We described our results from two gene therapy trials in patients with recurrent or newly diagnosed malignant glioma, UPCI 95-033 and 99-111, respectively. We identified common challenges in the administration of these protocols. Both protocols utilized participants' autologous fibroblasts that had to be stably transfected with the TFG-hIL4-Neo-TK retroviral vector, and expanded to sufficient numbers. A retroviral vector was chosen because of stable expression and positive efficacy data in preclinical models [6
]. However, these processes required at least 7 to 8 weeks, which posed a major feasibility concern. Especially in the UPCI 95-033 trial, 4 of 6 participants were withdrawn from the study due to tumor re-recurrence before they could receive the first vaccine. In addition, the first patient who received the vaccines had a rapid recurrence after the surgery for tumor debulking and vaccine cell harvesting, and underwent another tumor resection before beginning vaccination [23
]. Indeed, in the literature [24
], the median time to further tumor progression for patients with recurrent malignant glioma, even with therapy, is only 8 weeks. We performed skin harvesting during the craniotomy of recurred tumor for UPCI 95-033. As craniotomy and debulking surgery are usually performed as soon as these interventions are clinically indicated, there are no practical ways to perform skin harvesting and transfection earlier. In UPCI 99-111, this issue was not as significant as in UPCI 95-033 as only one of 6 participants was withdrawn due to tumor recurrence before the first vaccine. In this trial, standard FEBRT, which typically requires 6 weeks (2 Gy/dose × 30 doses), was used following surgical resection of primary tumors, and this may have temporarily suppressed tumor growth and provided an adequate time for vaccine generation and post-irradiation delivery.
The primary endpoint of these studies was evaluation of safety, and we did not observe any regimen-limiting toxicity. Because autologous whole glioma cells were used as the source of the vaccines, there was a theoretical concern for inducing autoimmune encephalitis. Although the lack of autoimmune events in our two trials is reassuring in this regard, it does not indicate the ultimate safety of such approaches, and continued vigilance for such toxicity is warranted, particularly because only 7 evaluable participants were accrued. It has been demonstrated that the clinical benefit of immunotherapy in other types of cancers is often associated with induction of autoimmunity [25
]. On the other hand, patients with advanced cancers and compromised immune systems may not be ideal candidates for assessing either the toxicity or efficacy of therapeutic cancer vaccines [26
]. Further evaluation of these important issues is warranted in larger numbers of glioma patients receiving current, more intensive, vaccine regimens.
With regard to induction of immune response, two participants in UPCI 95-033 who received vaccines composed of IL-4-transfected fibroblasts and irradiated autologous glioma cells demonstrated either local (i.e. immune cell infiltrate at the vaccine site) or systemic (i.e. IFN-γ ELISPOT response against EphA2-derived epitope) immune responses. These observations are consistent with previous clinical trials in our institute that local production of IL-4 by gene-transfected fibroblasts at the vaccine site caused local endothelial activation and recruitment of immune effectors [27
]. On the other hand, no evidence of systemic immune response was observed in five participants in UPCI 99-111 who received IL-4-transfected fibroblasts admixed with DCs loaded with tumor-lysate. These results were somewhat surprising because we had expected that the vaccine formulation using type-1 DCs would be more potent than the formulation in UPCI 95-033. In UPCI 99-111, each participant received only two intradermal injections of 1 × 106
lysate-loaded type-1 DCs. The dose of DC and number of vaccinations in this study were probably suboptimal based on other studies demonstrating encouraging clinical activity of DC-based glioma vaccines [20
]. In these studies, higher DC numbers and repeated vaccinations (1 × 107
/injection × 3 or more vaccinnations) were employed.
In addition to the intensity of vaccine regimens, careful consideration is required for optimal administration routes of DC vaccines. In a randomized study comparing direct intra-lymph-nodal, intravenous and intradermal administration of DC vaccines in patients with metastatic melanomas [29
], the intra-lymph-nodal route was well-tolerated for up to 5 × 107
DCs/injection, and induced significantly higher levels of specific CD8+
T cells based on cytokine secretion, when compared with other routes. The feasibility, safety and efficacy of intra-lymph-nodal administration of DC-based vaccines has also been demonstrated using the activation of antigen-specific CD4+
T cell responses as an endpoint [30
]. Therefore, in our currently ongoing type-1 DC-based trials (UPCI 04-136 and 05-115), we selected ultrasound-guided intra-lymph-nodal injections of type-1 DCs.
Novel type-1 DCs were employed in UPCI 99-111, and these DCs generated from each participant demonstrated enhanced IL-12 p70 producing capability in vitro
compared to standard DCs from the same donors. These results indicate that we have established standard operation procedures for GMP-grade type-1 DCs for our ongoing and future type-1 DC-based trials. In addition, further improvement of the DC culture protocol has lead to development of a serum-free culture protocol for type-1 DCs [15
]. Type-1 DCs are expected to induce type-1 immune effector cells, including T-helper (Th)1 and type-1 cytotoxic T-lymphocytes (Tc1). Our recent studies in murine brain tumor models have demonstrated that type-1 T-cell response is particularly favorable for anti-brain tumor immunotherapy owing to high level surface expression on Tc1 cells of a type-1 chemokine receptor CXCR3 [31
] and an integrin, very late antigen-4 [32
], both of which mediate critical roles in efficient trafficking of anti-tumor T-cells to the brain tumor site. Therefore, our ongoing and future development of glioma vaccine-based immunotherapy is directed towards induction of effective and safe type-1 immunity in patients suffering from brain tumors.
Although IL-4 is typically described as an inducer of a type-2 immune response [33
], local delivery of IL-4 at an immunization site induced IL-12 production by antigen-presenting cells, and Th1-type responses in our pre-clinical tumor vaccine model [34
] as well as in an infectious disease (Leishmania major) model [35
]. Interestingly, in this Leishmania major model [35
], systemic or prolonged recombinant (r)IL-4 delivery rather induced Th2-response, suggesting bi-functional roles of IL-4 depending upon target cell types (i.e. antigen-presenting cells vs. T-cells). Indeed, in vitro
exposure of DCs to IL-4 promotes IL-12 secretion from DCs [10
We also evaluated preliminary therapeutic activity of these regimens. In two participants in UPCI 95-033 who received scheduled vaccines, temporary stabilization of disease and radiological improvement was observed for 4 to 6 months. However, the clinical course of the first patient after vaccination was complex due to transiently increased gadolinium enhancement on MRI scans at 2 to 3 months after vaccination [23
]. This transient worsening may not have been true tumor progression, but pseudo-tumor progression representing inflammatory response at the tumor site. Biopsy of the tumor would have been the ultimate modality to distinguish these two scenarios. As we currently employ more intensified vaccine regimens (discussed in the next paragraph), the distinction between true- and pseudo-tumor progression is becoming a more critical issue. Biopsy may not always be clinically justifiable in patients with recurrent glioma. The utility of MR spectroscopy and metabolic positron emission tomography (PET) imaging for differentiating tumor progression from immune-mediated inflammatory treatment effect remains to be established. In 99-111, the median time to progression after surgical resection in 5 participants who received scheduled vaccines was 6 months (ranging 4 to 10 months); and this is not superior to the current standard regimen with FEBRT and temozolomide (6.9 months) or FEBRT alone (5.0 months) [36
]. The low intensity of the vaccine regimen in the UPCI 99-111 study, with only 1 × 106
DC/injection × two injections, may have limited the manifestation of any therapeutic immune response. Future studies will need to determine whether intensification of the vaccine approach can enhance therapeutic efficacy without a counterbalancing induction of detrimental autoimmunity.
Although our analyses of IL-4 tumor vaccine models have led us to understand the critical roles of type-1 immunity for brain tumor immunotherapy, effective type-1 immunity may also be achieved by clinically more feasible modalities than IL-4 transfected fibroblasts. Indeed, our recent preclinical study has demonstrated that a Toll-like receptor-3 ligand poly-ICLC, which has been extensively evaluated for safety in patients with glioma [37
], enhances type-1 anti-glioma immunity when combined with peripheral vaccine regimens [38
]. We have also isolated and characterized HLA-A2-restricted CD8+
CTL epitopes derived from human GAAs, such as IL-13Rα2 [18
] and EphA2 [19
] to develop vaccines using "off the shelf" synthetic peptides for these epitopes. Based on these studies, we recently implemented vaccine trials evaluating safety and immunological activity of type-1 inducing DCs loaded with synthetic peptides encoding these GAA-derived CTL epitopes in HLA-A2+
patients with recurrent glioma (UPCI 04-136 and 05-115). In these trials, to date, 7 of 8 participants successfully received 4 scheduled vaccinations with 2-week intervals without tumor recurrence, suggesting improved feasibility of these novel vaccine approaches.
Our ongoing progress in basic understanding in brain tumor immunology will allow us to develop more efficient and feasible immunotherapy strategies for patients suffering from these intractable diseases.