Reduced Breast Tumor Growth after Immunization with a Tumor-Restricted MUC1 Glycopeptide Conjugated to Tetanus Toxoid

Function of the humoral immune system drives prognoses in breast cancer (1, 2). However, a substantial proportion of breast cancer patients showed an incomplete response or failed to respond to common therapies (surgery, chemotherapy, radiotherapy), reflecting the absence of a curative endogenous immune reaction (3). These resistant cancers may require additional strategies, as for example, additional tumor-specific vaccinations to become responsive to tumor-associated antigens (TAAs; ref. 4). Early breast cancer vaccine trials have proven that the nontoxic therapeutic modality is able to induce specific, long-lasting, antitumor immune responses due to the establishment of immunologic memory but showed weak therapeutic effects (5). The majority of these trials have been carried out in the metastatic setting, which could have negatively influenced the outcome because of the large tumor burden and/or the severe pretreatment with chemotherapeutics that disable the immune system (6–10). The standard treatment for early breast cancer patients remains surgery to remove the primary tumor followed by adjuvant therapies (e.g., endocrine therapy or chemotherapy; ref. 11). Thus, the highest benefit of breast cancer vaccines appears to be in preventive therapy (12, 13) in the setting of minimal residual disease in which breast cancer patients are apparently cancer-free after surgery but are in danger of relapse due to the presence of micrometastatic tumors. By treating the operated breast cancer patients with a cancer vaccine as an additional adjuvant therapy, it is expected that the vaccine will destroy the remaining primary tumor cells, thereby improving long-term survival (14). The development of effective breast cancer vaccines depends on the identification of exactly specified TAAs that function as tumor rejection targets. Many endogenous proteins induce tumor-specific T cells, but these T cells do not induce tumor regression or rejection (11). An effective therapeutic vaccine must be able to break peripheral tolerance and to activate even low-affinity T cells that were not eliminated during selection in the thymus (15). The glycoprotein mucin1 (MUC1) (16, 17) has received attention as a TAA for vaccine formulations. Because cancer patients typically develop low levels of specific antibodies against MUC1, it should be possible to amplify these baseline immune responses to a therapeutically relevant level by proper construction of the vaccines (18–20). Tumor-associated MUC1 ((TA)MUC1) shows enhanced expression in breast cancers (90%; ref. 21), especially in triple-negative breast cancers (TNBC; 94%; ref. 22). The protein backbones of (TA)MUC1 are incompletely glycosylated, resulting in exposure of peptide antigen sequences to the immune system. In addition, the altered glycosylation influences the conformation of these glycopeptide epitopes (23). Biochemical studies have shown that the extracellular domain of MUC1 contains a variable number of tandem repeats (VNTR) of a 20-amino acid peptide sequence (PAHGVTSAPDTRPAPGSTAP). MUC1 is expressed fully glycosylated in epithelial cells of healthy mammary glands. The large glycans shield the peptide backbone. Therefore, targeting the aberrantly glycosylated (TA)MUC1 for active cancer immunotherapy can exploit the difference in glycosylation of MUC1 between cancer and healthy cells. The immune reaction should eliminate cancerous cells while leaving healthy mammary cells unharmed (24, 25). Autoantibodies in the sera of patients (26) bound specific (TA)MUC1-glycopeptides that are glycosylated in the GSTA domain of the tandem repeat region (Ser17, Thr18). The recognition of individual (TA)MUC1 epitopes, derived from the self-antigen MUC1, by the immune system may be diminished due to T-cell tolerance and thymic deletion, resulting in the absence of an antitumoral response (27, 28). Nevertheless, we and other groups showed that tolerance of human MUC1-transgenic (huMUC1-tg) mice to huMUC1 was overcome by hu(TA)MUC1-glycopeptide–derived antigen vaccination by using immune-stimulating protein carriers (29–31). The conjugation to Tetanus Toxoid (TTox) as a Th cell–stimulating component led to effective (TA)MUC1-based antitumor vaccines (32, 33).

Herein, a synthetic vaccine was used for preventive vaccination in a preclinical breast cancer mouse model expressing hu(TA)MUC1. The vaccine consists of the 22mer glycopeptide PAHGVTSAPDTRPAPGSTAPPA of the huMUC1 tandem repeat region, which contains the sialylated tumor-associated carbohydrate antigen ST_N at serine-17, in conjugation with TTox (hu(TA)MUC1-glycopeptide-TTox vaccine). Previous immunization studies of wild-type (WT) mice with this vaccine induced antisera that exhibited high IgG antibody titers that specifically bound to (TA)MUC1-expressing human breast cancer tissues (33). Subsequently, monoclonal antibodies (mAb) have been generated with the aid of the above-described hu(TA)MUC1-glycopeptide-TTox vaccine. Among those, the mAb GGSK-1/30 may represent a diagnostic marker for hu(TA)MUC1-expressing cancers that bind to its specific glycopeptide epitope based on the antigen structure of the hu(TA)MUC1-glycopeptide-TTox vaccine (34). In this study, GGSK-1/30 is used for the evaluation of the hu(TA)MUC1 expression in breast cancer specimens and in a preclinical breast cancer mouse model.

A panel of 35 TNBC patients was examined for expression of (TA)MUC1 by using GGSK-1/30 as a diagnostic tool. Patients’ characteristics are given in Table 1. IHC analyses were performed on 4-μm thick formalin-fixed paraffin-embedded (FFPE) sections according to standard procedures. In brief, FFPE slides were subsequently deparaffinized using graded alcohol and xylene. Antigen retrieval reactions were performed in a steamer in citrate buffer of pH10 for 30 minutes. Three percent H₂O₂ solution was applied to block endogenous peroxidase at room temperature for 5 minutes. The samples were stained with GGSK-1/30 (1 μg/mL), followed by a polymeric biotin-free visualization system reaction (EnVision, DAKO Diagnostic Company). In a next step, the sections were incubated with 3,3-diaminobenzidine (DAB; EnVision, DAKO Diagnostic Company) for 5 minutes and counterstained with Mayer's hematoxylin solution. Human breast cancer cells (MCF-7) expressing (TA)MUC1 served as a positive control. Paraffin sections of healthy breast tissue and paraffin sections of TNBC breast cancer tumors were examined. All slides were analyzed using a Leica light microscope (Leica Microsystem Vertrieb Company) by two of the authors (A.-S. Heimes and J. Jäkel) trained in histologic and IHC diagnostics, unaware of the clinical outcome. Additionally, the magnitude of expression of (TA)MUC1 was analyzed according to the scoring system of Sinn and colleagues (35). This work was approved by the Ethics Committee of University Medical Center Mainz. All patients gave written informed consent before participating in this study.

(TA)MUC1 expression was evaluated using an immunoreactivity score (IRS) as described by Sinn and colleagues (35). In brief, the percentage of GGSK-1/30 positive tumor cells (0% = 0, 1%–10% = 1, 11%–50% = 2, 51%–80% = 3, 81%–100% = 4) and the staining intensity (negative = 0, weak = 1 moderate = 2, strong = 3) were multiplied, resulting in an IRS from 0 to 12. Cases with IRS 0–4 were considered as low MUC1 expression, whereas cases with IRS 6–12 were considered as high MUC1 expression.

C57BL/6-TG(MUC1)79.24GEND/J (35) mice (short: huMUC1-tg, The Jackson Laboratory) express the human MUC1 gene on all epithelial cells and are crossed with mice of the breast cancer model strain PyMT (ref. 36; a kind gift from the group of Prof. Wolfram Ruf, Mainz, University Medical Center). Double-transgenic female mice (PyMTxhuMUC1-tg) which are positive for the breast cancer antigen and the huMUC1 transgene develop after 14 weeks several mammary gland tumors (4 to 5 tumors per mouse) expressing hu(TA)MUC1, comparable with human breast cancer stage IV after 14 weeks. All mice used for this study were bred and housed in a specific pathogen-free colony at the animal facility of Johannes Gutenberg-University Mainz following institutionally approved protocols (permission was obtained from the Landesuntersuchungsamt Koblenz, reference number: 23 177-07/G 08-1-019).

The synthetic glycopeptide vaccine that targets a specific MUC1 glycan pattern on human breast cancer cells was generated as described before (33). The glycopeptide moiety represents a 22-mer amino acid sequence from the VNTR region of huMUC1. It includes 2 immune dominant motifs, the PDTRP and the GSTA sequences. Serine at position 17 of the GSTA motif was glycosylated with an ST_N carbohydrate antigen. The synthetic hu(TA)MUC1-glycopeptide was conjugated to Tetanus Toxoid (TTox) carrier protein, which is known to induce T helper cell–specific immune responses in humans (37). The synthetic vaccine was dissolved in phospate buffered saline (PBS) and emulsified (ratio 1:1) in Incomplete Freund's Adjuvant (Sigma-Aldrich). This emulsion (10 μg/40 μL) was injected intraperitoneal (i.p.) per mouse per immunization.

GGSK-1/30 is an IgG1 mAb raised against the specific hu(TA)MUC1-glycopeptide contained in the vaccine (34). Antibodies were isolated from the supernatant of hybridoma cells with the aid of a protein G column, concentrated using 50% ammonium sulfate precipitation, and buffer exchange to PBS was performed with a PD-10 desalting column (GE Healthcare Life Science; 17085101).

In order to evaluate whether double-transgenic PyMTxhuMUC1-tg mice can serve as a preclinical breast cancer model for preventive vaccination studies, tumor tissues of such mice were first tested concerning their expression of hu(TA)MUC1 glycoproteins. To this end, binding of fluorescence labeled GGSK-1/30 (AF647N: far-red-fluorescent dye) was tested via fluorescence microscopy. Tumors were removed ex vivo, frozen in liquid nitrogen and tissue sections were prepared. The sections were transferred to slides, stained with DAPI (1/1,000) and AF647N-labeled GGSK-1/30 (5 μg/mL), washed 3 times, and mounted with coverslips. Mammary gland sections of huMUC1-tg mice and tumor sections of PyMT-tg mice served as controls. Analyses were performed applying confocal microscope LEICA TCS SP8 with DMi8 microscope and Zeiss LSM510 META laser scanning microscope.

With the aim to monitor breast tumor growth in detail after preventive vaccination, we inoculated WT mice with tumor cells from primary tumor cell lines generated from tumors of 18-week-old PyMTxhuMUC1-tg mice. For the generation of such tumor lines, biopsies from the developing tumor tissue were digested by collagenase A (Roche, 2 mg/mL) and RQ1 DNAse (Promega, 1:2,000) and cultured in IMDM + 10% FCS + 1% glutamine + 1% sodium pyruvate for 6 weeks. In the first 2 weeks, the cells were washed every third day to remove tissue residue. After that, primary tumor cells were cultured for an additional 4 weeks to obtain the outgrowing adherent tumor cells and were passaged every time at a confluency of 70%. The expression of huMUC1 was analyzed by quantitative real-time PCR (qRT-PCR), and binding of GGSK-1/30 to hu(TA)MUC1 glycoprotein was determined by FACS analyses. Primary tumor cells of PyMT-tg mice served as control cells.

Reverse-transcribed mRNA of 2 × 10⁶ tumor cells was prepared using TRIzol reagent (Invitrogen, Life Technologies) according to the manufacturer's instructions. The RNA concentration and quality was measured by photometry (Eppendorf BioPhotometer plus). Reverse transcription was performed using the MMLV reverse transcriptase (Thermo Scientific) according to the manufacturer's instructions and qRT-PCR data were obtained with the following primers: huMUC1 for: 5′-GTGCCCCCTAGCAGTACCG-3′, rev: 5′-GACGTGCCCCTACAAGTTGG-3′, and as reference gene: EF-1a for: 5′-TGGATGCTCCAGGCCATAAGGA-3′, rev: 5′-TGCTCTCGTGTTTGTCCTCCAG-3′ by using the 5 times Hot Start Taq EVA Green (no rocks) mix (Axon) in the MyIQ iCycler (Bio-Rad) and the provided software (Bio-Rad iQ5 Standard Version 2.0) for data analysis. Human breast cancer cell line T47D (HTB-133; ATCC) and human mammary epithelial cell line (HMEC, kindly provided by Prof. Dr. Nicole Teusch, TH Cologne, Germany) served as positive controls for huMUC1 expression. B16F10 murine melanoma cell line (kindly provided by Dr. Mustafa Diken, TRON Mainz, Germany) served as negative control for huMUC1 expression. After receiving and subsequent expansion of these cell lines, aliquots were frozen, thawed for the respective experiments, and cells were kept in culture for a maximum of 2 weeks. Therefore, no phenotyping or mycoplasma assays were performed.

PyMT×huMUC1 and PyMT tumor cells (2 × 10⁵) were each incubated with 1 μg/mL GGSK-1/30 for 20 minutes at 4°C. The cells were washed 2 times with 100 μL of PBS and then incubated for 20 minutes at 4°C with a secondary antibody, goat–anti-mouse-IgG Alexa Fluor 488 (dilution 1:1,000 in PBS), and with a fixable viability dye eFluor780 (dilution 1:1,000 in PBS) to exclude false-positive dead cells. The cells were washed again 2 times with 100 μL PBS and subsequently taken up in 100 μL of PBS and analyzed on a BD Biosciences FACS LSR II machine. For each sample, 10⁴ cells were analyzed.

One group of 5 WT mice was vaccinated with the hu(TA)MUC1-glycopeptide-TTox vaccine (i.p. 10 μg/40 μL) at age of 6, 8, and 10 weeks. Five days after each vaccination, blood was collected from the tail vein and serum was prepared therefrom. The induced antibody titers and their isotypes were determined by ELISA. At age of 12 weeks, 2 weeks after the last immunization, mice received 1 × 10⁶ PyMTxhuMUC1 tumor cells subcutaneously (s.c.) in the flank. Nonimmunized mice at age of 12 weeks received the same quantity of PyMTxhuMUC1 tumor cells and served as controls. PyMTxhuMUC1 cells were harvested in 0.25% Trypsin-EDTA (Gibco), washed 2 times with PBS, and were taken up in PBS (1 × 10⁷ cells/mL). The tumor cell suspension (1 × 10⁶ cells/100 μL) was injected s.c. into the right flank of each mouse. Tumor growth was monitored every third day. Because of the oval shape of the tumors, length × width were calculated (mm²).

PyMTxhuMUC1-tg and PyMT-tg mice (the latter as control) at age 6 weeks were immunized 3 times at intervals of 2 weeks (i.p. 10 μg/40 μL). Five days after each vaccination, blood was collected from the tail vein, and serum was prepared therefrom. The antibody titers and the isotype titers of the induced antibodies were determined by ELISA. Immunized mice and untreated mice were sacrificed after 20 weeks; all tumors were isolated and quantified with a caliper. Because of the oval shape of the tumors and the wavy surface, length × width were calculated (mm²). Every PyMTxhuMUC1-tg and PyMT-tg mouse developed 4 to 5 mammary gland tumors. Tumor size is given as average of the number of tumors for each mouse. Survival of immunized PyMTxhuMUC1-tg mice in comparison with untreated PyMTxhuMUC1-tg mice was given in days after birth. The survival of the mice was monitored according to ethical guidelines.

Titers of IgG antibodies specific for hu(TA)MUC1-glycopeptides were determined by ELISA. Ninety-six-well plates (Nunc MaxiSorp flat-bottom) were incubated with 50 μL per well of hu(TA)MUC1-glycopeptide–BSA conjugate (2.5 μg/mL) in coating buffer (0.1 mol/L Na₂HPO₄ in water, pH 9.3) at 37°C. Washing steps (3 times) were processed using 100 μL blocking buffer (1% BSA, 0.2% Tween-20 in PBS). In order to saturate free binding sites, the coated plates were subsequently incubated for 20 minutes with 50 μL blocking buffer at 37°C. Antisera were diluted in blocking buffer (1:100 for the first column of the 96-well plate and then serially diluted in a ratio of 1:1), pipetted on the ELISA-plate and incubated for 45 minutes at 37°C. After 3 additional washing steps, the samples were incubated with biotinylated sheep anti-mouse IgG (0.48 μg/mL, stock solution 250 μg/mL) for 45 minutes at 37°C in 50 μL blocking buffer per well. The plates were washed 3 times and incubated for 15 minutes with streptavidin–horseradish peroxidase (1:1,000) in 50 μL blocking buffer. After washing 3 times, each well was treated with 1 mg/mL 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS), 0.01% hydrogen peroxide (1:4,000) in citrate buffer (40 mmol/L citric acid, 60 nmol/L Na₂HPO₄ × H₂O, pH 4.5). The optical density of each well was measured at 410 nm with a spectrophotometer (Tecan Reader, Genios).

For isotype determination, hu(TA)MUC1-glycopeptide–specific antisera were analyzed after the third immunization, and ELISA was carried out using the protocol described above. The following secondary antibodies were used: biotinylated anti-mouse-IgM (eBioscience, clone eB121-15F9), biotinylated anti-mouse-IgG1 (BD Pharmingen, clone A85-1), biotinylated anti-mouse-IgG2b (BD Pharmingen, clone R19-15).

The half-maximal antibody titers were calculated over the inflection point of the nonlinear regression (R² > 0.97) of the sigmoidal curve (antisera dilution against OD at 410 nm) followed by a four-parameter logistic curve analysis applying the software GraphPad Prism 6.

In order to be able to use the synthetic anti-hu(TA)MUC1-glycopeptide vaccine in the treatment of breast cancer, particularly in TNBC, the breast cancer cells must express on their surface a (TA)MUC1 that is recognized by the vaccine-induced antibodies. Using IHC, FFPE tissue of 35 patients with TNBC and 10 sections of healthy human breast tissue were stained with mAb GGSK-1/30. This mAb was generated upon immunization with the identical glycopeptide used for the development of the synthetic anti-hu(TA)MUC1 vaccine. The staining of a human breast cancer cell line with GGSK-1/30 was positive, whereas healthy breast tissue was negative in all cases (see Fig. 1A and B). In contrast, 97% of the TNBC tissue sections were positive for GGSK-1/30. Semiquantitative analysis of TNBC samples revealed that the intensity of staining varied within the cohort. Seventy-four percent of the 35 TNBC tissue samples were scored as high (TA)MUC1 expression (IRS 6-12; see Fig. 1C), 26% were scored as low (TA)MUC1 expression (IRS 0-4; see Fig. 1D). IHC confirmed the high specificity of the (TA)MUC1 glycopeptides we used as antigens for the synthetic breast cancer vaccines. To evaluate the therapeutic effect of these vaccines in vivo, a preclinical breast cancer mouse model was established.

PyMT-transgenic (PyMT-tg) mice represent a preclinical tumor model useful for development of therapeutic strategies for the treatment of human breast cancer. The morphologic similarities and the expression of biomarkers associated with poor outcome (overexpression of Her2/neu and loss of estrogen and progesterone receptors) are consistent with those in humans (50). By crossing huMUC1-tg mice, expressing huMUC1 on every epithelial cell, with PyMT-tg mice, which develop aggressive palpable mammary gland tumors (breast tumors) 14 weeks after birth, we were able to generate a preclinical breast cancer model expressing hu(TA)MUC1, like 90% of all human epithelial breast cancers. Hence, this model offers the possibility to study an active immune therapy that targets hu(TA)MUC1 to treat breast cancer, leading to the opportunity to apply the same or a similar strategy in a clinical setting. An obstacle concerning the use of altered self-antigens such as hu(TA)MUC1 as a target for vaccinations is the autoimmune response. Nonspecific binding to healthy tissue by induced serum antibodies needs to be excluded. Therefore, binding of GGSK-1/30, initially generated with the aid of the same vaccine used for the following preventive vaccination studies, was evaluated using healthy mammary gland tissue of tumor-free huMUC1-tg mice via fluorescence microscopy. Binding of GGSK-1/30 to healthy mammary gland epithelial tissue, i.e., to fully glycosylated huMUC1 (see Fig. 2A; ref. 38), or to breast tumor cells from PyMT-tg mice (see Fig. 2B) that do not express huMUC1, could not be detected. In contrast, GGSK-1/30 bound to the tumor tissue of PyMTxhuMUC1-tg mice, indicating that a tumor-specific variant of huMUC1 was expressed from breast tumor cells of such mice (see Fig. 2C).

After demonstrating expression of the hu(TA)MUC1 on breast tumor cells of PyMTxhuMUC1-tg mice, a primary tumor cell line was established to develop a transplantable breast tumor model for preventive vaccination studies. As negative control, a primary tumor cell line from PyMT tumor–bearing mice which did not express hu(TA)MUC1 was established in parallel. The huMUC1 expression of tumor cells derived from PyMTxhuMUC1 and from PyMT primary tumor cell lines was comparatively analyzed via qRT-PCR, and the surface expression of the hu(TA)MUC1-glycoprotein was determined by FACS analyses using GGSK-1/30.

Figure 3A shows the expression of huMUC1 mRNA in PyMTxhuMUC1 cells compared with the expression in human breast cancer cells (T47D; ref. 39) and in HMEC (40), the latter expressing fully glycosylated MUC1 (34). The strongest expression of huMUC1 mRNA was found in T47D cells closely followed by PyMTxhuMUC1 cells. The healthy breast epithelial cell line HMEC expressed much less huMUC1 mRNA underlining the observation that cancer cells overexpress huMUC1 (21). Tumor cells derived from PyMT-tg mice (PyMT cells) do not express huMUC1 mRNA, similar to the murine B16F10 melanoma cells that were chosen as a negative control. GGSK-1/30 bound to PyMTxhuMUC1 breast tumor cells but failed to recognize PyMT breast tumor cells, indicating an exclusive expression of hu(TA)MUC1-glycoprotein on PyMTxhuMUC1 cells (see Fig. 3B). Thus, the latter were used to inoculate localized breast tumors under the skin of preventively vaccinated mice. With this first well-controlled transplantation model, we wanted to investigate whether preventive vaccination leads to mobilization of the immune system that is strong enough to reject emerging breast tumors which express hu(TA)MUC1. To this end, WT mice were immunized with the synthetic vaccine 3 times at intervals of 2 weeks. Five days after the last immunization, blood samples were taken and antibody titers and isotypes were determined. Fourteen days after the last immunization, 1 × 10⁶ PyMTxhuMUC1 cells were transferred s.c. into the right flank of WT mice. Untreated WT mice, as controls, received the same quantity of PyMTxhuMUC1 cells. Tumor progression was monitored every third day. Because of the oval shape, the size of the tumors was measured by length × width (mm²).

WT mice that received a preventive immunization before tumor transplantation showed a significantly inhibited tumor progression (see Fig. 4A). Antibody titer analyses via ELISA revealed a specific humoral immune response against hu(TA)MUC1 in vaccinated mice (see Fig. 4B). In order to characterize this response in more detail, we performed isotype-specific ELISA and detected strong titers of IgG1 and moderate titers of IgG2b in WT mice (see Fig. 4C). The predominant IgG1 antibodies prove the induction of a specific, T helper 2 cell–mediated immune response also resulting in the induction of an immunologic memory (41). IgG2b antibodies are able to activate innate antitumor mechanisms, as for example, antibody-dependent cellular cytotoxicity (ADCC) or complement-derived cytotoxicity (CDC; refs. 42, 43). Furthermore, a strong IgG1-based humoral immune response can also activate CDC. The binding of antisera to PyMTxhuMUC1 tumor cells was tested by FACS analyses in vitro. Serum antibodies of all 5 mice exhibited strong binding to hu(TA)MUC1 on the cell surface, indicating that the antibodies generated through vaccination bind to the transplanted PyMTxhuMUC1 tumors in vivo and potentially induce ADCC and CDC that result in slower tumor progression (see Fig. 4D).

Subcutaneous transplantation of tumor cells is a simple, time-controllable, and easily replaceable tumor model. Presumably, the short-term primary hu(TA)MUC1-expressing breast cancer line and the developing breast cancer tumor in vivo exhibited similar properties. In addition, the usage of primary murine tumor cells expressing hu(TA)MUC1 rendered xenograft models such as human MCF-7 tumor cells in immune-deficient mice unnecessary (44). Nevertheless, tumor transplantation models do not completely reflect the physiologic situation in patients. Consequently, we decided to directly immunize PyMTxhuMUC1-tg mice that develop spontaneous mammary tumors comparable to human breast tumors. PyMT-tg mice at the age of 6 weeks were found to have mammary glands similar in structure to the terminal duct lobular unit in a healthy adult human breast (see as described in refs. 36, 45). PyMT-tg mice at 9 weeks start to develop epithelial proliferation that becomes malignant (36) and is morphologically similar to florid ductal epithelial breast hyperplasia in humans (46). The early malignant transition occurs in such mice between 8 and 12 weeks of age. These tumors are morphologically similar to human ductal carcinoma in situ with early stromal invasion (36). Initial vaccination at the age of 6 weeks followed by a boost 2 weeks later should induce a strong humoral anti-hu(TA)MUC1 immune response precisely at the time when a carcinoma might develop in humans if surgery had not eradicated all tumor cells. Another boost immunization after an additional 2 weeks should further promote the establishment of an immune memory helping to minimize the risk of a tumor relapse. Thus, this vaccination schedule was aimed to provoke a tumor-specific immune response just at the time of transition from the noninvasive to the invasive situation preventing tumor progression. Untreated PyMTxhuMUC1-tg mice served as controls. To prove the antigen-specific antitumor immune response through vaccination, tumor growth of immunized PyMT-tg mice was determined compared with untreated PyMT-tg mice. Mice developed spontaneous, palpable tumors after 14 weeks on 4 to 5 mammary glands. Because the growing tumors of the mammary glands were difficult to measure because they were embedded in the adipose tissue of the mammary glands, we decided to sacrifice all mice at an age of 20 weeks after birth in order to determine the size of the tumors ex vivo. Because of the oval shape of the tumors and their wavy surface, the area of each tumor was calculated (length × width). For each mouse, the average tumor size of all tumors was determined. Antigen-specific antibody titers were determined after the third and last immunization.

Figure 5A demonstrates that immunized PyMTxhuMUC1-tg mice developed significantly smaller tumors than nonimmunized PyMTxhuMUC1-tg mice. On the contrary, immunized PyMT-tg mice showed no tumor reduction in comparison with untreated PyMT-tg mice. Although all vaccinated mice revealed similar hu(TA)MUC1 glycopeptide–specific antibody titers after the third vaccination (see Fig. 5B), the induced humoral immune response showed significant tumor-reducing efficiency only in PyMTxhuMUC1-tg mice. Analyses of the isotype antibodies (see Fig. 5C) confirmed that moderate IgG2b titers were generated. Immunized PyMTxhuMUC1-tg mice compared with immunized PyMT-tg mice showed elevated IgM titers, possibly as a result of a residual T-cell tolerance in the PyMTxhuMUC1 double-tg mice. Nevertheless, an overall strong humoral immune response in the absence of autoimmune reactions was observed, confirming the results of previous vaccination studies in huMUC1-tg mice (31).

In an additional experiment, we investigated the correlation of tumor reduction through preventive vaccination to the survival time. To this end, PyMTxhuMUC1-tg mice were immunized 3 times starting at an age of 6 weeks after birth. Five days after the last immunization, blood samples were taken to determine the IgG titers and the antibody isotypes. Untreated PyMTxhuMUC1-tg mice served as control. Mice were sacrificed when the survival of the animals was severely restricted according to ethical guidelines. Figure 6A shows that preventively vaccinated, breast cancer bearing mice lived approximately 30 days longer than untreated mice. ELISA analyses of total IgG titers (see Fig. 6B) and isotype titers (see Fig. 6C) demonstrated the induction of a strong and specific humoral immune response against hu(TA)MUC1 in agreement with the results of the previous experiment (see Fig. 5). Sera from nonimmunized tumor-bearing mice did not exhibit any hu(TA)MUC1-glycopeptide antibody titers, nor binding to the MUC1-expressing T47D human breast cancer cell line or PyMTxhuMUC1 primary tumor cell line.

Collectively, our results demonstrated that the unique synthetic hu(TA)MUC1 glycopeptide that we used as an antigenic determinant mimics a tumor-specific epitope of aberrantly glycosylated human MUC1, thus enabling a differentiation between harmless self and harmful tumor tissue.

Preventive vaccination with a synthetic hu(TA)MUC1-glycopeptide-TTox vaccine inhibited progression of breast tumors in a preclinical mouse model. Although such mice transgenically expressed huMUC1, high hu(TA)MUC1-glycopeptide antibody titers were induced in the absence of any autoimmune responses. This result indicated that our hu(TA)MUC1-glycopeptide-TTox vaccine evoked a specific humoral immune response targeting selectively tumor-associated, aberrantly glycosylated huMUC1. The strong but not complete reduction of the tumor load in vaccinated PyMTxhuMUC1 mice indicated that a further improvement of the hu(TA)MUC1-glycopeptide-TTox vaccine was necessary, e.g., via coupling to dimannose in order to more efficiently target accessory cells, as we did in WT mice (47). The same holds true for combinatorial approaches, especially vaccination with checkpoint inhibitors (PD-L1, CTLA-4), which should also enhance the impact of the humoral anti-breast tumor response (48). Hence, the finding that in breast cancer patients high anti-MUC1 IgG levels were positively correlated with improved overall survival (49, 50) is an indication that an immunization strategy based on the hu(TA)MUC1-glycopeptide-TTox has the potential to inhibit breast tumor progression and metastasis.

No potential conflicts of interest were disclosed.

Conception and design: N. Stergiou, N. Gaidzik, M. Schmidt, H. Kunz, E. Schmitt

Development of methodology: N. Stergiou, N. Gaidzik, H. Kunz, E. Schmitt

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): N. Stergiou, A.-S. Heimes, J. Jäkel, W. Brenner, M. Schmidt, E. Schmitt

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): N. Stergiou, J. Jäkel, W. Brenner, M. Schmidt, E. Schmitt

Writing, review, and/or revision of the manuscript: N. Stergiou, P. Besenius, M. Schmidt, H. Kunz, E. Schmitt

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): N. Gaidzik, S. Dietzen, M. Schmidt, H. Kunz

Study supervision: H. Kunz, E. Schmitt

This project was supported by the Deutsche Forschungsgemeinschaft, CRC 1066 project B13 (N. Stergiou, E. Schmitt, and P. Besenius) and the “Research Center for Immunotherapy (FZI)” of the University Medical Center (E. Schmitt).

We thank Susanne Gebhard for her support in the staining and evaluation of breast tissue specimens.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.