Multiple whole tumor cell, tumor cell lysate, and tumor cell RNA vaccines, delivered with various adjuvants or via dendritic cells (DCs), have previously been tested in the clinic with moderate and sometimes interesting results. In a recent article published in Science Translational Medicine, Tanyi et al. report results from a pilot study with a new whole-cell tumor lysate dendritic cell vaccine and use the window of neoantigen responses to evaluate immunological activity.
Patients selected for this trial had recurrent advanced epithelial ovarian cancer, had been previously treated with chemotherapy, and were naive to immunotherapy. To prepare each personalized vaccine, researchers apheresed each patient, and monocytes were prepared by elutriation and induced to differentiate into dendritic cells ex vivo. At the same time, surgically resected tumor samples from each patient were dissociated into single cells, killed through oxidation with hypochlorous acid, and lysed using freeze thaw cycles to generate tumor lysate. DCs were pulsed with autologous tumor lysate and the resulting vaccines (OCDC) were injected into patients intranodally, under the guidance of ultrasound.
Following a modified safety phase 1 design, patients were divided into three cohorts: cohort 1 (5 patients) received only OCDC, cohort 2 (10 patients) received OCDC in combination with bevacizumab (a VEGF inhibitor intended to promote T cell infiltration), and cohort 3 (10 patients) received OCDC with bevacizumab and were pretreated with low-dose cyclophosphamide (a chemotherapeutic agent intended to attenuate Tregs). Vaccine doses were administered until disease progressed or the vaccine supply ran out; all patients received at least 5 vaccine doses.
Meeting the safety endpoint of the trial, no adverse events over grade 2 were attributed to treatment, and most adverse events were transient and grade 1. To test whether the vaccine treatment induced an antitumor immune response, the researchers compared peripheral blood mononuclear cell (PBMC) samples from patients before and after 5 doses of treatment. Overall, 11 of 22 evaluable patients were deemed vaccine responders due to an increase in IFNγ-producing T cells responding to tumor lysate-loaded antigen-presenting DCs at week 12. Similarly, 9 of 13 evaluable patients had T cells that responded directly to short-term tumor cell lines and were deemed tumor responders. In representative vaccine responders, CD4+ and CD8+ T cells reactive to tumor antigen were readily detectable in on-treatment PBMCs; responses were low or undetectable prior to treatment. Most vaccine-reactive T cells were polyfunctional, with CD4+ T cells skewed towards producing TNFα and IL-2, and CD8+ T cells producing TNFα and IFNγ. A progressive increase in the frequency of tumor-specific T cells in peripheral blood was associated with remission inversion (longer disease-free intervals before relapse compared those seen in the same patient following prior chemotherapy treatments). Ex vivo, the researchers observed vaccine-induced T cells directly and specifically killing autologous tumor cells. Suppression of tumor growth by transplanted autologous T cells was also observed in a patient-derived xenograft mouse model.
To determine whether the vaccine may have induced T cell responses specific to neoantigens exclusive to the tumor, the researchers used whole exome sequencing and CD8+ neoepitope prediction to generate pools of candidate neoepitopes to test patient-derived CD8+ T cells. Neoepitope-specific CD8+ T cells could be detected in pre-vaccine samples from 4 of 6 patients tested, while they could be detected in all 6 samples taken after treatment; the total number of reactive neoepitopes increased from 9 to 16. Interestingly, in patients who demonstrated an increase in a pre-existing neoepitope response, the newly induced neoantigen-specific T cells demonstrated significantly increased functional avidity. TCR sequencing appeared to indicate priming by the vaccine of new high-avidity clones, and predictions from homology modeling support that TCRs on post-treatment CD8+ T cells would lead to more numerous favorable interactions with peptide:MHC residues compared to pre-treatment TCRs, which would account for the enhanced avidity that was observed.
Of the 25 patients treated, 2 patients experienced a partial response and 13 experienced stable disease persisting for a median of 14 months. Patients whose T cells recognized tumor antigen-presenting DCs or tumor cells after vaccination had significantly longer progression-free survival, and most achieved remission inversion. At two years, the overall survival rate of responder patients was 100%, while the overall survival of non-responders was only 25%. Their results could not be attributed to preexisting immunity.
In regards to the addition of bevacizumab or bevacizumab and pretreatment with low-dose cyclophosphamide to OCDC, Tanyi et al. observed that a significantly higher proportion of patients in cohort 3 (receiving all three treatments) responded to tumor antigen. Patients in this cohort had a higher overall survival rate at 12 and 24 months compared to patients in cohorts that did not receive cyclophosphamide before each vaccination. The survival rate was also higher (78% vs 44%) than that seen in matched historical control patients who had been treated with bevacizumab and cyclophosphamide, but not vaccine. Overall, Tanyi et al. conclude that use of oxidized whole-tumor lysate DC vaccine is safe and effective, and that it may be most effective in combination with low-dose cyclophosphamide pre-treatment.
by Lauren Hitchings
