Design, synthesis and immunological evaluation of CRM₁₉₇-based immunogens functionalized with synthetic scaffolds displaying a tumor-associated MUC1 glycopeptide

Abstract

The development of effective vaccines against tumor-associated MUC1 (taMUC1) glycopeptide antigens remains a significant challenge due to their poor intrinsic immunogenicity. A key limitation in conjugate vaccine design lies in the structural alterations that occur upon carrier protein functionalization, which can reduce the accessibility of surface-conjugated antigens, ultimately compromising antigen presentation. In this study, we present a semi-synthetic vaccine platform in which taMUC1 glycopeptides are displayed on synthetic cyclopeptide scaffolds—configured either as monovalent or clustered tetravalent platforms—and subsequently grafted onto solvent-exposed amine residues of the CRM₁₉₇ protein via squaramide linkages. These conjugates were purified under denaturing conditions via reverse phase HPLC and evaluated in vivo through mouse immunization studies. Despite differences in antigen valency and glycopeptide loading per protein, both conjugates induced comparable levels of antigen-specific IgGs and CD4⁺/CD8⁺ T-cell activation when co-administered with the QS-21 adjuvant. Notably, although antibody titers were similar, post-immunization sera from mice immunized with the tetravalent conjugate plus the QS-21 adjuvant showed enhanced reactivity toward native taMUC1 expressed on MCF7 cancer cells, suggesting improved epitope recognition. These results highlight the impact of scaffold design, antigen display and adjuvantation on vaccine efficacy and establish a promising platform for the development of conjugate vaccines targeting weak tumor-associated antigens.

---

Introduction

Mucin-1 (MUC1) is a membrane-associated glycoprotein found normally in epithelial cells but aberrantly overexpressed in most human epithelial tumors.¹ These include triple-negative breast cancers (>90% occurrence),² as well as gastrointestinal, pancreas, lung, ovary, prostate, bladder, and endometrium tumors. The human MUC1 extracellular domain contains 25 to 125 tandem repeats comprising 20-amino acids and five possible O-glycosylation sites (GVTSAPDTRPAPGSTAPPAH). MUC1 is heavily glycosylated in healthy cells,¹ but malignant transformation leads to prematurely-truncated glycans, exposing otherwise cryptic epitopes such as the Tn antigen (αGalNAc-O-Ser/Thr).³ Tumor-associated (ta)MUC1 qualifies as a high-ranked antigen by the National Cancer Institute pilot project,⁴ and thus has become a key target for active immunotherapy approaches over the last decades.^5,6 Vaccination strategies focused on taMUC1 have been extensively reviewed,^6–9 and we refer readers to these overviews for comprehensive background and additional examples. In brief, these strategies include: (i) dendritic-cell (DC) vaccines,¹⁰ in which autologous DCs are generated ex vivo, loaded with antigens, matured with defined stimuli, and re administered to prime antigen-specific immunity. (ii) Genetic and vectored platforms (DNA/mRNA or viral vectors)^11–13 that drive in situ antigen expression and are often combined with immune modulators to enhance priming. (iii) Subunit vaccines, an umbrella term for formulations delivering synthetic taMUC1 epitopes, covalently conjugated with additional components required to trigger effective immune responses. These include constructs based on biologically-derived carriers^14,15 as well as inorganic nanoparticles.¹⁶ In addition, fully-synthetic, multi-component vaccines are constructs in which all key components (e.g. antigen, T-helper epitope and/or other immunogenicity enhancers) are chemically synthesized and covalently integrated into a single, precisely defined molecule.^7,17–22

While boosting antigens’ immunogenicity is a persistent challenge in vaccine design, this need is even more pronounced for weakly immunogenic motifs such as taMUC1-derived glycopeptides.²³ Semi-synthetic conjugate vaccines typically combine carrier proteins²⁴ with synthetic antigens, such as glycans, (glyco)peptides or other small molecules. Carrier proteins are pivotal for augmenting the overall immune response, particularly when paired with poorly immunogenic antigens such as taMUC1. They provide a source of CD4⁺ epitopes that can activate helper T-cells, which in turn leads to higher production of high-affinity IgG antibodies against the antigen, an aspect particularly important for molecules that do not naturally elicit T-cell help (i.e. haptens).²⁵ Additionally, grafting multiple copies of antigens onto surface-exposed regions of carrier proteins enhances antigen display in conjugate vaccines. This promotes stronger B-cell receptor binding and increased B-cell activation, ultimately resulting in more robust antibody responses.^26,27 Conjugation to carrier proteins is typically carried out via chemoselective approaches in which nucleophilic, solvent-accessible cysteine or lysine residues react with synthetic antigens incorporating complementary functionalities, or via reactive homo/hetero-bifunctional linkers.²⁸ In principle, maximizing antigen incorporation onto the carrier is desirable for enhancing the immunogenicity of the conjugate vaccine. However, this can also bring about some conformational changes in the protein structure that may reduce the accessibility of the conjugated epitopes, rendering them buried or less surface exposed and thus hindering antigen presentation.

CRM₁₉₇, a clinically approved, non-toxic mutant of diphtheria toxin, has been widely used as a carrier protein in conjugate vaccines to increase the immunogenicity of small molecule haptens or polysaccharide antigens.²⁹ Previous studies have reported conformational changes in CRM₁₉₇ upon derivatization with flexible linkers and/or conjugation with haptens.^30–32 Another work investigated the relationship between site-selective saccharide conjugation and immunogenicity of CRM₁₉₇-glycoconjugates incorporating around 1–4 copies of the Salmonella O-antigen per protein.³³ The study demonstrated that the specific site of glycan attachment on CRM₁₉₇ influences the antibody response induced by the constructs. Notably, the activity of sera from mice immunized with mono-antigen conjugates functionalized at two cysteine residues (Cys186–Cys201) via disulfide rebridging was ten times higher than that observed with the administration of a mono-conjugate at a lysine position (Lys37/39). The authors proposed that distinct conjugation sites may plausibly lead to differences in the exposure and presentation of the glycoconjugate to antigen-presenting cells. These insights have motivated subsequent, recent efforts to develop mono-antigen CRM₁₉₇ conjugates using the disulfide Cys186–Cys201 as a functionalization site,³⁴ while also retaining CRM₁₉₇ native secondary structure.³⁵

In contrast to these approaches, which prioritize precise conjugations at permissive sites with a view to maintaining CRM₁₉₇ structural integrity, our strategy exploits solvent-exposed CRM₁₉₇ amine residues to attach multiple copies of a synthetic scaffold displaying clustered taMUC1 glycopeptide epitopes. This innovative design strategy offers several advantages. First, by interposing the cyclic scaffold between the glycopeptide antigens and the carrier protein, convenient antigen separation is preserved even if partial unfolding of the carrier occurs, guaranteeing a proper antigen presentation. Second, the rigidity of the cyclic framework and its versatility for polyvalent functionalization favor an optimal antigen exposure and clustered display, thereby overcoming the accessibility limitations often associated with non-optimized conjugation sites. Thus, we applied this new conjugate platform design to the weakly immunogenic taMUC1 glycopeptide, gaining access to CRM₁₉₇-based constructs incorporating monovalent or tetravalent scaffolded antigens that were tested preclinically in vivo. When co-administered with the QS-21 adjuvant in mice, the lead constructs elicited high levels of functional IgG antibodies reactive against cancer cells expressing the native taMUC1 glycoproteins. Overall, this work establishes a generalizable conjugate vaccine approach to enhance the immunogenicity of weak antigens through the implementation of a convenient, scaffold-assisted antigen presentation strategy.

Results and discussion

Design and synthesis

The design of our prototype conjugate vaccine is based on a 10-mer cyclic peptide scaffold comprising two β-turn-inducing Gly–Pro motifs that stabilize its conformation in solution, providing a structured platform for effective antigen presentation.³⁶ This core scaffold offers a relatively rigid structure with lysine side chains extending in opposite directions to form two distinct spatial domains (upper and lower), which can be functionalized in a regioselective manner,³⁷ while also supporting a clustered antigen display (Fig. 1A).³⁸ Thus, we designed and synthesized two antigen-presenting platforms, referred to as “tetraMUC1” (1) and “monoMUC1” (2) (Fig. 1B). Both molecules incorporated a terminal mixed squaramate group (i.e. an amide-linked squaric acid monoethyl ester) in their lower domain, which was connected via a triethylene glycol-based spacer. This functional group, commonly used for conjugate vaccine synthesis,^39,40 serves as a reactive handle to attach the antigen-presenting platform to the CRM₁₉₇ carrier through its solvent-accessible amino groups via an asymmetric squaramide linkage (vide infra).

Fig. 1 (A) 3D representation of a 10-mer cyclopeptide displaying two Gly–Pro motifs as β-turn inducers (positions ①–②, and ⑥–⑦), with side chains of amino acid residues in positions ③, ⑤, ⑧, ⑩ pointing towards the “upper domain”, and ⑨, ④ pointing towards the “lower domain”. Adapted from Peluso et al.⁴¹ (Cambridge Crystallographic Data Centre “147541”). (B) Antigen-presenting platforms 1 (tetraMUC1) and 2 (monoMUC1) used in our prototype conjugate vaccine. Cyclopeptide scaffold residues marked as “X” indicate L-azidonorleucine positions that have been derivatized with glycopeptides (see Scheme 1B).

The “upper domain” of the scaffold was functionalized using click chemistry with four copies (“tetraMUC1” unit 1) or one copy (“monoMUC1” unit 2), respectively, of a 22-mer taMUC1 glycopeptide fragment (“PAHGVTSAPDTRPAPGST̲*APPA”) incorporating the Tn antigen (αGalNAc-O-Ser/Thr) on Thr18.

Antigen-presenting platforms 1 and 2 were assembled using a modular approach (Scheme 1). On one hand, glycopeptide antigen 3 was prepared by solid-phase peptide synthesis on a 2-chlorotrityl-functionalized polystyrene resin using the Fmoc/tBu strategy (Scheme 1A and Scheme S1, SI). An alkyne handle was attached at the N-terminus through a short ethylene glycol spacer for subsequent copper(I)-catalyzed azide–alkyne cycloaddition (CuAAC). Cleavage from the resin with concomitant removal of acid-labile protecting groups was followed by Zemplén deacetylation, yielding clickable glycopeptide antigen 3 in an overall 47% yield. On the other hand, cyclopeptide scaffolds displaying on the upper domain either four (4) or one (5) azidonorleucine residues (synthesized according to previously reported methods⁴²) were functionalized on their lower domain by amidation of the lysine's side chain with Fmoc-N-amido-PEG₃-acid, affording spacer-containing derivatives 6 and 7 (Scheme 1B). Following piperidine-mediated Fmoc removal, the free amine was reacted with diethyl squarate ester, providing tetra-azido and mono-azido activated intermediates 8 and 9, respectively. Finally, these azide-containing cyclic scaffolds were conjugated with glycopeptide 3via CuAAC (Schemes S6 and S7, SI), yielding tetraMUC1 and monoMUC1 molecules 1 and 2, respectively (Scheme 1B).

Scheme 1 (A) Synthesis of clickable taMUC1 glycopeptide antigen 3. (B) Modular assembly of tetraMUC1 (1) and monoMUC1 (2) antigen-presenting platforms. Reagents and conditions: (i) TFA/TIS/H₂O (98 : 1 : 1), r.t., 1 h, 55%; (ii) 30 mM NaOMe in MeOH, r.t., overnight, 86%; (iii) 1-(9H-fluoren-9-yl)-3-oxo-2,8,11,14-tetraoxa-4-azaheptadecan-17-oic acid (Fmoc-N-amido-PEG₃-acid, 1.5 equiv.), PyBOP (1.5 equiv.), DIPEA (3 equiv.), DMF, r.t., 2 h, 81% (for 6) and 83% (for 7); (iv) 20% piperidine in DMF, r.t., 1 h; then, 3,4-diethoxycyclobut-3-ene-1,2-dione (1.5 equiv.), DIPEA (2.0 equiv.), DMF, r.t., 2 h, 68% (for 8) and 71% (for 9); (v) 3 (4.4 equiv.), CuSO₄ (0.5 equiv.), THPTA (1.0 equiv.), sodium ascorbate (3 equiv.), DMF, PBS pH 7.4, r.t., 2 h, 71% (for tetraMUC1 1) and 74% (for monoMUC1 2). (C) CRM₁₉₇ conjugation of antigen-presenting units 1 and 2, and SDS-PAGE characterization. Abbreviations: TFA (trifluoroacetic acid), TIS (triisopropylsilane), r.t. (room temperature), DIPEA (N,N-diisopropylethylamine), DMF (N,N-dimethylformamide), THPTA [tris(3-hydroxypropyltriazolylmethyl)amine], PBS (phosphate buffered saline). Cyclopeptide scaffold residues marked as “X” correspond to L-azidonorleucine.

With the synthetic antigen-presenting platforms in hand, we first validated the squaramide conjugation protocol by reacting compound 1 with bovine serum albumin (BSA) as a carrier protein, obtaining BSA–tetraMUC1 conjugate (S1, Scheme S8, SI), which was used for coating the ELISA plates in post-immunization ELISA assays (vide infra). Mixed squaramates in their ethyl ester form, such as those featured in molecules 1 and 2, are versatile cross-coupling handles that can undergo condensation with primary or secondary amines in basic media to yield asymmetric squaramides.⁴³ Our initial experimental conditions involved mixing BSA (16 µM final conc.) and 1 (20 equiv.) in a 100 mM sodium carbonate buffer (pH 9.6) at 40 °C.³⁹ While target BSA-conjugate S1 could be formed and isolated, after 24 hours compound 1 was completely hydrolyzed to its corresponding squaric acid monoamide (H1, Fig. S24B, SI). To address this, we tried a 0.07 M Na₂B₄O₇/0.035 M KHCO₃ buffer solution (pH 9.5), analogously to an early report by Kamath et al.,⁴⁴ with successful results. To increase the reaction rate, the transformations were carried out at five-fold higher concentration (80 µM), resulting in significantly reduced hydrolysis (Fig. S25, SI). Following the above protocol, CRM₁₉₇ was reacted with molecules 1 or 2 for 48 hours to yield CRM–tetraMUC1 10 and CRM–monoMUC1 11 constructs, respectively (Scheme 1C). Next, we purified the conjugation reaction crudes by RP-HPLC under denaturing conditions (H₂O/CH₃CN in the presence of 0.05% trifluoroacetic acid, pH ≈ 1.3). The collected fractions were lyophilized and subsequently analyzed by gel electrophoresis (Scheme 1C and Fig. S27, SI). Coomassie-stained SDS-PAGE gels showed incomplete CRM₁₉₇ conjugation for construct 10, with a number of 0-to-5 tetraMUC1 copies attached (i.e. 0 to 20 glycopeptide units). CRM-monoMUC1 conjugate 11 appeared as a broad band spanning approximately 75–150 kDa, indicating the incorporation of 5-to-25 copies of monoMUC1 (i.e. 5 to 25 glycopeptide units).

Immunological evaluation

Next, we assessed in vivo the immunogenicity of the final CRM–tetra/monoMUC1 conjugates 10 and 11, along with their unconjugated counterparts 1 and 2 for comparison. Four groups of five C57BL/6 mice were immunized via three bi-weekly subcutaneous injections of each construct (20 µg of immunogen) in the presence of QS-21 (20 µg dose) as an adjuvant (Fig. 2A): CRM–tetraMUC1 10 (Group A), CRM–monoMUC1 11 (Group B), tetraMUC1 1 (Group C), and monoMUC1 2 (Group D). Moreover, an additional group of mice was administered CRM-tetraMUC1 conjugate 10 alone, without adjuvant (Group E). All vaccine prototypes were well tolerated under our dosing regimen: animals were monitored daily throughout the study and we did not observe overt signs of distress or illness, abnormal cage behavior, or unexpected weight loss beyond normal variation. End-point (day 42) sera were probed by coating ELISA plates with BSA–tetraMUC1 S1 (Scheme S8, SI). Total IgG analysis (Fig. S30, SI) was used to monitor the course of immunization and assess the constructs’ immunogenicity. As expected, pre-immune sera displayed baseline IgG levels (Fig. S30A, SI). By day 21, titers in Groups A, B, and E were already substantially higher than those of Groups C and D corresponding to mice injected the unconjugated compounds 1 and 2 (Fig. S30B, SI). The same trend was observed at day 42 after the third immunization (Fig. S30C, SI). Group E sera exhibited a typical dose–response curve, with optical density values beginning to decline at dilutions around 10⁴. In contrast, sera from Groups A and B (coadministered QS-21 plus CRM–tetraMUC1 10 and CRM–monoMUC1 11, respectively) displayed a pronounced hook (prozone) effect across the entire dilution range (10²–2 × 10⁵),⁴⁵ where antibody excess can interfere with antigen–antibody complex formation, leading to reduced and variable OD₄₅₀ values. To enable a more accurate analysis of IgG isotypes, the dilution ranges for Groups A, B and E were therefore adjusted, and dose–response curves were plotted starting from the dilution at which signal decline was first observed.

Fig. 2 (A) Mouse groups given each vaccine candidate formulation and immunization schedule. (B–E) ELISA dilution curves for IgG subclasses (IgG1, IgG2b, IgG2c, and IgG3) obtained from end-point sera collected at day 42. ELISA plates were coated with 0.5 μg per well of BSA–tetraMUC1 construct S1 (Scheme S8, SI). Each curve represents the mean OD (450 nm) values from five mice per groups. Dose–response curves were plotted starting from the dilution at which signal decline was first observed. Error visualization was achieved by plotting the mean as a line graph with overlaid points, where shaded ribbons represent inter-individual variability (±standard deviation). Visualizations were performed in R version 4.5.1 using RStudio version 2025.05.1 (Posit, PBC). Insets show the area under the curve (AUC, calculated with GraphPad Prism using the trapezoidal rule with baseline set to zero). (F–K) Cell-surface reactivity of antisera against taMUC1-expressing MCF7 cells by confocal microscopy: detailed regions of interest (see the SI, p. S37–S39 for experimental procedure and full images). Goat anti-mouse IgG secondary antibody Alexa Fluor™ 488 conjugate was used to reveal antibody binding (green). Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (blue).

IgG isotype analysis revealed dose–response curves for Groups A, B and E, with Groups A and B exhibiting the highest antibody responses for all IgG isotypes (Fig. 2B–E).⁴⁶ In contrast, sera of Groups C and D only showed extremely reduced IgG antibody levels, with low OD₄₅₀ values across all dilutions.

The near-background signal (even at the lowest dilution) and lack of sigmoidal decay observed for both groups’ sera indicate little to no generation of antigen-specific IgG antibodies in these cases, which suggests a markedly weaker or absent humoral response for both Groups (C and D) under these experimental conditions (Fig. 2B–E).

Notably, administration of the CRM–tetraMUC1 conjugate 10 in the absence of QS-21 (Group E) elicited a stronger antibody response compared to the admixture of tetraMUC1 1 and QS-21 (Group C), highlighting the benefit of this CRM conjugate vaccine approach. This is remarkable considering that the glycopeptide antigen dose (per construct) injected in group C was estimated to be about five-fold higher than that in group E (Table S1, SI). The modest variation of glycopeptide content per conjugate administered to group A (≈2.6 µg per injection) and group B (≈5 µg per injection) mice did not account for appreciable differences in terms of the induced total IgG and isotype IgG levels; the same observations apply to groups C and D. We note that exact glycopeptide contents per conjugate are difficult to assign for these complex, polydisperse conjugates. The loading values provided are rough, gel-based estimates intended to contextualize dosing and our conclusions rely on qualitative comparisons between formulations within single production lots. In addition, co-administration of CRM conjugates 10 or 11 with QS-21 (Groups A and B, respectively) generated detectable IgG3 responses (Fig. 2E), an isotype typically associated with carbohydrate antigens,⁴⁷ consistent with taMUC1 glycopeptide presentation.

We next assessed the functionality of the elicited antibodies by testing antisera from all five groups for binding to the native taMUC1 glycoprotein expressed on the surface of human breast cancer MCF7 cells, with mouse pre-immune sera included as negative control (Fig. 2F–K). Confocal microscopy revealed a distinctive and strong surface reactivity for antibodies induced by CRM–tetraMUC1 conjugate 10 coadministered with QS-21 (Group A) (Fig. 2G). Group A image shows a strong green fluorescence localized at the cell surface, indicative of a robust antibody binding and high antigen recognition (see also Fig. S32A, SI). In contrast, Group B showed a weaker and more diffuse signal (Fig. 2H, see also Fig. S32B, SI), with minimal surface staining, suggesting reduced antigen-binding efficiency despite comparable IgG levels to those of Group A (Fig. 2B–E). This divergence suggests that antibodies induced by CRM–monoMUC1 conjugate 11 in the presence of QS-21 (Group B) may possess lower affinity for the target taMUC1 antigen. Despite comparable IgG responses between Groups A and B (Fig. 2B–E), one plausible explanation for the stronger cell-surface recognition by Group A is that its tetravalent antigen display enhances B-cell receptor cross-linking and drives more effective affinity/avidity maturation to taMUC1 epitopes, yielding superior recognition of native taMUC1 on cells; this is consistent with valency-dependent enhancement of humoral responses and prior multivalent vaccine studies.^27,48 Groups C and D injected with tetraMUC1 1 and monoMUC1 2, respectively, displayed weak, poorly localized staining without discernible membrane pattern (Fig. 2I and J, see also Fig. S32C and D, SI), consistent with the minimal antibody levels elicited by the unconjugated compounds 1 and 2. Finally, Group E (mice immunized with CRM–tetraMUC1 10 alone) displayed localized but weaker surface fluorescence than Group A (Fig. 2K, see also Fig. S32E, SI), with a signal that appeared more diffuse than that observed for Group B (cf. Fig. S32B, SI). Collectively, these findings highlight the superior functional antibody response elicited by CRM–tetraMUC1 conjugate 10 plus QS-21 (Group A) and suggest qualitative differences among the Groups (most notably between A and B), particularly in terms of the binding properties of the antibodies generated by each construct.

In parallel, we sought to characterize ex vivo the cellular immune response induced by the constructs, with a focus on Groups A and B. At the time of sacrifice (day 42), whole splenocytes were harvested and assayed for T cell restimulation in the presence of CRM–tetraMUC1 10 (Groups A and E), CRM–monoMUC1 11 (Group B), tetraMUC1 1 (Group C) and monoMUC1 2 (Group D) (Fig. 3). After 72 h stimulation, splenocytes were analyzed for activation markers by flow cytometry, assessing the percentage of effector helper T cells⁴⁹ (CD4⁺CD44^highCD45RB^low, Fig. 3A), as well as effector cytotoxic T cells⁵⁰ (CD8⁺CD44^highCD62L^lowFig. 3B). Both Groups A and B, immunized with CRM₁₉₇ conjugates 10 and 11 formulated with QS-21, showed significantly higher frequencies of effector CD4⁺ T cells compared to Groups C and D, which received the unconjugated glycopeptide-containing scaffolds 1 and 2 plus QS-21 (Fig. 3A). The inferior response for Groups C and D could be expected and underscores the importance of covalent conjugation to a carrier system for effective T-cell engagement. While immunogens 10 and 11 feature antigen-presenting units that differ in terms of design (clustered versus monovalent) and number of copies per protein (Scheme 1C), no statistically significant difference in the frequency of effector CD4⁺ T cells was observed between Groups A and B (Fig. 3A). One possible explanation is that differences in epitope valency (clustered versus monovalent) and total antigen loading can differentially influence antigen uptake/processing and the abundance of carrier-derived MHC-II peptides ultimately presented by APCs, yielding broadly similar net CD4⁺ effector frequencies when carrier and adjuvant are held constant. Therefore, regardless of the differences in scaffold architecture and conjugation density, it can be argued that carrier protein-derived CD4⁺ T-cell epitopes may be similarly processed and presented by antigen-presenting cells. Nevertheless, only the frequency of activated memory CD4⁺ T cells from Group B mice is significantly higher than that observed for Groups C, D, and E altogether. Overall, these results suggest that distinct vaccine design approaches—whether based on scaffolded clustering (i.e.10) or increased glycopeptide antigen loading (i.e.11)—are both capable of driving robust helper T-cell responses, provided that optimal antigen conjugation platforms and adjuvantation strategies are employed.

Fig. 3 Percentage of (A) effector CD4⁺, and (B) effector CD8⁺ T cells. Data are shown as mean ± SEM (n = 5 mice per group). Statistical differences between groups were determined by one-way ANOVA followed by Tukey's post hoc test (GraphPad Prism), with P < 0.05 (*), and P < 0.01 (**) considered significant.

The frequencies of effector CD8⁺ cells among restimulated splenocytes were significantly higher in both Groups A and B (immunized with the CRM–conjugates 10 or 11 in the presence of QS-21) compared to Group E (injected with CRM–conjugate 10 without adjuvant) (Fig. 3B). Moreover, both CRM₁₉₇-linked constructs coformulated with QS-21, tetraMUC1 10 and monoMUC1 11 (Groups A and B, respectively) induced comparable CD8⁺ T-cell effector responses. This suggests that, when coadministered with the potent QS-21 adjuvant, neither the scaffold design nor the antigen loading (vide supra) critically influence the magnitude of the CD8⁺ T cell response. This pattern is consistent with QS-21 being responsible for dendritic-cell cross-presentation and CD8⁺ priming.⁵¹ Indeed, across all groups coinjected with QS-21, including CRM-conjugates 10 and 11 (Groups A and B) and unconjugated glycopeptide scaffolds tetraMUC1 1 and monoMUC1 2 (Groups C and D), CD8⁺ T-cell responses were comparable. These data further imply that conjugation with CRM₁₉₇ does not enhance CD8⁺ T-cell activation beyond what is already achieved by the taMUC1 glycopeptide scaffold-plus-QS-21 formulation. These results underscore the key immunostimulatory role of QS-21 in these conjugate vaccine prototypes (i.e.10 and 11), not only in driving humoral immune responses (Fig. 2) but also in promoting CD8⁺ T cell activation.

Conclusions

In summary, this study presents a novel design strategy for semi-synthetic conjugate vaccines targeting the tumor-associated MUC1 (taMUC1) glycopeptide antigen—a class of weakly immunogenic target that has posed significant challenges in cancer vaccine development. This new approach involved first grafting the taMUC1 glycopeptide antigens onto rigid synthetic scaffolds (that function as antigen-presenting platforms) and then anchoring these to the CRM₁₉₇ carrier protein to increase vaccine immunogenicity. With this design, we aimed to enhance antigen accessibility/presentation and immune recognition, while also minimizing the negative impact arising from the conformational changes induced in the CRM₁₉₇ structure upon conjugation—even with conjugates being purified under denaturing conditions. CRM–tetraMUC1 10 and CRM–monoMUC1 11, which differ in scaffold valency and antigen loading, elicited comparable taMUC1-specific IgG levels and CD4⁺/CD8⁺ T cell responses when formulated with the QS-21 adjuvant, highlighting the power of this semi-synthetic platform. Notably, antisera induced by conjugate 10—featuring a clustered tetravalent antigen display—demonstrated superior recognition of native taMUC1 on MCF7 cancer cells, as revealed by confocal microscopy. Overall, these findings emphasize the critical roles of both antigen architecture and adjuvantation in shaping and enhancing vaccine efficacy. In conclusion, in this study we have developed and validated an innovative, semi-synthetic conjugate vaccine approach in which rigid, synthetic scaffolds are leveraged to facilitate effective taMUC1 antigen presentation for improved immune activation in vivo; future work will translate this framework into monodisperse constructs with enhanced analytical definition. The modularity and immunogenicity of this platform makes it potentially generalizable for the development of additional cancer-related vaccine candidates, offering a promising route to advance active immunotherapy strategies targeting challenging tumor-associated antigens.

Author contributions

Conceptualization, C. P., O. R., J. A., A. F. T.; chemical synthesis, D. G., C. P.; immunological evaluation, L. A., J. A.; data analysis, C. P., D. G., L. A., O. R., J. A., A. F. T.; schemes & figures, C. P., D. G.; project coordination, C. P., O. R., J. A., A. F. T.; funding acquisition, O. R., J. A., A. F. T.; writing – original draft, C. P.; writing – review & editing, C. P., D. G., O. R., J. A., A. F. T.

Conflicts of interest

There are no conflicts to declare.

Ethical statement

All animal procedures complied with the Guidelines for the Accommodation and Care of Animals (European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes) and institutional policies. All experimental protocols were reviewed and approved by the relevant local authorities. The CIC bioGUNE animal facility is fully accredited by AAALAC International.

Data availability

Supplementary information (SI): experimental details for chemical syntheses, including HPLC, MS and ¹H-NMR analytical data, and immunological evaluation procedures (i.e. immunization protocol, antibody quantification and subtyping, antibody reactivity against human MCF7 cancer cell line by confocal microscopy, and ex vivo cellular characterization) have been reported in the SI, together with additional figures, schemes and tables. See DOI: https://doi.org/10.1039/d5bm01393c.

Acknowledgements

We thank Felix Elortza and Ibon Iloro from the CIC bioGUNE Proteomics Platform, Javier Calvo from the CIC biomaGUNE Mass Spectrometry their support with MALDI and HRMS analyses. O.R. acknowledges the Université Grenoble Alpes, the ICMG FR 2607, the French ANR projects Glyco@Alps (ANR-15-IDEX-02), Labex ARCANE and CBH-EUR-GS (ANR-17-EURE-0003), the European Research Council for the Consolidator Grant “LEGO” (647938), the Proof-of-Concept Grants “THERA-LEGO” (963862) and “PATHO-LEGO” (101100924). J.A. acknowledges funding from the Spanish Minisry of Science, Innovation and Universities (AEI), PID2021-124328OB-I00. A.F.-T. acknowledges funding from the Spanish Research Agency AEI (PID2020-117911RB-I00, MCIN/AEI/10.13039/501100011033), the Scientific Foundation of the Spanish Association Against Cancer (LAB AECC 2022-LABAE223462FERN), and the European Research Council (ERC-2021-POC-101069368).

References

1. D. W. Kufe, Nat. Rev. Cancer, 2009, 9, 874–885.
2. A. Siroy, F. W. Abdul-Karim, J. Miedler, N. Fong, P. Fu, H. Gilmore and J. Baar, Hum. Pathol., 2013, 44, 2159–2166.
3. T. Ju, V. I. Otto and R. D. Cummings, Angew. Chem., Int. Ed., 2011, 50, 1770–1791.
4. M. A. Cheever, J. P. Allison, A. S. Ferris, O. J. Finn, B. M. Hastings, T. T. Hecht, I. Mellman, S. A. Prindiville, J. L. Viner, L. M. Weiner and L. M. Matrisian, Clin. Cancer Res., 2009, 15, 5323–5337.
5. D. M. Beckwith and M. Cudic, Semin. Immunol., 2020, 47, 101389.
6. T. Gao, Q. Cen and H. Lei, Biomed. Pharmacother., 2020, 132, 110888.
7. D. M. McDonald, S. N. Byrne and R. J. Payne, Front. Chem., 2015, 3, 60.
8. L. Yin, Y. Zhou, S. Hong, F. Ding and H. Cai, ChemBioChem, 2023, 24, e202200805.
9. U. S. Grewal and R. Kurzrock, npj Precis. Oncol., 2025, 9, 218.
10. J. Wierecky, M. Mueller and P. Brossart, Cancer Immunol. Immunother., 2006, 55, 63–67.
11. L. A. Snyder, T. J. Goletz, G. R. Gunn, F. F. Shi, M. C. Harris, K. Cochlin, C. McCauley, S. G. McCarthy, P. J. Branigan and D. M. Knight, Vaccine, 2006, 24, 3340–3352.
12. L. Liu, Y. Wang, L. Miao, Q. Liu, S. Musetti, J. Li and L. Huang, Mol. Ther., 2018, 26, 45–55.
13. R. Ramlau, E. Quoix, J. Rolski, M. Pless, H. Lena, E. Lévy, M. Krzakowski, D. Hess, E. Tartour, M. P. Chenard, J. M. Limacher, N. Bizouarne, B. Acres, C. Halluard and T. Velu, J. Thorac. Oncol., 2008, 3, 735–744.
14. X. Wu, Z. Yin, C. McKay, C. Pett, J. Yu, M. Schorlemer, T. Gohl, S. Sungsuwan, S. Ramadan, C. Baniel, A. Allmon, R. Das, U. Westerlind, M. G. Finn and X. Huang, J. Am. Chem. Soc., 2018, 140, 16596–16609.
15. J. J. Du, C. W. Wang, W. B. Xu, L. Zhang, Y. K. Tang, S. H. Zhou, X. F. Gao, G. F. Yang and J. Guo, iScience, 2020, 23, 100935.
16. M. Ghirardello, A. Guerreiro, C. Bretón, I. Compañón, J. Zabalza, V. Mangini, J. M. Peregrina, G. J. L. Bernardes and F. Corzana, Chem. – Eur. J., 2025, 31, e202501286.
17. N. Stergiou, M. Urschbach, A. Gabba, E. Schmitt, H. Kunz and P. Besenius, Chem. Rec., 2021, 21, 3313–3331.
18. V. Lakshminarayanan, P. Thompson, M. A. Wolfert, T. Buskas, J. M. Bradley, L. B. Pathangey, C. S. Madsen, P. A. Cohen, S. J. Gendler and G.-J. Boons, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 261–266.
19. Z.-H. Huang, Z.-Y. Sun, Y. Gao, P.-G. Chen, Y.-F. Liu, Y.-X. Chen and Y.-M. Li, Vaccines, 2014, 2, 549–562.
20. A. Asín, F. García-Martín, J. H. Busto, A. Avenoza, J. M. Peregrina and F. Corzana, Curr. Med. Chem., 2021, 29, 1258–1270.
21. C. Pifferi, A. Ruiz-De-Angulo, D. Goyard, C. Tiertant, N. Sacristán, D. Barriales, N. Berthet, J. Anguita, O. Renaudet and A. Fernández-Tejada, Chem. Sci., 2020, 11, 4488–4498.
22. C. Pifferi, L. Aguinagalde, A. Ruiz-De-Angulo, N. Sacristán, P. T. Baschirotto, A. Poveda, J. Jiménez-Barbero, J. Anguita and A. Fernández-Tejada, Chem. Sci., 2023, 14, 3501–3513.
23. I. A. Bermejo, A. Guerreiro, A. Eguskiza, N. Martínez-Sáez, F. S. Lazaris, A. Asín, V. J. Somovilla, I. Compañón, T. K. Raju, S. Tadic, P. Garrido, J. García-Sanmartín, V. Mangini, A. S. Grosso, F. Marcelo, A. Avenoza, J. H. Busto, F. García-Martín, R. Hurtado-Guerrero, J. M. Peregrina, G. J. L. Bernardes, A. Martínez, R. Fiammengo and F. Corzana, JACS Au, 2024, 4, 150–163.
24. M. E. Pichichero, Hum. Vaccines Immunother., 2013, 9, 2505–2523.
25. F. Y. Avci, X. Li, M. Tsuji and D. L. Kasper, Nat. Med., 2011, 17, 1602–1609.
26. N. R. Bennett, D. B. Zwick, A. H. Courtney and L. L. Kiessling, ACS Chem. Biol., 2015, 10, 1817–1824.
27. Y. Kato, R. K. Abbott, B. L. Freeman, S. Haupt, B. Groschel, M. Silva, S. Menis, D. J. Irvine, W. R. Schief and S. Crotty, Immunity, 2020, 53, 548–563.
28. O. Boutureira and G. J. L. Bernardes, Chem. Rev., 2015, 115, 2174–2195.
29. E. A. Khatuntseva and N. E. Nifantiev, Int. J. Biol. Macromol., 2022, 218, 775–798.
30. D. T. Crane, B. Bolgiano and C. Jones, Eur. J. Biochem., 1997, 246, 320–327.
31. S. Pecetta, P. Lo Surdo, M. Tontini, D. Proietti, C. Zambonelli, M. J. Bottomley, M. Biagini, F. Berti, P. Costantino, M. R. Romano, F. Buricchi, D. Donnarumma and N. Norais, Vaccine, 2015, 33, 314–320.
32. J. Jaffe, K. Wucherer, J. Sperry, Q. Zou, Q. Chang, M. A. Massa, K. Bhattacharya, S. Kumar, M. Caparon, D. Stead, P. Wright, A. Dirksen and M. B. Francis, Bioconjugate Chem., 2019, 30, 47–53.
33. G. Stefanetti, Q. Y. Hu, A. Usera, Z. Robinson, M. Allan, A. Singh, H. Imase, J. Cobb, H. Zhai, D. Quinn, M. Lei, A. Saul, R. Adamo, C. A. MacLennan and F. Micoli, Angew. Chem., Int. Ed., 2015, 54, 13198–13203.
34. N. Trattnig, Z. Li, G. P. Bosman, P. Kosma and G. J. Boons, ChemBioChem, 2022, 23, e202200408.
35. A. Guerreiro, I. Compañón, F. S. Lazaris, C. Labão-Almeida, P. Oroz, M. Ghirardello, M. C. Marques, F. Corzana and G. J. L. Bernardes, Angew. Chem., Int. Ed., 2024, 63, e202411009.
36. C. Pifferi, N. Berthet and O. Renaudet, Biomater. Sci., 2017, 5, 953–965.
37. M. Mutter, P. Dumy, P. Garrouste, C. Lehmann, M. Mathieu, C. Peggion, S. Peluso, A. Razaname and G. Tuchscherer, Angew. Chem., Int. Ed. Engl., 1996, 35, 1482–1485.
38. D. Goyard, A. M. S. Ortiz, D. Boturyn and O. Renaudet, Chem. Soc. Rev., 2022, 51, 8756–8783.
39. Z. Huang, S. Chen, X. Li and H. Liu, Chem.: Methods, 2024, 4, e202300054.
40. L. A. Marchetti, L. K. Kumawat, N. Mao, J. C. Stephens and R. B. P. Elmes, Chem, 2019, 5, 1398–1485.
41. S. Peluso, T. Rückle, C. Lehmann, M. Mutter, C. Peggion and M. Crisma, ChemBioChem, 2001, 2, 432–437.
42. I. Bossu, N. Berthet, P. Dumy and O. Renaudet, J. Carbohydr. Chem., 2011, 30, 458–468.
43. R. Ian Storer, C. Aciro and L. H. Jones, Chem. Soc. Rev., 2011, 40, 2330–2346.
44. V. P. Kamath, P. Diedrich and O. Hindsgaul, Glycoconjugate J., 1996, 13, 315–319.
45. D. Slavica, Biochem. Med., 2009, 19, 50–62.
46. D. Kao, A. Lux, A. Schaffert, R. Lang, F. Altmann and F. Nimmerjahn, Eur. J. Immunol., 2017, 47, 2070–2079.
47. R. M. Perlmutter, D. Hansburg, D. E. Briles, R. A. Nicolotti and J. M. David, J. Immunol., 1978, 121, 566–572.
48. E. C. Wamhoff, L. Ronsard, J. Feldman, G. A. Knappe, B. M. Hauser, A. Romanov, J. B. Case, S. Sanapala, E. C. Lam, K. J. S. Denis, J. Boucau, A. K. Barczak, A. B. Balazs, M. S. Diamond, A. G. Schmidt, D. Lingwood and M. Bathe, Nat. Commun., 2024, 15, 795.
49. C. Bunce and E. B. Bell, J. Exp. Med., 1997, 185, 767–776.
50. Y. Nakajima, K. Chamoto, T. Oura and T. Honjo, Proc. Natl. Acad. Sci. U. S. A., 2021, 118, e2103730118.
51. M. H. Den Brok, C. Büll, M. Wassink, A. M. De Graaf, J. A. Wagenaars, M. Minderman, M. Thakur, S. Amigorena, E. O. Rijke, C. C. Schrier and G. J. Adema, Nat. Commun., 2016, 7, 13324.

---

Footnote

† Present address: Centre de Biophysique Moléculaire, CNRS UPR 4301, Rue Charles Sadron, 45071 Orléans Cedex 2, France.

---