Therapeutic vaccines in cancer: moving from immunomonitoring to immunoguiding

Therapeutic vaccines for the treatment of cancer aim to induce or boost T-cell immunity against tumor-specific or tumor-associated antigens when it is has failed to develop naturally or insufficiently to control the outgrowing tumor. There have been multiple approaches to develop such vaccines, including the use of protein or peptides, together with adjuvants, naked DNA, RNA and viral or bacterial vectors encoding specific tumor antigens. In the last 20 years, more than 200 clinical trials of different antitumor vaccines aiming to induce tumor-specific T-cell immunity in cancer patients have been reported [1]. Most of these trials have shown evidence of immunogenicity and safety but clinical efficacy, in general, is low [1,2], which is one of the reasons that a direct correlation between clinical efficacy and T-cell reactivity is hard to establish [3].

However, it has become clear from animal models that both tumor-specific CD8⁺ and CD4⁺ T cells play an important role in the control and regression of tumors [4,5]. It is important to realize that in all these animal models the high success rate is associated with a relatively short disease-period in which disease-mediated immune suppression does not occur and therefore, these animal models hardly mimic the situation found in patients with cancer. Importantly, mouse models of sporadic cancer show that immunogenic tumors induce T-cell tolerance during tumor development [6]. Correspondingly, human tumors develop slowly over time and establish themselves in the face of immunity. Our own work on human papillomavirus-induced cervical cancer demonstrates that human immunogenic tumors critically influence the type and polarization of the tumor-specific T cells that are induced spontaneously following the presentation of tumor-derived immunogenic epitopes. In these patients, the systemic immune response against the tumor-specific viral antigens, E6 and E7, is either not detectable or exhibits a noninflammatory cytokine profile [7]. Furthermore, T cells isolated from cervical tumors and tumor-draining lymph nodes contained E6- and E7-specific CD4⁺ T-helper cells [8], as well as E6- and E7-specific CD4⁺ regulatory T cells, which were able to suppress proliferation and cytokine production (IFN-γ and IL-2) of both naive T cells and Th1 cells upon recognition of their cognate antigen [8,9]. Similarly, wrongly polarized tumor-specific CD4⁺ T cells have been observed in colorectal cancer [10,11] and oral squamous cell carcinoma [12], whereas in melanoma, regulatory T cells responding to the selfantigens, LAGE-1 and ARCT-1, were found [13–15]. This indicates that it is of great importance to fully characterize the spontaneous immune response to the tumor antigen of choice, because the presence of wrongly polarized T cells, and even tumor-specific regulatory T cells in the peritumoral milieu, may thwart the induction of an effective antitumor response in an antigen-specific manner and, as such, lower vaccine efficacy.

Immunomonitoring

Based on the concept established in animal models that T cells are key in the combat against tumors, cancer immunotherapy trials include measures of the number of subjects that mount a vaccine-induced T-cell response, as well as the strength of a detected T-cell response as surrogate markers for vaccine efficacy. A plethora of different immunological assays to monitor antigen-specific T-cell responses are available. Prime examples are the ELISPOT assay [16–18], intracellular cytokine staining [19,20], MHC–peptide multimer staining [21] and proliferation assays. Although there has been substantial effort made in setting up these cellular immune assays, the lack of assay standardization and validation has caused a high variability in the results, revealing yet another factor that contributes to the failure to show a correlation between immune responses and clinical end points. Harmonization of the use of popular immune assays, converting them from capricious measurements into reliable immune surrogates for clinical trials, could substantially accelerate the development of immunotherapeutic strategies. Harmonization aims to identify all sources of variation that influence the outcome of an applied assay, make crucial choices to prevent this and standardize those choices across laboratories throughout the entire field for optimal performance and consistent results. Fortunately, there is an increasing awareness of the necessity of such harmonization and assay validation [22–25]. Over the last 2 years, two nonprofit organizations in the cancer immunotherapy field, the Cancer Vaccine Consortium of the Sabin Vaccine Institute and the Cancer Immunotherapy Monitoring Panel, have initiated large international Proficiency panels [26,27], which served not only as validation programs for their participants, but also provided the first guidelines for the standardization of assays in order to become reliable and widely acceptable tools for monitoring of immune responses in clinical trials.

Evidently, the number and type of assays and samples used are determined by logistics. In most cases, a number of only relatively small (50 ml) blood samples are taken because patients have no problems consenting to this and because it is easy to isolate peripheral blood cells and store them for later studies. Therefore, most immunomonitoring efforts are entirely focused on the measurement of a vaccine’s capacity to induce a systemic T-cell response. The choice of assay is fairly simple, albeit dependent on the vaccine type tested. For instance, if the vaccine used consists of peptides that represent a CD8⁺ T-cell epitope, as is the case in many immunotherapeutic cancer trials, immunomonitoring will be focused on the detection of CD8⁺ T cells specific for this epitope only. In such a case, one assay would be enough; however, it is recommended that at least two validated assays are used in parallel – preferably one structural and one functional assay [27–30]. Vaccines that potentially trigger CD4⁺ and CD8⁺ T cells to a broad number of epitopes may require additional assays but the depth of monitoring will be limited by the size of the samples. While this approach is sufficient to answer the immunological objectives of vaccine trials, it will only provide limited insight with respect to our understanding of why vaccines fail to induce T-cell responses in patients or, presumably more importantly, why the systemically present vaccine-induced T cells fail to mediate clinical impact. In general, no direct correlations between vaccine-induced immunity and clinical effects on premalignancies [31–33] or cancers are found [1,34,35]. Importantly, assays that can screen multiple parameters are currently being set up and applied for the measurement of vaccine-induced immune responses. This new generation of assays bear the promise to become a future surrogate for a clinical correlate, as was recently demonstrated by protection against leishmaniasis in a mouse model [36]. It may well be that assays already available and running in the laboratories could function as a surrogate marker for clinical efficacy if conducted in a validated way and used in a wise combination. A recent report showing that not the appearance of a delayed-type hypersensitivity reaction per se, but only the detection of tumor-specific T cells in this skin test site – assayed by taking a biopsy from the delayed-type hypersensitivity reaction site – correlated with progression-free survival of vaccinated patients serves as a perfect example [37].

Immunoguiding

As mentioned previously, human cancers develop over time, and during this development they undergo a process that is called immunoediting, reflected by the escape of cancer cells through mechanisms that not only include the upregulation of intrinsic mechanisms to resist immune attack and the downregulation of HLA molecules at the cell surface, but also by mechanisms that regulate the immunological make-up of antitumor T-cell responses [38]. While the most effective antitumor response is believed to consist of a CD4⁺ Th1 (IFN-γ, TNF-α and IL-2) and a CD8⁺ cytotoxic T-cell response, spontaneously induced antitumor responses can be wrongly polarized or even of a regulatory T-cell type [9,13,14,39,40]. As a result, tumor-specific T cells may not be able to infiltrate tumors [41,42] and, in the case that T cells do infiltrate tumor masses, this response may consist of ineffective tumor-specific T cells and even, immunesuppressive T cells [42–49]. Notably, most therapeutic vaccines for the immunotherapy of cancer are designed to enhance CD4⁺ and CD8⁺ T-cell effector immunity against the same antigens that are recognized by T cells previously found to be induced by these tumors. This means it is possible that vaccination might also – or instead – result in activation and expansion of these wrongly polarized or suppressive pre-existing tumor-specific T cells. The effect of this on the clinical efficacy of the vaccine is unknown and could differ from patient to patient, but may take extreme forms. This is illustrated in an elegant mouse study, which showed that vaccination of tumor-bearing hosts harboring a mixture of antigen-specific responder and regulatory T cells resulted in concomitant expansion of both T-cell populations. Importantly, the net effect of vaccination in this setting was the failure of the antitumor immune response [40]. The efficacy of such adaptive regulatory T cells depends on the balance between antigen-specific T-helper cells and regulatory T cells because in mice, where antigen-specific Th1 cells prevailed, suppression was overcome [50]. A small Phase II vaccination trial in patients with cervical cancer, not designed to measure clinical efficacy, revealed that, while the vaccine-induced responses were dominated by effector-type CD4⁺CD25⁺FOXP3^- type 1 cytokine IFN-γ-producing T cells, there was also an expansion of nonpolarized T cells, as well as CD4⁺FOXP3⁺ T cells following vaccination [51]. In two cases with progressive disease, tumor-specific Th1 cells did not outnumber regulatory T cells [51]. Moreover, it has been clearly shown that the presence of a high number of circulating effector-type T cells is not enough to induce clinical efficacy [31,32,52–54].

Another important aspect is to get these T cells to migrate to the tumor and to work within this microenvironment [55–57]. Clearly, studies on spontaneously- and vaccine-induced tumor-specific immunity show that this is not always guaranteed [43] and may be determined by the status of the tumor microenvironment [33,58–61]. Apparently, the clinical efficacy of a therapeutic cancer vaccine is determined by many different immunological issues, of which single determinants do not necessarily play a role in each and every patient. Therefore, it will not be enough to measure only the presence and magnitude of the desired T-cell response, but it will be important to ensure that immunomonitoring encompasses the parallel use of several tests, which together assess the presence, magnitude and function of all the subsets of T cells expected to respond (desired or unwanted), not only in the circulation, but also within the target tissue. Although the logistical problems associated with this highly demanding process to obtain the necessary materials are numerous, it can be carried out in those settings where the monitoring laboratory and clinic are in close collaboration and where the the underlying reasons for analysing these different samples are understood. In the end, the data sets generated may be helpful in explaining for each individual patient why vaccination did or did not work and, consequently, point to weaknesses in the vaccine strategy. It is here where the transition from immunomonitoring to immunoguiding may occur. Immunoguiding is the process where the results obtained by immunomonitoring are systematically used to understand the strengths and weaknesses of a vaccine and where these results have a strong impact on decisions made with respect to developmental aspects of the vaccine, as well as on the design of the immunotherapeutic treatment. For example, the observation of immunogenic competition between two co-injected antigens was the rationale for us to separate them in a subsequent trial [62]. A good understanding of the effects of a vaccine on the patient’s immune system may encourage one to move forward into bigger randomized trials or go one step back to optimize the vaccine (e.g., addition of strong adjuvants) or strategy used (e.g., change the dose and schedule and depletion of regulatory T cells) followed by immunomonitoring. Although data from immunomonitoring – at its current developmental stage – still does not allow predictions to be made about clinical end points (e.g., survival, time to progression and progression-free survival), the data sets obtained by proper immunomonitoring will give a good idea about the positive and negative mechanisms that play a role when a specific vaccine is used, and this allows a good assessment of the financial risks associated with testing a vaccine in Phase III trials [63]. Altogether, it will be inevitable that immunoguiding will become an important aspect in the development of cancer vaccines and it is highly likely that the first real, clinically successful, immunotherapeutic vaccine strategies have been advanced in this way.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.