The cationic liposome CCS/C adjuvant induces immunity to influenza independently of the adaptor protein MyD88

ABSTRACT

Traditional non-living vaccines are often least effective in the populations that need them most, such as neonates and elderly adults. Vaccine adjuvants are one approach to boost the immunogenicity of antigens in populations with reduced immunity. Ideally, vaccine adjuvants will increase the seroconversion rates across the population, lead to stronger immune responses, and enable the administration of fewer vaccine doses. We previously demonstrated that a cationic liposomal formulation of the commercial influenza split virus vaccine (CCS/C-HA) enhanced cellular and humoral immunity to the virus, increased seroconversion rates, and improved survival after live virus challenge in a preclinical model, as compared to the commercial vaccine as is (F-HA). We now evaluated vaccine efficacy in different strains and sexes of mice and determined the role of innate immunity in the mechanism of action of the CCS/C adjuvant by testing the response of mice deficient in Toll-like receptors or the TLR/IL-1 adaptor protein MyD88 following immunization with CCS/C-HA vs. F-HA. Although TLR2- and TLR4-deficient mice responded to F-HA immunization, F-HA immunization failed to engender a significant immune response in the absence of MyD88. In contrast, immunization with the CCS/C-HA vaccine overcame the requirement for MyD88 in the response to the commercial vaccine and improved the immune responses and seroconversion rates in all strains of mice tested, including those deficient in TLR2 and TLR4.

KEYWORDS:

• Influenza
• cationic liposomes
• TLR
• MyD88
• adjuvant

Introduction

Influenza vaccines are the primary strategy to reduce the substantial health burden from seasonal influenza. Inactivated influenza vaccines have been available since the 1930s¹ and are administered via intramuscular (i.m.) injection. Live attenuated, cold-adapted influenza vaccines (LAIV) were licensed in the United States in 2003, and are administered via nasal spray.^2-4 The effectiveness of influenza vaccines correlates predominantly to the age and immune competence of the vaccinee and the antigenic relatedness of vaccine strains to circulating strains.^5-7 More effective seasonal influenza vaccines are required in order to provide long-lasting immunity and broad protection against strains that differ antigenically from vaccine viruses. Of particular need are vaccine strategies that address immunity in populations at most risk of influenza infection, such as the elderly and neonates.

We previously reported that novel cationic liposomes composed of the polycationic sphingolipid N-palmitoyl d-erythro-sphingosyl-carbamoyl-spermine triacetate salt (CCS) and cholesterol (CCS/C), and including influenza hemagglutinin (HA) antigen, stimulate increased humoral and cellular immune responses in different animal models and protect against influenza virus infection^8-10 A. Joseph, N. Itskovitz-Cooper, S. Samira, O. Flasterstein, H. Eliyahu and D. Simberg et al., A new intranasal | Cited By in Scopus (32). In mice, intranasal (i.n.) administration of the standard commercial trivalent influenza vaccine formulated in CCS/C liposomes evokes strong systemic (serum) and local mucosal (nose and lungs) antibody responses as well as cellular responses and protective immunity.¹¹ The CCS/C-influenza vaccine is significantly more immunogenic than the standard vaccine given i.m. or i.n. High antibody titers and protective immunity to viral challenge are sustained for >9 months following immunization.^8,10 Furthermore, the CCS/C-adjuvanted vaccine is highly efficacious following a single dose (i.m. or i.n.) in both young (2 mo) and old (18 mo) mice, and it elicits high titers of strain cross-reactive Hemagglutination Inhibiting (HI) antibodies.¹¹ In ferrets, the i.n. CCS/C-influenza vaccine elicits up to 10-fold higher HI antibody titers compared to the conventional i.m. vaccine. In rats, the CCS/C influenza vaccine is significantly more immunogenic by the i.m. or i.n. route than conventional vaccine.¹⁰ Hence, CCS/C has the potential to be a valuable adjuvant.

Most adjuvants in clinical or experimental use enhance immunity by activating innate immune signaling pathways in antigen-presenting cells.¹² Since innate immunity is diminished in populations most at risk of infection,^13,14 the reliance of adjuvants on innate signaling pathways is a potential limitation. In the current study, we further elucidate the mechanism of action of the CCS/C influenza vaccine by testing it in mice deficient in receptors or signaling molecules of the innate immune system. Toll-Like Receptors (TLRs) in humans and mice respond to a broad array of evolutionarily conserved microbial molecules.¹⁵ TLR4 is required for the response to lipopolysaccharide from most gram-negative bacteria, whereas TLR2 governs the response to peptidoglycan and lipopeptides from most gram-positive bacteria.¹⁶ Stimulation of TLRs by microbial structures leads to the activation of transcription factors that drive cytokine expression, proliferation, survival, and inflammatory mediator expression. MyD88 is a central adaptor protein that assembles a protein complex at the intracellular portion of all TLRs except TLR3, and members of the IL-1 receptor family, enabling the recruitment of kinases.¹⁷ Therefore, much of innate immune signaling converges on MyD88, and MyD88 deficiency in mice models a state of severe compromise in innate immunity.^15,18

We now demonstrate that the CCS/C liposomes as adjuvant improve the immune response of TLR2- and TLR4-deficient mice, to the influenza vaccine. Furthermore, whereas MyD88 is required for an effective response to the commercial influenza vaccine, incorporation of this vaccine within the CCS/C liposomes overcomes this requirement.

Materials and methods

Animals

Mouse experiments were performed in the animal facility of the Hebrew University, Jerusalem, Israel. Specific pathogen-free (SPF) female and male BALB/c mice, female C3 H/HeNHsd mice, and female C57BL/6JOlaHsd mice (8–9 weeks old) were obtained from Harlan (Jerusalem, Israel). C57/BL10 wild-type (used for experiments shown in Figure 2 and Table 1) and C57/BL10ScN (TLR4^lps-del), a spontaneous TLR4 knock-out strain, were from the Jackson Laboratory (JAX stock #003752, Maine, USA). Myd88^−/- and Tlr2^−/-, mice backcrossed to the C57BL/6JOlaHsd background were kindly provided by Prof. S. Akira (Osaka University, Japan) to GN. All animal studies were approved by the Institutional Animal Care and Use Committees of the Hebrew University.

Table 1. Proliferative response and cytokine production by spleen cells following two doses of F-HA vs. CCS/C-HA vaccination in WT, Tlr2-/-, and Tlr4-/- mice

		Spleen
Vaccine (n = 5)	Incubation Treatment	Proliferation (SI^a)	IFN γ (pg/ml)	IL-2 (pg/ml)	IL-5 (pg/ml)
Con A	0.25 µg HA/ml	8.2	>1000	70 ± 13	152 ± 21
PBS	0.5 µg HA/ml	1.4	0	3 ± 1	0
	0.05 µg HA/ml	1.5	0	2 ± 0.4	0
F-HA WT	0.5 µg HA/ml	2	182 ± 10	75 ± 28	439 ± 79
	0.05 µg HA/ml	1.5	154 ± 18	57 ± 20	391 ± 62
CCS/C-HA WT	0.5 µg HA/ml	4	400 ± 22	96 ± 21	660 ± 59
	0.05 µg HA/ml	2.4	422 ± 65	22 ± 13	620 ± 28
F-HA Tlr2^−/-	0.5 µg HA/ml	1.8	0	63 ± 11	547 ± 28
	0.05 µg HA/ml	1.8	0	53 ± 13	368 ± 25
CCS/C-HA Tlr2^−/-	0.5 µg HA/ml	3.2	593 ± 62	590 ± 29	292 ± 45
	0.05 µg HA/ml	5	581 ± 112	342 ± 29	217 ± 25
F-HA Tlr4^−/-	0.5 µg HA/ml	0.8	29 ± 18	85 ± 17	325 ± 57
	0.05 µg HA/ml	1	0	61 ± 6	200 ± 23
CCS/C-HA Tlr4^−/-	0.5 µg HA/ml	3.3	417 ± 26	316 ± 38	306 ± 41
	0.05 µg HA/ml	2.6	463 ± 18	318 ± 42	302 ± 35

C57BL/6 (WT), Tlr2^−/- and Tlr4^−/- female mice (n = 3–5) were vaccinated on d 0 and 28 with PBS only or F-HA vs. CCS/C-HA (Aventis Pasteur Vaxigrip 2008–09, 3 μg HA). Spleen cells were harvested 38 d after the first vaccination (10 d after boost). ^aSI>2, positive.

Influenza antigens (HA)

A commercial trivalent split vaccine (Vaxigrip, Aventis Pasteur or Sanofi Pasteur) from several vaccination seasons, based on availability, was used throughout the study. The three strains of the vaccine were those recommended by the World Health Organization for the respective influenza season: A/Solomon Islands/3/2006 (H1N1)-like virus; A/Wisconsin/67/2005 (H3N2)-like virus; B/Malaysia/2506/2004-like virus (season 2007/2008, used for the mice strains and gender comparison studies). A/Brisbane/59/2007 (H1N1) – like strain IVR-148 derived from A/Brisbane/59/2007, A/Brisbane/10/2007 (H3N2) – like strain NYMC X-175 C derived from A/Uruguay/716/2007 and B/Florida/4/2006 – like strain B/Florida/4/2006 (season 2008/2009, used for the TLR/MyD88 knock-out mouse experiments). The level of HA is based on HA quantification as described in the commercial vaccine insert. In all seasonal vaccines, the HA content was ~50% of total split vaccine protein (quantified by Lowry method¹⁹).

Lipids

The polycationic sphingolipid N-palmitoyl D-erythro sphingosyl carbamoyl-spermine (CCS) was obtained from Biolab Ltd (Jerusalem, Israel). High purity (HP) cholesterol was obtained from Minakem, France. Cholesterol was used as “helper lipid”. Helper lipids such as cholesterol are neutral lipids that contribute to the stability and delivery efficiency of cationic lipids. All lipids used were ≥98% pure.

CCS/C-HA vaccine preparation and physicochemical characterization

Vaccine was formulated as described previously¹¹ using a 3:2 CCS:C molar ratio in PBS containing Vaxigrip at the desired HA:lipid weight ratio. The physical characterization of the CCS/C-based vaccine including the level of association of the commercial split vaccine proteins to the CCS/C liposomes, mean particle size, size range, and zeta potential were determined as described previously.¹⁰ As determined by the Lowry method after the separation between the liposomes and their medium, >85% and in most cases ~95% of the commercial split protein is associated with the cationic liposomes. The CCS/C-based vaccine mean particle size, and size range were determined using a particle size analyzer (BeckmanCoulter LS I3 320). This instrument combines two methods of size distribution analyses: multi-wave light diffraction and polarization intensity differential scattering (PIDS polarization of the light). This combination allows the determination of broad size distributions in the range of 40 nm to 2 mm and therefore is more suitable for particles out of the nano-size range such as the cationic liposomes used in this study. The vaccine liposomes are a mixture of uni-, oligo- and multi-lamellar liposomes (lipid vesicles) of various shapes and broad size distribution with a reproducible mean size of 1.0 µm, and >85% of the particles are in the range of 0.1 to 3.0 µm. As determined by using a Zetasizer Nano Series ZEN2600 F (Malvern Instruments, Malvern, UK) these liposomes are cationic having a positive zeta potential measured in PBS (0.01 M phosphate buffer in 0.145 M sodium chloride, pH 7.2) of ~33 mV (for more details on the characterization see ref. 10).

Immunization of mice

In mouse experiments, the un-adjuvanted commercial Vaxigrip split vaccine (referred to as free HA [F-HA]) and the CCS/C-HA vaccine were used at 0.5–1.0 μg HA/strain/dose and 0.15–3 mg CCS/C/dose (HA:CCS/C weight ratio 1:100) and administered once or twice (28 d apart) intramuscularly (i.m.) in 50 μl volumes.

Evaluation of humoral immune responses in mice

Mice were bled from the facial vein, and sera were prepared as described previously.¹¹ In the case of one immunization, mice were bled and sacrificed on d30 after vaccination. In the case of two immunizations, mice were bled on d 26, re-vaccinated on d 28, and then bled again 9 or 10 d later. HI antibodies were determined on individual serum samples by the standard HI assay, starting at a 1:20 sample dilution.²⁰ HI titer was tested against the most immunogenic strain from the commercial vaccine used in each experiment, as indicated in the figure legend. Mice with HI titer ≥40 (considered a protective titer in humans) were defined as seroconverted.

Assessment of cellular responses in mice

Mice were sacrificed by cervical dislocation under deep anesthesia, and spleens were isolated. Splenocytes were tested for proliferative responses and for the production of IFNγ, IL-2, and IL-5 following in vitro stimulation with antigen. Cell cultures were performed as described previously.¹⁰ Briefly, to measure proliferation 0.5 × 10⁶ cells per well were incubated in triplicate, with or without the antigen (0.05–0.5ug/ml), in a final volume of 0.2 ml. After 96 hr, cultures were pulsed with 1 μCi 3 H-thymidine ON. Con A was used as positive control (0.25 μg/ml). Data shown in Tables 1–2 represent the stimulation index (SI) = mean cpm + antigen/mean cpm without the antigen. To measure cytokine production, 2.5 × 10⁶ cells per well were incubated in duplicates as above. Con A was used as positive control (0.25 μg/ml), in a final volume of 1 ml. The supernatants were collected 72 hr later and tested by ELISA (BioLegend Max Set Deluxe, according to manufacturer’s instructions). Cytokine levels were calculated after subtraction of background (cells incubated with medium alone). Data in Table 1–2 represent a single experiment with 3–5 mice per group.

Table 2. Proliferative response and cytokine production by spleen cells following vaccination with one or two doses of F-HA vs. CCS/C-HA vaccine in WT and Myd88-/- mice

		Spleen
		Proliferation (SI^a)		IFN γ (pg/ml)		IL-2 (pg/ml)		IL-5 (pg/ml)
Vaccine (n = 5)	Incubation Treatment	Group A	Group B	Group A	Group B	Group A	Group B	Group A	Group B
Con A	0.25 µg HA/ml	5.1	3.3	>1000	>1000	10 ± 3.5	40 ± 6	215 ± 11	95 ± 16
PBS	0.5 µg HA/ml	1.1	1.1	0	0	0	0	0	0
	0.05 µg HA/ml	1.4	1.4	0	0	0	0	0	0
F-HA WT	0.5 µg HA/ml	0.3	0.4	3313 ± 279	2068 ± 216	25 ± 3.5	33 ± 3.5	97 ± 30	494 ± 38
	0.05 µg HA/ml	0.3	0.6	1680 ± 382	825 ± 35	20 ± 1	30 ± 7	68 ± 9	403 ± 16
CCS/C-HA WT	0.5 µg HA/ml	3.8	3	1103 ± 138	9640 ± 105	130 ± 14	225 ± 7	0	39 ± 8
	0.05 µg HA/ml	2.4	1.8	325 ± 21	1585 ± 219	88 ± 18	133 ± 11	0	0
F-HA Myd88^−/-	0.5 µg HA/ml	0.5	0.6	0	0	15 ± 3.5	68 ± 11	15 ± 3.5	258 ± 30
	0.05 µg HA/ml	1.1	1.2	0	0	0	35 ± 11	0	82 ± 21
CCS/C-HA Myd88^−/-	0.5 µg HA/ml	2.1	2.7	1040 ± 113	2615 ± 290	413 ± 17	188 ± 18	299 ± 28	988 ± 58
	0.05 µg HA/ml	2.8	3	480 ± 40	1895 ± 290	253 ± 11	150 ± 14	188 ± 23	836 ± 40

C57BL/6 (WT) and Myd88^−/- female mice (n = 3), were vaccinated once (Group A) or twice (d 0 and 27, Group B) with PBS only, or F-HA vs. CCS/C-HA (Aventis Pasteur Vaxigrip 2008–09, 3 μg HA). Spleen cells were harvested 30 d (Group A) and 36 d (Group B, 9 d after boost) after vaccination. ^aSI>2, positive.

Statistical analysis

In all studies, the differences between groups were analyzed using a two-tailed Student’s t-test when two means were compared, and analysis of variance (ANOVA) when more than two means were compared. P-values <0.05 were considered significant.

Results

CCS/C-HA adjuvant effect in different mouse strains

Genetic background influences both murine and human responses to vaccines and vaccine adjuvants through effects on antigen presentation and co-stimulatory signals, including the types and levels of cytokines that are produced during T and B cell activation.^21-24 We therefore compared the immunogenicity of the F-HA native vaccine and the CCS/C-HA adjuvanted vaccine in three inbred mouse strains differing in MHC haplotype and in the character of their immune responses.^25-27 Vaccine efficacy was compared in female mice of all strains, and in male Balb/c mice. HI titers differed among strains in response to both vaccines (Figure 1). For both vaccines independently, one-way ANOVA across the four groups of mice demonstrated that C3 H mice responded significantly better than the other strains, and that differences observed between the other strains did not reach significance. In C3 H and C57BL/6 female mice, and in Balb/c male mice, the CCS/C-HA vaccine induced significantly higher HI antibody titers than the F-HA vaccine (Figure 1). The effect of CCS/C-HA vaccination on HI titer across strains was highly significant compared to F-HA vaccination (P < .0001, 2-way ANOVA).

Figure 1. Effect of mouse strain and gender on immunogenicity of F-HA vs. CCS/C-HA

HI titer following F-HA vs. CCS/C-HA i.m. immunization of different strains of female mice, and Balb/c male mice (all age matched, n = 5). Mice were immunized with 0.5 µg HA/mouse in a volume of 50 µl. Serum HI titer was tested 4 weeks after immunization against A/Solomon Islands/3/2006 (H1N1) virus and is shown as mean ±SD. *P ≤ 0.05, ns, not significant. ANOVA across strains indicated a highly significant effect of treatment with CCS/C-HA vs. F-HA (P < .0001).

Involvement of TLR2 and TLR4 in the response to F-HA and CCS/C-HA vaccination

Innate immune signaling pathways enhance MHC-directed antigen presentation and co-stimulatory molecule expression and influence the phenotype and strength of the response through induction of an inflammatory cytokine milieu during the encounter between APCs and responder T cells.¹² TLRs 2 and 4 govern the response to a broad array of microbial molecules and host damage associated ligands generated at the immunization site.²⁸ We therefore tested the effect of one or two immunizations with F-HA vs. CCS/C-HA in mice deficient in TLR2 or TLR4. For both vaccines, HI titers were similar between WT, TLR2 knock-out (KO), and TLR4 KO mice in response to one (Figure 2a), or two (Figure 2b), immunizations. After primary immunization, the CCS/C-HA vaccine led to significantly higher titers in all strains (Figure 2a). However, after the second immunization, the CCS/C-HA vaccine showed a significant improvement in HI titer in WT mice (P = .01), but not in either of the KO strains (Figure 2b). Splenic T cell responses (proliferation and IFNγ cytokine secretion) were enhanced by immunization with CCS/C adjuvant in WT mice in a similar manner to our previous findings. In both TLR-deficient strains the CCS/C adjuvant greatly enhanced the splenocyte response to HA, with a similar pattern to that in WT mice – proliferation and production of IFNγ a Th1-type cytokine, were upregulated but IL-5, a Th2-type cytokine, was not impacted (Table 1). Together, the results indicate that TLR2 and TLR4 play a very minimal role in the adjuvant effect of CCS/C liposomes.

Figure 2. TLR2 and TLR4 do not contribute to antiviral immunity induced by F-HA and CCS/C-HA

HI titer following one immunization (A) or two immunizations (B) with F-HA vs. CCS/C-HA in WT, TLR2 KO, and TLR4 KO mice. Mice (n = 6) were immunized with 1 µg HA/mouse of each viral strain (3 µg total) in a volume of 50 µl. The CCS/C formulation contained 0.3 mg lipid/mouse. Serum HI titers against A/Brisbane (H1N1) virus were tested individually 26 d after the first vaccination and 10 d after the second vaccination. For each vaccine, there were no significant differences between the responses of WT, TLR2 KO, and TLR4 KO mice. The effect of the F-HA vs. CCS/C-HA vaccine within each strain of mice in response to one or two immunizations, ***P ≤ 0.001, **P ≤ 0.01, **P ≤ 0.05, ns, not significant.

The CCS/C adjuvant overcomes the requirement for MyD88 which is required for the immune response to F-HA

Mice genetically deficient in MyD88 lack most TLR signaling,²⁹ but also lack responsiveness to major pro-inflammatory cytokines that contribute to the maturation and phenotype of T cells such as IL-1 and IL-18. To explore the role of these pathways in the antibody response to HA immunization and to test the augmenting role of the liposomal adjuvant in the response to HA antigen, we immunized WT vs. Myd88^−/- mice. Mice were immunized once (D 0, Group A) or twice (D 0 and 28, Group B). Blood samples were collected on d 30 in Group A, and on d 26 and 36 in Group B. Samples were tested for HI antibody titers, and after sacrificing the animals on d 30 (Group A) and 36 (Group B), the spleens were harvested and cells were tested for cellular responses, including cytokine release and proliferative responses.

F-HA immunization was unable to induce HI antibodies in MyD88-deficient mice following a single immunization (n = 6) (Figure 3a). Strikingly, the CCS/C-HA vaccination overcame the dependency of F-HA immunization on MyD88-signaling. In MyD88-deficient mice, one vaccination with CCS/C-HA led to 100% seroconversion with HI titers equivalent to the titers in WT mice vaccinated twice with the F-HA vaccine (Figure 3a-b). Boost-vaccination of Myd88^−/- mice with CCS/C-HA led to high titer HI antibodies in all mice tested. In WT mice, the CCS/C-HA vaccine outperformed F-HA immunization, consistent with our previous results (Figures 1–2 and⁸). We next tested the antigen-specific proliferative response of splenic T cells from WT and Myd88^−/- mice immunized with F-HA vs. CCS/C-HA (Table 2). We observed variability in the magnitude of the response of WT mice between vaccination experiments (e.g. IFNγ level between Tables 1 and 2), and the CCS/C-HA failed to improve WT splenocyte IFNγ production in group A (Table 2). As expected, F-HA failed to elicit proliferative responses and cytokine production by anti-HA T cells in the absence of MyD88. In stark contrast, splenic T cells from Myd88^−/- mice immunized with CCS/C-HA proliferated and secreted IL-2, indicating activation in response to HA. Interestingly, IFNγ production increased in the Myd88^−/- cultures, but these cells also showed enhanced production of IL-5, a Th2-type cytokine.

Figure 3. Adjuvant effect of CCS/C-HA does not require MyD88

HI titer following one immunization (A) or two immunizations (B) with F-HA vs. CCS/C-HA in WT, and Myd88^−/- mice. Mice were immunized with 1 µg HA/mouse of each viral strain (3 µg total) in a volume of 50 µl. The CCS/C formulation contained 0.3 mg lipid/mouse. Serum HI titers against A/Brisbane virus were tested individually 26 d after the first vaccination and 9 d after the second vaccination. Percentages shown represent seroconversion rates defined by HI titer ≥40.

Discussion

Liposomes remain at the forefront of drug and vaccine design owing to their well-documented abilities to act as delivery vehicles. Whereas most research has focused on the ability of liposomes to deliver genes, drugs, and vaccine antigens, more recently liposomes have been shown to function as vaccine adjuvants by enhancing the immunogenicity of peptide and protein antigens. We previously demonstrated that the CCS/C liposomal adjuvant/carrier vaccine exhibits several advantages for influenza immunization. In contrast to immunization with non-adjuvanted HA antigen, a single i.m. immunization with the CCS/C-HA vaccine elicits potent and long-term serum HI titers, and characteristics of both Th1- and Th2-T cell helper responses in mice, rats, and ferrets.^8,10,11

In North America, annual influenza vaccination is recommended for all individuals >6 months old,³⁰ although the vast majority of influenza-related hospitalizations and deaths occur in individuals >65 y old.^31,32 Vaccination of the elderly with higher doses of HA per strain improved the prevention of influenza,³³ however, retrospective cohort studies that examined the effectiveness of high-dose vs. standard vaccination in the > 65-y-old population differed in their findings.^34,35 Alternatively, vaccine adjuvants can improve immunogenicity in high-risk groups,³⁶ and the FDA recently approved an influenza vaccine containing the adjuvant MF59 for seasonal influenza vaccination in individuals >65 y old.³⁷ Adjuvants may be particularly important for pandemic vaccines in order to shorten the response time and broaden cross-reactivity.³⁸ MF59 is an oil in water emulsion of squalene oil that increases immune responses when mixed with the trivalent vaccine. However, the MF59-adjuvanted vaccine is not recommended over other available vaccines due to the lack of evidence to support its efficacy compared to standard vaccination.³⁰

To improve the success of vaccination in the elderly, the effects of age on innate and adaptive immunity must be addressed. Importantly, most adjuvants boost immunity by recruiting innate immune cells and enhancing innate signaling pathways leading to more potent antigen presentation.¹² Some adjuvants are natural³⁹ or synthetic⁴⁰ Toll-like receptor (TLR) ligands that target specific TLR pathways, such as the LPS derivative Monophosphoryl lipid A.⁴¹ Others, such as the approved MF59, also depend on the innate immune TLR/IL-1 R signaling protein MyD88 for their adjuvant effect.^42,43 However, a quandary arises since much evidence demonstrates impaired innate immunity in both aged mice, and elderly individuals.^13,14,44,45 Aging reduces the number of antigen-presenting cells and their endogenous expression of MHC molecules, co-stimulatory molecules, and TLRs.^46,47 Activation of innate immune cells from the elderly results in significantly lower cytokine and chemokine expression compared to the response of cells from younger individuals.⁴⁷ This “immunosenescence” of the innate immune system presents a barrier to the efficacy of vaccines and vaccine adjuvants that rely on canonical innate signaling pathways to enhance adaptive immunity.

We therefore examined the immunogenicity of the CCS/C-HA vaccine compared to the standard vaccine in the setting of MHC differences, and deficiencies in innate immunity. Whereas CCS/C-HA improved immune responses across strains, the most significant finding was that the CCS/C adjuvant induces immunity (HI titers) independently of MyD88. In contrast, the standard trivalent vaccine was highly dependent on MyD88, and even two immunizations of Myd88^−/- mice were unable to induce significant seroconversion and HI titers. MyD88 is an adaptor protein that assembles a signaling complex at the intracellular domains of both TLRs and receptors of the IL-1 superfamily,¹⁷ suggesting that immunization with the trivalent vaccine requires activation of one or both, of these systems. The fact that immunogenicity induced by the trivalent vaccine was not significantly impaired in TLR2 or TLR4-deficient mice compared to WT mice does not rule out the involvement of other TLRs, especially TLR7 that recognizes single-stranded viral RNA and signals through MyD88.^40,48 The cytokine profile of splenic T cells from MyD88-deficient mice immunized with CCS/C-HA secreted a mixed pattern of Th1 and Th2 cytokines. Using pooled sera, we found that the CCS/C-HA vaccine-induced primarily an IgG1 response in MyD88-deficient mice (data not shown), suggestive of Th2-type immunity. Others have shown a shift to Th2-immunity following immunization or infection of MyD88-deficient mice,^49,50 and we demonstrated that pharmacologic inhibition of MyD88 induces a shift to Th2-type immunity.⁵¹ Of note, Alum, an adjuvant approved for human use, also induces Th2-immunity.⁵² Since the NLRP3 inflammasome was shown to mediate the adjuvant effect of alum,⁵³ future studies should determine the role of NLRP3 and other inflammasomes in the mechanism of action of the CCS/C adjuvant.

There are several limitations to the current study, including, but not limited to, small sample sizes and lack of repeat comparisons for some of the experiments, and the use of surrogate measures of vaccine efficacy (HI titer). We also used the most immunogenic strain to determine the HI titer of the immunized mice; however, to further the development of this adjuvant for human use, the HI titer against all strains, especially the least immunogenic, should be determined.

Our previous findings established that the CCS/C-HA vaccine is effective in young and aged mice, enables administration of fewer vaccine doses, and is protective in ferrets.^8,10,11 The current study adds to the benefit profile of the CCS/C-HA influenza vaccine by demonstrating its ability to induce immunity in the setting of impaired innate immune signaling. Future studies should focus on the protective efficacy of the CCS/C-HA vaccine in animal models of influenza infection, in particular in the setting of impaired immunity.

Abbreviations

APC	=	Antigen Presenting Cells
C	=	Cholesterol
CCS	=	Ceramide Carbamoyl Spermine
F-HA	=	free HA (non-liposomal associated)
HA	=	Hemagglutinin
HI	=	hemagglutination inhibition
i.m.	=	intramuscular
INFγ	=	interferon γ
IL-2/5	=	interleukin 2/5
SPF	=	specific pathogen free
KO	=	knockout
MyD88	=	myeloid differentiation factor 88
TLR	=	Toll Like Receptor
TNFα	=	tumor necrosis factor α.

Disclosure of potential conflicts of interest

All authors declare that they have no competing interests to disclose.

Acknowledgments

This work was supported by the Israel Science Foundation (grants 1396/12 and 1391/17), and by the Barenholz Research Fund (the Barenholz Research Fund was established by Professor Barenholz from royalties obtained from various Barenholz licensed patents, and these funds are dedicated to support research in the Barenholz Lab). OE was supported by a Hebrew University scholarship for outstanding students.