Citations

Despite the remarkable activity of CD19 directed chimeric antigen receptor T cell (CART19) therapy in the treatment of B cell malignancies, the therapy is limited by the development of severe life-threatening complications such as neurotoxicity (NT) and cytokine release syndrome (CRS). Additionally, durable efficacy following CART19 therapy is not optimal. Emerging literature suggests that inhibitory myeloid cells and their cytokines play an important role in inducing CAR-T cell toxicities and also contribute to the inhibition of their effector functions. Specifically, GM-CSF was identified as a critical cytokine in the development of NT and CRS after CART19 therapy. Neutralization of GM-CSF in preclinical models has been shown to prevent CRS and enhance CART anti-tumor activity through modulation of myeloid cell behavior, resulting in reduction of tumor associated macrophages. In addition to the predominant effect of GM-CSF on myeloid cells, there appears to be a direct effect on CART19 cells. In this study, we aimed to evaluate the direct effect of GM-CSF on CART cells. Our initial finding of enhanced anti-tumor activity of CART19 cells after GM-CSF inhibition suggested a direct effect of GM-CSF on CART cells (Sterner et al. 2019, Blood). In these experiments, a guide RNA (gRNA) targeting exon 3 of GM-CSF in a CRISPR-Cas9 lentiviral vector was used to knock out GM-CSF during CART cell manufacturing. This resulted in a disruption efficiency of approximately 70% of the GM-CSF gene. Using a high tumor burden xenograft model for relapsed acute lymphoblastic leukemia established through the engraftment of the CD19+ luciferase+ NALM6 cell line (1x106 cells intravenously) in immunocompromised NOD-SCID-γ-/- mice, treatment with low doses of GM-CSFk/o CART19 resulted in improved anti-tumor activity and overall survival compared to GM-CSFwt CART19. The lack of myeloid cells in this model pointed to an intrinsic effect of GM-CSF on CAR-T cells. To ensure that this was not related to an off-target effect of the gRNA, whole exome sequencing (WES) of the modified cells was performed. There was no difference in the single nucleotide variants or indel counts between GM-CSFk/o CART19 and GM-CSFwt CART19 (Figure 1A). WES was significant for only two alterations in the exon 3 targeted by the gRNA (Figure 1B). The high efficiency and accuracy of targeting exon 3 of GM-CSF indicated that the improvement in CART function is unlikely related to an off-target effect of the gRNA and suggested the possibility of a direct interaction between GM-CSF and CART cells as a potential mechanism behind the improved anti-tumor activity. To investigate this interaction, we first assessed the expression of GM-CSF receptors on CART cells. While resting CART cells do not express any GM-CSF receptors, our analysis robustly indicates that activated CART cells significantly upregulate both α and β subunits of the GM-CSF receptor. This finding was significant both when CART cells are activated through their T cell receptor with CD3/CD28 beads (Figure 1C) or through the CAR with irradiated NALM6 cells (Figure 1D). Additionally, activated GM-CSFk/o CART19 cells also upregulated GM-CSF receptors, indicating this upregulation is induced by T cell stimulation. These results suggest a direct interaction between GM-CSF and upregulated GM-CSFR on activated CART cells. Having demonstrated that 1) GM-CSF depletion enhances CART19 efficacy in xenograft models in the absence of monocytes and 2) T cell activation increases GM-CSF receptor expression, we sought to uncover the downstream changes resulting from this effect. Transcriptome interrogation of GM-CSFk/o CART19 revealed a distinct signature including a significant inhibition of the Fas death pathway, a known critical pathway in inducing CART cell apoptosis. This suggests a potential mechanism for enhanced CART19 activity following GM-CSF depletion (Figure 1E). In summary, our results strongly indicate that CART cells increase expression of GM-CSF receptor subunits when activated, resulting in modulation of CART cell functions. Furthermore, GM-CSFk/o CART19 revealed a distinct transcriptome signature compared to GM-CSFwt CART19. These results illuminate a novel mechanism for a direct modulatory effect of GM-CSF on activated CART cells. Disclosures Cox: Humanigen: Patents & Royalties. Sterner:Humanigen: Patents & Royalties. Sakemura:Humanigen: Patents & Royalties. Ahmed:Humanigen: Employment. Chappell:Humanigen: Employment. Durrant:Humanigen: Employment. Parikh:Acerta Pharma: Research Funding; MorphoSys: Research Funding; AbbVie: Honoraria, Research Funding; Genentech: Honoraria; Janssen: Research Funding; AstraZeneca: Honoraria, Research Funding; Pharmacyclics: Honoraria, Research Funding; Ascentage Pharma: Research Funding. Kay:MorphoSys: Other: Data Safety Monitoring Board; Infinity Pharmaceuticals: Other: DSMB; Celgene: Other: Data Safety Monitoring Board; Agios: Other: DSMB. Kenderian:Novartis: Patents & Royalties, Research Funding; Tolero: Research Funding; Lentigen: Research Funding; Humanigen: Other: Scientific advisory board , Patents & Royalties, Research Funding; Kite/Gilead: Research Funding; Morphosys: Research Funding.

CD19 directed chimeric antigen receptor T cell (CART) therapy has shown remarkable activity in B cell lymphoma and acute lymphoblastic leukemia leading to the approval of two CART therapies. With the emergence of therapeutic anti-CD19 antibodies for the treatment of B cell malignancies, it remains to be elucidated whether such antibodies would interfere with the ability of CD19 targeting CARTs to exert their anti-tumor effect in a subsequent therapy. To address a part of this question, we investigated the potential for functional interference between the monoclonal anti-CD19 antibody tafasitamab (MOR208) and CD19 directed CART cells (CART19). CART19 cells were generated through lentiviral transduction of healthy donor T cells with a second generation CD19 CAR construct (FMC63-CD8h-CD8TM-41BBζ) which is similar to the construct used for the FDA-approved CART tisagenlecleucel. Tafasitamab, is an Fc-enhanced humanized monoclonal antibody which mediates antibody-dependent cellular toxicity (ADCC), antibody-dependent cellular phagocytosis (ADCP) and direct cytotoxicity. It is currently being studied in phase 2 and 3 clinical trials in diffuse large B-cell lymphoma (DLBCL) in combination with the immunomodulatory agent lenalidomide (L-MIND) and the chemotherapeutic drug bendamustine (B-MIND). As a first step we confirmed the relevance of the tested CD19-positive target cell lines, JEKO (mantel cell lymphoma), Ly7 (DLBCL) and NALM-6 (ALL) based on functional activity of tafasitamab and CART19. In a 24 hours ADCC (tafasitamab titration plus natural killer (NK) cells; Figure 1A) and T cell cytotoxicity assays (CART19, E:T titrations; data not shown) distinct activity was observed for both therapies on all tested cell lines. Secondly, we studied whether the observed CART19 activity may be influenced by tafasitamab in case of a direct CD19 binding competition between tafasitamab and the CAR. To test for such binding competition we incubated the CD19+ cell lines NALM6 or JEKO with up to 100 µg/ml tafasitamab, to saturate the receptors. Subsequent flow cytometry analysis using the FMC63 antibody (carrying the same CD19 binding domain as contained in CART19) failed to detect CD19 expression, indicating a direct binding competition between FMC63 and tafasitamab (Figure 1B). Next, to investigate the potential impact of such binding competition on CART19 cell effector functions, we co-cultured tafasitamab CD19+ JEKO cell line at increasing concentrations of up to 100µg/ml, and then added CART19 cells at different effector to target ratios to the cell culture. The presence of tafasitamab, binding to the CD19 antigen, did not affect important CART cell effector functions such as antigen specific killing (Figure 1C), degranulation (Figure 1D), cytokine production or proliferation of CART19 (Figure 1E). In summary, our studies indicate that CART19 continue to exhibit potent antigen specific effector functions despite presence of tafasitamab and the related competition for CD19 binding. Besides the presented in vitro work the questions of therapeutic sequencing of tafasitamab and CART19 is being studied in xenograft models and will be presented at the meeting. Disclosures Sakemura: Humanigen: Patents & Royalties. Cox:Humanigen: Patents & Royalties. Schanzer:MorphoSys AG: Employment. Endell:MorphoSys AG: Employment, Patents & Royalties. Nowakowski:Selvita: Membership on an entity's Board of Directors or advisory committees; NanoString: Research Funding; MorphoSys: Consultancy, Research Funding; Genentech, Inc.: Research Funding; F. Hoffmann-La Roche Ltd: Research Funding; Curis: Research Funding; Bayer: Consultancy, Research Funding; Celgene: Consultancy, Research Funding. Kay:MorphoSys: Other: Data Safety Monitoring Board; Infinity Pharmaceuticals: Other: DSMB; Celgene: Other: Data Safety Monitoring Board; Agios: Other: DSMB. Kenderian:Novartis: Patents & Royalties, Research Funding; Tolero: Research Funding; Humanigen: Other: Scientific advisory board , Patents & Royalties, Research Funding; Lentigen: Research Funding; Morphosys: Research Funding; Kite/Gilead: Research Funding.

Background and purposeThe long‐term outcomes and stroke recurrence after basilar artery occlusion (BAO) are largely unknown. We aimed to assess these variables in a comparatively large series of consecutive patients.MethodsAdults with acute BAO were retrospectively identified from 1976 to 2011. Post‐discharge records were reviewed to assess for stroke recurrences, mortality and disability. Exploratory analysis of survival was carried out using Kaplan–Meier and log‐rank tests. Factors associated with survival time were determined using Cox models.ResultsA total of 86 patients (34% female, median age 72 [interquartile range (IQR), 60–79] years) with a median National Institutes of Health Stroke Scale score of 11 (IQR, 6–27) were included. Twenty‐nine patients (34%) died during the initial hospitalization. Median modified Rankin Scale (mRS) score at discharge among survivors was 4 (IQR, 2.5–5.5). At 1 and 5 years, 70% of survivors ad a mRS ≤3. Seventeen patients had recurrent strokes during the hospitalization and 12 patients had 19 recurrent strokes after discharge. The median survival time was 52 days (IQR, 6–1846). Older age per decade on admission [adjusted hazard ratios (aHR), 1.32; 95% confidence interval (CI), 1.05–1.66, P = 0.02] and a higher mRS at discharge (aHR, 4.48; 95% CI, 2.72–7.39, P < 0.0001) were associated with mortality. Patients who were not treated with any reperfusion therapy had a trend towards reduced mortality (aHR, 0.39; 95% CI, 0.14–1.08, P = 0.07).ConclusionsSurvivors from BAO had severe short‐term functional disability. Most deaths and stroke recurrences occurred within the first year following the initial event. The risk of death was higher in older and more disabled survivors. However, favorable long‐term recovery was possible.

Abstract BACKGROUND Efficacy of chimeric antigen receptor T cell (CART) therapy remains limited in solid tumors. Given the heterogeneity of surface receptor expression and immunosuppressive stromal microenvironment, strategies to target tumor neovasculature and tumor stromal cells are needed to help overcome CART inhibition by GBM. Eph receptors are the largest family of receptor tyrosine kinases and are integral to cell adhesion, migration, and axon guidance, during development and homeostasis. EphA3 is a receptor tyrosine kinase which is lowly expressed in adult tissues but is highly expressed in tumor neovasculature and tumor stromal cells in GBM and other solid tumors. EphA3 is over-expressed in up to 40% of GBM samples. We aimed to develop CART cells directed against EphA3 to use in targeting tumor neovasculature and tumor stromal cells in GBM. METHODS we developed a second generation CD28 co-stimulated CAR construct in a third generation lentivirus backbone to generate EphA3 CART cells using the single chain variable fragment of ifabotuzumab, a monoclonal antibody directed against EphA3. Patient derived GBM xenograft cell lines were used in these experiments. RESULTS EphA3 directed CART cells exhibited specific and potent antitumor activity against EphA3+ GBM cell lines with variable transcriptome EphA3 expression indicating its broader applicability in patients with GBM. Killing over 24-hour incubation was significant at low effector: target ratio: 52.5% killing at 1.25:1 against cell lines with 25.05% EphA3 expression and 37.1% killing at 1.25:1and 90% killing at 5:1 ratio against a cell line with 19.28% expression. Conversely, when co-cultured with UTD controls there was significantly lower killing, or growth of tumor cells. CONCLUSION We demonstrate for the first time that targeting EphA3 with CART cells is feasible, efficacious and represents a novel therapeutics strategy to target GBM. Data from in vivo and combinatorial CART experiments will be reported at the meeting.

Abstract Panobinostat (PBST) is a histone deacetylase inhibitor with biologic rationale in the treatment of gliomas harboring H3 K27M mutation. METHODS We performed a retrospective medical record review of adult patients with H3 K27M mutant diffuse midline glioma treated with PBST at Mayo Clinic (Rochester, MN). RESULTS Four patients were identified with a mean age 40 years (range 22–62 years). Tumor location was: spinal cord (thoracic n=1, cervical n=1), brainstem (n=1), and thalamus (n=1).Two patients underwent biopsy alone while two underwent a partial resection. Three patients were treated with radiation (36–54 Gy) immediately prior to PBST monotherapy, and one patient was treated with PBST monotherapy without preceding radiation. All patients were otherwise chemotherapy naïve and did not receive any concurrent chemotherapy with PBST. PBST was dosed at either 10 mg (n=1) or 20mg (n=3) administered on days 1, 3, 5, 8, 10 and 12 of 21 day cycles. The mean duration of PBST therapy was 5 months (range 2–11 months). PBST was well tolerated overall. One patient experienced an objective response per RANO criteria. Two patients continue on therapy (5 cycles, 12 cycles) with stable disease. One patient rapidly progressed after 2 cycles of PBST therapy. In contrast with the patients who derived benefit from PBST, this patient was younger (22) and had multifocal disease with leptomeningeal involvement at treatment onset. The average progression free survival post PBST initiation was 8 months. CONCLUSION This single institution case series shows promise that PBST may have therapeutic benefit in adult patients with H3 K27M mutant diffuse midline glioma.

Abstract BACKGROUND The Tumor Treating Fields (TTF) mechanism of action is hypothesized act upon mitosis by disrupting the cell spindle apparatus. However, biomarkers and human tissue correlation of the effect of the TTF device are lacking. CASE: A 41 year-old previously healthy woman was found to have a right frontoparietal glioblastoma after she developed subacute onset of left hemiparesis and headache. The tumor was IDH wild type MGMT promoter unmethylated and TERT mutated. Chromosomal Microarray (CMA) was performed and demonstrated genome wide aneuploidy primarily characterized by hyperdiploidy. She underwent a gross total resection of the enhancing portion of the tumor and subsequent standard chemoradiotherapy (60 Gy IMRT with concomitant temozolomide) followed by initiation of the TTF device and adjuvant temozolomide. Six months after diagnosis, after 72 days of using the TTF device with 75% compliance and completion of the third cycle of temozolomide the patient had radiographic recurrence. She underwent a repeat resection. Pathology demonstrated recurrent tumor. Genetic analysis no longer demonstrated chromosomal losses or gains: the tumor was diploid. DISCUSSION This case may be a demonstration of effect of the TTF device applying a selective pressure to cells with multiple chromosomal gains and losses as the efficiency of mitosis in these cells may already be severely compromised. We are currently performing CMA on all patients with available tissue treated with the TTF device at our institution, with a planned retrospective analysis to explore this hypothesis further. CONCLUSION Repeat resection with detailed genetic analysis of GBM recurrence after therapies may help generate hypotheses regarding mechanisms of resistance, and in this case, mechanisms of action of the TTF.

Abstract BACKGROUND Gliomas are a known risk factor for the development of spontaneous intracranial hemorrhage (ICH) independent of therapies directed against them, with studies reporting an incidence of 1.9–3.8%. Bevacizumab (BEV) is a humanized monoclonal antibody targeting vascular endothelial growth factor (VEGF) FDA approved for recurrent glioblastoma. ICH is commonly considered to be a potential adverse effect of BEV use in patients with glioma, with previous retrospective studies describing incidence of intracerebral hemorrhage ranging from 1.9–3.3%. Material and METHODS We performed a single center (Mayo Clinic, Rochester, MN) retrospective chart review of all patients who received BEV therapy at our institution for a diagnosis of primary CNS malignancy. We used ICD-9 and ICD-10 codes to identify adult patients with primary CNS neoplasms that subsequently developed ICH. RESULTS We screened 10,507 adult patients with a diagnosis of primary CNS neoplasm treated at our center from 01/31/2008 to 12/31/2018 and found 644 patients treated with bevacizumab. Of these, 23 (3.6%) suffered an ICH within 12 weeks of a BEV infusion (median 9 days post infusion, 1–24 IQR). The dose of BEV in all hemorrhages was either 10 mg/kg every two weeks (n = 21) or 15 mg/kg every three weeks (n = 2). No patients treated with infusions at 7.5 mg/kg every three weeks (n = 79) or 5 mg/kg every two weeks (n = 17) suffered ICH events, however this did not meet significance (P = 0.06) when compared to patients on 10 mg/kg or above. The median ICH score was 1 (0–2 IQR). There was an increase in ECOG scores documented at clinical follow up following the hemorrhage compared to the pre-hemorrhage functional status (ECOG 3.28 versus 1.2, p = < 0.0001), with 13 patients (56%) having developed severe debility or death (ECOG 4 n = 3, ECOG 5 n = 10). CONCLUSION In our single-center cohort of patients with a primary diagnosis of CNS neoplasms treated with BEV, ICH occurred in 3.6% of patients. Those who experienced an ICH in proximity to BEV infusion had significant morbidity and mortality with a clear decline in functional status. There was a signal of dose response as far as ICH incidence in our cohort with dosages 10 mg/kg per infusion or above.

e13509 Background: Gliomas are a known risk factor for the development of spontaneous intracranial hemorrhage (ICH) independent of therapies directed against them, with studies reporting an incidence of 1.9-3.8%. Bevacizumab (BEV) is a humanized monoclonal antibody targeting vascular endothelial growth factor (VEGF) FDA approved for recurrent glioblastoma. ICH is commonly considered to be a potential adverse effect of BEV use in patients with glioma, with previous retrospective studies describing incidence of intracerebral hemorrhage ranging from 1.9-3.3%. Methods: We performed a single center (Mayo Clinic, Rochester, MN) retrospective chart review of all patients who received BEV therapy at our institution for a diagnosis of primary CNS malignancy. We used ICD-9 and ICD-10 codes to identify adult patients with primary CNS neoplasms that subsequently developed ICH. Results: We screened 10,507 adult patients with a diagnosis of primary CNS neoplasm treated at our center from 01/31/2008 to 12/31/2018 and found 644 patients treated with bevacizumab. Of these, 23 (3.6%) suffered an ICH within 12 weeks of a BEV infusion (median 9 days post infusion, 1-24 IQR). The dose of BEV in all hemorrhages was either 10 mg/kg every two weeks (n = 21) or 15 mg/kg every three weeks (n = 2). No patients treated with infusions at 7.5 mg/kg every three weeks (n = 79) or 5 mg/kg every two weeks (n = 17) suffered ICH events, however this did not meet significance (P = 0.06) when compared to patients on 10 mg/kg or above. The median ICH score was 1 (0-2 IQR). There was an increase in ECOG scores documented at clinical follow up following the hemorrhage compared to the pre-hemorrhage functional status (ECOG 3.28 versus 1.2, p = < 0.0001), with 13 patients (56%) having developed severe debility or death (ECOG 4 n = 3, ECOG 5 n = 10). Conclusions: In our single-center cohort of patients with a primary diagnosis of CNS neoplasms treated with BEV, ICH occurred in 3.6% of patients. Those who experienced an ICH in proximity to BEV infusion had significant morbidity and mortality with a clear decline in functional status. There was a signal of dose response as far as ICH incidence in our cohort with dosages 10 mg/kg per infusion or above.

Pilocytic astrocytomas (PA) are highly vascular tumors with vascular endothelial growth factor (VEGF)–vascular endothelial growth factor receptor-2 (VEGFR-2) signaling present in the tumor vasculature. PA may, therefore, be responsive to VEGF blockade with bevacizumab (BEV). Data regarding the use of BEV in refractory PA in adults are limited primarily to case reports and case series of patients with recurrent PA. We conducted a single-center, retrospective cohort study from 2009 to 2018. We screened 426 patients with pathologically confirmed PA. We identified 5 adult patients with PA who received BEV at our institution with sufficient clinical follow-up to derive evidence of the efficacy and toxicity. All 5 patients experienced tumor progression after initial therapies which included surgery, radiation, and chemotherapy. Four patients received BEV as monotherapy, whereas 1 received BEV with the continuation of previously initiated alkylating chemotherapy (temozolomide). The average duration of BEV therapy was 10.2 months (range, 1 to 20 mo) with an average follow-up of 47 months (range, 6 to 112 mo). One patient had a severe necrotizing rash in areas of skin contact and discontinued after 1 cycle of BEV. All patients had stabilization per RANO criteria, with 1 patient experiencing progression after 10 months on treatment. One patient had disease progression 5 years after completion of BEV, but the tumor responded to repeat treatment with BEV. Our institution’s experience with the use of BEV in recurrent PA is in line with previous reports of therapeutic benefit in recurrent adult PA.