Cell-Penetrating Peptides-Based Liposomal Delivery System Enhanced Immunogenicity of Peptide-Based Vaccine against Group A Streptococcus

Peptide-based vaccine development represents a highly promising strategy for preventing Group A Streptococcus (GAS) infection. However, these vaccines need to be administered with the help of a delivery system and/or immune adjuvant. Cell-penetrating peptides (CPPs) have been used as a powerful tool for delivering various therapeutic agents, including peptides, as they can overcome the permeability barrier of cell membranes. Here, we used CPPs to deliver our lead lipopeptide-based vaccine (LCP-1). CPPs were anchored through a spacer to LCP-1-bearing multilamellar and unilamellar liposomes and administered to Swiss outbred mice. Tat47–57 conjugated to two palmitic acids via a (Gly)6 spacer (to form a liposome-anchoring moiety) was the most efficient system for triggering immune responses when combined with multilamellar liposomes bearing LCP-1. The immunostimulatory potential of a variety of other CPPs was examined following intranasal administration in mice. Among them, LCP-1/liposomes/Tat47–57 and LCP-1/liposomes/KALA induced the highest antibody titers. The antibodies produced showed high opsonic activity against clinically isolated GAS strains D3840 and GC2 203. The use of the CPP-liposome delivery system is a promising strategy for liposome-based GAS vaccine development.


Introduction
Peptide-based subunit vaccines exhibit an improved safety profile in comparison to conventional vaccines because they utilize only small antigens derived from a target pathogen. As such, they are free from redundant pathogen components, which reduce the possibility of inducing allergic or autoimmune responses. Moreover, peptide-based vaccines use synthetic peptides as antigens; these can be produced relatively easily and cost-effectively without any risk of biological contamination. However, peptide antigens are poorly immunogenic. Therefore, immune stimulators, namely adjuvants, are required in peptide-based vaccine formulations to improve efficacy.
Unfortunately, the rate of discovery in terms of effective and safe adjuvants is slower than that of discovering antigens. Currently, only a limited number of adjuvants are licensed for human vaccines, and those are not always effective when used with specific antigenic peptides [1,2]. Producing a suitable adjuvant remains an undeniable challenge in the development of peptide-based vaccines. Most adjuvants are designed to target pattern recognition receptors Vaccines 2021, 9,499 2 of 17 (PRRs), particularly toll-like receptors (TLRs), expressed on antigen-presenting cells (APCs) to achieve enhanced uptake of co-delivered antigen. However, other strategies to enhance antigen uptake by APCs have also been examined [3][4][5]. CPPs are a group of short peptides that have the special ability to overcome the permeability barrier of cell membranes and enter the cell interior in a non-invasive manner without assistance from membrane proteins [6]. Thus, they have been thoroughly explored for delivering various cargos, such as peptides, nucleic acids, proteins, nanoparticles, and liposomes, into cells [7][8][9]. CPPs have also been investigated in the vaccine delivery field over the past decade to facilitate antigen delivery into cells [10]. Although the internalization mechanisms of CPPs remain unclear, CPPs have already been employed to enhance both cellular immune responses via the delivery of antigen directly into the cytoplasm, and humoral immune responses, where antigens are delivered through the endocytic pathway [11][12][13].
Group A Streptococcus (GAS) is a Gram-positive human pathogen responsible for a plethora of diseases, ranging from non-invasive (e.g., pharyngitis and impetigo) to postinfectious diseases (e.g., rheumatic fever (RF) and rheumatic heart disease (RHD)). Pharyngitis is the most common GAS infectious disease, with over 600 million people treated worldwide, annually [14]. It is estimated that up to 3% of pharyngitis patients eventually develop acute RF which can evolve into RHD [15][16][17]. In 2015, over 30 million cases of rheumatic heart disease were observed globally [15]. Both RF and RHD are autoimmune diseases trigger by untreated or frequent GAS infections. These autoimmune responses are most likely generated due to the similar structure of human and GAS proteins, which activate antibodies and/or T cell responses against human proteins [18]. There are no licensed GAS vaccines currently available on the market despite nearly a century of work in this space. The earliest attempts to develop a GAS vaccine can be tracked back to the early 20th century, when live-attenuated or inactivated whole organisms were investigated [19,20]. However, these vaccine candidates stimulated allergic and autoimmune responses and provide protection against a narrow range of GAS strains [21,22]. As GAS infection is associated with an M-protein activity, further vaccine development focused mostly on this protein. However, the strategy of utilizing the whole M protein as an antigen has been abandoned due to the risk of autoimmune response [23], which triggered RF among the vaccinated children in the clinical trial [24]. Besides, a large number of studies reported that the M protein comprises the cross-reactive B and T cell epitopes with human tissues [25][26][27], and the immunization with M protein may not produce robust immune responses against a variety of GAS serotypes. Consequently, non-cross-reactive peptide fragments of the M-protein have been used instead as a suitable alternative for GAS vaccine development [28]. For example, an M-protein-derived B-cell epitope (J8 peptide: QAEDKVKQSREAKKQVEKALKQLEDKVQ) has been used in peptide-based GAS vaccine design [29] and recently reached phase I clinical trial [30]. Furthermore, lipid-core-peptide (LCP) systems have been frequently utilized in GAS vaccine studies [31][32][33]. LCP is a selfadjuvanting carrier system consisting of lipoamino acids (LAAs), a branching moiety and conjugated peptide. LAAs can be recognized by toll-like receptor-2 (TLR-2) on dendritic cells, thus they act as self-adjuvanting moieties, and conjugation between lipids and peptides can prevent enzymatic degradation of the peptides [33].

Characterization of LCP-1 Loaded Liposomes
Liposome properties, including surface charge, particle size, size distribution, and polydispersity index (PDI), were characterized with dynamic light scattering (DLS). All measurements were performed at 25 • C with a back-scattering angle of 173 • ; eleven runs were performed per measurement and each measurement was repeated five times. The mean ± standard deviation was calculated based on the five measurements and the results were analyzed by Malvern Zetasizer software.

In-Vivo Immunization
Outbred female Swiss (CD-1) mice (7-8 weeks old) obtained from the Animal Resource Centre (Perth, Western Australia) were used for the immunization study. The mice were housed in cages under sterile conditions and allowed to acclimatize for 7 days prior to experimentation. The mice were divided into experimental groups of five per group. All immunization protocols were approved by The University of Queensland Ethics Committee (Animal Ethics Unit, Office of Research Ethics, The University of Queensland; approval number: SCMB/AIBN/069/17) and conducted in compliance with the guidelines from the Australian National Health and Medical Research Council (NHMRC).
Immunization study 1: On primary immunization (day 0), mice in the negative control group were intranasally administered with 30 µL (15 µL/nare) of PBS, while mice in the positive control group were intranasally immunized with P25-J8 (30 µg) and CTB (10 µg) dissolved in 30 µL (15 µL/nare) of endotoxin-free water. Mice in the seven test groups were given 30 µL (15 µL/nare) of freshly prepared L1-L7 solution, equating to 30 µg of LCP-1 per mouse, respectively. Boosts (two total) were performed on days 14 and 28, with the same doses. Serum was collected via tail bleed on day −1, 13, and 27 and by cardiac puncture on day 38. The clear supernatant serum was collected after centrifugation for 10 min at 956 × g (3600 rpm). Serum samples were stored at −80 • C.
Immunization study 2: Initially, intranasal immunizations were performed under anesthesia (isoflurane). However, the resulting IgG levels were highly inconsistent, even within groups, after three doses (see Supporting Information, Figure S14). Thus, a second immunization study was performed without anesthesia. On day 0, mice were intranasally immunized with freshly prepared L1, L3 and L8-L15 solutions at a dose of 30 µL (15 µL/nare) or PBS for the control group, as described above. Boosts were performed on days 21 and 42. Blood was collected via tail bleed on day −1, 20 and 41 and by cardiac puncture on day 52, and processed as detailed above to produce clear supernatant serum. Serum samples were stored at −80 • C.

Determination of Antibody Titers
Enzyme-linked immunosorbent assays (ELISA) were used to determine the presence of J8-specific antibody (IgG, IgG1 and IgG2a) titers from the collected sera. J8 peptide (0.52 µg/well) was dissolved in 0.1 M sodium carbonate/bicarbonate (pH 9.6) coating buffer. Microtiter plates were coated with J8 peptide solution (100 µL/well) for 2 h at 37 • C, then blocked with 5% skim milk overnight at 4 • C to reduce non-specific binding. Serum samples were assessed based on serial two-fold dilutions, starting at a 1:100 dilution for serum IgG. Horseradish peroxidase-conjugated secondary antibodies (IgG, IgG1, and IgG2a) were added to the microtiter plates, followed by OPD substrate. The plates were incubated for 20 min in the dark at RT, then the optical density was measured at 450 nm. The antibody titers were described as the lowest possible dilution providing an absorbance of three standard deviations (SD) above the average absorbance of the control wells (serum from naive or PBS mice). Variation between the groups was assessed for statistical significance using one-way ANOVA followed by Tukey's post hoc test with GraphPad Prism software.

Indirect Bactericidal Assay
Opsonization assays were performed as previously reported [36,37]. Bacteria were streaked onto Todd Hewitt Broth (THB) agar plates supplemented with 5% yeast extract and incubated at 37 • C for 24 h. A single bacterial colony was transferred to THB (5 mL) supplemented with 5% yeast extract and incubated for another 24 h at 37 • C to obtaiñ 10 7 colony forming units (CFU)/mL. Inactivated serum was heated in a water bath at 50 • C for 30 min. An aliquot (10 µL) of the diluted culture solution was mixed with heatinactivated serum (10 µL) and horse blood (80 µL). The bacteria were incubated with the serum on a 96-well plate at 37 • C for 3 h. An aliquot (10 µL) from the culture material was analyzed based on CFUs counted from the plates. The plates were subsequently incubated at 37 • C for 24 h, and CFUs were counted. The opsonization activity of antibody serum (percent reduction in mean CFUs) was calculated (1 − [CFU in the presence of antipeptide sera]/[mean CFU in the presence of untreated media]) × 100). Opsonization assays were completed in duplicate.

Immunization
Tat 47-57 is the most popular and intensively investigated CPP in vaccine delivery studies [10]. Thus, it was selected as a model CPP to optimize our CPP-liposome anchoring strategy. Outbred Swiss mice were intranasally administered with PBS, LCP-1 alone, CTBadjuvanted antigen (P25-J8), or L1-L7. As a mucosal adjuvant, CTB was physically mixed with antigen P25-J8 to serve as the positive control, while PBS was used as the negative control. All groups received the same concentration of antigen (30 µg of LCP-1/mouse) in water (30 µL). The presence of J8-specific IgG antibodies in the serum was determined  Table 2), LCP-1, and the positive control (CTB+P25-J8) elicited significant J8-specific antibody titers in comparison to PBS in immunized mice (Figure 2). Mice administered with L3, which carried LCP-1 and compound 2 with a short spacer (Gly) 6 , triggered the highest antibody production among the tested liposomes. L2 bearing 1 (lipidated Tat 47-57 without spacer) and L6 bearing 10 (lipidated Tat 47-57 with a long PEG 3400 spacer) both produced significantly lower antibody titers than L3. Interestingly, when the concentration of the targeting moiety on the liposome's surface was increased (L5), lower antibody titers were observed compared to those triggered by L3. Moreover, although multilamellar L3 and unilamellar L4 had the exact same composition, mice receiving L3 produced significantly higher IgG titers than mice receiving unilamellar L4. No difference in IgG expression levels was detected among mice immunized with L1 and L7 (L1 physically mixed with Tat 47-57 ). To determine which CPPs had the highest adjuvating potential, Swiss mice were intranasally immunized with CPP/LCP-1/liposome systems. The lead vaccine candidate from the anchoring strategy study (L3) was selected as a positive control. All liposome formulations tested produced higher serum IgG titers than mice immunized with the negative control (PBS) after three immunizations. Mice vaccinated with L14, bearing lipidated KALA, produced the highest J8-specific IgG titers in comparison to the other vaccinated groups (Figure 3). The L14 antibody titers were higher even than L3; however, the difference was not statistically significant.

L3
Tat47-57 (Palmitic acid)2 (Gly)6 M L4 Tat47-57 (Palmitic acid)2 (Gly)6 U L5 Tat47-57 a (Palmitic acid)2 (Gly)6 M L6 Tat47-57 Tat47 To determine which CPPs had the highest adjuvating potential, Swiss mice were intranasally immunized with CPP/LCP-1/liposome systems. The lead vaccine candidate from the anchoring strategy study (L3) was selected as a positive control. All liposome formulations tested produced higher serum IgG titers than mice immunized with the negative control (PBS) after three immunizations. Mice vaccinated with L14, bearing lipidated KALA, produced the highest J8-specific IgG titers in comparison to the other vaccinated groups (Figure 3). The L14 antibody titers were higher even than L3; however, the difference was not statistically significant.
In addition, IgG1 and IgG2a titers were measured for mice immunized with L3 and L14 by ELISA, and compared to L1 (CPP free) and PBS (Figure 4). Similar IgG1/IgG2a levels were observed in L3 and L14 immunized mice, which correlates to balanced Th2/Th1 responses, respectively. In L1 immunized mice only IgG1 were detected (Th2 responses).
In addition, IgG1 and IgG2a titers were measured for mice immunized with L3 and L14 by ELISA, and compared to L1 (CPP free) and PBS (Figure 4). Similar IgG1/IgG2a levels were observed in L3 and L14 immunized mice, which correlates to balanced Th2/Th1 responses, respectively. In L1 immunized mice only IgG1 were detected (Th2 responses).

Evaluation of an Outbred Mouse Model for Opsonic Immune Responses Against GAS
To evaluate the quality of antibodies produced, an in vitro opsonization assay was performed against two clinically isolated GAS strains: D3840 and GC2 203. Five representative sera (PBS, L3, L12, L14, L15) were selected for the study. The sera obtained from mice vaccinated with L3 and L14 showed significantly higher opsonic activity against GC2 203 strains compared to the PBS group ( Figure 5). Furthermore, serum from L3-vaccinated mice showed significant opsonic activity against D3840, and was slightly more effective than L14 serum. Other tested sera were not opsonic. by serum taken at day 52 after primary immunization in Swiss mice (n = 5 per group). Statistical analysis was performed using one-way ANOVA followed by Tukey's post hoc test to compare with PBS-administered mice (ns, p > 0.05; * p < 0.05; ** p < 0.01;). Statistical analyses were performed using GraphPad Prism version 8.03.

Evaluation of an Outbred Mouse Model for Opsonic Immune Responses Against GAS
To evaluate the quality of antibodies produced, an in vitro opsonization assay was performed against two clinically isolated GAS strains: D3840 and GC2 203. Five representative sera (PBS, L3, L12, L14, L15) were selected for the study. The sera obtained from mice vaccinated with L3 and L14 showed significantly higher opsonic activity against GC2 203 strains compared to the PBS group ( Figure 5). Furthermore, serum from L3-vaccinated mice showed significant opsonic activity against D3840, and was slightly more effective than L14 serum. Other tested sera were not opsonic. levels were observed in L3 and L14 immunized mice, which correlates to balanced Th2/Th1 responses, respectively. In L1 immunized mice only IgG1 were detected (Th2 responses).

Evaluation of an Outbred Mouse Model for Opsonic Immune Responses Against GAS
To evaluate the quality of antibodies produced, an in vitro opsonization assay was performed against two clinically isolated GAS strains: D3840 and GC2 203. Five representative sera (PBS, L3, L12, L14, L15) were selected for the study. The sera obtained from mice vaccinated with L3 and L14 showed significantly higher opsonic activity against GC2 203 strains compared to the PBS group ( Figure 5). Furthermore, serum from L3-vaccinated mice showed significant opsonic activity against D3840, and was slightly more effective than L14 serum. Other tested sera were not opsonic. by serum taken at day 52 after primary immunization in Swiss mice (n = 5 per group). Statistical analysis was performed using one-way ANOVA followed by Tukey's post hoc test to compare with PBS-administered mice (ns, p > 0.05; * p < 0.05; ** p < 0.01;). Statistical analyses were performed using GraphPad Prism version 8.03. Figure 5. Average opsonization percentage of Group A Streptococcus strains (D3840 and GC2 203) by serum taken at day 52 after primary immunization in Swiss mice (n = 5 per group). Statistical analysis was performed using one-way ANOVA followed by Tukey's post hoc test to compare with PBS-administered mice (ns, p > 0.05; * p < 0.05; ** p < 0.01;). Statistical analyses were performed using GraphPad Prism version 8.03.

Discussion
GAS pathogens infect humans primarily through the upper respiratory tract [28,43]. Thus, intranasal administration presents an attractive vaccine delivery route, as it mimics the natural route of the infection. Intranasal vaccines can also be self-administered and require no specialized equipment [44,45]. Nasal mucosal surfaces are highly vascularized, which could facilitate rapid antigen absorption into systemic circulation. Moreover, proteolytic enzyme activity associated with nasal mucosal tissue is low, significantly reducing the risk of antigen degradation. Importantly, the tissue primarily responsible for inducing immune responses following intranasal vaccine delivery is the nasal-associated lymphoid tissue (NALT). The similarities between mouse and human NALT substantially simplifies vaccine development and progression to clinical trials.
The use of peptide-based antigens is the most promising strategy for the development of GAS vaccines, as it eliminates the risk of inducing GAS-associated autoimmune responses [1]. Indeed, all GAS vaccines that have reached clinical trials in the past decades have been peptide-based. However, peptide vaccines require the use of strong, but safe, adjuvants. Here, we propose the use of CPPs as a self-adjuvanting moiety for GAS vaccines.
Over the past decade, CPPs have been investigated for antigen delivery to cells [10]. In vaccine constructs, CPPs are usually conjugated directly to the antigen. While this ensures co-delivery, it also increases the complexity of vaccine design. Most conjugations have been achieved through construction of a recombinant plasmid that expressed both antigen and CPP genes (for protein-based vaccines) [11,[46][47][48][49][50], or a CPP/plasmid DNA complex (for DNA-based vaccines) [51,52]. On the contrary, here we designed a liposome-based delivery system that anchors both CPP and antigen/vaccine on the same liposomal nanoparticles to ensure co-delivery. To achieve this, the N-terminal of Tat 47-57 and other CPPs were lipidated. This delivery system can be easily customized by: (a) introducing/modifying linkers between the lipid/s and CPPs; (b) modifying the anchoring lipids; and (c) regulating CPP concentration on the particle surface. Particle size and properties (e.g., through change of liposome lamellarity) are additionally controllable.
Previously, we demonstrated that LCP-1 bearing two C16 lipoamino acids (2-amino-D,L-hexadecanoic acid) can be effectively anchored into liposomes [53,54]. Therefore, CPPs were modified with two C16 fatty acids (palmitic acids) to achieve the same anchoring properties. To assess the necessity of such an anchoring strategy, a physical mixture of LCP-1/liposomes and Tat 47-57 (L7) was prepared. The immune response induced by L7 was similar to that of the CPP-free liposomes (L1: LPC-1/liposomes) and, as expected, was significantly lower than that of the CPP-anchored equivalent, L3 (Figure 2). This suggests that lipidation of CPP is essential for the efficacy of liposome-based delivery systems. Additionally, lower IgG production was detected in mice treated with the positive control (physical mixture P25-J8/CTB) compared to mice receiving L3, suggesting an advantage of the CPP/liposome system over the standard adjuvating strategy.
The presentation of lipidated CPPs on the liposome surface can be regulated by the introduction of a spacer between the anchoring lipid moieties and the CPPs. Spacers with variable lengths were examined (see compounds 1, 2, 3, 10, Figure 1). Initially, simple lipidated Tat 47-57 without spacer (1), with a short spacer (2), and with a long spacer (10) were examined. Among them, vaccine candidate L3 bearing lipidated CPP 2 (with short linker) induced the highest IgG antibody titers against J8 (Figure 2). Extension of the polyGly spacer to 12 units (3) reduced the immunogenicity of the corresponding liposomes (L8) (Figure 3). Therefore, the presence of a short spacer (Gly) 6 between the CPP and liposome was found to be optimal for vaccine efficacy. Similarly, Daudey and co-workers demonstrated that placing short spacers (equivalent of 4-8 glycines) between a model peptide, K 3 (KIAALKE) 3 , and the anchoring lipid resulted in more effective interaction of K 3 with biological membranes [55]; longer spacers were less effective.
It is well-known that CPP concentration has an impact on how these peptides and their cargo interact with the cell surface [56,57]. Thus, we explored two levels of CPP concentration in our vaccine delivery system. L5 comprised a five-fold higher quantity of 2 compared to L3 and all other liposomal formulations. Interestingly, mice treated with L5 produced less IgG than mice treated with L3 (Figure 2), which indicated that increasing CPP concentration might not necessary enhance CPP-mediated uptake of LCP-1-loaded liposomes. Low CPP concentration is generally regarded as preferable, as higher concentrations can be associated with adverse effect. For example, Transportan (GWTLNSAGYLL-GKINLKALAALAKKIL) is toxic when its concentration exceeds 5 µM [58,59], and MAP (KLALKLALKALKAALKLA) has a strong toxic effect on various cell lines at concentrations over 1 µM [58,60]. Therefore, we also examined the potential toxicity of selected formulations (L1, L3, and L14); none were toxic to non-or cancerous human cell lines (see Supporting Information, Figure S14) and no noticeable side effects were observed in the vaccinated mice.
Multilamellar liposomes have been reported to be as effective as small-sized unilamellar liposomes in inducing humoral immunity [53]. We tested both to identify which structure was more efficient in improving immune responses. Unilamellar L4 bearing Tat 47-57 produced a significantly lower IgG immune titer than multilamellar L3 (Figure 2). To further confirm this observation, liposomes bearing KALA were also produced in two forms; again, multilamellar liposomes (L14) were more efficient in stimulating antibody production than unilamellar liposomes (L15) (Figure 3).
Induction of high antibody titers by a vaccine does not necessarily equate to high vaccine efficacy, as the antibodies produced may be inadequate at stopping bacterial growth. Therefore, an opsonization experiment was performed using serum from the mice immunized by the most effective formulations (L3 and L14, bearing Tat 47-57 and KALA, respectively), as well as selected low-efficacy liposomes (L12 and unilamellar L15, bearing LAH4 and KALA, respectively) and PBS. Serum from mice treated with L3 and L14 induced high opsonic IgG antibody titers, while serum from the other three formulations failed to kill GAS at the tested concentration. This dramatic difference in opsonization capacity corresponds to the differences in antibody response; for example, L14 induced 15-fold higher antibody titers than L15 (Figure 3). Given these findings, both Tat 47-57and KALA-based liposomes are promising delivery systems for intranasally administered lipopeptide-based vaccines.

Conclusions
We demonstrated that the incorporation of Tat 47-57 in an LCP-1-based liposomal delivery system enhanced J8-specific antibody production when: (a) the CPPs were anchored to liposomes; (b) an appropriate spacer was used between CPPs and the lipidic anchoring moieties; (c) CPPs were presented in relatively low concertation; and (d) multilamellar liposomes were employed. Moreover, the vaccine candidate bearing Tat 47-57 produced a significantly higher immune response compared with a physical mixture of peptide antigen and the commercial adjuvant, CTB, which suggested that lipidated Tat 45-57 could be considered as a potential alternative to classic adjuvants. We also compared a variety of CPPs for their immunostimulating/delivery potency for the first time. We found that of those tested, only Tat 47-57 and KALA, upon incorporation into liposomes, induced the production of high levels of opsonic antibodies against GAS. In addition, Tat 47-57 and KALA-based formulations induced well-balanced Th1/Th2 responses, desirable for vac-cine efficacy and safety. In summary, CPPs have the potential to improve humoral immune responses and, in combination with liposomes, provide more potent immune stimulation than the classical adjuvant CTB.