Breast cancer is a particularly frequent malignant disorder striking the female population worldwide with an incidence of 1.3 million per year, bearing a life time risk of 1:8 and being responsible for around 380,000 deaths [1]. In Europe, around 400,000 new cancer cases are diagnosed, and more than 130,000 deaths are reported every year [2].

Treatment of breast cancer has been impressively diversified ever since the introduction of antiendocrine treatment in the 1970s, and at approximately the same period, the anti Her-2/neu antibody trastuzumab. The latter proved that targeting a biologically important epitope in combination with chemotherapy produced significantly better outcomes compared to chemotherapy alone.

However, such monoclonal antibodies require frequent applications, and their half-life limits the duration of therapy. Therefore, generating cancer vaccines which stimulate the patient’s own immune system to produce long-lasting anti-tumor responses has become an attractive treatment strategy.

The general topic of vaccination for malignant disease in general and breast cancer in particular includes prophylaxis as well as using vaccine against recurrence of early disease, and as therapy for metastatic breast cancer.

Creating a successful tumor vaccine involves the following considerations:

  1. 1.

    The complexity of the tumors, their environments, and their mechanisms for evading immunotherapy, which a successful vaccine must overcome;

  2. 2.

    The selection of a tumor-associated antigen (TAA) as vaccine target;

  3. 3.

    The delivery of these self-antigens, for example, in the form of peptides which may increase the immunogenicity of the respective TAAs;

  4. 4.

    The use of certain adjuvant systems to promote anti-tumor immune responses and counteract the immunosuppressive capacity of the tumor microenvironment.

These aspects are covered in the discussion below, focused on the current status of vaccine development in various stages of breast cancer and concentrating on Her-2/neu-overexpressing cancers.

Immunological modulation of the tumor microenvironment

As reviewed by Hannahan and Weinberg, the tumor with its cells and stroma resembles an organ—though functionally and structurally abnormal. It contains various cells such as fibroblasts, cells of the vascular system or immune cells producing cytokines, chemokines, growth factors and antibodies, which determine the outcome of the clinical disease [3, 4]. The tumor growth and dissemination of tumors depend on such factors as sustained proliferation signaling, evasion of growth suppression, activation of invasion and metastases, enabling of replicative immortality, induction of angiogenesis, and resistance to cell death [4]. These developments in turn depend upon the tumor microenvironment in which the inflammatory process may rather promote than inhibit tumor growth [5]: The infiltration of cells that leads to potential proliferation of tumor cells consists preferentially of Th2 cells, B cells and regulatory cells, which inhibit the generation of cytotoxic T cells directed against the tumor cells.

Induction of regulatory cells (Tregs; CD4+CD25+Foxp3+) may be directly induced by self-tumor antigens to limit acute inflammation and prevent autoimmunity. Very recently, the involvement of a previously undescribed subset of tumor-evoked regulatory B cells (CD19+CD25high CD69high) has been described in a mouse model of breast cancer [6]. These B reg cells induce TGF-β-dependent conversion of Foxp3 Tregs from resting CD4+ T cells and seem to be involved in lung metastases. At the same time, Treg cell infiltration seems to be an independent predictive marker of poor prognosis and of increased likelihood of relapse of some tumor types, including breast cancer [7]. Some clinical studies have shown that depletion or inactivation of Treg cells in combination with tumor-specific vaccination increased the tumor-specific immune responses [8, 9].

There is, however, also the possibility to control or potentially eliminate tumor cells by preferential infiltration of Th-1 like cells. These cells penetrate through the tumor and infiltrate into the parenchyma and produce an environment rich in IL-12, IFN-γ, TNF-α, and IL-2. Such a Th1-biased cytokine network may increase antigen presentation and generation of cytotoxic T cells that favor tissue destruction. Gene expression studies identified several genes associated with such Th1 immunity and CTL generation to be associated with improved survival and prognosis [5, 10]. In a subset of patients with triple-negative breast cancer who tend to remain free of metastases, such an IFN-based gene signature has been identified as a predictor for relapse-free interval and increased overall survival [11, 12].

Major immunomodulatory pathways to be induced by vaccination may therefore focus on the induction of Th1-biased responses, along with suppression of regulatory cells.

Therapeutic and prophylactic vaccines: pros and cons

Prophylactic and therapeutic vaccines represent intriguing elements of multidisciplinary treatment of cancer patients. However, translating these promising preclinical vaccination concepts into the clinical setting has met several hurdles [13].

Tumor cell aggregates can escape immunoregulation through downregulation of MHC I in about 70–90 % of all human tumors [14]. It has been shown that oncogene-induced downregulation of MHC-I (e.g., by c-Ras or c-Myc, or by Her-2 overexpression) leads to defects in the antigen-presenting pathways [15, 16] with inhibition of tumor recognition by NK and CTL cells [17].

Moreover, tolerance to the self-tumor antigens must be accomplished by the use of immunodominant epitopes in a different molecular environment to the tolerized host. Vaccines evaluated in clinical trials so far are based on autologous tumor cells delivered by dendritic cells; [1821] or DNA [22]; (ongoing trials, proteins [2326] or peptides inducing CTL, Th1 responses, and/or humoral responses [2732].

In addition, in the case of a rapidly growing tumor, a therapeutic vaccine might achieve significant effect only in combination with chemotherapy [3335], radiotherapy [36, 37], or other approaches, such as passive immunotherapy with monoclonal antibodies or kinase inhibitors [26, 38, 39] that can produce synergistic or additive effects. These combinations might reduce tumor growth and progression, without, however, achieving complete control of the disease process.

With respect to preventive vaccination, precancerous lesions may not yet be detected by the immune system or organized as tissues that induce immunosuppressive signals. Therefore, vaccination in this early stage of disease with modified potential self-antigens would induce an effective immune response compared to an advanced stage where an organized, solid tumor mass presents all the previously described cell characteristics and immunologic interactions [40].

However, a major long term concern with prophylactic vaccines is the induction of autoimmunity after vaccination with TAA. Clinical data have shown that cancer vaccines are of low toxicity and do not induce autoimmunity, but, given the limited number of subjects in current trials, such signals might be missed. In light of the—relatively low—risk of cardiotoxicity related to treatment with trastuzumab, there is also concern that Her-2 neu-targeting vaccines might induce clinical toxicity. This toxicity could persist simultaneously with the vaccine-induced immune response. Using such vaccines in the preventive or adjuvant setting might represent excessive risk for clinically potentially healthy individuals.

Nevertheless, the greatest promise of cancer vaccines might lie in preventing tumor recurrence or progression and prolonging survival in patients with minimal residual disease [40, 41].

Tumor-specific antigens/tumor-associated antigens as vaccine targets

For the use of cancer vaccine antigens, two categories of tumor-related antigens come into question: (i) tumor-specific antigens (TSA) and (ii) TAA. TSAs are unique tumor antigens, essential for tumorigenesis and cancer progression, caused by somatic mutation, activated by human retroviruses, or induced and expressed on tumor cells by viral antigens, but not found on any normal adult somatic tissues [42]. As these antigens are novel for the immune system and consequently no tolerance being induced, TSAs would be ideal candidates for tumor vaccines. However, as they are expressed on individual patient’s tumor cells, such vaccine development would be limited to a personalized approach, thus being too costly and time-consuming to be practical. TAAs, in contrast, are molecules appearing on various cancer and normal cells, though with different expression levels. They are commonly over- or aberrantly expressed on tumors with the same histology and shared among tumors of different origin. When found at low levels also on normal cells, TAAs are weakly immunogenic due to tolerance against self-antigen. This presents obstacles to successful vaccine development.

Examples of TAAs expressed on breast cancer cells

A team of experts assembled by the US National Cancer Institute (NCI) reviewed a prioritization of cancer antigens in a ranking list of 75 antigens. Consideration included therapeutic function, immunogenicity, specificity, oncogenicity, expression level and percentage of positive cells, stem cell expression, number of patients with antigen positive cancers, and number of epitopes and cellular location of expression [43]. While none of the listed cancer antigens fulfilled the criteria of an optimal vaccine antigen, 46 of them were immunogenic in clinical trials and 20 were suggested to have clinical efficiency. This ranking list contains cancer antigens in general and also the most relevant breast cancer antigens. With respect to the breast cancer antigens, the following eight are listed. This review places special emphasis on Her-2/neu:

MUC-1, a membrane-associated glycoprotein, is expressed on many ductal epithelia in pancreas, lung, gastrointestinal tract, and breast. Because it is overexpressed in more than 70 % of cancers, it is an attractive target for broad cancer vaccination, including metastatic breast and ovarian cancer. [4447]. This antigen is ranked as No. 2 according to NCI.

The p53 antigen, a tumor suppressor antigen (non-mutant and mutant form, ranked as No 9 and 17, respectively, by the NCI) has been used as dendritic cell vaccine in several phase I studies with breast cancer patients [48, 49].

The carcinoembryonic antigen (CEA) is overexpressed in most carcinomas of the colon, rectum, lung, pancreas, gastrointestinal tract, and breast. As an adhesion molecule, it promotes metastatic processes. CEA is ranked by the NCI as No 13 and has been used in phase I studies with breast and ovarian cancer patients [50].

The human telomerase reverse transcriptase (hTERT), has been considered as a target antigen for vaccination due to its nearly universal expression in human cancer cells and its important role in tumor growth. Vaccination with this antigen has shown limited success, and it has been ranked No 23 [51, 52].

Sialyl-Tn Ag is an immature glycosylation product of serine and threonine protein core (usually masked by the complete glycosylate chain), which is highly expressed on tumor cells and associated with tumor progression and metastasis [53]. Clinical trials with this antigen, ranked No 56, so far have shown limited success [5456].

In addition, BORIS and Fos-related antigens were ranked as breast cancer antigens with low priority by the NCI [57, 58].

Her-2/neu is one of the most investigated tumor antigens in breast cancer (No 6 of the NCI ranking list), a 185 kD gene product of erbB2/neu proto-oncogene belonging to the epidermal growth factor receptor family. By 1998, numerous studies, encompassing more than 15,000 patients, had been published on the role of Her-2/neu in breast cancer [59]. Her-2/neu is weakly detectable on epithelial cells of normal tissues. It is overexpressed in 20–30 % of primary and metastatic breast cancer, and also in ovarian, pancreas, or stomach cancer [60]. It is linked with a poor prognosis and high risk of relapse [61]. Her-2/neu consists of a cysteine-rich extracellular domain (ECD) with several glycosylation sites, a hydrophobic transmembrane domain, and an intracellular conserved tyrosine kinase domain. Crystallographic studies have provided significant insight into functions and regulation of dimerization of the ErbB (Her) receptors. EGF receptor activation and dimerization are ligand-dependent, and binding of the ligand results in a conformational change of the receptor monomer that enables receptor dimerization [62]. As Her-2/neu (ErbB2) has no direct ligand, it is suggested that it may function as a co-receptor (or heterodimerization partner), and its overexpression can cause cell transformation even in the absence of added ligands [63]. The arrangement of the domains (stabilizing the extended conformation) suggests that Her-2/neu may be auto-activated with the dimerization arm constantly exposed. Activated ErbB3 molecules preferentially heterodimerize with Her-2/neu (ErbB2) forming the most prevalent and mitogenically potent complex [62].

Based on these characteristics and properties, Her-2/neu has become an attractive target for immunotherapy. Numerous antibodies directed against the ECD have been generated by immunizing mice with cells expressing Her-2/neu. One of the monoclonal antibodies has been shown to have particular properties for reducing growth of Her-2/neu-expressing tumors in mice, by mechanisms such as cell-mediated cytotoxicity (ADCC) or complement-dependent cell (CDC) lysis. Based on these results, a humanized form of this antibody, trastuzumab, has been generated. Trastuzumab has been shown to mediate its antitumor effects in humans, apart from ADCC, also via NK cell activation, internalization, and degradation of Her-2/neu and disruption of receptor dimerization [64]. Thus, it is currently used with great efficacy as monotherapy or even more successfully in combination with chemotherapy [65], kinase inhibitors [66, 67] and in combination with pertuzumab which inhibits the heterodimerization of Her-2 with Her-3 resulting in a persisting clinical benefit in the case of emergence of trastuzumab resistance [68, 69].

However, clinical use of such monoclonal antibodies as trastuzumab is cost-intensive and thus not routinely available in many countries, and particularly in the third world. Moreover, therapy requires regular applications at short intervals, can only target single epitopes, and is of limited therapeutic duration due to the agent’s half-life. Therefore, induction of Her-2 specific immune responses with long-lasting anti-tumor effects is regarded as an alternative treatment strategy.

Peptides derived from Her-2/neu for vaccination

Several systems for antigen delivery have been recently categorized by Mittendorf et al. [70], such as whole tumor cell vaccines, dendritic cell vaccines, viral vector vaccines, and peptide vaccines, the latter being the focus of the present review. Three different preparations of peptide vaccines can be considered, i.e., genetic vaccines encoding Her-2 neu peptides, MHC I or MHC II-dependent T cell peptide vaccines, and MHC-independent B cell vaccines.

Genetic vaccines encoding for single or multiple Her-2/neu peptides

Early studies in animals have shown that DNA vaccines are of rather low immunogenicity. In order to overcome this problem, several approaches have been investigated; these include the plasmid design, routes of administration (e.g., i.m./i.d. electroporation, gene gun) [71], use of immunomodulators and adjuvants (TLR-agonist such as CpGs, GMCSF etc.) [72], and prime-boost strategies (plasmid DNA/recombinant protein, plasmid DNA/viral vector, or viral vector/plasmid DNA) [73]. With respect to the plasmid design, it was shown that DNA vaccines without the use of adjuvants encoding only for the ECD of Her-2/neu did not provide any protection, while DNA vaccines coding for the extracellular and transmembrane domains of Her-2/neu showed a transient effect [74]. Addition of certain adjuvants and interleukins (e.g., IL-18, IL-12, IL-15, IL-23, Flt3L) has been shown to improve the efficacy of Her-2/neu DNA vaccines. Furthermore, bicistronic vaccines coexpressing Her-2/neu and cytokines improved the antitumor effects in comparison to coadministration of monocistronic DNA vaccines [75]. Along these lines, a chimeric DNA vaccine encoding for three HLA-A2 restricted epitopes of the extracellular and intracellular domain (ICD) of Her-2/neu was shown to prevent/reduce tumor growth as well as seeding of lung metastases in a model of metastatic breast cancer [76].

The only recently published pilot clinical trial in patients with metastatic breast cancer described the use of a full-length signaling deficient version of Her-2/neu together with low doses of GM-CSF and IL-2 [77]. This vaccine, which was administered in conjunction with trastuzumab, was safe and induced longlasting cellular and humoral immune responses. Further clinical trials are needed to demonstrate the clinical efficacy of such vaccines.

T-cell peptide vaccines

Regarding T cell peptides, synthetic small CD8+ T cell epitopes, primarily to induce cytotoxic immune responses, have been tested, and the use of synthetic long hybrid peptides including T-helper and T-cytotoxic epitopes has been described. However, vaccination with short peptides (8–10 aa) was reported to induce tolerance rather than immune responses, because they can bind directly to MHC I on all cells/non-professional antigen-presenting cells. Such activation of CD8 T cells without co-stimulatory signals induces tolerance most likely via anergy; those antigen-specific CD8+ memory T cells which can be transiently activated possess low cytotoxic activity [78, 79].

Therefore, helper peptides have been added to CD8+ epitopes in the vaccines to improve the immune responses by presentation of both CD8+ and CD4+ T cell epitopes by dendritic cells (DC). T helper cells will thereby be activated, which in turn can signal CD8+ cells to become effector cells with killing capacities.

It should be noted that such peptides can also exogenously load T and B cells, which might then rather function as tolerogenic antigen-presenting cells. Thus, further improvement was made by production of long synthetic hybrid peptides (27–35 aa) to ensure that only DCs process and present the peptides to result in stronger and more effective immune responses elicited by both CD4+ and CD8+ T cells. This effect might be induced by an increased duration of epitope presentation in draining lymph nodes resulting in effective clonal expansion and increased IFN-γ production by the effector cells [80].

Selected peptides will induce only a narrow T cell response due to a restricted HLA type. Therefore, T cell-based vaccines covering the peptides of the complete protein sequence have been proposed to cover multiple HLA I and II peptide epitopes for the induction of CD4+ and CD8+ T cells able to induce anti-tumor responses to all available tumor antigen epitopes. Several clinical trials have been performed with different T cell-peptide based vaccines:

Clinical trials with T cell peptide vaccines

Early clinical trials using T helper peptides from the ECD, the ICD as well as peptides from a region restricted to MHC-class I (i.e., HLA-A2) binding areas showed that intradermal immunizations up to six times every month led to peptide-specific T helper responses and CD8 responses along with anti-Her-2 specific cellular immune responses [27]. Kinetic studies up to 1 year from the start of immunization revealed that peptide-specific responses were induced early, about 2–4 months after the start of immunization, while for development of Her-2 neu specific immune responses, the majority of patients needed to complete the immunization schedule with six vaccinations. In around 13 % of these patients, the immune responses lasted up to 1 year after the start of immunization [81].

Using GM-CSF as adjuvant for immunization, the phenomenon of epitope spreading, i.e., extension of T cell immune responses to regions of the Her-2/neu protein which are not included in the vaccine, was described and was believed to be related to a more effective presentation of subdominant epitopes induced by GM-CSF. However, compared to the significant T helper and CD8 T cell responses, humoral responses were less induced by this vaccine [28] since the use of the adjuvant GM-CSF along with HLA-restricted T cell epitopes preferentially seems to induce Th-1 immune responses and IFN-γ secretion. Rather recently it was shown that the combination of this tri-peptide vaccine with trastuzumab is well tolerated and that Her-2-specific T cell responses were induced upon vaccination in the presence of the monoclonal antibody application. Follow-up studies addressing overall survival as clinical endpoint will clarify the therapeutic efficacy of this combination treatment [39].

Another T cell peptide vaccine approach is based on the use of a MHC-class II binding peptide from the ICD of Her-2/neu (AE37), which stimulates CD4-T cell responses. To enhance this effect, the peptide was linked to the li-key moiety of the MHC class II associated invariant chain. A phase I trial showed that vaccination with this hybrid-peptide induced T cell responses without additional use of an adjuvant [82, 83]. Very recently, a phase II study with a HLA-A2/A3 restricted peptide from the ECD (E75) was used for immunization with GM-CSF, and showed that the peptide vaccine was immunogenic in HLA-A2+ patients [84] and prevented recurrence in these selected high risk patients [85, 30]. The authors state that a phase III study is planned, most likely with a combination of several peptides representing MHC-class I epitopes from the ECD and the transmembrane domain (GP2) as well as a promiscuous HLA class II binding peptide from the ICD to facilitate a broader use of the vaccine [86].

B cell peptide vaccines

Another approach for the development of peptide vaccines against Her-2/neu positive breast cancers has focused on the use of Her-2-specific B cell epitopes to preferentially induce antibody responses with anti-tumor activity against Her-2/neu, as described for the monoclonal antibody trastuzumab [2]. The rationale behind this approach is that antibodies bind to surface-expressed tumor antigens independent of MHC-I expression. Thus, down-regulation of MHC-I molecules as tumor evasion mechanisms does not affect the epitope accessibility for antibodies. The only possibility of tumor escape would be through the selection of antigen-loss variants, which, however, has not, or has only rarely been observed to occur with Her-2/neu. The problem of HLA restriction of T cell epitopes, which the production of overlapping poly-peptide constructs attempted to overcome, is not of concern for B cell epitopes which allows a broad application across all HLA-types.

On the one hand, the selection of B cell epitopes has been based on phage display technology using different anti-Her-2 monoclonal antibodies, including trastuzumab with known antitumor effects for the selection of epitopes located in the binding sites of these Abs [8790]. Recently, a new antibody epitope was identified by monoclonal abs raised against Her-2/neu expressing tumor cells. This peptide included B cell epitopes and both helper and cytotoxic T cell epitopes, indicating that in case of adoption for cancer therapy, anti-tumor antibodies and cytotoxic T cells might be induced [91].

Alternatively, computer aided programs for selection of peptides that give rise to anti-peptide and anti-Her-2 humoral responses have been described. Her-2 B cell epitopes were linked to promiscuous T cell epitopes from a measles virus fusion protein which induced humoral immune responses against the peptides and the Her-2 protein [31]. The authors further demonstrated that the combination of the peptides was more effective than immunization with single peptides, and that additional application of IL-12 decreased the number of lung metastases in a murine model of transplantation of syngeneic Her-2 overexpressing tumor cells. Along these lines, we previously identified three B cell peptides at the ECD of Her-2/neu by computer-aided algorithms, two of them located within the binding region of trastuzumab, the other one within the region of the dimerization loop of Her-2/neu [92]. Preclinical development and evaluation as well as translation into a Phase I clinical trial will be described in detail in the next paragraph.

Development of a B cell-based vaccine against Her-2/neu overexpressing breast cancer

The mechanisms by which trastuzumab elicits its anti-cancer effects have been well-described, and include antibody-dependent cytotoxicity (ADCC), activation of natural killer cells, internalization and degradation of Her-2/neu resulting in disruption of receptor dimerization and suppression of angiogenesis [64]. It was, therefore, our intention to induce a polyclonal response according to the action of Herceptin by selection of epitopes located within the Herceptin binding region and/or the dimerization loop to inhibit receptor dimerization. Of the seven selected putative B cell peptides, three (aa 378–394; aa 545–560; aa 610–623) gave rise to peptide-specific as well as Her-2/neu specific antibody responses when conjugated to tetanus toxoid [92]. The anti-tumor effect of these antibodies were tested in vitro, showing a reduction of tumor cell proliferation of Her-2/neu overexpressing tumor cells as well as tumor lysis by CDC lysis.

Importantly, we have demonstrated that antibodies derived after immunization with a mixture of the three peptides mediated ADCC in a manner comparable with Herceptin, indicating that the combination of the peptides was superior to single peptide immunization due to eliciting synergistic anti-tumor effects. Using this multi-peptide approach for immunization in a transgenic mouse model spontaneously developing c-neu overexpressing breast cancers revealed that the tumor-free interval was prolonged and the tumor growth progression was delayed [93]. As it was proposed that polarization of the immune response toward a Th1 type has an important impact on tumor prevention (via IFN-γ-induced antibody isotype switching with enhanced ADCC-mediating properties, activation of NK cells and macrophages, and inhibition of angiogenesis within tumor lesions), the vaccination strategy was improved by addition of IL-12. Indeed, the induced Th-1-biased anti-Her-2 immune responses (increased IgG2a and IFN-γ levels) were associated with a significantly prolonged tumor-free interval and delayed tumor growth progression, as compared with peptide immunization without Th1 immunomodulation. Based on these results, a Phase I clinical trial in metastatic breast cancer (MBC) patients was initiated [32].

Clinical trial with the B cell multi-epitope vaccine

In order also to enhance patients’ immune responses, the peptides were coupled to immunopotentiating reconstituted influenza virosomes which are used as carrier systems in licenced vaccines against infectious diseases with known Th1 and Th2-inducing properties and high safety profile [94]. Ten patients were chosen, with metastatic disease without overexpression of Her-2/neu (Her-2 +, ++, FISH negative; er/PGr positive, life expectancy >4 months) This patient collective was chosen for ethical reasons in order to avoid withholding standard treatment with trastuzumab in Her-2 driven disease. Patients were immunized three times in monthly intervals [32]. The results from the primary endpoint on safety revealed that only minimal local side effects at the injection site, but no systemic effects, occurred in a few patients (4/10). The majority of patients (8/10) developed anti-peptide antibodies, and all but one also showed Her-2/neu specific antibody levels.

Due to conjugation of the peptides to the virosome carrier system bystander CD4+ T cell responses, necessary to induce antibody isotype switching and induction of memory responses to the B cell peptides, were seen in IL-2, TNF-α and IFN-γ production of PBMCs upon re-stimulation with virosomes in vitro.

It should be noted that the humoral and cellular immune responses were induced in patients across all HLA types, indicating that all patients (except for 2 non-responders) were fully susceptible to vaccination. The immunocompetence of these patients with advanced metastatic breast cancer was further demonstrated by the characterization of lymphocyte subpopulations in comparison to age and gender-matched healthy controls. No differences in the distribution of lymphocyte subpopulations were detected between stage IV cancer patients and healthy controls before and after immunization.

It is worth mentioning that patients displayed a significantly higher number of regulatory cells (Tregs) in the peripheral blood compared to healthy controls before vaccination. Our results are confirmed by a recent publication showing that particularly in Her-2/neu positive patients, the number of Tregs in the blood is increased [95]. Previous studies have demonstrated that tumor infiltration by Tregs is associated with reduced survival of breast cancer patients [96, 7]. Importantly, our results showed that the number of Tregs significantly declined after the vaccination course suggesting that a decrease of Treg cells might be important for induction of proper immune responses as well as for vaccine effectiveness.

The fact that similar effects were also described after treatment with trastuzumab [97], aromatase inhibitor [96], or certain chemotherapeutic treatments [33] argues for a combined Treg cell inhibition strategy along with vaccination to broaden the anti-tumor treatment effects.

With respect to the clinical outcome, this study was not designed to focus on the efficacy of the vaccine, since out of ethical reasons patients enrolled in this Phase I trial did not harbour Her-2/neu overexpressing tumors. Thus, in order to further evaluate the clinical response upon vaccination with the B cell multi-peptide vaccine a randomized phase II trial in patients with overexpressing Her-2/neu breast but also gastric cancer [65] is currently in preparation.

In summary, the B cell multi-epitope vaccine induces several characteristics which make this vaccine an interesting candidate for further exploitation of the anti-tumor treatment approach in Her-2/neu overexpressing malignancies: (i) induction of antibodies with anti-tumor activity and induction of immunological memory; (ii) polyclonal humoral responses (advantage over treatment with monoclonal antibodies); (iii) HLA independent peptides (advantage over CD8+ and CD4+ peptide vaccines); (iv) induction of bystander Th1 immune responses by selection of proper carrier/adjuvant system; (v) reduction of regulatory T cells (possibly counteracting tumor evasion mechanisms and improving vaccine responsiveness).

Adjuvants and immunomodulators for breast cancer vaccines

As already emphasized earlier in this review, the use of adjuvants in almost any vaccine is important to enhance the magnitude, the coverage, the quality and the permanence of vaccine-specific immune responses, while inducing minimal toxicity and immunity towards the adjuvant substance itself. All these requirements are of particular importance for cancer vaccines due to the low immunogenicity of the TAAs, circumvention of tolerance and the obvious necessity to induce strong Th1/cellular responses [98].

Adjuvants with Th1 promoting properties

The new generation of adjuvants in recently licenced vaccines include virosomes, oil in water emulsions, such as MF59® and ASO3®, or the TLR-4 agonist monophospharyl lipid A (MPL+ alum; ASO4®), which have been developed to enhance weak immune responses in certain target populations (elderly, immunocompromised etc.) by induction of strong humoral and cellular immune responses. Therefore, the following adjuvants might be suited to be used in conjunction with breast cancer peptide vaccines as well:

As described, immunostimulatory reconstituted influenza virosomes (IRIVs) are spherical, unilamelar vesicles prepared by detergent removal from a mixture of natural and synthetic phospholipids and the influenza surface glycoprotein hemagglutinin (HA), which facilitates antigen delivery to immune cells [94]. Based on specific fusion mechanisms of vaccine antigens (i.e. antigen linked either to the surface or encapsulated in the virosomes) targeting of MHCI and/or MHC class II pathways is possible. Additionally, different proteins or peptides can be cross-linked to the vesicles thereby constituting carrier and delivery system for induction of B and T cell immune responses to the linked antigens [99]. Accordingly, we selected this adjuvant system for our B cell peptide vaccine and demonstrated—as previously described—that significant Her-2/neu antibody levels along with Th-1polarized bystander responses induced by the virosome components were generated in MBC patients.

Oil-in water (O/W) emulsions, using squalen in the adjuvants MF59® or ASO3® (+α-tocopherol) have been primarily developed and used for improvement of vaccine responses to influenza antigens [100]. The mechanisms of action have been described for MF59®, leading to increased antigen uptake, increase of cytokine production by APCs and increased chemokine induced migration of APC into draining lymph nodes. Thereby antibody titers are increased with a balanced Th1/Th2 immune profile. In pilot studies, comparing different adjuvants (Al(OH)3, Freund’s adjuvant, GERBU, virosomes, squalen-O/W emulsions) for the B cell epitopes we registered that, in contrast to immunization with AL(OH)3, peptide/Her-2-specific antibody levels and cellular responses (IFN-γ, IL-2, TNF-α) were comparable in mice immunized with either virosome preparations or the O/W emulsions.

ASO4®, an aqueous formulation of the TLR-4 agonist monophospharyl lipid A and alum is included in a licenced vaccine against HPV-induced cervical cancer [101]. Further development of the combination of TLR-4 and TLR-9 agonists is currently ongoing in clinical cancer vaccine trials [102], such as a phase III trial using MAGE-A3 as antigen against non-small cell lung cancer [103]. Recent preclinical studies describe the use of TLR-9 agonists in conjunction with genetic vaccines against breast cancer (71). Other adjuvant formulations including further TLR- agonists, such as TLR-2, 3, 4 or 5, also seem to be interesting candidates for adjuvants in cancer vaccines. However clinical investigations have so far demonstrated only limited efficacy against cancers, partly because of some concerns regarding severe toxicity and vaccination failure due to non-specific activation of different immune cells including regulatory T cells [104, 105].

Cytokines and chemokines for immunomodulation

Th-1 promoting cytokines, including IL-2, IL-12, IL-15, IL-23, IL-18 or GM-CSF have been tested with varying success in preclinical and some clinical trials in conjunction with different treatment approaches against various tumor entities [106, 107, 75]. In combination with peptide vaccines against breast cancer, mainly GM-CSF has been used in clinical trials as adjuvant in conjunction with T cell peptide vaccines to enhance Th1 and CD8+ T cell responses and to a lesser extent also humoral responses, as previously described [28, 79].

Immunomodulators for suppression of regulatory cells

As already mentioned, an increased presence of T regulatory cells in patients with certain cancers, including breast cancer, has been suggested as one of the prediction markers of tumor progression or tumor recurrence, while at the same time being also associated with low responsiveness to vaccination. In several mouse models it was shown that depletion of Treg cells by anti-CD25+ antibody treatment significantly enhanced the immune responses after vaccination leading to improvement of the anti-tumor effects [108]. Very recently it was shown in Her-2/neu transgenic mice that low dose treatment with cyclophosphamide led to selective depletion of CD4+Foxp3+CD25low T regulatory cells allowing activation of tumor-clearing high avidity T cells [109]. Along these lines it was shown in patients with colon cancer as well as breast cancer that treatment with certain chemotherapeutic substances, such as cyclophosphamide, doxorubicin or 5-fluorouracil [33, 110], aromatase inhibitor [96] or trastuzumab [97] resulted in reduction of circulating Treg cells. Thus, having in mind that vaccination per se can also reduce regulatory cells in circulation [32], application of one of these treatment substances with Treg-suppressing properties in combination with an anti-tumor vaccine seems to be a promising strategy to induce additive/synergist anti-tumor effects, leading to improvement of the clinical outcome.

Change of application route and mucosal adjuvants: future aspects

Based on the fact that Her-2/neu overexpression occurs also in about 20 % of gastric cancers, which are shown to be highly treatment resistant, it is tempting to speculate whether vaccination strategies that lead not only to systemic but also to local anti-tumor responses could improve the efficacy of the therapeutic treatment approaches. We therefore also investigated the immune responses to the described B cell peptides after mucosal (nasal, oral) immunization. As potent mucosal adjuvant, which is also approved for human use in one of the existing oral cholera vaccines, the cholera toxin B subunit was used. Indeed, our preliminary data indicate that the mucosal vaccination approach gave rise to peptide/Her-2 specific antibody levels in serum along with mucosal specific IgA antibody levels. Concomitant induction of IgG and IgA might be a promising new concept for certain cancer vaccines, as it has been recently shown that therapeutic IgA and IgG antibodies target different effector cells engaged in tumor cell killing; importantly the superiority of FcαRI induced neutrophil-mediated tumor cell killing has been recently demonstrated for several TAAs, including Her-2/neu [111]. Thus, further in vitro and in vivo studies on the anti-tumor efficacy of mucosal vaccination are warranted to advance the concept of a change of application route in certain tumor entities.

Concluding remarks

Increasing knowledge about immunosuppressive evasion mechanisms in the microenvironment of tumors explains the difficulties in the development of counteracting strategies and suboptimal efficacy of anti-tumor vaccines in comparison to vaccines against infections.

One of the most investigated tumor antigens in breast cancer is Her-2/neu. Among the several delivery systems for this antigen the generation of peptide vaccines is the most simple and cost effective method. The more complex DNA vaccines encoding Her-2/neu peptides are, so far, the ones least developed for clinical application. The most widely tested peptide vaccines in clinical settings are T cell epitope vaccines, which however present certain limitations, mainly due to their HLA restriction. It is thus an obvious advantage of B cell peptide vaccines that the lack of HLA restriction allows a broad use among cancer patients. Nevertheless, further clinical trials (phase II and later phase III studies) will have to prove their clinical efficacy and also their advantage over the current use of monoclonal antibodies due to reduced application and longevity of immune responses.

Ideally, therapeutic vaccines should be used as early as possible in disease to prevent recurrence and dissemination of tumors, eventually even leading to prophylactic application in risk patients. A future aspect of vaccine application may lie not only in the systemic way but also via the mucosal route particularly in tumors at mucosal surfaces, such as gastric cancers.