A Therapeutic Her2/neu Vaccine Targeting Dendritic Cells Preferentially Inhibits the Growth of Low Her2/neu–Expressing Tumor in HLA-A2 Transgenic Mice

The Her2/neu proto-oncogene, is a validated target in cancer, as various anti-Her2/neu antibodies (trastuzumab, pertuzumab, adotrastuzumab emtansine, etc.) or tyrosine kinase inhibitors (lapatinib) demonstrated their clinical efficacy in patients with Her2/neu–overexpressing breast and gastric cancers. Unfortunately, even after the best combination of Her2/neu–directed treatments and chemotherapy, the progression-free survival in patients with metastatic breast cancer is 18.7 months (1).

Intrinsic properties of Her2/neu may explain the clinical efficacy of these approved drugs and the interest to develop other therapeutic approaches against this molecule. Indeed, its overexpression is required to maintain the malignant phenotype of these cancers and tumor escape by downregulation of Her2/neu is thus more difficult to achieve. Cancer stem cells, which contribute to tumor metastasis and treatment resistance, express increased Her2/neu levels (2). Other arguments support the development of active immunotherapy such as anti-Her2/neu T-cell vaccines to complement and improve the clinical activity of commercially available anti-Her2/neu antibodies. Her2/neu is immunogenic and elicits natural Her2/neu–specific T-cell responses in patients overexpressing this antigen (3, 4). Intratumoral T-cell infiltration is associated with a good prognosis in these patients (5) and the presence of high levels of tumor-infiltrating lymphocytes (TIL) were associated with increased trastuzumab benefit in Her2/neu⁺ breast cancer patients (6).

The role of T cells in controlling Her2/neu–expressing tumors was also emphasized by the demonstration that intratumoral injection of human effector T cells genetically modified to express chimeric antigen receptor (CAR) against Her2/neu completely blocked the growth and metastasis of Her2/neu–expressing pancreatic adenocarcinoma xenografts in SCID mice (7).

Various preclinical and clinical studies showed that the combination of anti-Her2/neu antibodies and vaccines capable to inducing anti-Her2/neu T cells synergized in promoting the regression of Her2/neu–expressing tumors (8–10). In mice and humans, anti-Her2/neu antibodies enhanced the induction of Her2/neu–specific CD8⁺ T cells and favor Th1 cell infiltration (11–14). In accordance with these results, it has been reported that effective Her2/neu antibody treatment also requires an adaptive immune response involving CD8⁺ T cells (13). These data argue for boosting T-cell responses against Her2/neu–positive breast cancer and various groups clearly demonstrated the ability of Her2/neu cancer vaccines based on dendritic cells, peptides, recombinant proteins, recombinant virus, or DNA to induce specific CD4⁺ and CD8⁺ T-cell responses in humans (15–17). However, up until now, no Her2/neu vaccine has been approved for the clinics.

E75, a nine amino acid peptide derived from the extracellular domain of Her2/neu protein (KIFGSLAFL, Her2/neu residues 369–377; ref. 18) is the best studied Her2/neu–derived peptide to date. It binds to HLA-A2 and HLA-A3 expressed in 60%–75% of the general population and is processed and presented by tumor cells (19, 20). Although the exact immunodominant peptide (Her2/neu 369–377 or Her2/neu 373–382) presented by tumor cells is debated, anti-E75 CD8⁺T cells recognized these two epitopes (21).

The E75 peptide–based vaccine (nelipepimut-S) has been shown to be immunogenic in humans (22, 23), but induced suboptimal results in clinical trials designed for the prevention of clinical recurrence in high-risk disease-free breast cancer patients (24). Nevertheless, a phase III clinical trial based on the administration of the E75 peptide vaccine combined with GM-CSF is ongoing, making it the most clinically advanced Her2/neu targeting breast cancer vaccine (NCT01479244).

In parallel, recent impressive clinical results using immunomodulators blocking the PD-1/PD-L1 pathway underscore the major role of antitumor CD8⁺ T cells in tumor regression (25). The E75 peptide vaccine elicited specific CD8⁺ T cells, but their levels were weak and the increase of E75-specific CD8⁺ T cells after vaccine administration was most often lower than 2-fold (26–28). This CD8⁺T cells waned rapidly and required repetitive booster doses (29), as already observed with other Her2/neu peptide–based vaccines (30).

We have developed and validated a nonreplicative vector that targets dendritic cells, the B subunit of the Shiga toxin (STxB). When coupled to various tumor antigens, STxB elicits a strong induction of specific CD8⁺ T cells, which can be further enhanced by the addition of adjuvants (31, 32). Compared with other vectors, STxB targets the same glycolipid receptor, gb3 which are identical in all species. In addition, we already demonstrated its efficacy to favor crosspresentation in human dendritic cells (33). STxB displays intrinsic immunogenicity demonstrated by a weak induction of anti-STxB antibodies, but which did not interfere with repetitive immunization (31).

The aim of this study was to improve and optimize the E75 peptide vaccine using this delivery vector and to assess the role of various parameters (Her2/neu expression, combination with trastuzumab) in the efficacy of the cancer vaccine in a relevant preclinical model.

Humanized HLA-A2 transgenic (TG) mice were obtained from Dr. Y.C. Lone (INSERM U1014, Hopital Paul Brousse, Villejuif, France) and from Dr. F. Lemmonier (Institut Pasteur, Paris, France).

To set up an hHer2/neu-HLA-A2 tumor model, we transfected the B16-HLA-A2 melanoma cell line with a cDNA encoding hHer2/neu. A clone coexpressing HLA-A2 and hHer2 was selected (Supplementary Fig. S1A). We then subcutaneously grafted this B16-HLA-A2-hHer2/neu clone and monitored the tumor growth in HLA-A2 TG mice. Steady tumor growth was reproducibly achieved using at least 5 × 10⁵ cells (Supplementary Fig. S1B).

All mice were kept under specific pathogen–free conditions at the INSERM U970 animal facility. All experiments have been reviewed and approved by Paris-Descartes Ethical Committee for Animal Experimentation (CEEA34.ET.154.12).

B16-A2 cells were obtained from Dr. Y.C. Lone (INSERM U1014). B16-A2-Her2/neu cells were generated by stable transfection after electroporation with a plasmid encoding human Her2/neu provided by P. Bruhns (Institut Pasteur) and selected with 0.5 μγ/mL of G418 (Invivogen) and 7 μg/mL of blasticidine (Invivogen) supplemented medium.

To obtain the low and high Her2/neu–expressing B16-A2 cells, stable B16-A2-Her2/neu transfectants were sorted based on the Her2/neu expression.

SKBR3-A2 breast tumor cells were obtained from T. Dubois (Institut Curie, Paris, France) and UACC812 cells were bought from ATCC. SK-Mel 37 melanoma cells and MCF-7 breast tumor cells were obtained from B. Maillere (CEA, Saclay).

For ELISpot experiments, tumor cell lines were first cultured for 48 hours with 100 U/mL of rhIFNγ (Peprotech) or Universal IFN type I (R&D Systems), and then treated with 1 μg/mL mitomycin C for 1 hour. After washing, cells were used as stimulatory cells in IFNγ ELISpot assay (5,000 cells/well).

The anti-mouse CD8 (clone 53–6.7) and anti-human Her2/neu (CD340, clone 24D2) mAb were purchased from Biolegend (Distributor Ozyme) and the anti-human HLA-A2 (clone BB7.2) from BD Biosciences.

The HLA-A2-Her2/neu_369–377 dextramer was obtained from Immudex. For dextramer analysis, cells were incubated with PE-labeled dextramer (45 minutes at 4°C in the dark). After incubation and washes, labeled anti-CD8 mAbs were used to phenotype the positive dextramer CD8⁺ T cells. An irrelevant dextramer was used in each experiment to assure the specificity of the reaction against Her2/neu. The background values obtained with cells labeled with the irrelevant dextramer were deduced from the percentage shown.

Acquisitions were performed on BD LSRII, and data were analyzed with FlowJo Software (Tree Star).

CD8⁺ splenic T cells were isolated from immunized HLA-A2 TG mice by magnetic beads (Miltenyi Biotec). Specific HLA-A2-Her2/neu_369–377 CD8⁺ T cells were sorted by cell sorter (FACSAria, Becton Dickinson) after dextramer staining.

IFNγ ELISpot kits were purchased from Diaclone and used according to the manufacturer's recommendations. Briefly, in ELISpot IFNγ mAb–coated plates, CD8⁺ T cells were cultured in duplicate with antigen-presenting cells (APC) pulsed with the E75 peptide (10 μg/mL) or with tumor cells (after mitomycin C treatment and prestimulated or not with rhIFNγ or type I IFN). Plates were cultured for 18 hours at 37°C, and spots were revealed following the manufacturer's instructions. Positive controls included cells stimulated with 100 ng/mL of phorbolmyristate acetate (PMA) and 500 ng/mL of ionomycin (Sigma Aldrich). Negative controls included cells cultured in the absence of the E75 peptide. IFNγ spot-forming cells were counted on a C.T.L-ELISpot system (C.T.L). A response was considered positive if the number of spots in the well stimulated with specific peptide was 2-fold higher than the number of spots in the well without peptide, with a cut-off at 10 spot-forming cells.

For details on these techniques, see Supplementary Data.

In the therapeutic setting, HLA-A2 TG mice were subcutaneously injected with 1.5 × 10⁶ B16-HLA-A2-Her2/neu tumor cells in the right flank. Once the tumors were palpable (around 3–4 days after challenge), mice were vaccinated by the subcutaneous route with the STxB-E75 (0.5 nmol) or E75 peptide (0.5 nmol) combined with granulocyte macrophage colony-stimulating factor (GM-CSF; 20 μg.). One day later, CpG (50 μg) was subcutaneously injected at the same site. Seven days after the first immunization, mice were boosted with the vaccine alone.

In the prophylactic setting, mice were firstly immunized with the vaccine at day 0 and day 14 and one week later grafted with the tumor cells.

Tumor sizes were measured twice per week by using a caliper (mm² = length × width). For the survival analysis, a tumor size of at least 300 mm² was defined as the experimental endpoint.

For in vivo experiment with anti-Her2/neu mAb (4D5, Bioinvest Int), 100 μg of antibody were injected intraperitoneally at the same day as the vaccine.

For in vivo CD8⁺ T-cell depletion, 100 μg of anti-CD8 mAb (YTS 1694, Proteogenix) was injected intraperitoneally once per week, beginning 2 days before the first immunization dose.

Statistical analyses were performed with GraphPad Prism software (GraphPad Software Inc.). Data were expressed as means ± SD and are representative of at least two independent experiments. Significance was assessed with the Mann–Whitney test to compare two different groups, and Kaplan–Meier curves to compare the survival of the different groups of mice.

We have previously demonstrated a synergistic effect between the STxB delivery system and the aGalCer adjuvant (31), which binds to CD1d. As HLA-A2 TG mice did not express CD1d, we had to evaluate new adjuvants in combination with STxB. We showed that HLA-A2 TG mice immunized with STxB-E75 alone did not elicit a significant anti-E75 CD8⁺ T-cell response, as detected by dextramer staining (<0.1%) or ELISpot (Supplementary Fig. S2 and data not shown). In contrast, 0.2% or 0.75% of anti-E75 CD8⁺ T cell were detected, when STxB-E75 was coadministered with GM-CSF or CpG at the prime dose, respectively (Supplementary Fig. S2). Interestingly, a synergy for the induction of anti-E75 CD8⁺ T cell was demonstrated, when GM-CSF and CpG were combined during the prime with CpG given 24 hours after the vaccine and GM-CSF administration (Supplementary Fig. S2). Indeed, more than 5% of E75-specific CD8⁺ T cells could be observed with the vaccine system comprising STxB-E75 plus GM-CSF and CpG (Supplementary Fig. S2).

HLA-A2 TG mice (n = 4) were immunized with 0.5 nmol (20 μg) of STxB-E75, or with 0.5 nmol (3 μg) of the E75-derived peptide, as the optimal 9 mer peptide (Her2/neu_369–377), or the optimal 9 mer peptide plus the spacer sequence known to favor cleavage (RRAR), also present as a spacer between the Her2/neu_369–377 peptide and STxB in the STxB-E75 protein. All the vaccines were combined with GM-CSF and CpG. Immunization of mice with the E75 peptide alone or the E75 peptide flanked with the RRAR sequence did not induce anti-E75-CD8⁺ T cells, as detected by dextramer analysis (Fig. 1A) or ELISpot (Fig. 1C). In contrast, vaccination of mice with the STxB-E75 fusion protein at the same molarity (0.5 nmol) elicited high levels of dextramer-positive cells (more than 5%; Fig. 1A and B) and significant levels of IFNγ-producing CD8⁺ T cells (Fig. 1C).

To test for the ability of the E75 peptide to induce specific CD8⁺ T-cell responses, mice were immunized with 100 μg (25 nmol) of the E75 peptide combined with GM-CSF and CpG, which represents a 50-fold molar excess over the dose used for STxB-E75. Under these conditions, an induction of anti-E75 CD8⁺ T cells, reaching a mean value of 0.5% of total CD8⁺ T cells, could be observed (Fig. 1A) and they produce IFNγ (Fig. 1C). However, their levels were much lower than those observed in mice immunized with the STxB-E75 vaccine administered at only 0.5 nmol (Fig. 1).

It is well recognized that the quality of CD8⁺ T-cell response influences the clinical efficacy of antitumor vaccine. More specifically, the presence of multifunctional CD8⁺ T cells is considered as a relevant read-out to predict a positive clinical outcome (34). The E75 peptide vaccine was used at 50-fold higher molar concentrations than STxB-E75, as no induction of specific CD8⁺ T cells was observed at low E75 peptide concentrations (Fig. 1). The anti-E75 CD8⁺ T cells detected in mice immunized with STxB-E75 produced more chemokines (Rantes, MIP1α) than those induced by the E75 peptide, but we could not conclude from these results about the physiologic consequence of these differences (Fig. 2A). Moreover, except for the production of IL13, anti-E75 CD8⁺ T cells induced by STxB-E75 or E75 did not produce significant levels (>10 pg/mL) of the main Th2 (IL4, IL5) or Th17 cytokines (IL17; Fig. 2A). Interestingly, regardless of the vaccine formulation used (E75 peptide or STxB-E75), E75-specific CD8⁺T cells also produced a panel of chemokines (i.e., MIP1β, Rantes, eotaxin, IL9) that play a role in the recruitment of myeloid and T cells, which could amplify locally the antitumor immune response (Fig. 2A and data not shown). Therefore, the anti-E75 CD8⁺ T cells induced by STxB-E75 or E75 are bona fide multifunctional T cells.

We next evaluated the cytotoxic activity of E75-specific CD8⁺ T cells elicited in mice immunized with the different vaccine formulations, as this feature is considered a major hallmark of antitumor CD8⁺T cells. We demonstrated that the in vivo cytotoxicity effect of anti-E75 CD8⁺ T cells induced by STxB-E75 was higher (53% ± 20%) than that elicited by the E75 peptide (12% ± 15%; Fig. 2B).

Various groups reported that the avidity of CD8⁺ T cells is an important parameter for their antitumor activity (35). We therefore assessed the avidity of the anti-E75 CD8⁺ T cells induced after immunization of HLA-A2 TG mice with STxB-E75 or E75. As shown in Fig. 2C, the E75-specific CD8⁺T cells induced with STxB-E75 possess a higher avidity than those induced by the E75 peptide. The anti-E75 CD8⁺ T cells induced by STxB-E75 immunization can be activated by APCs sensitized with as low as 0.001 μg/mL of peptide (Fig. 2C).

The sustained induction of antitumor CD8⁺ T cells is a prerequisite for the achievement of a clinical antitumor response and the rapid waning of the antigen-specific CD8⁺ T cells may explain failure of cancer vaccines (29). To address this point, we measured the presence of E75 -specific CD8⁺ T cells, 30 days after the booster dose administration. At this later timepoint, the anti-E75 CD8⁺ T cells that were elicited by STxB-E75 are present at higher levels and produced more IFNγ than those induced by the E75 peptides (Fig. 2D).

The efficacy of the STxB-E75 vaccine in controlling hHer2-expressing tumors was tested in both prophylactic and therapeutic setting in HLA-A2 TG mice.

In a therapeutic setting, STxB-E75 clearly delayed tumor growth, when compared with nonimmunized mice (P < 0.001), or mice vaccinated with the E75 peptide (P < 0.001; Fig. 3A). In agreement with these results, 50% of mice vaccinated with STxB-E75 were still alive 50 days after the tumor cell engraftment (Fig. 3B). Immunization with the E75 peptide alone also had a significant albeit weaker effect on tumor growth, when compared with nonimmunized mice (P < 0.05), and 16% of mice immunized with this peptide were alive 50 days after tumor graft (Fig. 3B).

In prophylactic experiments, mice were immunized with STxB-E75 or the nonvectorized E75 peptide, both in combination with GM-CSF and CpG. A control group (not vaccinated mice) was also included. One week after the second immunization, mice were grafted with the HLA-A2-hHer2/neu B16 tumor cells. Immunization with STxB-E75 significantly inhibited tumor growth when compared with nonvaccinated mice (P < 0.05) or mice immunized with the E75 peptide (P < 0.01; Fig. 3C). The E75 peptide had no antitumor activity when compared with nonimmunized mice (Fig. 3D). These results were confirmed, when using survival as a read out. Indeed, 50% of mice vaccinated with STxB-E75 were still alive 40 days after tumor grafting, whereas all nonvaccinated mice or those immunized with the E75 peptide had died at day 37 (P < 0.05; Fig. 3D).

To confirm the role of CD8⁺ T cells in the clinical efficacy of the STxB-E75 vaccine, we demonstrated that the therapeutic tumor protection was lost when the CD8⁺ T cells were depleted (Fig. 3E and F). In addition, after vaccination with STxB-E75, we could detect intratumoral anti-E75 CD8⁺ T cells (Supplementary Fig. S3).

Preclinical results have pointed out the role of Her2/neu expression in the downregulation of HLA class I molecules, and preliminary clinical results suggest that the E75 vaccine is less potent in breast cancer patients whose tumors express high levels of Her2/neu (23, 36). On the basis of these evidences, we therefore assessed the ability of anti-E75 CD8⁺ T cells generated by the STxB-E75 immunization to recognize tumor cell lines expressing variable levels of Her2/neu.

Surprisingly, we found that anti-E75 CD8⁺ T cells were more readily activated by B16-HLA-A2 tumor cells expressing low levels of Her2/neu, compared with the high Her2/neu HLA-A2 B16 tumors (Fig. 4A). In line with these results, human breast tumor cells expressing the highest levels of Her-2 (SKBR-3 and UACC812; Supplementary Fig. S4B) were not recognized by the anti-E75 CD8⁺ T cells, whereas the SK-Mel37 melanoma cells and the MCF-7 breast tumor cells which express moderate levels of Her2/neu were recognized directly, or in the presence of IFNγ (Fig. 4B). IFNγ increases the HLA-A2 expression in the SKBR-3 breast tumor cells (Supplementary Fig. S4), but when cocultured with the anti-E75 CD8⁺T cells, the number of spots (n = 6) by the IFNγ ELISpot assay remains below the threshold of positivity set at 10 spots (Fig. 4).

To build on these results, we sorted B16-HLA-A2-Her2/neu^low or B16-HLA-A2-Her2/neu^high tumor cells (Supplementary Fig. S4). In the therapeutic setting, the STxB-E75 vaccine was more efficient to control the growth of B16-HLA-A2 Her2^low tumors than the B16-HLA-A2-Her2/neu^high tumors (Fig. 5A and C). In the survival analysis, 35% of the B16-HLA-A2-Her2/neu^low tumor-bearing mice treated with the STxB-E75 vaccine were alive at day 50, while none of the B16-HLA-A2-Her2/neu^lhigh tumor bearing mice were alive at the same time point (Fig. 5B and D). As a future cancer vaccine targeting Her2/neu has to position considering the current standard therapies, we assessed whether a synergy could be observed between the STxB-E75 vaccine and the anti-Her2/neu mAb. Interestingly, the use of anti-Her2/neu mAb alone was inefficient to inhibit the growth of tumors at any level of Her2/neu expression (Fig. 5A and C). In combination with the STxB-E75 vaccine, anti-Her2/neu mAb potentiated the effect of the vaccine on tumor growth in both models, but when survival was analyzed, anti-Her2/neu mAb potentiated the efficacy of the vaccine only in the B16-HLA-A2-Her2/neu^low tumors (Fig. 5B).

In this study, we demonstrated that the STxB-E75 vaccine efficiently elicits high levels of long-lasting, multifunctional, cytotoxic, and high avidity E75-specific CD8⁺ T cells in HLA-A2 TG mice, which were able to recognize human tumors expressing Her2/neu. Both in terms of concentrations and quality of the induced anti-E75 CD8⁺T cells, the potency of the STxB-E75 vaccine was far greater that of the E75 peptide alone, which is currently tested in phase III clinical trials in patients with breast cancer. This difference persisted even when the E75 peptide was used in 50-fold molar excess. This weak immunogenicity of E75 in preclinical model was already reported which supports our results (37).

The Peoples group has been able to correlate the magnitude of the E75-specific T-cell responses that were elicited by the E75 vaccine with clinical benefit (23). This observation suggests that the higher potency of the STxB-E75 vaccine to elicit antigen-specific CD8⁺ T cells, may improve the clinical response. To support this possible extrapolation, we could show in a preclinical HLA-A2-hHer2/neu tumor model, that the efficacy of STxB-E75 was significantly higher than that of nonvectorized E75 peptide in its capacity to inhibit tumor growth and to improve overall survival. In previous clinical trials, GM-CSF was used in combination with E75 peptide (36). We found that the addition of CpG to GM-CSF greatly improved the magnitude of the anti-E75 CD8⁺ T cells induced by STxB-E75, while this effect was less pronounced for the E75 peptide vaccine, similar to what has previously been reported for other peptide vaccines (38). The use of GM-CSF and CpG to recruit dendritic cells and induce maturation of dendritic cells, respectively, may be particularly useful for delivery vectors targeting dendritic cells such as STxB or anti-Dec205 (39).

To better position the E75-based vaccine from the clinical perspective, we showed that induced anti-E75 CD8⁺ T cells better recognized murine tumor cells expressing low Her2/neu levels. These results could be explained by the interference of Her2/neu with T-cell recognition via an inhibition of HLA class I molecules mediated by an activation of the MAPK pathway (40, 41). An inverse correlation between Her2/neu and MHC class I antigen-processing machinery (APM) component levels in human mammary carcinoma lesions has also been reported supporting this hypothesis (41). Interestingly, patients who benefited most from vaccination with the E75 peptide included those who did not express high levels of Her2/neu (3+), but rather low or moderate levels (1+, 2+; ref. 23). Beyond the Her2/neu cancer vaccine, high Her2/neu–expressing tumors from breast cancer patients will be associated with impaired antigen presentation leading to a poor recognition of tumor antigen by CD8⁺ T cells and this group of patients is unlikely to respond to cancer vaccines and immunotherapy (40). In line with these results, we observed that the STxB-E75 vaccine was efficient in a therapeutic setting to inhibit the growth of low-Her2/neu expressing HLA-A2-B16 tumors in HLA-A2 mice, while it showed no significant effect in the control of high-Her2/neu-expressing HLA-A2-B16 tumors both in term of tumor growth and survival analysis. Thus unexpectedly, low Her2–expressing tumors are more responsive to the vaccine than high Her2/neu–expressing tumors.

When the STxB-E75 vaccine was combined with anti-Her2/neu mAb, a clear synergy was observed on the control of growth of tumors expressing low or high levels of Her2/neu. However, an impact on survival was mainly observed with the low Her2/neu–expressing tumor model. The lack of effect with the anti-Her2/neu mAb when used alone could be explained by the fact that in this model, Her2/neu is not a natural driver of the tumor, which expresses this antigen only after transfection.

In humans, combination of trastuzumab with Her2/neu–based vaccines showed a synergistic effect in the induction of potent T-cell responses and encouraging clinical results were observed (8). Two phase II clinical trials are ongoing based on these preliminary results (NCT02297698 and NCT01570036). A mechanism that could account for this synergy is the ability of the anti-Her2/neu antibody to induce Her2/neu internalization and degradation, leading to the generation of Her2/neu–derived peptides presented by MHC class I molecules to correspondingly specific lymphocytes (42). Indeed, trastuzumab-pretreated breast cancer cells were lyzed more efficiently by E75 primed CTL, when compared with untreated breast cancer cells (43). In the current study, we did not observe increased levels of anti-E75 CD8⁺ T cells when the STxB-E75 vaccine was combined with anti-Her2/neu mAb (data not shown).

A possible limitation of the current study resides in the fact that hHer2/neu is not a self-antigen in humanized HLA-A2 TG mice. To the best of our knowledge, HLA-A2-hHer2/neu TG mice are not available. However, we have already demonstrated the ability of STxB coupled to murine Her2/neu–derived peptide to break tolerance against Her2/neu as self-antigen (44). In addition, spontaneous Her2/neu–specific T-cell responses and antibody responses have been documented in patients with Her2/neu⁺ tumors (3), and patients treated with peptide-based vaccines have mounted HER2/neu-specific immune responses (23), implying that overcoming immunologic tolerance against Her2/neu in patients with breast cancer is feasible.

It might also be argued that the use of a single CD8 peptide exerts less immune pressure than polyepitopic vaccines, and that the absence of CD4-associated peptides as in the case of STxB-E75, usually does not favor the generation of persistent CD8⁺ T cells. However, encouraging clinical responses have already been obtained with the E75 peptide alone, as well as with other CD8 peptide-based vaccine strategies (45). Although the use of long peptides incorporating CD4 and CD8 epitopes is often privileged to achieve significant CD8⁺ T-cell responses (17, 46), this study and previously published results from our group clearly document that the STxB delivery vector coupled to various CD8 peptides induces functional and persistent CD8⁺ T-cell responses, thereby bypassing the requirement of CD4 helper signal (44). In addition, epitope spreading has been reported after Her2/neu vaccines administration and the secondary recruitment of a broad antitumor T-cell repertoire may participate in the clinical response of the monoepitopic vaccines (8, 47, 48).

Cardiac toxicity has initially represented a clinical concern for the targeting of Her2/neu by mAb and a serious adverse event was reported following the administration of chimeric antigen receptor T-cell recognizing Her2/neu in a patient with colorectal cancer (49). However, all Her2/neu vaccines that were tested in humans alone or in combination with trastuzumab were devoid of significant clinical toxicity.

Regarding the toxicity of STxB, in vitro STxB did not exhibit cytotoxic effect on various normal cell lines and after repetitive administration of high dose of STxB, no toxicity was observed in mice (50).

The analysis of the dynamics of spontaneous or vaccine-induced antitumor CD8⁺ T cells clearly showed that in the course of tumor progression, they are anergized via the upregulation of checkpoint inhibitors (PD-1, CTLA-4) or blocked by various immunosuppressive mechanisms including regulatory T cells (51). Reversal of these blocking factors improved the efficacy of a Her2/neu cancer vaccine by reactivating intratumoral T cells leading to improved clinical results (52).

As from a clinical perspective, Her2/neu cancer vaccines would be combined with anti-Her2/neu mAb, especially breast cancers expressing high levels of Her2/neu, it is important to notice that T cells and immune signatures are crucial for the efficacy of anti-Her2/neu mAb treatments (13, 53). Patients with low intratumoral immune infiltrates will thus likely benefit from a combination of Her2/neu mAb with Her2/neu cancer vaccine to favor T-cell infiltration and ideally combined with a blockade of PD-1/PD-L1 pathway (54). In addition, in low Her2/neu–expressing tumors, cancer vaccines alone may overcome the clinical resistance to anti-Her2/neu mAb likely because T cells could be activated by a low amount of the HLA-peptide complex, while high levels of Her2/neu are required for the efficacy of antibodies.

The body of experiments presented in this work clearly demonstrates that the STxB-E75 vaccine is a potent candidate to be tested in the clinics especially in patients with tumors expressing low levels of Her2/neu. In association with an anti-Her2/neu mAb, the Her2/neu cancer vaccine may be particularly useful in patients with low intratumoral immune infiltrates, either alone, or in combination with blockers of checkpoint inhibitors.

C. Sibley reports receiving commercial research grants from ImmunoTargets SAS. No potential conflicts of interest were disclosed by the other authors.

Conception and design: T. Tran, M.O. Diniz, J. Medioni, E. Tartour

Development of methodology: T. Tran, M.O. Diniz, A. Gey, J. Medioni, E. Tartour

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T. Tran, M.O. Diniz, E. Dransart, A. Gey, N. Merillon, Y.C. Lone, E. Tartour

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T. Tran, M.O. Diniz, A. Gey, N. Merillon, S. Oudard, E. Tartour

Writing, review, and/or revision of the manuscript: T. Tran, M.O. Diniz, S. Godefroy, C. Sibley, J. Medioni, S. Oudard, L. Johannes, L.C.S. Ferreira, E. Tartour

Study supervision: S. Godefroy, L. Johannes, E. Tartour

Other (financial support for research conducted leading to this manuscript): C. Sibley

This work was supported by grant from Institut National du Cancer (INCA) (PLBIO11-022- IDF-JOHANNES; to L. Johannes and E. Tartour), Immuno Target SAS (to E. Tartour and L. Johannes), Ligue contre le Cancer (to E. Tartour), Université Sorbonne Paris Cité (to E. Tartour), Labex Immuno-Oncology (to E. Tartour), SIRIC CARPEM (to E. Tartour), and ERC advanced grant (project 340485; to L. Johannes).

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.