Synthesis and Structure Optimization of Star Copolymers as Tunable Macromolecular Carriers for Minimal Immunogen Vaccine Delivery

Minimal immunogen vaccines are being developed to focus antibody responses against otherwise challenging targets, including human immunodeficiency virus (HIV), but multimerization of the minimal peptide immunogen on a carrier platform is required for activity. Star copolymers comprising multiple hydrophilic polymer chains (“arms”) radiating from a central dendrimer unit (“core”) were recently reported to be an effective platform for arraying minimal immunogens for inducing antibody responses in mice and primates. However, the impact of different parameters of the star copolymer (e.g., minimal immunogen density and hydrodynamic size) on antibody responses and the optimal synthetic route for controlling those parameters remains to be fully explored. We synthesized a library of star copolymers composed of poly[N-(2-hydroxypropyl)methacrylamide] hydrophilic arms extending from poly(amidoamine) dendrimer cores with the aim of identifying the optimal composition for use as minimal immunogen vaccines. Our results show that the length of the polymer arms has a crucial impact on the star copolymer hydrodynamic size and is precisely tunable over a range of 20–50 nm diameter, while the dendrimer generation affects the maximum number of arms (and therefore minimal immunogens) that can be attached to the surface of the dendrimer. In addition, high-resolution images of selected star copolymer taken by a custom-modified environmental scanning electron microscope enabled the acquisition of high-resolution images, providing new insights into the star copolymer structure. Finally, in vivo studies assessing a star copolymer vaccine comprising an HIV minimal immunogen showed the criticality of polymer arm length in promoting antibody responses and highlighting the importance of composition tunability to yield the desired biological effect.


■ INTRODUCTION
−3 While this strategy has been highly effective in preventing many diseases, full-length proteins have two key limitations.First, use of full-length proteins that are typically composed of hundreds of amino acids can lead to antibody responses against highly variable sites of the pathogen that distracts antibody responses away from conserved sites needed for efficacy. 4−7 An emerging strategy to overcome these limitations is to use only an active part of the full-length protein, referred to as a minimal immunogen, which allows antibody responses to be focused against key sites while avoiding those that could lead to ADE. 8,9 However, minimal immunogens, in isolation, are poorly immunogenic. 10This is due to their relatively low molecular weight (<10 kg•mol −1 ) resulting in rapid clearance and poor drainage to lymph nodes (LNs) 11 and the inability of soluble, minimal immunogens to sufficiently cross-link B cell receptors needed for antibody induction. 12−15 Some of the most advanced minimal immunogen vaccines in development have relied on recombinant nanoparticles that fuse the minimal immunogen to a structural protein that selfassembles into a multimer displaying multiple copies of the minimal immunogen. 16Approaches under development include those using bacteriophages, 17 keyhole limpet hemocyanin (KLH, a naturally occurring multimeric protein nanoparticle), 18 ferritin nanoparticles, and various virus-like particles. 19,20While these recombinant approaches have shown promise, certain minimal immunogens contain post-translation modifications (PTMs), including glycosylation sites key for antibody recognition that cannot be readily encoded using recombinant approaches; 21 additionally, the number and density of immunogens that can affect antibody responses cannot be readily controlled. 22,23−26 As an alternative, fully synthetic star copolymer-based vaccines comprising minimal immunogens linked to hydrophilic polymer arms radiating from dendrimer cores provide the potential advantages that minimal immunogen density, orientation, and scaffold size can be precisely tuned to modulate antibody responses.−32 More recently, star copolymer-based minimal immunogen vaccines were shown to be safe for administration to mice and primates for inducing neutralizing antibody responses; however, the impact of various structural parameters (minimal immunogen density and hydrodynamic diameter) on antibody responses and the synthetic approach to optimize such parameters needs further investigation. 33ndeed, ligand density, linker length, and hydrodynamic size have been shown to impact antibody responses with other platforms. 12−14 Furthermore, optimizing these parameters for glycosylated minimal immunogens, which are notoriously poorly immunogenic may be even more critical to ensuring effective antibody responses. 34herefore, the objectives of this work were to investigate how various parameters of star copolymer vaccines (polymer arm length, minimal peptide immunogen density, etc.) affect star copolymer physicochemical properties (e.g., hydrodynamic size) and capacity for inducing antibody responses against a glycosylated minimal immunogen ("Man9 V3") derived from human immunodeficiency virus (HIV). 35A custom-modified environmental scanning electron microscope (ESEM) equipped with a scanning transmission electron microscopy (STEM) detector was used to provide unique ultrahigh-resolution images of selected star copolymers that document their size and shape in the unsolvated state.After optimizing synthetic conditions for controlling polymer arm loading and star copolymer vaccine size, we examined how minimal immunogen density and hydrodynamic size impact antibody responses.The results show how the synthetic approach can be used to control the minimal peptide immunogen density and hydrodynamic size of star copolymer vaccines for promoting antibody responses.

Synthesis and Characterization of Star Copolymers.
The preparation of the star copolymer was performed in two synthetic steps, including the synthesis of heterobifunctional polymer arms, followed by conjugation of the arms to the dendrimer cores (for the reaction scheme, see Scheme 1).
The polymer arms P1−P5 were synthesized by the reversible addition−fragmentation chain-transfer (RAFT) polymerization of HPMA in the presence of different molar amounts of the thiazolidine-2-thione (TT)-(for polymers P1− P3) or azide group (N 3 )-functionalized (for polymers P4 and P5) chain-transfer agent and initiator.The polymers of number-average molecular weights (M n ) of ∼10−80 kg• mol −1 were distinguished by a narrow molecular weight distribution (Đ ≤ 1.1, Đ = M w /M n ) and high functionalities (f ≥ 0.8, f is defined as the number of functional end groups per polymer chain calculated as the ratio between M n obtained from the size exclusion chromatography (SEC) measurement and M n calculated from the end group analysis) of either TT or azide end groups.The dithiobenzoate (DTB) groups of the heterobifunctional polymers were subsequently replaced with propargyl (for polymers P1−P3) or TT groups (for polymers P4 and P5) by the homolytic reaction with a high molar excess of the functionalized azoinitiators to introduce the reactive moieties to the ends of the polymer arms for the biorthogonal attachment of the peptide vaccine.No significant changes in M n and Đ nor a decrease in the reactive groups at opposite ends of the chains were observed in this substitution reaction.The characteristics for heterobifunctional polymers P1−P5 are summarized in Table 1, and their SEC profiles are depicted in Figure S2.
The heterobifunctional polymer arms P1−P3 with TT and propargyl end groups were then grafted onto PAMAM dendrimers of different generations, varying in molecular weight and number of surface ∼NH 2 groups (for PAMAM dendrimer characteristics, see Table S1), using three different molar ratios of the PAMAM surface ∼NH 2 groups to the TT  terminal group on the polymer arms (1:1, 2:1, and 3:1).The conjugation reaction was carried out in dry methanol to prevent competitive hydrolysis of the TT groups and hence lower conjugation recovery.Based on the reaction conditions and the composition of the reactants, the resulting product was a mixture of star copolymers of various physicochemical parameters (S1−S35) and unreacted heterobifunctional polymer.Although the resulting product consisted of a binary mixture of two polymers, suitably selected chromatographic conditions in combination with light scattering (LS) and refractive index (RI) detectors (for details, see Materials and Methods section) allowed precise characterization of both types of materials without the need for purification.Only the star copolymer S17, which was further used for covalent coupling with the electron microscopy (EM) contrast label and star copolymers S28−S31, which were further used for minimal peptide immunogen binding, were purified from unreacted linear polymer using centrifugal filter units (for SEC chromatograms of the star copolymer S17 before and after the purification, see Figure S1).All star copolymers, irrespective of the method of preparation, exhibited a narrow distribution of molecular weights (Đ ≤ 1.35), suggesting the preparation of defined materials.The influence of polymer arm length, dendrimer generation, and reactant functional group ratio on the size, molecular weight, morphology, and yield of the star copolymer was investigated in detail (for star copolymer characteristics, see Table 2).
Polymer Arm Length Is a Key Determinant of Star Copolymer Size.The size of the carrier used to deliver minimal immunogens has been shown to be a key factor for influencing immune responses.Therefore, we sought to investigate whether polymer arm length could be used to modulate the star copolymer hydrodynamic size.
Our results showed that the length of the polymer arm was found to play a key role in designing star copolymers of a certain size and the number of grafts that can be linked to the dendrimer core.In general, the gyration radius (R g ) and the number-average molecular weight (M n ) of the star copolymers increase with an increasing M n of the polymer arm using a dendrimer of the same generation and the same molar ratio of reactants (see Table 2 and Figure 1).For example, the star copolymer S8 consisting of the arms of polymer P1 (9.8 kg• mol −1 ) has R g and M n values of 10.2 nm and 332.3 kg•mol −1 , respectively; R g and M n of the star copolymer S17 with the P2 arms (15.8 kg•mol −1 ) are 14.5 nm and 501.7 kg•mol −1 ; and the star copolymer S26 formed by the P3 arms (41.1 kg•mol −1 ) has R g and M n of 19.0 nm and 1045.0 kg•mol −1 .These data show that the star copolymer size can be precisely tuned by controlling the polymer arm M n .
Perhaps not unexpectedly, the number of arms that can be attached to the dendrimer core (N) decreases with increasing polymer arm M n (see Table 2 and Figure 1).Accordingly, for star copolymers S4, S13, and S22 comprising polymer arms P1, P2, and P3 with increasing molecular weight, the number of polymer arms attached (N) is 30, 28, and 21.As size and minimal immunogen density can impact biological activity, these competing factors must therefore be balanced in the design of star copolymers for use as minimal vaccines.
Dendrimer Core Generation Impacts Arm Density but not Star Copolymer Size.The dendrimer core generation (G3−G5) was another parameter assessed for its impact on size, molecular weight, and number of polymer arms attached.Our results showed that the R g values of star copolymers were almost independent of the dendrimer generation, while the number of attached polymer arms and hence the M n of the star copolymers increased linearly (see Table 2 and Figure 1).For example, attaching polymer arms P1 to PAMAM, G3 resulted in a 9.6 nm star copolymer S1 with 24 arms, whereas the star copolymer S4 composed of PAMAM, G4 had a R g of 9.6 nm and 30 arms, and the star copolymer S7 measured 10.2 nm and contained 38 arms.These results suggest that the PAMAM dendrimer generation (i.e., degree of dendrimer branching) affects the density of the arms that can be attached, but minimally impact the size of the star copolymer, and that increasing the dendrimer generation can be used as a means for accommodating higher densities of polymers chains.Given the higher density of arm loading that can be achieved with higher generations of PAMAM dendrimers, G5 PAMAM dendrimers were selected as the preferred core for further evaluation as vaccines later.Though, as the studies herein only evaluated G3-G5 dendrimers, further assessment may be warranted to understand if the observed trends apply to lower and higher dendrimer generations (<G3, > G5).

Impact of Polymer to Dendrimer Molar Ratio on Star
Copolymer Size and Yield.We also assessed whether the synthetic scheme had an impact on the star copolymer properties.Our results showed that star copolymer size and yield are impacted by the reactant molar ratio.The data showed that both the N and the M n of the star copolymers decreased linearly with increasing molar ratio of PAMAM ̃NH 2 groups to PHPMA ̃TT groups, regardless of the polymer arm length and PAMAM generation used (see Table 2 and Figure 1).For example, star copolymers S13, S14, and S15 were synthesized by mixing PAMAM, G4 with polymer arm P2 at functional group ratios (n( ̃NH 2 )/n( ̃TT)) of 1:1, 2:1, and 3:1, respectively.The resulting star copolymer S13 had an M n of 452.6 kg•mol −1 and 28 polymer arms; the copolymer S14 had 344.5 kg•mol −1 and 21 polymer arms; and the copolymer S15 had 251.0 kg•mol −1 and 15 polymer arms.This suggests that at lower ratios (n( ̃NH 2 )/n( ̃TT)) when there are fewer amines available on the dendrimer core, there is steric hindrance that results in competition for binding, which is also consistent observed yields of 43, 67, and 69% for S13, S14, and S15, respectively.Notably, increasing the n( ̃NH 2 )/n( ̃TT) ratio has minimal to no impact on the size of the resulting star copolymers, with S13, S14, and S15, all having R g values between ∼12 and 13 nm.This can be partly attributed to the low dependence of the star copolymer size on the number of polymer arms and partly to the different sensitivity of the characterization technique (static light scattering) to changes in the monitored quantities (while the weight-average molecular weight (M w ), according to the Rayleigh's approximation for the spherical particle, increases linearly with the scattering intensity (I LS ), the diameter is proportional to the sixth root of the I LS ).Based on these data, the reaction scheme using a 2:1 ratio of core and polymer chain functional groups, n( ̃NH 2 )/n( ̃TT), was selected as the preferred route as it affords star copolymers with high polymer arm density at an acceptable yield.
Ultrahigh-Resolution Imaging of Star Copolymers.To our knowledge, the size and shape of star copolymer vaccines have not been assessed in detail.Therefore, ultrahighresolution images of selected star copolymers were provided to document their size and shape.For these purposes, the purified star copolymer S17 was chosen as a representative carrier with satisfactory size (R g = 29 nm) and a high number of polymer arms (N = 30).To increase the signal-to-noise ratio and improve the contrast of the image from the STEM detector, the copolymer was labeled with Au-based nanoparticles (mono-sulfo-NHS-undecagold, Au 11 ).In this work, the Au 11 nanoparticles were attached to either (i) the surface of the PAMAM core, (ii) the end groups of the PHPMA arms, or (iii) both the PAMAM core and the ends of the PHPMA arms.We found that the best imaging was provided by a copolymer that was labeled only at the ends of the polymer arms, with the contrast of the micrographs also being affected by the number of attached Au 11 nanoparticles.From the group of star copolymers decorated with approximately 30, 15, and 5 contrast agent molecules, the one with the smallest number of Au 11 nanoparticles (5) proved to be the most suitable for STEM analysis.Several examples of micrographs of star copolymer S17 containing ∼5 Au 11 nanoparticles at the end of its PHPMA arms are shown in Figure 2.
The observed local decrease in the intensity of the detected signal (or the contrast, respectively) in the analyzed sample (Figure 2A−E) can be attributed to the absence of Au 11 nanoparticles on PHPMA arms (only ∼17% of the arms are modified with Au 11 ) or the spatial arrangement of individual PHPMA arms (some ends of the arms may be hidden inside the polymer coils).In Figure 2 D (white arrow), it can be seen − due to in situ freeze-drying sample preparation − that there is local change of shape or length of the polymer arms.In contrast, a sample with an intact structure and a homogeneous distribution of arms is shown in Figure 2E.The more contrasting area around the star copolymer core (compared to the outer edges) may indicate that the dense dendrimer core (∼6 nm in diameter for G5) is surrounded by expanded chains of hydrophilic polymer arms (Figure 2B,C).Although the obtained micrographs of star copolymers are two-dimensional projections into the X and Y planes, based on changes in intensity in image contrast, star copolymers can be described as spatially symmetric structures of approximately a spherical shape.The most frequently observed star copolymers of a spherical shape with a typical radius of about 7.5 nm are depicted in Figure 2A−D.The difference between the radius of the samples measured by EM and dynamic light scattering (DLS) (R h ∼ 15 nm) is about 50%.This can be attributed to the different measurement conditions of the two methods; while in the case of electron microscopy, the sample is measured in a completely dried state (using the freeze-drying method), with DLS, the sample is analyzed in a fully solvated state (aqueous solution), which causes the expansion (swelling) of the arms.In addition, the DLS provides a zaverage size that is strongly influenced by the presence (even of low populations) of large particles, while the resulting particle size from the EM measurement is calculated by arithmetic averaging, giving the number-average size.Spatially asymmetric formations (Figure 2E) were observed very rarely.
Since the resolution required for convincing imaging of samples is at the limit of the possibilities of the microscope used, the obtained micrographs had to be modified by postprocessing methods.Brightness and contrast were optimized, in particular.The effect of micrograph modification can be seen in Figure 2A, where the background of the micrograph is the original raw micrograph, whereas the postprocessed micrograph is depicted in a white frame with bar.During the detailed analysis of the sample, it was possible to find also an area in which a cluster of free (unbound or released) Au 11 nanoparticles can be seen (Figure 2 F).The measured size of Au nanoparticles in the cluster (0.6−0.8 nm) corresponds to the size of the undecagold (Au 11 ) contrast agent, which is reported by the manufacturer (Nanoprobes Inc.).Star copolymer micrographs confirm a notably uneven surface structure (Figure 2A−E), which is consistent with the intended application of these materials as carriers of minimal peptide immunogens mimicking viruses.
Synthesis and Characterization of Star Copolymer Vaccines.After having optimized a synthetic route for modulating polymer arm density and particle size, we next sought to investigate how these parameters impact biological activity of a star copolymer vaccine delivering a representative minimal immunogen consisting of a glycopeptide derived from the V3 loop of HIV Envelope protein, referred to as Man 9 V3.
Star copolymer vaccines V1−V4 were synthesized by reacting purified star copolymers S27−S30 having different lengths and densities of azide-terminated polymer arms with the dibenzocyclooctyne (DBCO)-functionalized V3 glycopeptide (Man 9 V3-DBCO) via a strain-promoted cycloaddition reaction (see Scheme 2).The coupling of (macro)molecules mediated by the interaction between azides and DBCO groups was chosen because of their biorthogonality, fast reaction kinetics, and lack of catalyst and byproducts simplifying purification.The conjugation of Man 9 V3-DBCO to the polymer platform was manifested by an increase in the M n , from which the number of bound glycopeptide units was calculated.The degree of substitution of the polymer arms by glycopeptide units ranged from ∼66 to 100%, with higher conjugation efficiencies achieved for star copolymers with longer polymer arms.Binding of the glycopeptide also led to an increase in the hydrodynamic radii (R h ), although it should be noted that the radii of gyration (R g ) of the star copolymer vaccines were similar to those of the unmodified star copolymers within experimental measurement error, which is likely due to the lower sensitivity of static light scattering used to determine R g .Changes in the slope of the Zimm plot, from which R g is calculated, are very small for particles <30 nm, while changes in the diffusive motion of the particles, from which R h is calculated, are more sensitive to be reliably detected.The Đ values of the star copolymer vaccines were very low (Đ < 1.3), indicating that conjugation was not accompanied by cross-linking or other undesirable side reactions (see Table 3).
Ultrahigh-Resolution Imaging of Star Copolymer Vaccines.To document the size and shape of the star copolymer vaccines not only in solution but also in the dry state, we subjected them to detailed EM analysis.However, while in the case of unmodified star copolymers, it was possible to decorate their surfaces with gold nanoparticles (Au 11 ) to enhance the image contrast, this strategy could not be implemented in the case of vaccines because the end groups of the polymer arms were occupied by minimal peptide immunogen molecules.For this reason, the direct visualization of unlabeled star copolymer vaccines by EM was a major challenge.However, as can be seen from the acquired micrographs of a representative star copolymer vaccine with ∼70 kg•mol −1 polymer arms (Figure 3), by using a suitable method and optimizing the measurement conditions, it was possible to reliably capture the analyzed subjects.The analyzed vaccine nanoparticles had a spherical shape with a radius of ∼25 nm, which is in close agreement with the values obtained from the DLS measurement.
Impact of Star Copolymer Arm Density and Molecular Weight on Vaccine Activity.We next assessed whether modular properties of the star copolymer could be tuned to impact immune responses when used as a vaccine delivering the minimal immunogen, Man 9 V3.Star copolymers with two different arm lengths (M n target ∼10 or 80 kg•mol −1 ) and number of polymer arms (N ∼ 10 or 30) were prepared to generate star copolymer vaccines with varying size and minimal immunogen density as summarized in Table 3 and then assessed for the capacity to induce antibody responses in mice following vaccination.
Whereas vaccination with the soluble monomeric Man 9 V3 ("Free Man 9 V3") did not induce antibodies above background levels in nai ̈ve, untreated animals, all the star copolymer compositions induced detectable antibody responses (Figure 4).Furthermore, there was a clear trend between increasing arm M n and increased antibody responses with the star copolymers with 80 kg•mol −1 arms (V2 and V4) inducing   antibody titers that were greater than 10-fold higher than the star copolymers with 10 kg•mol −1 arms (V1 and V3) and nearly 100-fold and statistically significantly higher than background levels observed in nai ̈ve animals.No differences in antibody responses were observed with star copolymers having 10 or 30 arms, which suggests that the star copolymer size may be more critical than the polymer arm density (number of polymer arms) in promoting antibody responses.
Of note, the responses observed at 2 weeks after the final vaccination (day 70), corresponding to the peak of the antibody response, are comparable to the levels observed against Man 9 V3 in other studies that were considered biologically meaningful. 33,36hile these preliminary data highlight some potential advantages of the star copolymer as a tunable platform for delivering minimal immunogens to modulate antibody responses, further immunological assessment characterizing how such parameters impact pharmacokinetics and biodistribution as well as uptake and processing by innate immune cell populations to influence antibody and T cell responses will be needed.−39 Therefore, future studies assessing how star copolymer arm length (and hydrodynamic size) impacts distribution to and uptake by immune cells in lymph nodes may be helpful for understanding the mechanistic basis for the results presented herein.
Star Copolymer Vaccines Have Excellent Recovery Following Sterile Filtration.Sterile filtration to ensure product sterility is a key step in vaccine manufacturing.A potential advantage of star copolymers is that their size can be tuned and tightly controlled between about 10 and 50 nm diameter, which is well below most sterile filter pore sizes of about 0.2 μm.To assess the capacity of the star copolymer vaccines to undergo sterile filtration, a representative star-Man 9 V3 vaccine with 30 fully conjugated ∼10 kg•mol −1 polymer arms was filtered through 0.2 μm pore filters comprising either polytetrafluoroethylene (PTFE), nylon, polyethlysulfone (PES), or cellulose acetate (CA) and then assessed for material properties and recovery by SEC and highperformance liquid chromatography (HPLC).Notably, the star copolymer vaccine showed excellent recovery with no impact on particle size and molecular weight following filtration from PTFE, nylon, and PES, whereas no star copolymer vaccine was recoverable following filtration through CA (see Table S2).These data show that star copolymer vaccines can be efficiently and nondestructively filtered through conventional filtration membranes (except CA), thereby ensuring sterility required for injectables intended for human use.

Stability of Star Copolymers and Star Copolymer
Vaccines in Aqueous Buffer.For commercial utility, star copolymer vaccines should have sufficient stability in aqueous buffers during manufacturing, storage, and solutions prior to administration in the clinical or hospital setting.First, we assessed the stability of a representative star copolymer with ∼10 PHPMA arms of M n ∼ 30 kg•mol −1 placed at 37 °C in a temperature-controlled environment and then sampled at up to 16 weeks to assess for degradation by DLS and SEC.Notably, the star copolymer showed excellent stability at 37 °C; 10% was degraded through 2 weeks, which increased to 15% degraded by 4 weeks and more than 40% degraded by 16 weeks (Table S3A and Figure S4).The gradual degradation of the star copolymer was probably due to the retro-Michael reaction occurring in the PAMAM dendrimers, which is especially favored in aqueous solutions at elevated temperatures.As a consequence, a branching methacrylamide chain (conjugated with PHPMA arm) is cleaved, and a defective dendrimer with one missing branch is formed. 40We further  evaluated the stability of the representative star copolymer vaccine comprising Man 9 V3 glycopeptide (∼10 units) stored under various conditions, including 25 °C for 5 days and 4 °C and −20 °C for up to 16 weeks either as a lyophilized solid or in PBS solution.The results of the DLS and SEC analyses showed that the values of all monitored quantities (R h , M w , and Đ) remained within the experimental error practically unchanged (Table S3B and Figure S5).Overall, these data clearly demonstrate that the star copolymers based on PAMAM and PHPMA as well as their conjugates with Man 9 V3 have excellent stability under manufacturing, handling, and storage conditions, which bodes well for their potential use in commercial applications.

■ CONCLUSIONS
The presented article discusses the synthesis and optimization of the structure of star copolymers as tunable nanoparticle carriers of minimal immunogens for use as vaccines.Specifically, we studied how the length of PHPMA arms, dendrimer generation, and molar ratio of arms to dendrimer functional groups impact star copolymer size, arm density, and yield.In addition, ultrahigh-resolution images of a selected star copolymer were provided to document its size and shape in the unsolvated state.Three important conclusions can be drawn from the measured data: (i) the polymer arms with a higher M n form larger-sized star copolymers, but they attach to the surface of the dendrimer to a lower extent than the polymer arms with lower M n ; (ii) the branching degree (generation) of the dendrimer can control the density of the arms extending from the dendrimer core, but not the size of the star copolymer; and (iii) the use of a higher molar ratio of PAMAM ̃NH 2 groups to PHPMA ̃TT groups leads to a lower polymer arm density but higher star copolymer yield, regardless of the polymer arm length and PAMAM generation used.Finally, with an eye to clinical translation, we investigated how the modular parameters of star copolymers impact their activity for use as minimal immunogen vaccines using a glycopeptide derived from the envelope protein of the HIV-1 virus as a representative minimal immunogen.The results showed that all star copolymer vaccine compositions elicited a detectable antibody response, with the effect being more pronounced for the larger star copolymers with longer arms, while the number of polymer arms had little effect on the antibody titer.Considering the long-term storage stability as refrigerated (4 °C) or frozen (−20 °C) solutions or even better as lyophilized solids and excellent yields after sterile filtration, star copolymer vaccines represent a very promising platform for minimal immunogen vaccines.Synthesis of Monomer, Functionalized Chain Transfer Agent, and Initiators.N-(2-Hydroxypropyl)methacrylamide (HPMA) was synthesized by reacting methacryloyl chloride with (RS)-1-aminopropan-2-ol in dichloromethane in the presence of sodium carbonate as described in reference. 41ithiobenzoic acid 1-cyano-1-methyl-4-oxo-4-(2-thioxothiazolidin-3-yl)butyl ester (CPP-TT) was prepared by the reaction of CPP with TT in dichloromethane in the presence of DCC and DMAP. 42ithiobenzoic acid 3-(3-azidopropylcarbamoyl)-1-cyano-1methylpropyl ester (CPP-N 3 ) was produced by the reaction of CPP with 3-azidopropylamine in ethyl acetate in the presence of EDC. 43-[1-Cyano-1-methyl-4-oxo-4-(2-thioxo-thiazolidin-3-yl)-butylazo]-2-methyl-5-oxo-5-(2-thioxothiazolidin-3-yl)-pentanenitrile (ACVA-(TT) 2 ) was prepared by the reaction of ACVA with TT in tetrahydrofuran in the presence of DCC and DMAP.44 4-Cyano-4-(1-cyano-3-ethynylcarbamoyl-1-methylpropylazo)-N-ethynyl-4-methylbutyramide (ACVA-(Pg) 2 ) was synthesized by reacting ACVA with propargylamine in dichloromethane in the presence of EDC and DMAP.43 N-(3-Azidopropyl)-4-[3-(3-azidopropylcarbamoyl)-1-cyano-1-methylpropylazo]-4-cyano-4-methylbutyramide (ACVA-(N 3 ) 2 ) was synthesized by reacting ACVA with 3-azidopropylamine in dichloromethane in the presence of EDC and DMAP.45 Synthesis of Functionalized Immunostimulant and Minimal Peptide Immunogen.1-(4-(Aminomethyl)benzyl)-2-butyl-1H-imidazo[4,5-c]quinolin-4-amine (2Bxy) was prepared by a multistep reaction starting from quinoline-2,4-diol as previously described.11 Man 9 V3, a glycosylated HIV V3 minimal peptide immunogen of the composition EINCTRPNNNTRPGEIIGDIR-QAHCNISRA with a C-terminal PEG 3 -DBCO linker, was synthesized following a procedure similar to that described in ref 21, except that the C-terminus of the glycopeptide was linked to a Boc-amine-PEG 3 linker instead of biotin as shown in Scheme S1.Briefly, Man 9 GlcNAc 2 −NH 2 (5) was originally synthesized using complex glycosyl coupling chemistry while N-terminal fragment (4) and Boc-amine-PEG 3 -linked Cterminal fragment (1) were acquired through solid-phase peptide synthesis.Man 9 GlcNAc 2 glycosyl amine (5) was attached to the free carboxylic acid side chain at position 301 on the N-terminal fragment (4) and at position 332 on the Boc-amine-PEG 3 linked C-terminal fragment (1) to provide glycopeptide thioester (7) and N-terminal cysteinyl glycopeptide (8), respectively.These two fragments were then coupled using native chemical ligation, immediately followed by cyclization via disulfide formation to yield the Man 9 V3-PEG 3 -amine.DBCO-NHS ester was then introduced to react with the amine group, affording the final product Man 9 V3-DBCO (11).

Synthesis of Heterobifunctional Polymer Arms.
Polymer arms P1−P3 were synthesized by RAFT polymerization of HPMA in a tert-butyl alcohol/DMSO mixture in the presence of various amounts of CPP-TT and ACVA-(TT) 2 (see Table 1).Polymer arms P4 and P5 were prepared under the same conditions but in the presence of CPP-N 3 and ACVA-(N 3 ) 2 .In the second step, the dithiobenzoate (DTB) end groups of the polymers were capped by the homolytic reaction with an excess of functionalized initiators, ACVA-(Pg) 2 in the case of polymer arms P1−P3, and ACVA-(TT) 2 in the case of polymer arms P4 and P5, to form propargyl (Pg) or carbonylthiazolidine-2-thione (TT) groups, respectively.Below is given a typical procedure for the synthesis of heterobifunctional polymer P1.
A mixture of CPP-TT (39.4 mg, 103.0 μmol) and ACVA-(TT) 2 (24.9 mg, 51.5 μmol) was dissolved in 1.6 mL of DMSO and added to a solution of HPMA (2.0 g, 14.0 mmol) in 14.1 mL of tert-butanol.The reaction mixture was thoroughly bubbled with argon and polymerized in a sealed glass ampule at 70 °C for 16 h.Then, the mixture was precipitated into 300 mL of acetone/diethyl ether (3:1), and the precipitate formed was filtered, redissolved in methanol, and precipitated again into the same precipitant.After drying under vacuum, 828 mg (41%) of the product was obtained as an orange, amorphous powder.The number-average molecular weight (M n ) and the dispersity (Đ) of the polymer precursor were 9.2 kg•mol −1 and 1.03, respectively.
A mixture of polymer precursor (200.0 mg, 21.7 μmol of DTB gr.) and ACVA-(Pg) 2 (231.1 mg, 652.2 μmol) was dissolved in DMSO (2 mL), thoroughly bubbled with argon, and incubated at 80 °C for 3 h.Then, the mixture was precipitated into 40 mL of acetone/diethyl ether (3:1), the precipitate formed was filtered, redissolved in methanol, and precipitated again into the same precipitant.After drying under vacuum, 154 mg (77%) of the P1 polymer was obtained as a pale-yellow amorphous powder.The M n and Đ of the P1 polymer were 9.8 kg•mol −1 and 1.03, respectively.The molar content of the thiazolidine-2-thione (TT) end group of the polymer was 98.9 μmol/g, corresponding to an average functionality of 0.97.
Synthesis of Star Copolymers.Star copolymers were synthesized by acylation of the PAMAM dendrimer of various generations (G3, G4, and G5) with a heterobifunctional polymer arm of different molecular weights (10, 16, 41, and 96 kDa) at different molar ratios of the PAMAM amino groups to the TT terminal group on the polymer arms (1:1, 2:1, and 3:1).Below is given a typical procedure for the synthesis of star copolymer S1.
Methanolic solution of PAMAM, G3 (1.1 μL of 20 wt %, 30.3 nmol of ̃NH 2 gr.) was added to a solution of linear polymer P1 (9.8 mg, 0.97 μmol of TT gr.) in 100 μL of methanol, and the reaction mixture was shaken at 25 °C for 16 h.Then, the mixture was precipitated into 2.0 mL of diethyl ether; the precipitate formed was collected by centrifugation, redissolved in water, and freeze-dried to give 9.5 mg of a mixture of star copolymer S1 and linear polymer P1.The M n , Đ, radius of gyration (R g ) and weight fraction of star copolymer S1 in the isolated product were 246.6 kg•mol −1 , 1.12, 9.6 nm, and 48.7%, respectively.
A selected star copolymer S17 was further purified from the unreacted linear polymers by membrane filtration using RC centrifugal filter units with molecular weight cutoff (MWCO) 100 kg•mol −1 (Sigma-Aldrich, Czech Republic) in PBS (4×) and in H 2 O (2×) and isolated by lyophilization.For SEC chromatograms of star copolymer S17 before and after the purification, see Figure S1.
Labeling of Star Copolymers with the Contrast Agent.The purified star copolymer S17 was labeled by reaction with an Au-based contrast agent (Au 11 ).The Au 11 nanoparticles were covalently linked to either (i) the surface of the PAMAM core, (ii) the end groups of the PHPMA arms, or (iii) both the PAMAM core and the ends of the PHPMA arms.In addition, in the case of approach (i), different numbers (30,  15, and 5) of Au 11 nanoparticles were attached to the ends of the polymer arms.An example of the preparation of a star copolymer labeled with approximately 5 Au 11 nanoparticles attached to the ends of PHPMA arms is described below: The purified star copolymer S17 (171.8 mg, 47 μmol of −NH 2 gr.) was dissolved in 3.436 mL of N,N'-dimethylacetamide (DMAc) and mixed with 45 μL of acetic anhydride (470 μmol), and the solution was shaken at 25 °C for 2 h.The reaction mixture was diluted with methanol (1:1) and separated on a column filled with Sephadex LH-20 (Sigma-Aldrich, Czech Republic) in methanol.The polymer fraction was precipitated into diethyl ether yielding 165 mg of the star copolymer with acetylated amino groups on the surface of PAMAM.
Afterward, the copolymer (155.0 mg, 11 μmol of Pg gr.) was dissolved in 3.2 mL of DMAc, 3-azidopropylamine (2.2 μL, 22 μmol), TBTA (5.9 mg, 11 μmol), and CuBr (1.6 mg, 11 μmol) were added, and the reaction mixture was thoroughly bubbled with Ar.After 24 h, the reaction mixture was diluted by the addition of 8-hydroxyquinoline and separated on a Sephadex LH-20 column in methanol.The polymer fraction was precipitated into diethyl ether, yielding 147 mg of the star copolymer with the primary amino groups at the ends of the PHPMA arms.
Finally, an aqueous solution of the star copolymer (0.64 mg, 46 nmol of ̃NH 2 gr.) was mixed with the Au 11 nanoparticles (8 nmol) dissolved in 1 mL of 10% isopropanol in water and the reaction mixture was shaken overnight at 25 °C.The resulting product was purified on a PD-10 column (Sigma-Aldrich, Czech Republic) in H 2 O and lyophilized to give 0.58 mg of the star copolymer with approximately 5 Au 11 molecules attached to the ends of the PHPMA arms.
Synthesis of Star Copolymer Vaccines.Star copolymer vaccines comprising a G5 PAMAM dendrimer core with different lengths and densities of PHPMA polymer arms were prepared as described above, except the polymer arms (P4 and P5) were terminated with an azide group used to link the HIV minimal peptide immunogen, Man 9 V3 bearing a dibenzylcyclooctyl (DBCO) group, to the star copolymer via strainpromoted azide−alkyne cycloaddition.For example, star copolymer vaccine V3 was prepared by mixing 1.09 mg of star copolymer S29 (0.116 μmol of ̃N3 group) dissolved in 1.1 μL of DMSO/DMF (1/1 v/v) and 0.91 mg of Man 9 V3-DBCO (0.119 μmol) in 18.3 μL of DMSO/DMF (1/1 v/v).The reaction of star copolymer with a slight, 3 mol % excess of Man 9 V3 (or 1:1.03 molar ratio) was allowed to proceed at 25 °C overnight.An aliquot was then injected onto the HPLC instrument, and conversion was evaluated by comparing the areas under the curve of unreacted Man 9 V3-DBCO before and after the reaction.If free Man 9 V3-DBCO (>3%) was present in the reaction mixture, additional star copolymer was added to consume the remaining glycopeptide until its content was below 3%.The star copolymer vaccine was lyophilized, Bioconjugate Chemistry redissolved in 1× PBS, and stored at −20 °C before use for animal studies.The M n and Đ of star copolymer vaccine V3 were 135.9 kg•mol −1 and 1.18, respectively.
Size-Exclusion Chromatography.The number-and weight-average molecular weights (M n and M w ), dispersities (Đ), and gyration radii (R g ) of all linear polymer arms as well as the star copolymers were determined by size-exclusion chromatography (SEC) on an HPLC system (Shimadzu Corp., Japan) equipped with internal UV−VIS diode array detector (SPD-M20A) and external differential refractometer (Optilab T-rEX) and multiangle light scattering detector (DAWN HELEOS II, both Wyatt Technology Corp., CA, USA).TSKgel SuperAW3000 and SuperAW4000 columns (Tosoh Bioscience, PA, USA) in series were used to analyze samples in a mobile phase of 80% methanol/20% sodium acetate buffer (0.3 M, pH 6.5) at a flow rate of 0.6 mL•min −1 .The dn/dc values of 0.168 and 0.176 mL•g −1 were used to calculate the molecular weights of linear/star copolymers and star copolymer vaccines, respectively.
High-Performance Liquid Chromatography.The purity of low-molecular-weight compounds, including monomer, chain transfer agents, initiators, and immunostimulant, was verified on a high-performance liquid chromatography (HPLC) system (Shimadzu, Japan) equipped with an internal UV−vis diode array (SPD-M20A) and ELSD (LTII) detectors using the Chromolith HighResolution RP-18e reverse-phase column (Merck, USA), with a linear gradient (0−100%) of water−acetonitrile mixture containing 0.1% TFA at a flow rate of 2.5 mL•min −1 .The course of conjugation of Man 9 V3-DBCO glycopeptide to the star copolymers was monitored by an HPLC system (Agilent 1260 Infinity II, CA, USA) equipped with an internal UV−vis diode array detector using an Agilent InfinityLab Poroshell 120 EC-C18 reverse-phase column, with a linear gradient (5−65%) of water−acetonitrile mixture containing 0.05% TFA at a flow rate of 1.25 mL•min −1 .
UV−vis Spectrophotometry.The spectrophotometric analyses of functionalized linear polymers were performed in quartz glass cuvettes on a Specord Plus UV−vis spectrophotometer (Analytik Jena, Jena, Germany).The molar content of the terminal DTB and TT groups in the polymers was determined at 302 and 305 nm in methanol using the molar absorption coefficient of 12,100 and 10,300 L•mol −1 •cm −1 , respectively.
Dynamic Light Scattering.The hydrodynamic sizes of purified star copolymers were determined by the dynamic light scattering (DLS) technique at a scattering angle of 173°using a Nano-ZS instrument (Malvern Instruments, UK) equipped with a 4 mW, 633 nm laser.Measurements were performed at a sample concentration of 1.0 mg•mL −1 in PBS buffers (0.15 M, pH 7.4) at 37 °C.For the evaluation of the dynamic light scattering data, the DTS (Nano) program was used.The resulting hydrodynamic sizes were arithmetic means of at least 10 independent measurements.
Electron Microscopy.Ultrahigh-resolution electron microscopy imaging of Au-labeled star copolymers and unlabeled star copolymer vaccines was performed on a custom-modified environmental scanning electron microscope (ESEM) Quanta 650 FEG (Thermo Fisher Scientific, MA, USA) 46 equipped with a detector for scanning transmission electron microscopy (STEM).The lyophilized samples were dissolved in distilled water and applied to a lacey carbon film on a copper grid.The solution of star copolymers in water (2 μL) was applied on a TEM grid covered with holey carbon film. 47,48Then, the samples were in situ freeze-dried at 20 °C and 10 Pa in the ESEM specimen chamber 49 (operated under environmental mode).Observation was performed at a beam energy of 30 keV, beam current of 5 pA, and working distance of 5.3 mm in high vacuum mode using a dark-field STEM detector.Micrographs were postprocessed using MountainsSEM software (Digital Surf, France).
In Vivo Vaccination.Animals were housed and cared for in accordance with the American Association for Accreditation of Laboratory Animal Care standards in accredited facilities at the Vaccine Research Center, and all animal procedures were performed according to a protocol approved by the Institutional Animal Care and Use Committees of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.
Female Balb/c mice, 8−12 weeks of age, were obtained from The Jackson Laboratory (Bar Harbor, ME, USA) and maintained at the Vaccine Research Center's (VRC) Animal Care Facility (Bethesda, MD, USA) under pathogen-free conditions.Star copolymers for vaccination were prepared as 25 μg of Man 9 V3 minimal peptide immunogen equivalent star copolymer in PBS buffer with adjuvant comprising 5 μg PADRE and 5 nmol of the TLR-7/8 agonist, 2BXy, as previously described. 33Immunizations were given intramuscularly at days 0, 28, and 56 and blood was sampled at day 70 to isolate serum for assessment of antibody responses.
Antibody Measurements.Pierce Streptavidin Coated Plates (Thermo Scientific) were coated overnight at 4 °C with a Man 9 V3-biotin probe in PBS buffer.Plates were then blocked with PBS + FCS, and serum was applied in serial 10-fold dilutions and incubated at 37 °C.Detection was performed at room temperature with total antimouse IgG HRP-conjugated secondary antibodies (1:6000, Southern Biotech) followed by TMB + substrate−chromogen (Dako) and a 2N sulfuric acid stop solution.Washing was performed between steps with PBS + 0.05% Tween 20.Plates were read on a spectrophotometer (and data were analyzed in Prism (GraphPad).End point titers were determined by fitting data using a four-parameter dose− response curve.

Data Availability Statement
The data underlying this study are available in the published article and its Supporting Information.
Reaction scheme for the preparation of Man 9 V3-DBCO (Scheme S1), characteristics of PAMAM dendrimers (Table S1), SEC chromatograms of unpurified and purified star copolymer S17 (Figure S1), SEC chromatograms of heterobifunctional polymer arms P1−P3 (Figure S2), R h -distribution function of purified star copolymer S17 (Figure S3), size parameters and mass recoveries of star copolymer vaccine solutions before and after filtration (Table S2), size parameters of star copolymers and star copolymer vaccines stored under different conditions (Table S3), SEC chromatograms of star copolymers stored under different conditions (Figure S4), and SEC chromatograms of star copolymer vaccines stored under different conditions (Figure S5) (PDF)

Figure 1 .
Figure 1.Influence of PAMAM dendrimer generation and molar ratio of the PAMAM surface.̃NH 2 groups to the TT terminal group on the polymer arms on the gyration radius (R g ) and number (#) of polymer arms of star copolymers.

Figure 2 .
Figure 2. Examples of ultrahigh-resolution images of star copolymers S17: (A−E) STEM analysis of star copolymers.Indication of the inner arrangement is viewed.The differences between the images are due to the different distribution of Au 11 nanoparticles in the sample.(F) Cluster of 8 free Au-based nanoparticles, each with a diameter of 0.8 nm, corresponding to the Au 11 .

Scheme 2 .
Scheme 2. Cartoon Depiction (A) and Reaction Scheme (B) for the Preparation of Star Copolymer Vaccines (II) Synthesized by Conjugation of Man 9 V3 Glycopeptide to Star Copolymers (I)

Figure 3 .
Figure 3. Ultrahigh-resolution micrographs of unlabeled star copolymer vaccine construct with ∼70 kg•mol −1 polymer arms obtained by ESEM.The overall image at the top left shows the arrangement of star copolymer vaccine of similar shapes and sizes, while the remaining images show a zoomed-in view of individual ∼50 nm star copolymer vaccines emphasizing their spherical shape and entangled structure of the polymer coils.

Figure 4 .
Figure 4. Antibody responses following vaccination with different star copolymer vaccine compositions, free (unformulated Man 9 V3) or nai ̈ve control.Balb/c mice (n = 5/group) were immunized at days 0, 28, and 56 and serum was collected on day 70 and assessed for anti-Man 9 V3 IgG antibodies by enzyme-linked immunosorbent assay (ELISA).Data are presented as the mean ± standard deviation end point titer.Differences between each experimental group and the nai ̈ve control were assessed for statistical significance using one-way ANOVA with Bonferroni correction for multiple comparisons; asterisks (*) indicate statistical significance (P < 0.05) between the indicated group and nai ̈ve.

Table 1 .
Characteristics of the Heterobifunctional PHPMA Polymer Arms with TT Groups at One Side and Propargyl (P1−P3) or Azide Groups (P4 and P5) on the Other Side of the Polymer Chain

Table 2 .
Characteristics of Star Copolymers S1−S26 Composed of Polymer Arms P1−P3 Extending from PAMAM Dendrimers G3−G5, Prepared at Three Different Molar Ratios (1:1, 2:1, and 3:1) of the PAMAM Surface ÑH 2 Groups to the ̃TT Terminal Group on the Polymer Arms Molar ratio of PAMAM∼NH 2 groups to PHPMA∼TT groups in the reaction mixture.ii Number-average molecular weight of the star copolymer determined by SEC.iii Star copolymer dispersity defined as the ratio of weight-average (M w ) to number-average (M n ) molecular weight determined by SEC.iv Number of polymer arms attached to the PAMAM dendrimer core evaluated by SEC.
i v Yield of the conjugation reaction evaluated by SEC.vi Radius of gyration of the star copolymer determined by SEC.

Table 3 .
Characteristics of Star Copolymers S27−S30 (A) and Their Conjugates with the Man 9 V3 Glycopeptide and the Star Copolymer Vaccines V1−V4 (B) Number-average molecular weight of the star copolymer/star copolymer vaccine determined by SEC.ii Star copolymer/star copolymer vaccine dispersity defined as the ratio of weight-average (M w ) to number-average (M n ) molecular weight determined by SEC.iii Radius of gyration of the star copolymer/star copolymer vaccine determined by SEC.iv Hydrodynamic radius of the star copolymer/star copolymer vaccine determined by DLS.v Number of polymer arms attached to the PAMAM dendrimer core evaluated by SEC.vi Number of minimal peptide immunogen units attached to the PHPMA arms of the star copolymer evaluated by SEC. i