3 wk after prime, the influenza vaccine (WT-Vac) and Alum (WT-Alum) groups of WT mice showed low levels of virus-specific IgG Abs (Fig. 1B). CD4KO mice with vaccine (KO-Vac) only or Alum-adjuvanted (KO-Alum) vaccination did not induce virus-specific IgG Abs after prime (Fig. 1B), suggesting that vaccine is a T-dependent Ag and that Alum adjuvant effects are dependent on CD4⁺ T cells. Surprisingly, the MPL+Alum and MPL adjuvant groups induced IgG Abs in CD4KO (KO-Vac+MPL+Alum, KO-Vac+MPL) mice at high levels comparable to the corresponding WT groups (WT-Vac+MPL+Alum, WT-Vac+MPL) after prime immunization (Fig. 1B). Therefore, MPL+Alum combination and MPL adjuvants can overcome a defect of CD4⁺ T cells in priming IgG Abs to T-dependent influenza vaccine in CD4KO mice.

When we determined IgG Abs in boost-vaccinated WT and CD4KO mice (Fig. 1C), the vaccine alone KO group did not induce IgG Abs whereas the WT-Vac group showed substantial levels of IgG with dominant IgG1 (Th2 type) isotype (Fig. 1D, 1E). Also, the KO-Vac+Alum group showed a significantly lower level of IgG and IgG1 Abs compared with those in WT-Alum mice (Fig. 1C, 1D), indicating that CD4⁺ T cells are required for Alum adjuvant effects. The KO-MPL+Alum group exhibited high levels of IgG and IgG1 isotype Abs, which are comparable to those in WT-MPL+Alum mice and higher than the KO-MPL group (Fig. 1C, 1D). The KO-Vac+MPL group displayed boost IgG Abs comparable to those in WT mice (Fig. 1C) but a trend of lower levels of IgG1 Abs compared with those in WT mice (Fig. 1D). Induction of low IgG2c (Th1 type) levels was observed in the WT-Vac and WT-Vac+Alum groups but not the corresponding KO groups (Fig. 1E). The KO-Vac+MPL+Alum group induced IgG2c Abs at substantial but lower levels compared with those in the corresponding WT mice (Fig. 1E). The KO-Vac+MPL group was effective in inducing IgG2c isotype Abs to an equivalent level as observed in the corresponding WT mice (Fig. 1E). We observed that IgG Ab levels in KO-MPL and particularly in KO-MPL+Alum mice kept increasing when determined at 3 mo after boost vaccination (Table I), which were still maintained for >9 mo (data not shown). Importantly, note that MPL+Alum– and MPL-adjuvanted vaccination of CD4KO mice induced IgG Abs to comparable levels in the corresponding WT mouse vaccination.

Serum HAI titers are used as a measure of functional Abs predicting the efficacy of protection against influenza virus. The WT-Vac and WT-Alum groups showed a trend of lower levels of HAI titers compared with those in the WT-MPL+Alum and WT-MPL groups. Vaccine only and Vac+Alum–immunized CD4KO mice did not induce detectable levels of HAI titers (Fig. 1F). In contrast, the KO-MPL+Alum and KO-MPL groups showed high levels of HAI titers similar to those in WT-MPL+Alum and WT-MPL immune mice, and there were no statistical significances in the corresponding WT and KO groups (Fig. 1F). Thus, MPL+Alum or MPL adjuvant in influenza vaccination can induce protective Abs in CD4KO mice at comparable levels as observed in corresponding WT-adjuvant immune mice.

To determine adjuvant effects of MPL+Alum in primed mice, we primed WT mice with vaccine only and then boosted the mice with vaccine only or Vac+MPL+Alum. The boosted mice with MPL+Alum showed significantly higher levels of IgG Ab responses (Fig. 1G, Vac+MPL+Alum boost; 13.6 ± 2.47 μg/ml) compared with the vaccine only primed and boosted mice (Fig. 1G, Vac boost; 3.71 ± 0.48 μg/ml).

To determine adjuvant effects on improving the efficacy of protection, naive and immunized WT and CD4KO mice were intranasally challenged with a lethal dose of A/California/04/2009 (H1N1) virus after 5 mo of vaccination (Fig. 2). Both CD4KO and WT naive mice all died by days 8 and 10 postinfection, respectively. The WT-Vac group showed severe weight loss of ∼18%, resulting in 60% of survival rates, whereas the KO-Vac group did not show any protection, similar to naive infection mice (Fig. 2A, 2B). The WT-Alum group displayed moderate weight loss of 9–10 and 100% survival rates. In contrast, KO-Alum immune mice were very sick, as indicated by 18% weight loss at a peak point, and thus resulted in only 40% survival rates. Both the WT-MPL+Alum and WT-MPL groups showed 100% protection without weight loss (Fig. 2A). Also, KO-MPL+Alum and KO-MPL mice were similarly well protected against the same lethal dose challenge as used in WT mice, although a transient weight loss was observed in these KO mice at 6–9 d after challenge, probably due to the lack of CD4⁺ T cells (Fig. 2A, 2B).

At day 5 after challenge, lung viral titers were measured to determine the efficacy of clearing lung viral loads (Fig. 2C). The naive, vaccine only, and Vac+Alum groups showed high levels of viral loads in WT and particularly in CD4KO mice, which are consistent with low or no HAI titers, severe weight loss, and low survival rates in these groups. MPL+Alum and MPL adjuvant in the vaccination led to low lung viral titers, and there were no statistical differences between WT and CD4KO mice, although the MPL+Alum group showed a trend of increasing viral titers in CD4KO mice.

To further determine the protective roles of immune sera from CD4KO mice, we infected naive mice with a mixture of a lethal dose of A/California/04/2009 (H1N1) virus and immune sera or naive sera. All naive mice that were infected with virus plus naive or vaccine only sera died at day 8 or 9 postinfection (Fig. 2D). The naive mice infected with virus- and Alum-adjuvanted vaccination sera showed 10% body weight loss, but naive mice that received MPL+Alum– or MPL-adjuvanted immune sera did not show any weight loss (Fig. 2D). Overall, these results provide evidence that MPL+Alum– or MPL-adjuvanted vaccination induces protective immunity against lethal infection in CD4KO mice at a comparable level to that observed in WT mice.

Influenza virus can cause severe lung inflammation, leading to pneumonia and high mortality. To better assess the protective effects of adjuvant vaccination on alleviating disease symptoms, blood oxygen saturation (SpO₂) levels were measured in live animals at day 4 after influenza virus infection by using oximetry (Fig. 3A). In both WT and CD4KO mice, the naive infection and vaccine only groups exhibited significantly lower SpO₂ levels of ∼90% (Fig. 3A) compared with uninfected control mice. Low SpO₂ levels would indicate a severe breathing disorder caused by blocking the airways due to respiratory virus infection. In contrast, MPL+Alum–adjuvanted vaccination resulted in maintaining normal SpO₂ levels >98% in both WT and CD4KO mice.

Upon lethal influenza virus challenge, highest levels of inflammatory IL-6 and TNF-α cytokines were observed in WT-naive mice (Fig. 3B, 3C). Also, substantially high levels of IL-6 and TNF-α were detected in BALF from WT-naive, KO-naive, WT-Vac, KO-Vac, and KO-Alum mice at day 5 after challenge. In contrast, IL-6 and TNF-α cytokines were undetected or detected at low levels in BALF of WT-MPL+Alum and KO-MPL+Alum mice (Fig. 3B, 3C). KO-MPL mice displayed IL-6 and TNF-α cytokines in BALF at moderate levels (Fig. 3B, 3C). Thus, these results suggest that MPL+Alum–adjuvanted vaccination prevents the induction of proinflammatory cytokines even in CD4KO mice.

As for further evidence of lung inflammation, examination of lung histology showed that severe immune cell infiltration around the airways, alveolar septa, and interstitial spaces as well as narrowing or collapsing airways were observed in WT and CD4KO naive mice at day 5 postinfection (Fig. 3D), which are consistent with low SpO₂ levels. Significant lung histopathology was displayed in the WT-Vac, KO-Vac, and KO-Alum groups (Fig. 3D). Low to moderate histopathology, including thickening in the airway endothelial layers, was revealed in the WT-Alum, WT-MPL, and KO-MPL groups (Fig. 3D). Both WT-MPL+Alum and KO-MPL+Alum mice did not exhibit an overt sign of lung histopathology, indicating protection from lung inflammation due to influenza viral infection (Fig. 3D).

A goal of vaccination is to induce long-lived Ab-secreting cell (ASC) responses. Vaccinated mice were challenged after 5 mo of vaccination, and then BM and spleen cells were harvested at day 5 after challenge for analysis of IgG Abs secreted from ASCs in in vitro culture supernatants using ELISA (Fig. 4) as described (20). WT-Vac and WT-Alum mice showed only low levels of vaccine-specific IgG Ab production in BM cell cultures (Fig. 4A). BM cells from WT-MPL mice showed moderate levels of in vitro IgG Ab secretion whereas WT-MPL+Alum immune mouse BM cells produced significantly higher levels of IgG Abs after 1 d in vitro culture (Fig. 4A). As a measure of memory B cells, spleen cells were cultured for 5 d in the presence of inactivated virus (A/California/04/2009 H1N1). Spleen cells from WT-Alum mice produced low levels of IgG Abs. In contrast, WT-MPL+Alum groups of mice induced high levels of IgG Ab production in spleen cells (Fig. 4B).

CD4⁺ T cells are required to induce long-lived isotype-switched IgG ASCs in BM after viral infection (21, 22). We determined whether certain adjuvants could mediate the in vitro production of IgG Abs in BM and spleen cells from CD4KO mice after 5 mo of vaccination at day 5 after challenge (Fig. 4). KO-Vac mice did not show in vitro IgG Ab production in BM and spleen cells. The KO-MPL+Alum and MPL groups showed high levels of in vitro IgG Ab production in BM cells, which are comparable to those in corresponding WT mice after 1 d in vitro culture (Fig. 4). MPL+Alum adjuvant was more effective in in vitro IgG protection in 5-d culture of CD4KO spleen cells compared with MPL. Consistent with in vivo IgG responses, MPL+Alum adjuvant may effectively mediate the generation of long-lived IgG ASCs in CD4KO mice comparable to those in WT mice.

CD4KO mice might have developed compensatory immune components contributing to the alternative B cell help. To test this possibility, CD4⁺ T cells in WT mice were acutely depleted up to >99% by CD4-depleting Ab treatment before and during adjuvanted prime and boost vaccination (Fig. 5A). Approximately 3-fold lower levels of IgG (mostly IgG1 isotype) Ab responses were induced in acute CD4-depleted WT mice with MPL-adjuvanted vaccination compared with those in CD4KO mice with the same adjuvanted vaccination (Fig. 5B, Table I). Also, acute CD4-depleted WT mice with MPL-adjuvanted vaccination displayed weight loss of ∼8–12% (Fig. 5C), which is lower protective efficacy than that in CD4KO mice (Fig. 2A). CD4-depleted WT mice with MPL+Alum–adjuvanted vaccination showed 10-fold lower levels of IgG Abs compared with those in CD4KO mice (Table I). MHCII KO mice deficient in CD4⁺ T cells in addition to a genetic defect in MHCII expression (16, 23) were used for adjuvanted influenza vaccination. The levels of IgG Abs in MHCII KO mice with MPL- or MPL+Alum–adjuvanted influenza vaccination were significantly lower by 32-fold compared with those in CD4KO mice (Table I). These experimental data suggest that compensatory immune components, particularly MHCII-expressing APCs, developed in CD4KO mice contribute to the generation of isotype-switched IgG Abs, probably by providing an alternative form of T cell help.

To determine adjuvant effects on innate responses in CD4KO mice, we injected adjuvants (MPL+Alum, MPL, Alum) i.p. to WT and CD4KO mice. Sera were taken at 1.5, 6, and 24 h after the injection and then the mice were sacrificed to collect peritoneal exudates. Cytokines and chemokines in peritoneal exudates and sera were determined by ELISA. MPL injection of WT mice acutely induced extreme high levels of proinflammatory cytokines (IL-6, TNF-α) and MCP-1 chemokine within 1.5 h systemically in blood, which were rapidly reduced within 6 h (Fig. 6A). In contrast, MPL+Alum did not acutely raise inflammatory cytokines and MCP-1 chemokine but induced a low level of TNF-α and MCP-1 at 6 h postinjection in sera of WT mice (Fig. 6A). Alum injection did not induce cytokines or chemokines. Therefore, these results suggest that Alum in the combination MPL+Alum adjuvant plays a role of attenuating acute inflammation due to MPL injection in WT mice.

A different pattern was observed after adjuvant injection in CD4KO mice (Fig. 6B). Serum IL-6 and TNF-α cytokines were induced to a moderate and low level, respectively, at 1.5 h and then lowered to a basal level within 6 h after MPL+Alum or MPL injection of CD4KO mice (Fig. 6B). MPL or MPL+Alum injection induced a high level of serum MCP-1 in CD4KO mice. These results indicate that MPL+Alum may have an effect on acutely inducing cytokines and MCP-1 chemokine in CD4KO mice, which is different from WT mice.

In contrast to acute cytokine levels in blood, we observed a different profile of cytokines and chemokines at the injection site after 24 h (Fig. 6C). MPL injection of WT or CD4KO mice induced low levels of chemokines (MCP-1, RANTES) in the peritoneal cavity after 24 h (Fig. 6C). Interestingly, MPL+Alum injection of CD4KO mice induced moderate levels of IL-6, TNF-α, MCP-1, and RANTES in the peritoneal cavity after 24 h, which are higher than those in MPL+Alum i.p. injection of WT mice (Fig. 6C).

Immune cell types and cellularity at the injection site might provide insight into mechanisms of adjuvant effects in CD4KO mice. We analyzed cell types recruited in the peritoneal cavity at 24 h after i.p. injection of mice with adjuvants. Naive WT mice maintain high cellularity of macrophages in the peritoneal cavity (Fig. 7A, 7B). Interestingly, Alum injection of WT mice resulted in almost complete depletion of macrophages (Fig. 7A, 7B) and lower cellularity of plasmacytoid DCs (pDCs; 17%, Fig. 7I) and CD11b^low DCs (8%, Fig. 7H) in the peritoneal cavity, whereas CD4KO Alum maintained CD11b^low/high DC populations compared with PBS controls (Fig. 7G, 7H). In contrast, WT-MPL+Alum mice had low levels (20–30%) of macrophages (CD11b⁺F4/80⁺, CD11b⁺F4/80⁺MHCII^high, Fig. 7A, 7B), pDCs (11%, Fig. 7I), and CD11b^low DCs (10%, Fig. 7H) in the peritoneal cavity compared with those of PBS mock control mice (Fig. 7A, 7B, 7G–I). Differing from a profile in WT mice, MPL+Alum injection of CD4KO mice maintained the cellularity of macrophages (100%, Fig. 7A, 7B) and resulted in increasing CD11b^low DCs and CD11b^high DCs (Fig. 7G, 7H). Additionally, CD4KO MPL+Alum mice displayed significantly increased recruitment of neutrophils, monocytes, and NK cells by 76-, 23-, and 5-fold (Fig. 7C, 7D, 7F), respectively, which is a similar profile observed in WT-MPL+Alum mice. WT-MPL and CD4KO MPL mice showed a similar pattern of cellular changes in macrophages (reduced to 25%, Fig. 7A, 7B), monocytes (4-fold up, Fig. 6C), neutrophils (8-fold up, Fig. 7D), and NK cells (4-fold up, Fig. 7F). Uniquely, CD4KO MPL mice showed 2- to 4-fold increased levels of CD11b^high DCs, CD11b^low DCs, and pDCs compared with those in WT-MPL mice (Fig. 7G, 7H, 7I). KO-Alum mice recruited the highest level of eosinophils (4-fold) while reducing the levels of macrophages (<10%) and pDCs (<25%) in the peritoneal cavity compared with KO-PBS control mice (Fig. 7A, 7B, 7I). Additionally, CD4KO mice showed higher cellularity of the DN T cell (CD3⁺CD4⁻CD8⁻) population, and MPL injection in both WT and CD4KO mice increased the DN T cells (2-fold compared with the PBS group) at the site of injection (Fig. 7J).

Overall, MPL+Alum KO mice maintained ∼4- to 6-fold higher levels of MHCII^high macrophages and CD11b^high/low DC populations compared with those in MPL+Alum WT mice. Also, 2- to 4-fold higher levels of CD11b^high/low DC and DN T cell populations were observed in KO-MPL mice than in WT-MPL mice. These results suggest that differential cellularity of macrophages and DC populations together with DN T cells might be at least partially contributing to alternative B cell help in CD4KO mice.

DCs are important cells to link between innate and adaptive immune responses. After 2-d cultures of DCs in the presence of adjuvants, cytokine levels in culture supernatants and activation marker expression on BMDCs were measured (Fig. 8A, 8B). MPL+Alum and MPL induced proinflammatory cytokine production by DCs, but Alum-treated DCs did not produce cytokines (Fig. 8A). MPL+Alum showed high levels of IL-6 and TNF-α but displayed a lower IL-12 cytokine level compared with the MPL alone adjuvant. In terms of DC activation markers, MPL showed significantly higher levels of CD40, CD80, and CD86 expression than did MPL+Alum on DCs (Fig. 8B). All adjuvants, that is, MPL+Alum, MPL, and Alum, increased MHCII^high DC populations (data not shown). These data indicate that Alum in MPL+Alum combination appears to attenuate stimulatory effects on DCs by MPL, a major player in in vitro activation of DCs.

To gain better understanding of in vivo adjuvant effects on reducing the cellularity of macrophages and DCs, we further tested in vitro cell death. Alum and MPL+Alum adjuvants were found to induce significant loss in cell viability after in vitro cultures of BM-derived primary cells (Fig. 8C).

To investigate a further possible mechanism of combination MPL+Alum adjuvant, we determined in vitro IgG production by B cells and proliferation of DN T cells after culture with or without adjuvant-pretreated BMDCs (Fig. 8D–F). MPL+Alum–pretreated BMDCs could support CD4KO mouse splenic B cells for IgG production in vitro, but splenic B cells with adjuvant only could not (Fig. 8D). Additionally, proliferation of DN T cells in CD4KO mouse splenocytes was stimulated by coculture with MPL+Alum–pretreated BMDCs, and MHCII mAb treatment suppressed its proliferation effects (Fig. 8E). Additionally, to determine cognate or noncognate help, we determined proliferation of DN T cells after incubation with MHCII⁺ BMDCs stimulated with vaccine or nonvaccine Ags (Fig. 8F). Vaccine-treated BMDCs significantly increased the proliferation of DN T cells, but a different virus (A/Philippine H3N2) or OVA-treated BMDCs did not (Fig. 8F). These results provide evidence that MHCII⁺ APCs contribute to stimulating Ag-specific DN T cells and providing alternative B cell help in a genetically CD4-deficient condition.

Subunit vaccines provide a safe alternative to live-attenuated virus vaccines but have poor immunogenicity, requiring effective adjuvants to enhance the vaccine efficacy. Immune-competent mouse models are often highly responsive to experimental vaccines and adjuvants; however, they may not represent the efficacy expected in humans (24, 25). It is also thought that adjuvant effects are mediated by specific types of activated CD4⁺ T cells that would be educated via MHCII-expressing APCs through multiple innate immune cellular interactions and the production of inflammatory cytokines resulting from adjuvant stimulation (7, 26–28). That is, vaccine adjuvants are important for modulating the types of CD4⁺ T cells to induce the observed outcomes of adaptive immune responses in a conventional model. In contrast, the data in this study provide evidence that MPL+Alum adjuvant combination can mediate the induction of isotype-switched IgG Abs, conferring protective immunity in CD4KO mice comparable to those in WT mice even after 5 mo of vaccination. Therefore, this study suggests an alternative pathway and/or cells in providing help to the B cells for IgG production in CD4KO mice in the context of MPL+Alum– and MPL-adjuvanted influenza vaccination.

Alum adjuvants were unable to induce IgG Abs against split vaccine in CD4KO mice after prime. IgG1 Abs after boost immunization of CD4KO mice with Alum progressively waned to a further lower level after 3 mo (Table I). In contrast to Alum, MPL+Alum and MPL adjuvant effects were potent in WT and CD4KO mice even with prime only. The major difference between Alum and MPL+Alum is the induction of inflammatory cytokines, which was evident in vitro and in vivo. As expected, the results revealed that MPL- or MPL+Alum–activated BMDCs secrete IL-6 and TNF-α inflammatory cytokines, whereas Alum by itself did not. MPL+Alum did not directly stimulate CD4⁺ T or B cells in vitro but was shown to directly activate DCs in vitro primarily due to MPL, leading to the induction of Ag-specific T cells in vivo (29). Therefore, the microenvironment creating inflammatory cytokines by MPL+Alum and MPL is likely to play a major role in priming IgG isotype–switched B cell response in CD4KO mice. Previous studies demonstrated that mice lacking CD40 or CD4⁺ T cells during sublethal primary infection induced only short-lived IgG Abs waning within 60 d and no Ab-secreting plasma cells (21, 22). Therefore, the results of long-lived IgG Ab responses mediated by MPL+Alum and MPL adjuvant in CD4KO mice are particularly notable.

The mechanisms of adjuvant effects on enhancing vaccine efficacy remain poorly understood, particularly in a CD4-deficient condition, although CD4-independent activation of CD8⁺ T cells and B cells has been reported. DN αβ T cells were shown to play a role in generating CD4-independent isotype-switched IgG Abs using CD4KO and TCRβ KO mouse models (30). Another study demonstrated that both B7-1 (CD80) and B7-2 (CD86) costimulatory molecules on DCs were required for IgG1 and IgG2a responses (31). The TNF-α pathway was important for activating cytotoxic effector CD8⁺ T cells in the absence of CD4 T cells (32). This study provides evidence that a CD4 genetic defect led to developing other compensatory immune components even in a mock (PBS) treatment condition. PBS treatment of CD4KO mice showed higher levels of MHCII^high macrophages, CD11b^high DCs, DN T cells, and a lower level of pDCs compared with those in WT mice. Upon treatment with combination MPL+Alum adjuvant, CD4KO mice further increased cellularity of MHCII^high macrophages and CD11b^high and CD11b^low DCs compared with those in WT mice, in addition to comparable increases in monocytes, neutrophils, NK cells, and eosinophils. These cellular increases at the site of injection appeared to be correlated with high levels of cytokines (IL-6, TNF-α) and chemokines (MCP-1, RANTES) upon treatment of CD4KO mice with combination MPL+Alum adjuvant. Meanwhile, MPL-treated CD4KO mice showed increased levels in DC populations (CD11b^high, CD11b^low DCs, pDCs) and DN T cells compared with those in MPL-treated WT mice. We also observed further increased levels of DCs, but not macrophages, compared with those in MPL+Alum–treated CD4KO mice. Thus, it is possible that cellular components contributing to alternative B cell help in CD4KO mice are likely to be different between MPL (mostly various DC subsets, DN T cells) and combination MPL+Alum (mostly macrophages, minimally CD11b^low DCs) adjuvants.

To gain further insight into possible roles of these compensatory immune components in providing alternative B cell help in CD4KO mice, we determined IgG responses in acutely CD4-depleted WT mice. We observed lower levels of IgG responses in CD4-depleted WT mice with MPL and MPL+Alum compared with those in CD4KO mice with the same adjuvanted vaccination. Thus, it is possible that compensatory immune components, including MHCII^high macrophages, DC populations, and DN T cells, developed in CD4KO mice are contributing to overcoming defects in CD4 help to B cells for the generation of isotype-switched Abs by MPL and combination MPL+Alum vaccination. In line with these results, MHCII KO mice with MPL+Alum adjuvant vaccination were found to induce lower levels of IgG responses by 32-fold than those in CD4KO mice with the same adjuvant vaccination (Table I). Additionally, MPL+Alum–activated BMDCs might have the capability to stimulate naive splenic B cells to secrete IgG Abs in vitro. Because efficient naive CD4⁺ T cell priming does not require B cells expressing MHCII (33), MHCII-expressing APCs such as DCs and macrophages significantly contribute to the generation of isotype-switched IgG Abs, probably via an alternative pathway different from conventional CD4⁺ T cell help.

Adjuvant roles of Alum and MPL in MPL+Alum combination are yet to be fully understood. Differences in serum IgG Ab levels between MPL+Alum and MPL appeared to be greater at a later time point of 3 mo boost in CD4KO mice (Table I), which are further supported by ASC responses in spleens. Alum in AS04 (MPL+Alum) appeared to have a property of attenuating innate immune-stimulating activities by MPL in vitro and in vivo. MPL was found to highly upregulate the expression of costimulatory molecules (CD40, CD80, and CD86) on BMDCs, which can provide alternative B cell help to produce isotype-switched Abs (34). Meanwhile, MPL+Alum appeared to moderately suppress the stimulatory effects of MPL on upregulating costimulatory molecules and IL-12 cytokine production during BMDC in vitro cultures. Attenuating acute induction of inflammatory cytokines would improve the safety for AS04 adjuvant–formulated human vaccination.

Adjuvant-induced cell death has been known to be a mechanism of adjuvanticity for more than a decade (35), although cell death is often considered an undesirable side effect. TNF family cytokines are classic inducers of programmed necrosis via receptor-interacting protein kinases, which promotes inflammation (36). Alum was demonstrated to induce uric acid by causing sterile cell death (37). Also, there is a well-known link between TLR activation and cell death leading to proapoptotic activities (38). We found that in vitro cultures of BMDCs with Alum or MPL+Alum resulted in cell death. In in vivo studies, WT mice with Alum, MPL+Alum, or MPL injection resulted in a significant loss of macrophages, CD11b^low DCs, and pDCs in the peritoneal cavity. CD4KO mice with Alum or MPL injection exhibited a significant loss in macrophages in the peritoneal cavity whereas MPL+Alum retained up to 80% of macrophages. Thus, MPL+Alum adjuvant combination might be contributing to protect macrophages from a severe cellular loss in the in vivo injection site in WT mice compared with Alum, and more prominently in CD4KO mice compared with Alum and MPL.

In summary, MPL+Alum showed effective adjuvant effects on inducing IgG isotype-switched Abs and conferring protective immunity in CD4KO mice, which was more effective compared with those in WT mice with influenza vaccine only or Alum-adjuvanted vaccination. MPL+Alum showed moderate levels of cytokines (IL-6, TNF-α) and chemokines (MCP-1, RANTES) in the peritoneal cavity at 24 h after injection of CD4KO mice. MPL+Alum appears to have differential effects on generating a local inflammatory microenvironment, maintaining macrophages, and attenuating acute inflammation compared with those of MPL and Alum. MHCII-expressing cellular components, DN T cells, and soluble cytokines and chemokines at the site of injection are likely to be the major contributing factors in providing alternative help to B cells for inducing IgG Ab responses in the context of MPL+Alum–adjuvanted influenza vaccination of CD4KO mice.

The authors have no financial conflicts of interest.