Significant strides have been made in immunomodulatory approaches to treating patients with either primary or recurring glioblastoma. During the Central Nervous System Clinical Science Symposium “Immunotherapy for Central Nervous System Tumors: Biomarkers and Novel Data,” held on Sunday, May 31, researchers discussed three distinct approaches being investigated for this indication.
Dr. David A. Reardon
Immunotherapy in Recurrent Glioblastoma
In the first presentation, David A. Reardon, MD, of the Dana-Farber Cancer Institute, presented data from the ReACT study, which looked at rindopepimut (CDX-10) plus bevacizumab in the setting of recurrent glioblastoma (Abstract 2009).
The rindopepimut molecule, which is composed of an EGFRvIII peptide conjugated to keyhole limpet hemocyanin (KLH) and administered via intradermal injection, targets the EGFRvIII variant known to be associated with poor prognostic outcomes in patients with glioblastoma.
The rationale for studying rindopepimut in patients with relapsed EGFRvIII-positive glioblastoma derives from multiple studies demonstrating improved overall survival (OS) and progression-free survival (PFS), Dr. Reardon said.
The ReACT study enrolled patients to receive rindopepimut plus bevacizumab (BV; 36 patients) or rindopepimut plus placebo vaccine (unconjugated KLH proteins; 37 patients). All patients were EGFRvIII positive and were enrolled after a first or second recurrence of disease; patients had no prior exposure to an anti-VEGF agent. Although there is no standard of care for recurrent glioblastoma, an antiangiogenic agent was used in both groups because it may be additive to immunosuppressive therapy, Dr. Reardon said.
“There is a strong rationale for combining an immunotherapeutic with an antiangiogenic [agent],” Dr. Reardon said. “VEGF, the most important driver of tumor-associated angiogenesis, also importantly regulates immunosuppression in the environment of the tumor, so by blocking VEGF, we may also significantly enhance the immunogenicity of the immune therapeutic.”
Overall, rindopepimut was well tolerated and was associated with only minor grade 1/2 events at the injection site; rindopepimut use did not present any additional safety issues with bevacizumab use, and there were no serious adverse events in the active treatment arm, Dr. Reardon said.
Dr. Reardon noted that the ReACT study met its primary outcome of improvement in PFS at 6 months (28% for rindopepimut plus BV vs. 16% for control plus BV in the intention-to-treat population; p = 0.1163) and demonstrated effects on numerous secondary endpoints.
An analysis of radiographic response demonstrated evidence of durable response, and there was a significant reduction in corticosteroid use among patients in the rindopepimut group, Dr. Reardon said.
“This is the first immunotherapeutic to show a survival benefit for glioblastoma evaluating a randomized, placebo-controlled study in the intention-to-treat population,” Dr. Reardon said. A phase III registration has completed accrual and the first data release is expected in about 2 months.
Discussing the paper, Fabio M. Iwamoto, MD, of Columbia University Medical Center, noted that, to date, rindopepimut has demonstrated remarkably consistent and beneficial results for patients with glioblastoma.
However, Dr. Iwamoto said, two important methodologic issues with the study should be considered. First, the distribution of patients after first and second relapse was unbalanced (92% of patients were enrolled after a first relapse in the rindopepimut group, compared with 76% in the placebo arm). He noted that previous studies have suggested differences in response rates after first or second relapse.
There were also imbalances in the number of patients with surgery or gross tumor resection after their last relapse, although these differences likely did not affect outcomes, Dr. Iwamoto said.
In a second presentation, David S. Baskin, MD, of Houston Methodist Hospital, shared results from a phase II study that explored gene-mediated cytotoxic immunotherapy (GMCI) as adjuvant to surgical resection for patients with newly diagnosed glioblastoma (Abstract 2010).
For the study, an adeno-associated viral vector expressing thymidine kinase was introduced into the tumor. This was followed by oral valacyclovir. This strategy offered a two-pronged mechanism of action, according to Dr. Baskin. Thymidine kinase phosphorylates prodrug in the tumor to form a cytotoxic nucleotide analog that functions to kill tumor cells by stopping DNA replication and shutting down DNA repair.
“As the tumor dies, we believe tumor-associated antigens are released,” Dr. Baskin said. “As the [tumor-associated antigens] are released and uptake occurs in antigen-presenting cells, we also know from work in the lab that thymine kinase has separate pro-inflammatory stimulating effects, acting like a super antigen protein, particularly with IL-2 and IL-12, giving a very powerful systemic immune response.”
In a combination phase Ib (13 patients) and phase II (52 patients) study, patients were matched to historical controls used as a benchmark for outcomes. A total of 48 patients completed the study and were matched against 134 evaluable matched control patients.
In the GMCI treatment groups, patients received the vaccine and surgery on day zero, started valacyclovir on day 1, and received radiation and temozolomide starting at week 1 and temozolomide alone starting at week 8.
Overall, GMCI was well tolerated, Dr. Baskin said, and was associated with only a few possibly associated grade 3 events (one each of insomnia, speech impairment, wound complication, and headache; two each of hyponatremia and lymphopenia; and three cases of ALT/AST elevation) and a single grade 4 event (neuropathy-motor event).
There was a 27% relative improvement in OS and a 25% relative improvement in PFS among patients in the GMCI study group, Dr. Baskin said. The intervention appeared to be particularly beneficial when combined with gross total resection, improving median OS from 16.9 months to 25.0 months.
John Sampson, MD, PhD, of the Preston Robert Tisch Brain Tumor Center, Duke University, lauded the investigators for demonstrating a positive benefit in OS and PFS.
“I think the greatest strength of the study is the overall survival and progression-free survival where there is a very convincing difference between [patients receiving the] study drug and the control group,” Dr. Sampson said.
As well, he said, the use of a preplanned match, although not a substitute for randomization, provided a worthy comparator—and perhaps something that should be considered for future clinical trials—in what would otherwise be a single-arm study.
The study raised some questions about standard of care for patients with newly diagnosed glioblastoma, according to Dr. Sampson. For example, radiation therapy ablates the solid mass tumor, but it also upregulates a number of surface proteins on the tumor, thus setting the stage for continued systemic immune activity—a phenomenon known as the abscopal effect. The approach to attacking glioblastoma used in the current study paradigm may offer a mechanism to parse out whether this is actually occurring and how beneficial it is, Dr. Sampson said.
The use of temozolomide is also intriguing, Dr. Sampson added, because it functions to kill T cells. A vaccine offered at the nadir of the apoptotic cycle may stimulate homeostatic reproduction of T cells and, over time, the continued use of vaccination may create an environment where vaccine-induced T cells proliferate and fill the void created by the death of T cells secondary to chemotherapy.
Finally, Dr. Sampson said that future immunotherapy studies will ideally employ reliable immune monitoring assays to help understand exactly where immune responses occur (i.e., locally at the site of tumor versus in the lymph node) and the signaling pathways involved in communication with the brain; greater understanding of these could supply targets for future interventions and mechanisms to boost the response.
Dr. Orin Bloch
Heat Shock Protein-based Vaccine
Orin Bloch, MD, of Northwestern University, presented an alternative approach to immunotherapy for patients with newly diagnosed glioblastoma (Abstract 2011). Dr. Bloch is part of a research effort looking at the use of heat shock proteins (HSPs) for antigen delivery.
“If you extract heat shock proteins with their bound antigens and deliver them in naked form into the systemic circulation, there is uptake into antigen-presenting cells through the CD-91 receptor, where the peptide is dissociated, cleaved, and cross-presented on MHC class 1 and 2 [peptides] for stimulation of both the CD8- and CD4-positive T-cell response,” Dr. Bloch said.
“Additionally, these heat shock proteins can stimulate toll-like receptors, and actually activate proinflammatory cytokine release, so they act as their own adjuvant and give you both innate immunity activity, as well as adaptive immune activity,” he said.
In the study, 46 patients with newly diagnosed glioblastoma from eight centers were enrolled to undergo surgery (90% resection) followed by a course of radiotherapy; patients with stable disease on MRI were offered biweekly vaccination until tumor depletion or progression.
In an intention-to-treat analysis, median OS (the primary endpoint) was 23.8 months and PFS was 18.0 months, both of which were extremely favorable compared with historical findings, Dr. Bloch said.
There were no serious vaccine-associated grade 3/4 events, and most adverse events were related to localized injection site reactions, Dr. Bloch said.
In further analysis, the study investigators noted that survival was significantly higher among patients with low levels of circulating PD-L1, defined as 54.5% expression or lower, compared with patients with high levels (OS 44.7 vs. 18.0 months, respectively; hazard ratio 4.0, 95% CI [1.4, 12.7]; p = 0.008).
According to Dr. Bloch, PD-L stimulation induces T-cell apoptosis and the PD-L pathway thus limits overexpression of the immune system; patients with glioblastoma tend to have low levels of systemic T cells and high circulating PD-L levels.
“This suggests there is a role for checkpoint inhibition, even on top of vaccination, to modulate … patients [with high PD-L1 expression] and perhaps push their survival towards that of the low PD-L1 group,” Dr. Bloch said.
In discussing the abstract, Amy Heimberger, MD, of MD Anderson Cancer Center, noted that PD-L1 expression may be important, but might not be an appropriate biomarker for risk stratification when used alone. She suggested that it might be necessary to monitor for PD-1 in addition to its ligand to predict which patients might benefit from checkpoint inhibition.
“You have to have the ligand and target—PD-1 and PD-L1—for risk stratification,” Dr. Heimberger said.
Watch the session: Visit the ASCO Virtual Meeting website.