Cancer immunity and clinical oncology - Strategies to improve immunotherapy in cancer

Cancer immunity and clinical oncology - Strategies to improve immunotherapy in cancer


The ideal antigen for immunotherapy is exclusively expressed by tumour cells and is highly immunogenic, such as the CT-antigen NY-ESO-1. In addition, the majority, or even better, all cells of the tumour should express the antigen and present its epitopes on their surface in the context of MHC molecules in sufficient amounts.

Multiple injections of the entire antigenic protein in a free form, as long peptides, as mRNAs, or as cDNA encoding the antigen are considered a useful approach, as all forms will preferentially be taken up and presented by dendritic cells. More efficiently, the antigen can be coupled to substances targeting dendritic cells (e.g. anti-CD205 antibodies). Independent of the form of the antigen, it must be administered together with an appropriate signal to induce dendritic cell maturation.

However, which signals are needed for appropriate dendritic cell maturation in vivo is not yet fully understood. Because the immune system developed to fight pathogens and is therefore efficiently activated by those, vaccination with recombinant (attenuated) pathogens that express tumour antigens may be superior to vaccination with recombinant protein, peptides, or tumour cell lysates.

The amount and the biophysical nature of the antigen determine the duration of the antigenic stimulus and there is evidence that antigen persistence is required for the development and maintenance of effector function by responding T cells. This may require multiple injections in most cases, as most vaccines will be short lived in vivo.

The relatively short half-life of endogenous dendritic cells (estimated to be <1 week) also argues for multiple vaccinations. If recombinant pathogens are used as a vaccine, it may be necessary to avoid neutralizing antibody activity to use another pathogen for each consecutive immunization (so-called prime–boost protocols), such as recombinant poxvirus, adenovirus, and herpesvirus or modified strains of salmonella or mycobacteria after cancer antigen encoding DNA transfection.

T cells and modulation of T cell function by antibodies

The activation of both naive and antigen-experienced T cells requires multiple signals and improving the quality of each of these signals may contribute to a more efficient response to persisting antigens such as tumours or some pathogens.

The quality of signal 1, the interaction between the TCR and the peptide/MHC complex, is determined by ligand density and TCR avidity. The current opinion is that this interaction must have a minimal duration and strength for adequate T-cell activation. As high-avidity T cells may have been purged from the repertoire of tumour-specific T cells, increasing the numbers of peptide/MHC complexes per APC may be considered during vaccine development.

This may be achieved by targeting the relevant antigen to the compartments in which MHC class I- or class II-loading takes place. Signal 2 is delivered to the responding T cell as the net effect of costimulatory (CD80/CD28, CD86/CD28, CD70/CD27, B7RP1/ICOS, 4-1BBL/4-1BB, OX40L/OX40) and of coinhibitory (CD80/CTLA4, CD86/CTLA4, B7H1/PD-1, B7DC/PD-1) interactions that are mediated by surface molecules on the APC and T cell respectively.

Costimulatory and coinhibitory molecules belong to the B7 or the TNF family and the number of known costimulatory and coinhibitory molecules is still increasing. Interference with these signals such that costimulation is increased and/or coinhibition is diminished is a promising approach for enhancing the magnitude and the quality of the tumour-specific T cell response. Diminishing coinhibition is presumably the better option, as mature dendritic cells usually express sufficient levels of costimulatory molecules.

Nevertheless, positive effects of agonistic antibodies to 4-1BB and OX40 on tumour immunity were reported. PD-1 was shown to be important in peripheral tolerance induction and accordingly, PD-1-deficient mice spontaneously develop autoimmunity. In addition, it was shown that anergy of specific CD8+ T cells as a result of overwhelming antigen could be prevented by blocking PD-1/PD-1L interactions in mice and in humans. In cancer patients, the expression of PD-1L in the tumour or the expression of PD-1 by tumour-infiltrating lymphocytes was found to correlate with poor prognosis.

The interaction of CD152 (CTLA4) with its ligands CD80 (B7.1) and CD86 (B7.2) has been studied more extensively in the context of tumour-specific immunity. CTLA4 is transiently expressed on T cells upon activation and plays a critical role in down-regulation of T cell responses as well as in peripheral T-cell tolerance, which is illustrated by the lethal generalized autoimmunity in CTLA4-deficient mice.

Treatment with anti-CTLA4 antibodies resulted in rejection of established tumours in different mouse models, but in order to be effective in the poorly immunogenic B16 melanoma model it had to be combined with GM-CSF treatment. The beneficial effect in experimental systems encouraged the use of anti-CTLA4 treatment as a single agent or together with therapeutic vaccination in patients with melanoma.

Major clinical responses were seen in these trials, also in patients with advanced disease. It has been suggested that anti-CTLA4 treatment acts through selective inhibition of Treg, which constitutively express CTLA4. This idea was challenged recently: First, it was shown that CTLA4 blockade had no impact on suppressive function in vitro.

Second, Allison and co-workers used anti-CTLA4 and GM-CSF treatment in the B16 rejection model and elegantly showed that anti-CTLA4 had no effect on the number or function of tumour-infiltrating Tregs, but that instead the number of effectors in the tumour greatly increased resulting in an higher ratio of effector T cells to Tregs, which may explain the beneficial effect of anti-CTLA4/GM-CSF treatment.

Antigen-presenting cells (APCs)

Mature or activated dendritic cells have the unique capacity to activate naive CD4+ and CD8+ T cells and to reactivate memory CD8+ T cells in vivo. On the other hand, steady-state dendritic cells have been shown to induce peripheral tolerance in both CD4+ and CD8+ T cells. Thus, it is crucial that the vaccine is presented to T cells by mature, activated dendritic cells and that presentation by other cell types or by steady-state dendritic cells is avoided as much as possible.

In addition and ideally, the vaccine should overcome peripheral tolerance that may have been induced by the tumour. In principle, two approaches may be considered:

  1. loading of in vitro generated autologous dendritic cells with antigen (peptide, protein, tumour lysate, lentivirus, or nucleic acids encoding tumour antigens) followed by administration of dendritic cells to the patient;
  2. in vivo targeting of the vaccine to dendritic cells with a simultaneous dendritic cell-maturation signal.

The major drawback for the therapeutic injection of antigen-presenting dendritic cells is their limited migration. Most dendritic cells were found to remain at the injection site and less than 1% were homing to draining lymphoid tissue. Nevertheless, many studies documented successful induction of immunity against viral and tumour antigens in healthy volunteers and in patients, respectively, although without significant clinical effects.

A more promising alternative is targeting of antigen to dendritic cells in vivo, which circumvents problems of migration. Injection of proteins or long peptides together with adjuvants is presumably superior to the injection of short peptide epitopes, as dendritic cells are the only cells that efficiently take up and cross-present proteins or long peptides, and their in vivo half-life is probably longer.

However, the process of cross-presentation is rather inefficient and may require the injection of relatively large amounts of antigen. Efficient targeting of antigen to dendritic cells can be achieved by coupling the antigen to an antibody specific to endocytic receptors on dendritic cells:

It has been shown that antigens coupled to anti-DEC205 (CD205) potently induce CD4+ and CD8+ T cells in murine model systems and in humans when given together with a dendritic cell-maturation stimulus. The use of modified and attenuated pathogens as vectors may combine targeting to dendritic cells and delivery of correct maturation stimuli because many, if not all, pathogens infect dendritic cells or are efficiently taken up by them and because they provide ‘danger’ signals for dendritic cell maturation.

Many vectors, including adenovirus, herpesvirus, and different poxviruses have been shown to efficiently induce the desired responses in patients. Pre-existing immunity to the vector due to previous immunization or infection, and safety issues especially in immunocompromised patients, limit the choice of viral vectors.

Cancer immunity and clinical oncology - Concluding remarks

Significant advances have been made in the field of cancer immunotherapy over the last decade. The most important step forward has been the identification of tumour antigens with immunogenic potential. The utilization and improvement of cancer immunotherapy in the clinic demands carefully conducted and coordinated but discovery-oriented translational research in the form of clinical trials that include thorough monitoring of immune and clinical responses.

As selective outgrowth of antigen-loss variants due to immunoediting may occur during vaccination of cancer patients, antigen expression patterns of the tumour and metastases should be monitored during immunotherapy, if possible, followed by surgical resection of antigen-negative tumours.

At this rather early stage of development cancer immunotherapy should be offered to cancer patients only within carefully monitored clinical trials of experienced clinical research teams. In addition, it may be rewarding to include patients with early-stage disease in immunotherapy trials, because immunoediting of the tumour and subversion of the immune response by the tumour is presumably less pronounced in those patients.

The keys to efficient cancer vaccination are:

  1. availability of an antigen known to be a strong immunogen in cancer patients;
  2. the route, schedule, and packaging of antigen to induce an optimal immune response in vivo;
  3. combination of antigen with an adjuvant, to support the induction of a strong immune response in vivo;
  4. long-term immunizations, and
  5. inhibition of regulatory signals, such as CTLA-4, Tregs, and PD-1.