**4. Lack of standards: Disadvantage or opportunity for further improvement?**

The 'dendritic cell therapy for cancer'-world today is still characterized by a large variety of similar but not identical approaches. Even the definition of dendritic cell can slightly vary according to their progenitors with different resulting markers on their surface. Although direct harvesting from the peripheral blood is possible, some DC are differentiated out of stem cells or cord blood, but for the vast majority of clinical grade DC today, monocytes are harvested out of the peripheral blood and differentiated into DC. Different culture protocols and conditions result in different phenotypes, but the minimal criteria should be respected before one can claim the cells to be dendritic cells for clinical use: they should display clear cytoplasmic veils, have a high expression of MHC class II molecules as well as costimulatory molecules like CD86 and have lost their 'monocyte' markers like CD14. Growing consensus is being reached about the mature DC being the preferred state of the cells to re-inject into the patients, rather than immature cells being able to tolerize rather than immunize the patient. Several maturation cocktails are being used, none of them however have been proven to be superior to the other variants in clinical use, although some evidence exists for the critical involvement of TNFα in the cocktail(25). As the serum of cancer patients might contain identified and unidentified immune suppressive agents inhibiting a good DC differentiation in ex vivo cultures, some favor the use of serum free culture conditions.

Dendritic cells should be loaded with relevant glioma-associated antigens. The sources of these antigens differ widely from well-defined, possibly acid-eluted peptides, proteins, whole tumor cell lysates and homogenates, total tumor RNA, vector constructs, apoptotic

Immunotherapy for Malignant Gliomas: A Roadmap for the Future 419

Not only the state of the disease, but also the patient's age and possible timing of vaccination in the course of the treatment are relevant items to consider in the global treatment paradigm using DC-based vaccinations. Indeed, like in many types of oncological treatments, younger age seems to be correlated to a better response and survival after tumor vaccination. This can easily be understood as basic immunological features like thymic involution over time are very likely to influence the potential of active specific immunotherapy approaches in general. As for timing of the implementation of immunotherapy, the question remains whether an upfront integration of the DC-based vaccines in the standard postoperative radiochemotherapy regimen(27) will result in a netto benefit over the alternative approach to apply DC-based tumor vaccination after reoperation for recurrent disease as a single postoperative adjuvant treatment modality without interference of e.g. chemotherapy. Even more basic questions like the best frequency, dose, administration route and boost vaccine frequency and content are unanswered as yet. We do know that in DC-based tumor vaccination, no dose-response, nor dose-toxicity phenomenon is involved, but the optimal 'pharmacological' vaccine

characteristics can only be concluded after further comparative clinical data.

itself.

**and the patient** 

Although clinical trials with the current standards of tumor vaccines are mandatory to define the presumed position of cancer immunotherapy in the global oncological treatment regimens, we should realize ourselves that further improvements of the products itself will continue to be made in the next decade. This is of utmost importance to estimate the future potentials of the therapy without losing credits for further innovations of the technology

**5. How to learn more? A difficult interplay between the technique, the tumor** 

Three main areas of elucidation arise at the moment. First of all, intrinsic improvements of the vaccine details itself are to be monitored by advanced immune monitoring techniques(28). Apart from the classical monitoring assays aimed at detecting a specific anti-tumoral immune response, we should move towards a more global appreciation of the immune system and changes under therapy, both conventional and vaccination therapy. Indeed, preliminary evidence is emerging about the importance of the global immune status of the patient at time of diagnosis and treatment steps. Radiochemotherapy leads to a 're-setting' of the patient's immune system, possibly already then priming it towards a favorable or unfavorable starting position for subsequent immunotherapy. The vaccine itself might cause quantitative and qualitative shifts in immune cell subtypes, not only regulatory or effector T cells, but also natural killer cells or natural killer T cells. We should start to include monitoring of local immune cells in the target environment, i.e. the brain and brain tumor micro-environment, rather than only in the blood, which is often not (at all) correlating with the local immune conditions in the target organs. Probably coinciding with shifts in cellular compartments of the immune system, cytokine environments might mimic the underlying changes and are good candidates to represent relevant switches in micro-environment facilitating or suppressing an effective tumor rejection by the immune system. Usually, a Th-1 proinflammatory environment is believed to be beneficial for tumor rejection, although other

and necrotic bodies. The same lack of data about which source of antigen leads to the most effective vaccine construct is blurring a uniformal approach today. However, many theoretical considerations, could rather support the use of whole tumor cell antigens rather than well-defined single peptides. The most important argument against the exclusive use of a single peptide is the well-established phenomenon of selecting antigen-loss variants of the tumor, leading to an inevitable tumor escape of probably less immune susceptible tumor clones after vaccination. Moreover, the broad repertoire of whole tumor cell derived tumor associated antigens, will lead to presentation of the relevant processed antigens in any type of HLA constitution of the patient, both in an MCH class I and II context. An efficient immunization in these cases is more prone to result in a comprehensive polyclonal T and B cell activation able to provide a better coverage of different tumor clones of the intrinsic heterogenous malignant glioma cells. Nevertheless, further improvements might be expected if a more immunogenic apoptosis pathway could be used to create the source of tumor antigens. Nanoparticles could maybe improve uptake and processing efficiency in the DC and justify further investigation.

Even harder to estimate today, is the optimal use of the appropriate danger signals (PAMP's or pathogen associated molecular patterns or DAMP's being danger associated molecular patterns) to further increase the potency of the vaccine preparation(26). Several candidates, most of them being Toll-like receptor agonists, could substantially potentiate the clinical impact of the dendritic cell-based vaccine. To date, only preliminary data exist to support the use of e.g. imiquimod , poly I:C or clinical grade LPS in the clinical applications of DC vaccines. Several others like ssRNA even have to start being explored for this application. It is evident that with so many parameters being undecided in terms of the most clinically potent DC-based vaccination approach, that a large spectrum of DC-based vaccines result, going from the crude 'dirty' vaccines to the highly elaborated, often genetically engineered cell constructs.

The lack of standards is not only to be found in the vaccine production arm, but even in the target population of patients to treat with DC-based vaccines. Most immunotherapy trials, not only those in brain cancer patients, have been performed in end-stage patients, often heavily pre-treated. Anergic states, either induced by myelo-ablative regimens or by the advanced state of the cancer itself, compromise the theoretical potential of tumor vaccines. Mounting evidence exists nowadays, that DC-based cancer vaccines should be applied to patients with a minimal residual disease status rather than in end-stage cancer patients. Some researchers even advocate the use of cancer vaccines in pre-cancerous lesions, as such referring to the historical prophylactic nature of vaccines and the abundant evidence of the prophylactic efficacy of tumor vaccines in rodent models. In terms of glioma patients, this would imply a shift of the focus to patients with low grade gliomas, as these tumors tend to dedifferentiate over time to secondary high grade gliomas. It would however be very hard, if possible at all in this state of knowledge, to show any survival benefit of this approach in low grade gliomas as the natural course of this condition is so variable that one would need very extensive patient numbers to reach an adequate statistical power in any trial design. Malignant gliomas on the other hand, have the disputable statistical advantage of being rapidly progressive lesions with an almost invariable fatal outcome: a substantial deviation from this natural evolution in vaccinated patients will rapidly result in a broad clinical awareness and possible recognition of the value of cancer vaccines in this setting.

and necrotic bodies. The same lack of data about which source of antigen leads to the most effective vaccine construct is blurring a uniformal approach today. However, many theoretical considerations, could rather support the use of whole tumor cell antigens rather than well-defined single peptides. The most important argument against the exclusive use of a single peptide is the well-established phenomenon of selecting antigen-loss variants of the tumor, leading to an inevitable tumor escape of probably less immune susceptible tumor clones after vaccination. Moreover, the broad repertoire of whole tumor cell derived tumor associated antigens, will lead to presentation of the relevant processed antigens in any type of HLA constitution of the patient, both in an MCH class I and II context. An efficient immunization in these cases is more prone to result in a comprehensive polyclonal T and B cell activation able to provide a better coverage of different tumor clones of the intrinsic heterogenous malignant glioma cells. Nevertheless, further improvements might be expected if a more immunogenic apoptosis pathway could be used to create the source of tumor antigens. Nanoparticles could maybe improve uptake and processing efficiency in the

Even harder to estimate today, is the optimal use of the appropriate danger signals (PAMP's or pathogen associated molecular patterns or DAMP's being danger associated molecular patterns) to further increase the potency of the vaccine preparation(26). Several candidates, most of them being Toll-like receptor agonists, could substantially potentiate the clinical impact of the dendritic cell-based vaccine. To date, only preliminary data exist to support the use of e.g. imiquimod , poly I:C or clinical grade LPS in the clinical applications of DC vaccines. Several others like ssRNA even have to start being explored for this application. It is evident that with so many parameters being undecided in terms of the most clinically potent DC-based vaccination approach, that a large spectrum of DC-based vaccines result, going from the crude 'dirty' vaccines to the highly elaborated, often genetically engineered

The lack of standards is not only to be found in the vaccine production arm, but even in the target population of patients to treat with DC-based vaccines. Most immunotherapy trials, not only those in brain cancer patients, have been performed in end-stage patients, often heavily pre-treated. Anergic states, either induced by myelo-ablative regimens or by the advanced state of the cancer itself, compromise the theoretical potential of tumor vaccines. Mounting evidence exists nowadays, that DC-based cancer vaccines should be applied to patients with a minimal residual disease status rather than in end-stage cancer patients. Some researchers even advocate the use of cancer vaccines in pre-cancerous lesions, as such referring to the historical prophylactic nature of vaccines and the abundant evidence of the prophylactic efficacy of tumor vaccines in rodent models. In terms of glioma patients, this would imply a shift of the focus to patients with low grade gliomas, as these tumors tend to dedifferentiate over time to secondary high grade gliomas. It would however be very hard, if possible at all in this state of knowledge, to show any survival benefit of this approach in low grade gliomas as the natural course of this condition is so variable that one would need very extensive patient numbers to reach an adequate statistical power in any trial design. Malignant gliomas on the other hand, have the disputable statistical advantage of being rapidly progressive lesions with an almost invariable fatal outcome: a substantial deviation from this natural evolution in vaccinated patients will rapidly result in a broad clinical awareness and possible

recognition of the value of cancer vaccines in this setting.

DC and justify further investigation.

cell constructs.

Not only the state of the disease, but also the patient's age and possible timing of vaccination in the course of the treatment are relevant items to consider in the global treatment paradigm using DC-based vaccinations. Indeed, like in many types of oncological treatments, younger age seems to be correlated to a better response and survival after tumor vaccination. This can easily be understood as basic immunological features like thymic involution over time are very likely to influence the potential of active specific immunotherapy approaches in general. As for timing of the implementation of immunotherapy, the question remains whether an upfront integration of the DC-based vaccines in the standard postoperative radiochemotherapy regimen(27) will result in a netto benefit over the alternative approach to apply DC-based tumor vaccination after reoperation for recurrent disease as a single postoperative adjuvant treatment modality without interference of e.g. chemotherapy. Even more basic questions like the best frequency, dose, administration route and boost vaccine frequency and content are unanswered as yet. We do know that in DC-based tumor vaccination, no dose-response, nor dose-toxicity phenomenon is involved, but the optimal 'pharmacological' vaccine characteristics can only be concluded after further comparative clinical data.

Although clinical trials with the current standards of tumor vaccines are mandatory to define the presumed position of cancer immunotherapy in the global oncological treatment regimens, we should realize ourselves that further improvements of the products itself will continue to be made in the next decade. This is of utmost importance to estimate the future potentials of the therapy without losing credits for further innovations of the technology itself.
