Synthesis of an Artificial MUC1 Glycopeptide and Disulfide-Specific Bioconjugation to the Carrier Protein CRM₁₉₇

We started by designing and synthesizing glycopeptide 1’ (Figure S1), which closely resembles the previously utilized derivative 1 used in the development of a nano-vaccine.¹² Compound 1 was developed using a structure-guided design approach with the aim to improve the binding affinity toward anti-MUC1 antibodies. This features ensures that the antibodies generated in mice using a non-natural MUC1 derivative can effectively target naturally occurring TA-MUC1 proteins present on the surface of cancer cells.¹² The use of non-natural amino acids in 1 also has the potential to increase its immunogenicity, making this non-natural MUC1 derivative more readily recognized and responsive by the immune system than its natural counterpart.¹² The new glycopeptide 1’ used in the current study differs from 1 by the absence of the cysteine-histidine (Cys-His) sequence in the N-terminal region. Instead, it features a cyclooctyne handle for disulfide-specific conjugation with the protein carrier CRM₁₉₇.

Concerning the protein carrier, treatment of the protein CRM₁₉₇ with an excess of tris(2-carboxyethyl)phosphine (TCEP) at 25 °C for 2 h selectively reduced C186–C201 bond (Figure 2a).¹⁸ Subsequent reaction with 20 equiv. of 2 (see Supporting Information for details) in phosphate-buffered saline (PBS) buffer (pH 7.4) at 21 °C for 5 days gave the desired azide-tagged protein upon purification using a Zeba spin column (CRM₁₉₇-linker-azide). Liquid chromatography–mass spectrometry (LC–MS) analysis revealed a single peak at 58705 Da, corresponding to the modification of one disulfide and subsequent re-bridging to introduce the azido-tag (Figure 2b, see also the obtained SDS-PAGE in Figure S2). Circular dichroism (CD) confirmed that the installation of 2 retained the secondary structure of native CRM₁₉₇ (Figure 2c).^{18, 19} The conjugate CRM₁₉₇-linker-azide was subsequently treated with glycopeptide 1’ (1 : 20 protein/glycopeptide ratio) at 25 °C for 20 h (Figure 2d). The reaction mixture was then purified on a Zeba spin column and analyzed by LC–MS confirming successful conjugation and formation of homogeneous CRM₁₉₇-linker-1’ (Figure 2e). Molecular dynamics (MD) simulations performed with this conjugate showed that the addition of the glycopeptide has no impact on the 3D structure of the protein (Figure 2f and Figure S3). These calculations suggested that the linker between the carrier and the glycopeptide is flexible, which is reflected in optimal exposure of the glycopeptide (antigen) to the solvent. This exposure is essential for the efficient antigen presentation to the immune system. It is worth noting that although only one copy of glycopeptide 1’ was presented by the protein, up to 4 possible isomers could be expected (Z/E-oxime and 1,4/1,5-triazole adducts). Importantly, the conjugation procedure was amenable to scale up and enabled production of homogeneous vaccine CRM₁₉₇-linker-1’ for immunization studies in mice. Therefore, we successfully produced a chemically defined homogeneous vaccine by site-specific conjugation of a single copy of an artificial antigen to the CRM₁₉₇ that preserved the 3D structure of the protein and is ready for mouse immunization studies.

Specific Antibodies Against the Unnatural MUC1 are Elicited by CRM₁₉₇-Linker-1’

Once verified that the endotoxin levels in the samples were safe for mouse administration [below 1 endotoxin units per milliliter (EU/mL), Figures S4 and S5],²¹ we tested the immunogenic potential of CRM₁₉₇-linker-1’ in mice. For this purpose, a group of five MUC1 transgenic (MUC1.Tg) mice was immunized with an initial subcutaneous (SC) dose followed by three equal booster doses of the vaccine candidate with 21-day intervals, while one control group was treated with unconjugated CRM₁₉₇ and another with only PBS. Serum from the immunized mice was collected 5 days after each injection to examine the levels of anti-MUC1 antibodies (Figure 3a). For this, an ELISA assay was performed loading the plates with glycopeptide 1’ (Figure 1). As shown in Figure 3b, anti-MUC1 antibodies were detected after the 2^nd injection. Antibody titers were higher after both the 3^rd and 4^th immunizations, which suggests a potential boost effect. Importantly, the antibodies elicited by CRM₁₉₇-linker-1’ in mice were mainly IgG antibodies (Figure 3b, c), while the levels of IgM antibodies were significantly lower (Figure 4d). This result is indicative of a more mature response that involved class-switch recombination, as the immunoglobulin heavy chain class switching is known to occur rapidly after activation of mature naive B cells, resulting in a switch from expressing IgM to IgG. Thus, having higher levels of IgG in the sera indicate that activation of mature B cells have occurred.^{7b-7d, 7f, 22} The IgG antibodies were further examined with an additional ELISA assay to determine their isotypes (Figure 3c). While all IgG isotypes were detected, the triggered antibodies were mostly of the IgG2 type, which suggests a triggered Th1 response. These findings differed from our previous studies in which a MUC1 glycopeptide with the same non-natural substitutions (glycopeptide 1, Figure 1) but displayed heterogeneously on the surface of a gold nanoparticle, where we observed elevated levels of IgG1 typical of a Th2 response.¹² In mice, high levels of IgG1 are associated with a Th2 response, a humoral immune response type that promotes B cell proliferation and induction of antibodies production. Conversely, high levels of IgG2 are related to a Th1 response, a cellular mediated immune response that stimulates B cells to produce IgM and IgG1 antibodies and promotes the activation of CD8⁺ CTLs.²³ Moreover, antibodies from the IgG2a and IgG2b subclasses are known to activate the complement. Thus, Th1-associated antigens are considered good candidates for vaccine development. In addition, we also detected antibodies against CRM_197, which indicated that CRM₁₉₇ is promoting an immune response as predicted, leading to a stronger effect (Figures S6, S7 and S8).

When comparing total IgG titers of the vaccine candidate CRM₁₉₇-linker-1’ with those of the previously reported formulation, where unnatural glycopeptide 1 (Figure 1) was conjugated to gold nanoparticles (AuNP-1) in a heterogeneous formulation,¹² and using the same immunization protocol, the homogeneous vaccine described here induces statistically higher total IgG titers. This outcome is observed despite differences in vaccine design and the distinct mouse models used for immunization (MUC1.Tg for CRM₁₉₇-linker-1’ and Balb/c for AuNP-1) (Figure S9).

Next, we checked whether the generated antibodies were able to recognize human TA-MUC1, which is overexpressed in cancer cells. Serum from the immunized mice was incubated with T47D cell line, which displays TA-MUC1 on its surface²⁴ (used as a positive control) and with HEK293T²⁵ (as a negative control), and a binding analysis was performed using flow cytometry. This experiment showed that the antibodies generated by our vaccine in mice were able to bind to T47D cells. In contrast, negligible binding was observed when using HEK293T cells, which is consistent with the absence of TA-MUC1 on their surface (compare Figure 3e and 3f). These results are in good agreement with those obtained from confocal microscopy (Figure 3g) that showed the presence of the MUC1 antigen on the surface of T47D cells (green color) but not on HEK293T cells. Hence, we could clearly demonstrate that CRM₁₉₇-linker-1’ vaccine successfully triggers the production of antibodies that are able to recognize the natural human TA-MUC1.

Vaccine CRM₁₉₇-Linker-1’ Vaccine Delays Tumor Growth when Administered Prophylactically

To evaluate the potential of the vaccine-induced antibodies in providing protection against cancer, we induced tumors 6 days after the last injection in immunized mice. We inoculated MC38-MUC1 cells which derive from a murine colon adenocarcinoma engineered to express human MUC1.²⁶ Tumor growth was followed for 15 days after inoculation at which time mice were sacrificed. A delay in tumor growth was observed in animals immunized with CRM₁₉₇-linker-1’ relative to controls immunized with CRM₁₉₇ or PBS (Figures 4a, b). The tumor size of the treated mice was decreased by approximately 55 % at day 15 relative to control. Of note, one mouse of the vaccine-treated group did not exhibit detectable tumor growth up to day 15 following the inoculation with cancer cells. Circulating cytokine levels were also analyzed after tumor inoculation. We observed increased levels of Th1 related cytokines, such as IL-2 and IFN-γ and TNF-α (Figure 4c), which correlates with low amounts of Th2 associated cytokines, such as IL-4, IL-10 and IL-13 (Figure S10). This data reinforces a Th1 type response to CRM₁₉₇-linker-1′ also suggested by the isotype of the IgGs detected in the sera (Figure 3c).

Next, we evaluated the efficacy of T-cells to promote CTL responses. Briefly, CD3⁺ T-cells (which include CD4⁺ and CD8⁺ T-cells) were isolated from the spleens of the mice from both the treatment group and the controls (Figures S11 and S12) and used in a CTL assay. We observed a ~20 % reduction in the viability of MC38-MUC1 cells when they were co-cultured with T-cells isolated from the spleens of treated mice compared to those obtained from the control group (Figure 4d). The experiment suggests that T-cells from the treatment group have the capacity to recognize TA-MUC1 expressed in the MC38-MUC1 cells, which triggers cancer cell death. Hence, CRM₁₉₇-linker-1’ can retard tumor proliferation, potentially by boosting the activity of cytotoxic T-cells targeting cancer cells that express human TA-MUC1, along with other contributing factors.

CRM₁₉₇-Linker-1’ Vaccine Delays Tumor Growth and Prolongs Survival when Administered Therapeutically

Next, we evaluated the efficacy of our homogeneous conjugate as potential therapeutic vaccine for reducing tumor growth. For this purpose, we administered MC38-MUC1 or Panc02-MUC1 cancer cells to untreated mice (Figures 5 and 6). As previously mentioned, the former cell line represents a murine colon adenocarcinoma that expresses TA-MUC1 on its surface. The latter cell line corresponds to a murine pancreatic cancer model that expresses the human MUC1 sequence on its surface.²⁷ Once tumors reached an average size of 100 mm³, treatments were initiated with the CRM₁₉₇-linker-1’ either as sole therapy (Figure 5) or in combination with a checkpoint inhibitor (Figure 6).

When using the CRM₁₉₇-linker-1’ vaccine alone (Figure 5), four identical SC doses as in the previous experiments were administered 2 days apart. In the MC38-MUC1 cancer model, tumor size in treated mice decreased by approximately 40 % by day 7, while in the Panc02-MUC1 model it decreased by approximately 75 % by day 25, days when the first control animals reached the endpoint (Figure 5, upper and lower panels, respectively). In terms of survival probability, all vaccine-treated mice in the MC38-MUC1 model survived beyond day 10, in contrast to the non-treated mice, whose tumors reached 1000 mm³ by day 10. The survival probability of the treated mice was about 40 % until day 15 (Figure 5, upper panel). Although tumors generally grow more slowly in the Panc02-MUC1 model, similar results were observed, with delayed tumor growth and prolonged survival in the treated animals compared to the controls. In this case, the survival probability of treated mice on day 54 was 25 % (Figure 5, lower panel). These results indicate that our homogeneous vaccine has a potential therapeutic effect on tumor growth, as evidenced by the prolonged survival time.

Immune checkpoint inhibitors employed in immune checkpoint blockade therapies have shown great results and revolutionized the field of cancer treatment.²⁸ Therefore, we chose to administer our homogeneous vaccine in combination with checkpoint inhibitors to study the combined effect of both treatments in one therapeutic formulation. Typically, antibodies targeting the PD-1/PD-L1 axis or CTLA-4 have demonstrated remarkable effectiveness in generating anti-tumor responses across a spectrum of cancers, characterized by durable outcomes and wide-ranging biological effects.²⁹ As a PD-1 inhibitor, we chose an anti PD-1 antibody. PD-1, also known as CD279, is expressed on activated T-cells and interacts with one of its two ligands, PD-L1 or PD-L2, which are often located on tumor cells. This interaction often leads to T cell dysfunction and allows the tumor to evade the immune system.³⁰ This innovative approach to immunotherapy has recently been shown to be effective in enhancing the immune response in mice.³¹ Moreover, anti-PD1 immunotherapy has shown a great success in treating cancer patients. However, it is notable that only a minority of patients, approximately 20 %, exhibit significant responses to these therapies likely due to tumor heterogeneity and others immune evasion mechanisms employed by tumors.³² Therefore, we chose to administer our homogeneous vaccine in combination with checkpoint inhibitors to study the combined effect of both treatments in one therapeutic formulation.

In this combined therapeutic approach, four identical SC doses of the vaccine were administered 2 days apart, regardless of the cancer cell line used (Figure 6), while the anti-PD1 antibody was administered via three intraperitoneal (IP) injections 3 days apart, with an IgG2a isotype antibody used as a control. The mice were closely monitored and sacrificed once tumors reached a size of 1000 mm³. Of note, this therapy resulted in a significant reduction in tumor size. In the MC38-MUC1 cancer model, there was an approximately 70 % reduction in tumor size in the treated mice group compared to the PBS-treated mice by day 10 (when the first control mouse reached the endpoint). Additionally, mice treated with CRM197-linker-1’ and anti-PD1 showed a survival rate of approximately 80 % by day 20. Remarkably, one mouse showed a complete regression of the tumor at day 65 when it was euthanized (Figure 6, upper panel). Similar results were obtained with the pancreatic cancer model. In this case, tumor growth was reduced approximately 84 % by day 25 (day of death of the first control animal) in mice treated with the combination therapy, and more than 20 % of the mice survived at day 60 (Figure 6, lower panel).

Together our data demonstrates that the homogeneous conjugate CRM197-linker-1’ is a promising vaccine candidate for cancer treatment. This candidate elicits a robust Th1 immune response and generates antibodies that selectively recognize tumor-associated antigens. Furthermore, it significantly delays tumor growth and enhances survival in mouse models of both colon adenocarcinoma and pancreatic cancer. Importantly, its efficacy is further amplified when combined with the PD-1 immune checkpoint inhibitor, suggesting a synergistic potential that leads to improved survival rates in murine models.