Background

Immunotherapy is a promising alternative strategy for patients with glioblastoma (GBM). Here, we conducted a phase I/II trial to address the safety and immunogenicity of a multipeptide therapeutic vaccine (IMA950) adjuvanted with Poly-ICLC in patients with newly diagnosed GBM.

Methods

Sixteen HLA-A2⁺ GBM patients were included. Patients received the standard of care treatment including surgery, 6-week chemo-radiation therapy, and 6-12 adjuvant cycles of temozolomide. IMA950 (composed of nine glioma-associated CD8 peptides and two tumor-associated CD4 peptides) was injected starting one week after the end of chemo-radiation therapy. Primary endpoints were safety and immunogenicity; secondary endpoints were OS and PFS at six and nine months.

Results

In the first six patients, peptides were injected intradermally (i.d.) and Poly-ICLC intramuscularly (i.m.). Low levels of vaccine-induced CD8 T cell responses were detected following this vaccination schedule, leading us to modify the vaccination protocol. In the modified schedule, peptides and adjuvant were mixed before injection, which was given i.m. for six patients and subcutaneously (s.c.) for seven patients, in an effort to determine the optimal injection route. Clinically, the IMA950/Poly-ICLC vaccine was well tolerated, the most common side effect being local inflammation at the injection site. A few patients experienced cerebral edema, which was manageable with steroids. Analysis of vaccine-induced T cell responses revealed that 63% of patients responded to one CD8 peptide and 37% responded to two or more CD8 peptides. Modifying the vaccination protocol resulted in detection of multipeptide responses in 46% of patients, as compared with 17% in the initial protocol. Vaccine-induced CD4 T cell responses were detected in the majority of patients and were of Th1 phenotype. Median OS was 21.2 months.

Conclusions

Overall, the IMA950/Poly-ICLC vaccine is safe, immunogenic and preliminary median OS is encouraging. Definitive clinical results and correlation with immunological data will indicate the most efficacious route of vaccination for use in further trials in glioma patients.

Clinical trial identification

NCT01920191

Legal entity responsible for the study

Pierre-Yves Dietrich

Funding

The Gateway for cancer research, Rising Tides (USA)

Disclosure

All authors have declared no conflicts of interest.

Background

CD47-targeted immune checkpoint inhibitors have shown potent anti-tumor efficacy in several malignant tumors. However, glioblastoma, one of the most malignant brain tumors in the central nervous system (CNS), was still a challenge for CD47-targeted immunotherapy.

Methods

SIRPα-Fc, a CD47-targeted fusion protein, was employed to investigate the anti-tumor efficacy of blocking CD47-SIRPα axis alone or combined with autophagy depletion in glioblastoma. Confocal microscopy, transmission electron microscopy and western blot were used to detect autophagy. Glioblastoma xenograft tumors were established to investigate the synergistic anti-tumor effect of simultaneously blocking CD47-SIRPα axis and autophagy in glioblastoma. Syngeneic immunocompetent glioblastoma models were established to detect the role of adaptive immune response in SIRPα-Fc-treated glioblastoma.

Results

SIRPα-Fc has potent anti-glioblastoma efficacy via increasing macrophages phagocytosis and cytotoxicity against glioblastoma cells. Meanwhile, autophagy and autophagic flux were markedly triggered by SIRPα-Fc in glioblastoma cells through inactivation of Akt/mTOR. Importantly, abolishment of the autophagy by pharmacological inhibitors or siRNA enhanced SIRPα-Fc-induced macrophage phagocytosis and cytotoxicity against glioblastoma cells. Simultaneously blocking CD47-SIRPα axis and autophagy increased macrophages infiltration and tumor cell apoptosis, eliciting enhanced anti-glioblastoma efficacy and prolonged survival of tumor-bearing mice compared with treatment with SIRPα-Fc alone. Furthermore, adaptive immune responses, mainly including CD8⁺ T cells, were also involved in the anti-glioblastoma efficacy of SIRPα-Fc.

Expected Conclusions

Our data revealed that SIRPα-Fc could elicit potent anti-glioblastoma efficacy, combined with autophagy inhibition significantly enhanced the anti-glioblastoma efficacy, which highlighted a potential therapeutic approach for glioblastoma through disrupting CD47-SIRPα axis alone or combined with autophagy depletion.

Legal entity responsible for the study

Fudan University

Funding

None

Disclosure

All authors have declared no conflicts of interest.

Background

The majority of patients with metastatic melanoma obtain only short-term or no benefit at all from checkpoint inhibitor (CPI) immunotherapy. In this study, we investigated whether the immune system of patients progressing on or following treatment with CPI was able to generate functional tumor-specific immune responses.

Methods

Tumor-infiltrating lymphocytes (TILs) were isolated and expanded from metastatic melanoma lesions in progression on or after anti-PD-1 and anti-CTLA-4. Tumor-specific immune responses were assessed with co-culture assays of TILs and autologous tumor cells. In a phase I/II clinical trial, TILs were administered with lymphodepleting chemotherapy, pegIFNa2b and IL-2 to 12 patients with CPI-resistant melanoma.

Results

TILs from 19 metastases of an equal number of patients could be assessed for T cell recognition of autologous tumor cells. All metastases were progressive on or following anti-PD-1 (19/19) and anti-CTLA-4 (16/19). Functional anti-tumor immune responses were detected in 15/19 patients (79%). Both CD8⁺ (in 14/19 patients, 74%) and CD4⁺ (in 14/19 patients, 74%) TILs were able to recognize autologous tumors. On average, a large fraction of CD8+ TILs (mean 25%, range 1.1-84%) recognized tumor cells. This is similar or higher than in cohorts of unselected patient populations with metastatic melanoma presented in previous studies. Additional histology analyses to identify the location of the immune infiltrates are ongoing. Out of 12 patients with anti-CTLA-4/anti-PD-1-resistant melanoma treated with TILs, two partial responses were observed, of which one is ongoing.

Conclusions

Tumor-reactive T cells appear to infiltrate heavily the tumor microenvironment of patients who failed previous CPIs. These patients can still respond to infusion of unselected autologous TILs. Our results warrant further testing of novel immune re-activation strategies in anti-PD-1-/anti-CTLA-4-resistant melanoma.

Clinical trial identification

NCT02379195

Legal entity responsible for the study

Center for Cancer Immune Therapy, Herlev Hospital, Herlev, Denmark

Funding

None

Disclosure

I.M. Svane: Advisory Board and lectures: Roche, Novartis, Merck, MSD, Celgene, Incyte, TILT bio, Pfizer, BMS, Astra Zeneca; Limited grants to institution for translational research from: BMS, Roche, Novartis. M. Donia: Honorarium for lectures: MSD, BMS, Genzyme; Travel support: Novartis, MSD, BMS, Roche, Pfizer. All other authors have declared no conflicts of interest.

Background

Tumor specific TIL can be in vitro expanded and have the ability to induce complete and durable tumor regression in some patients with melanoma following ACT. In this preclinical study, we investigated the feasibility of expanding TIL from sarcomas, as well as performing functional in vitro analyses on these.

Methods

Fresh tumor samples were obtained, and TIL were isolated and expanded in growth medium containing IL-2. In a sub study, we investigated the effect of adding an agonistic CD137 antibody (Urelumab, BMS) and/or an anti-CD3 antibody (OKT3) to the medium. Phenotype and functional analyses was performed using flow cytometry and IFNγ-Elispot. Cytotoxicity analyses were performed using Xcelligence.

Results

Tumor samples from 28 patients with 8 different sarcoma subtypes were obtained, and we were able to expand a minimum of 40 million TILs from 25 of these. Mean expansion times were 32 days (14 - 61). 87,7% (36,4 – 99,1) of these cells were CD3+, and of these, 66,7% (16,3 – 99,1) were CD4+, and 21,8% (0,1 – 50,6) were CD8+. Adding anti-CD137 and/or OKT3 increased total yield of TILs; anti-CD137 skewed the phenotype significantly towards more CD8+ TILs and in some cases NK cells and γδ cells. TILs from 11 of 22 tested tumor samples from 7 of 8 different sarcoma subtypes demonstrated reactivity against autologous tumor cells using IFNγ-Elispot. These results were verified in an intracellular cytokine release assay using flow cytometry, and in cytotoxicity assays. In some cases the fraction of reactive cells was more than 20%. In TILs stimulated with anti-CD137 the reactivity increased in four of four tested samples.

Conclusions

We were able to expand TIL from 90% of tested tumor samples. Expanded TIL were a mix of CD4+ and CD8+ with CD4+ being predominant. Half of the TIL cultures showed in vitro tumor reactivity, which in some cases was as high as previously only seen in melanoma samples. Early analyses suggest that addition of anti-CD137 could influence expansion time, phenotype, and functional capacity of the expanded TIL. Based on these results, we plan to initiate clinical testing of TIL based ACT in sarcoma patients.

Legal entity responsible for the study

Center for Cancer Immune Therapy, Herlev Hospital

Funding

Center for Cancer Immune Therapy, Herlev Hospital

Disclosure

All authors have declared no conflicts of interest.

Background

Chimeric Antigen Receptor (CAR)-based therapy gained full attention when the common B-cell marker CD19 was targeted. Parallel trials were run and demonstrated a remission rate ranging from 70 to 90% in paediatric acute lymphoid leukaemia (ALL) patients, and around 50% in adult Non-Hodgkin Lymphoma (NHL) patients. CD19 can be considered as an ideal target since its expression is restricted to B cells, however, alternative targets should be defined for B-cell malignancies showing a CD19CAR success rate lower than in ALL. A candidate is CD37. It is a tetraspanin protein that is widely expressed on the surface of mature B cells. It is also a marker specific for NHL and thus represents an attractive alternative target.

Methods

We have used flow cytometry to test the expression of CD37 in different cell lines and also in lymphoma patient biopsies. We have isolated the coding sequence of a specific anti-CD37 antibody from an hybridoma and have designed a CAR. This CD37CAR was expressed in different effector cell types and was tested for its ability to stimulate in functional assays (killing and cytokine release). We have established two human xenograft model in NSG mice and validated the efficacy of mRNA electroporated CD37CAR redirected Tc.

Results

We confirmed that CD37 could be found in different types of lymphoma and further observed a clustered expression as opposed to some classical markers. The CD37CAR design generated a functional product although the single chain variable domain (scFv) was sensitive to the composition of the hinge region. Redirected effector cells (Tc from human peripheral blood and the NK cell line NK-92) could selectively recognize to CD37+ targets, be stimulated and become cytotoxic. CD37CAR Tc were as efficient as CD19CAR Tc to kill double positive cells, but revealed superior against U-2932 cell line which is CD37^high and CD19^low. Finally, transiently redirected Tc were shown to reduce or completely inhibit tumour progression into engrafted NGS mice.

Conclusions

Our findings suggest that CD37CAR Tc can be used as an alternative to CD19CAR Tc, especially when CD19 expression is lost or reduced in patients’ tumour cells.

Legal entity responsible for the study

Oslo University Hospital

Funding

Norwegian Cancer Society, South-Eastern Norway Regional Health Authority, Research Council of Norway

Disclosure

All authors have declared no conflicts of interest.

Background

At therapeutic concentrations, PEGylated IL-10 (AM0010) stimulates the cytotoxicity, survival and proliferation of intratumoral antigen activated CD8+ T cells in pre-clinical cancer models and in patients. AM0010 activates antigen stimulated CD8 T cells while PD-1 inhibits them, providing a rationale for combining AM0010 with PD-1 inhibitors.

Methods

34 NSCLC pts received AM0010 (10-20ug/kg QD, SC) with pembrolizumab (2mg/kg, q3wk IV; n = 5) or nivolumab (3mg/kg, q2wk IV; n = 29). Pts had a median of 2 prior therapies. The median follow-up is 14.9 mo (range 5.6-30.3). Tumor responses were assessed by irRC. 20 patients were analyzed for PD-L1 expression (22C3). Immune responses were measured by analysis of serum cytokines (Luminex), activation of blood derived CD8 T cells (FACS) and peripheral T cell clonality (TCR sequencing). Tumor tissue was analyzed for tumor mutational burden by WES and mRNA expression of IO markers (Nanostring).

Results

AM0010 plus anti-PD-1 was well tolerated. All TrAEs were reversible and transient. G3/4 TrAEs included thrombocytopenia (8), anemia (7), fatigue (6), rash (4), pyrexia (2), hypertriglyceridemia (3) and pneumonitis (1). As of Sept 15, 2017, 27 pts had at least 1 tumor assessment. 10 pts (36.4%) had a partial response (PR) and 12 pts (44.4%) had stable disease (SD). Notably, of five patients with PD-L1 + ≥ 50% NSCLC, four had a PR, 1 had SD with 47% reduction of tumor burden.Table: 9PD

Preliminary response data stratified for PD-L1

The mOS was 19.7 months, 1-year survival was 71%. Biomarker analysis exploring the correlation of the tumor mutational burden, mRNA profiling and the systemic immune response with tumor response and overall survival will be presented.

Conclusions

AM0010 in combination with an anti-PD-1 is well-tolerated in advanced NSCLC pts. The improved efficacy regardless of PD-L1 status and the observed CD8 T cell activation is promising and encourages the continued study of AM0010 in combination with an anti-PD-1.

Clinical trial identification

NCT02009449

Legal entity responsible for the study

ARMO BioSciences

Funding

ARMO BioSciences

Disclosure

N. Ratti, A. Hung, M. Oft: Employee of ARMO BioSciences. P. Van Vlasselaer: Employee and stock owner of ARMO BioSciences. All other authors have declared no conflicts of interest.

Background

Intra-arterial therapies, including transarterial chemoembolization (TACE) and selective internal radiation therapy with ⁹⁰Yttrium (SIRT), are effective locoregional treatments for hepatocellular carcinoma (HCC). High levels of tumor-infiltrating lymphocytes (TILs) are associated with a better prognosis in HCC. We hypothesized that TACE and SIRT may modify immunogenic tumor microenvironment. To analyze this question, we evaluated TILs in patients who underwent partial hepatectomy for HCC after preoperative TACE or SIRT and in patients operated without preoperative treatment (SURG).

Methods

Epidemiological, clinical and pathological data were analyzed in SIRT (n = 12), TACE (n = 16) and SURG (n = 32) groups. Sections for digital image analysis (DIA) were prepared from paraffin blocks. Immunohistochemistry stains were performed for CD3, CD4, CD8, CD20 and Granzyme B. The slides were scanned and analyzed with DIA Visiopharm software. After exclusion of the necrotic zones, TILs were quantified as percentages of positive cells/analyzed area.

Results

Baseline patient and tumor characteristics were similar in the 3 groups. Median times between SIRT or TACE and partial hepatectomy were not significantly different (16 versus 11 weeks). In resected HCC, preoperative SIRT significantly increased CD3+ TILs, including both CD4+ and CD8+ subpopulations, as compared to TACE and SURG groups. Preoperative TACE did not significantly affect TILs, neither for CD3+ nor for CD4+ or CD8+ subpopulations, as compared to the SURG group. CD20+ B lymphocyte infiltrates were similar in the 3 groups. Significantly increased expression of Granzyme B was demonstrated in SIRT patients while no modification of Granzyme B levels was observed between TACE and SURG patients.

Conclusions

In contrast with TACE, SIRT increases TILs levels and intratumor cytotoxic activity in HCC. These results suggest that local attraction/activation of effector T cells may contribute to the anti-tumor effect of SIRT, supporting the clinical development of therapeutic approaches combining SIRT and immunotherapy.

Legal entity responsible for the study

Vincent Donckier

Funding

None

Disclosure

All authors have declared no conflicts of interest.

Background

The functional role and clinical utility of measuring PD-1, LAG-3 and TIM-3 in NSCLC tumor tissue remain poorly understood even though corresponding clinical trials with blocking agents are ongoing. We analyzed the expression of these targets in association with key functional immune metrics and outcome after treatment with PD-1 axis blockers in human NSCLC.

Methods

We performed CyTOF on immune cell suspensions from 20 primary human NSCLCs to map the distribution of PD-1, LAG-3 and TIM-3 and explore their function. We analyzed RNA levels of these markers in TCGA NSCLC datasets and their association with CD4/CD8 mRNA and mutational burden. Using multiplex quantitative immunofluorescence (QIF) we measured the levels of CD3, PD-1, LAG-3 and TIM-3 in 66 pre-treatment tumor samples from NSCLC patients receiving PD-1 axis blockers in 2 independent cohorts (Cohort #1, Yale, N = 42 cases [Training set] and cohort #2, Cleveland Clinic/University of Navarra, N = 24 [Validation set].

Results

In primary NSCLCs, PD-1 was predominantly expressed on T- and NKT cells. LAG-3 expression was higher in CD8⁺, CD4⁺CD25⁺FOXP3⁺ and NKT cell subsets, but low/absent in antigen-presenting cells (APCs). TIM-3 was broadly expressed in adaptive and innate immune cells, with the highest levels in APCs. Expression of all 3 markers in T-cells was associated with lymphocyte activation (CD69/HLA-DR), effector function (Granzyme-B) and proliferation (Ki-67). LAG-3-expressing T-cells showed higher association with early activation and effector function than TILs expressing PD-1 or TIM-3. In TCGA, PD-1 and LAG-3 transcripts strongly correlated with CD8, while TIM-3 was associated with CD4 mRNA. There was limited association between the markers and tumor mutational burden. In pre-treatment specimens from both cohorts of patients treated with PD-1 blockers, elevated LAG-3 but not PD-1 or TIM-3 protein were significantly associated with shorter OS.

Conclusions

PD-1, LAG-3 and TIM-3 show variable expression and are associated with T-cell activation and effector function in NSCLC. Elevated T-cell LAG-3 in baseline tumor samples predicts primary resistance to PD-1-axis blockers.

Legal entity responsible for the study

Kurt A. Schalper- Yale University School of Medicine

Funding

NIH/NCI, DOD-LCRP, SU2C, LCRF

Disclosure

B.S. Henick: Participated in a consulting program for Boehringer Ingelheim. I. Melero: Consultancy: BMS, Roche, Bayer, Lily, AstraZeneca, Genmab, Aliigator, Tusk; Grants: Roche, BMS, Alligator, Pfizer. K.A. Schalper: Merck, Tesaro, Takeda, Onkaido/Moderna, Navigate/Novartis, Surface Oncology, Celgene, Vasculox. All other authors have declared no conflicts of interest.

Background

INO-5150 (PSA and PSMA) +/- INO-9012 (IL-12) immunotherapy was administered to biochemically recurrent prostate cancer patients (pts) to demonstrate breaking tolerance and assess antigen-specific humoral and cellular immune responses for stabilization of disease progression.

Methods

Phase I, open-label, multi-center study in pts post-definitive therapy with a rising PSA after surgery and/or RT and PSA doubling time (DT)> 3 months (mos), testosterone > 150 ng/dL, no concomitant ADT and no evidence of metastases. Safety, immunogenicity and efficacy were evaluated in 4 treatment arms in 60 planned pts (A: 16, 2mg INO-5150; B: 15, 8.5 mg INO-5150; C: 15, 2mg INO-5150 + 1mg INO-9012; D: 16, 8.5mg INO-5150 + 1mg INO-9012) treated with 4 IM doses followed by electroporation on day 0, wks 3, 12 and 24 who were followed for a total of 72 Wks.

Results

Across all cohorts, among evaluable pts, 35/58 (60%) had IFN-γ reactivity by ELISPOT. Antibody titers against PSA and PSMA were observed in 6/61 (10%) and 5/61 (8%) pts respectively. 19/50 (38%) pts across the trial had CD38 and Perforin + CD8 T cell responses with the greatest proportion in arm A, 8/14 (57%) as well as a high PD-1 expression of 68.6% in this arm at week 27. Pts with immune reactivity had dampening of % PSA rise (median: 28.0 vs 53.4, p = 0.02) compared to non-reactive pts. Myeloid derived stem cell (MDSC) assessment showed that the frequency of MDSCs is decreased in pts with >70% PSA decline compared to pts with radiographically confirmed disease progression. Median Day 0 PSADT improved from 8.2 to 19.0 months (mos) (p = 0.002) for all cohorts combined at last observation with highest numerical improvement in Cohort A (26.6 mos). 56/61 (90%) pts are metastasis free at median follow-up of 436 days (85-582). Safety summary: 11 Grade (Gr) 3 SAEs in 6 pts reported. 82% of pts had Gr 1-3 AE’s primarily related to injection site reactions.

Conclusions

INO-5150 +/- INO-9012 was safe and well tolerated. Data demonstrated both PSA and PSMA are immunogenic and INO-5150 induced cellular immune responses. Additional analyses and follow-up are ongoing to further elucidate the correlation of immunologic efficacy and clinical benefit that has been initially observed.

Clinical trial identification

NCT02514213

Legal entity responsible for the study

Inovio Pharmaceuticals

Funding

Inovio Pharmaceuticals

Disclosure

I. Csiki, K. Bhatt, M. Morrow, K. Kraynyak, L. Liu, T. McMullan, J. Lee, B. Sachetta, S. Rosencranz, M. Bagarazzi: Employee of Inovio Pharmaceuticals. All other authors have declared no conflicts of interest.