Citations

Background Stroke or Cerebrovascular Accident (CVA) is one of the leading causes of mortality in the world whether it be ischemic or hemorrhagic 1,2. There has been an association of increased incidence of stroke in patients with cancer due to multiple factors 3. Regarding hematologic malignancies, there have been studies to assess the risk of developing stroke in patients with Hodgkin lymphoma (HL), however the information on non-Hodgkin lymphoma (NHL) is limited and not well understood4. We conducted a retrospective case-control study to assess the incidence, risks, and complications and length of stay of patients with current or prior history of NHL who developed an ischemic stroke. Methods This case-control study identified 29,632 patients older than 18 years, with or without CVA and with or without active or history of Non-Hodgkin Lymphoma (NHL) who were admitted to diverse private hospitals between 2018-2022. The patients were divided into two groups, one with current or prior NHL (n=10,699) and a control group of patients (n=18,933) with similar characteristics. Patients with active hemorrhagic stroke, pregnancy, malignancies other than NHL, prior hemorrhagic or ischemic stroke, infective endocarditis, atrial fibrillation, and acute coronary syndrome on admission were excluded. We compared patient's characteristics and the risk of developing a stroke based on the presence or absence of NHL as well as the factors that increased hospital mortality. Length of stay was evaluated as a secondary outcome. Bivariate statistical analysis was used to calculate the odds ratios using RStudio 4.2.2 and JMP Pro 16.0 for all data analysis. Results Of the group with NHL 46.39 % (n=4958) were female, and 53.61% (n=5729) were male, 42% were current or former smokers (n=4497), 27.49% (n=2941) had diabetes, 44.68% (n=4780) had hypertension; 3.34% (357) of them passed away. Of the control group 57.51 % (n=10876) were female, and 42.49% (n=8036) were male, 35% were current or former smokers (n=6763), 16.43% (n=3111) had diabetes, 28.89% (n=5469) had hypertension; 0.47% (n=88) of them passed away. The mean length of stay of patients with stroke was 7.99 days in the NHL group, and 5.99 days in the control group. Regarding the admissions due to stroke, there were 214 patients in the group with NHL, and in the control group where there were 113. Patients with NHL are more likely to have a CVA compared to the control group (OR 2.39 p<0.0001 CI 1.47~3.85) Patients who have NHL are more likely to die (OR=6.08 p<0.001) CI 4.77~7.74) compared to patients without NHL from all causes. Conclusion Patients with Non-Hodgkin lymphoma have an increased risk of developing a stroke compared to patients with similar characteristics but without NHL (or other malignancies). It was also noted that patients had an increased length of stay. Other variables and characteristics of patients that were identified to be more likely to have a CVA and statistically significant for this sample are already well-known risk factors for CVA including diabetes, hypertension, and cardiovascular disease. However, this study aimed to isolate those variables to identify the risk for stroke in this unique population. Our study limitation is the broad spectrum of NHL and possible confounders between the different variants of NHL. More studies are needed to investigate this. References Navi B, Reiner A, Kamel H, et al. Risk of Arterial Thromboembolism in Patients with Cancer. J Am Coll Cardiol. 2017 Aug, 70 (8) 926-938. Turner, M., Murchie, P., Derby, S. et al. Is stroke incidence increased in survivors of adult cancers? A systematic review and meta-analysis. J Cancer Surviv 16, 1414-1448 (2022). https://doi.org/10.1007/s11764-021-01122-7 Dardiotis E, Aloizou AM, Markoula S, et al. Cancer-associated stroke: Pathophysiology, detection, and management (Review). Int J Oncol. 2019;54(3):779-796. doi:10.3892/ijo.2019.4669 Wei Y, Yang Q, Qin Q, et al. Profiling of the Risk Factors and Designing of a Model to Identify Ischemic Stroke in Patients with Non-Hodgkin Lymphoma: A Multicenter Retrospective Study. Eur Neurol. 2020;83(1):41-48. doi:10.1159/000506046

Our group has reported that transient CD19 antigen masking of leukemic cells with the CD19 monoclonal antibody (mAb), tafasitamab, resulted in reduction of CD19 chimeric antigen receptor T cell (CART19) apoptosis, enhanced CAR T proliferation, improved anti-tumor activity, diminished tumor pyroptosis, and reduced severity of cytokine release syndrome (CRS) in preclinical models. In this report, we aimed to study the mAb characteristics responsible for these observations. Specifically, we hypothesized that binding affinity of CD19 mAbs might affect the subsequent activation and apoptosis of CART19 as well as tumor cell pyroptosis in preclinical models. To test our hypothesis, we first determined the affinities of the CD19-targeting antibodies tafasitamab, HD37 and FMC63 to recombinant human CD19 protein. Tafasitamab demonstrated the strongest binding affinity among the three tested antibodies [tafasitamab ( K D: 1.0±0.2 nM), HD37 ( K D: 4.1±0.3 nM), FMC63 ( K D: 3.4±1.2 nM)]. Next, we studied CART19 apoptosis in the presence of these CD19 mAbs with different binding affinities. Here, we co-cultured CART19, CD19 + mantle cell lymphoma JeKo, and 400 μg/mL of CD19 mAb for 1 hour and measured CD3 + 7-AAD - Annexin V + T cells with flow cytometry. The number of apoptotic CART19 cells was significantly reduced by tafasitamab co-treatment compared to HD37, FMC63, or control IgG [control IgG (mean = 19.1, standard deviation (SD) = 0.8), tafasitamab (mean = 13.1, SD = 2.9), HD37 (mean = 17.8, SD = 0.8), FMC63 (mean = 18.4, SD = 1.0)]. We also performed an in vitro high mobility group box 1 (HMGB1) assay to assess if different affinities of the CD19 mAbs affect tumor cell pyroptosis triggered by CART19. Here, we co-cultured CART19, CD19 + JeKo, and 400 μg/mL of CD19 mAb for 24 hours and assessed HMGB1 on JeKo. In analogy to the apoptosis data, tafasitamab was the only mAb that prevented tumor cell pyroptosis, indicated by reduced HMGB1 expression (Fig 1A). Given these results, we aimed to study the impact of CD19 occupancy with different CD19 mAbs prior to CART19 infusion on tumor pyroptosis in vivo. We utilized a mouse model that mimics CAR T toxicities. Immunocompromised NSG mice were inoculated via tail vein with 5x10 6 leukemic cells derived from patients with relapsed/refractory acute lymphoblastic leukemia. Two weeks later (day -7), tumor burden was assessed by peripheral blood sampling. Leukemic cells were defined as murine CD45 -, human CD45 + CD19 + CD20 + via flow cytometry, and volumetric absolute number was calculated. Mice were then randomized according to tumor burden to receive IgG control [Group 1 (G1)], tafasitamab (G2), HD37 (G3), or FMC63 (G4) (5 mg/kg, ip) at day -7. CRS was determined by weight loss, development of motor weakness, poor body posture, and elevated serum cytokine levels. At day 0, all mice received 3x10 6 luciferase + CART19 via tail vein. Mice were monitored with daily bioluminescent imaging (BLI) to track and quantify CART19 expansion and physical assessment to detect toxicity. BLI on day 4 revealed that CART19 expansion in the tafasitamab-treated group (G2, mean = 1x10 7, SD = 2.3x10 7) was significantly lower than in the control group (mean = 3.5x10 7, SD = 9.3x10 7), the HD37-treated group (G3, mean = 4x10 7, SD = 6.6x10 7), and the FMC63-treated group (G4, mean = 2x10 7, SD = 3.59x10 7), [G1 vs G2: 95% confidence interval (CI) 3.5x10 7 to 1.1x10 8, p <0.0001, G1 vs G3: 95% CI -7.5x10 7 to -1.1x10 8, p=0.005, G1 vs G4: 95% CI -4.5x10 7 to 1.9x10 7, p=0.71] indicating that tafasitamab pretreatment significantly reduced early activation of CART19 compared to IgG control and HD37 in this model. Mice were bled on day 4 to analyze plasma cytokine. Tafasitamab-treated mice (G2) secreted significantly less IFNγ (Fig 1B), IL10 (Mean G1: 78, G2: 2.52, G3: 12.2, G4: 13.9 pg/mL), MDC (Mean G1: 4072, G2: 445, G3: 921, G4: 1364 pg/mL), and MIP1β (Mean G1: 283, G2: 59.1, G3: 148, G4: 123 pg/mL) compared to G1, G3, and G4. Satellite mice were euthanized on day 1 and spleens were harvested. Western blot analysis of splenocytes revealed that G1, 3, and 4 showed significantly lower expression of apoptosis-inducing protein gasdermin E on leukemic cells compared to G2, indicating reduced tumor pyroptosis in G2. In summary, application of the high affinity mAb tafasitamab prior to CART19 resulted in reduced CAR T apoptosis, diminished tumor cell pyroptosis, and reduction of CRS compared to lower affinity CD19 mAbs in preclinical models.

Abstract Introduction: Irinotecan is an inhibitor of the topoisomerase I enzyme that is essential in DNA transcription, replication, and repair. It is considered a “backbone” regime in colorectal cancer in adults and in pediatric gliomas to which other antineoplastic agents are added to improve the outcome. It is used as well in the treatment of many other types of adult and pediatric cancers. Irinotecan acts as a prodrug of SN-38 which has approximately 100-1000-fold greater cytotoxic activity. The metabolic and pharmacokinetic behavior of irinotecan is very complex and currently not completely elucidated. Clinical data have shown high variability in treated patients in the efficacy, toxicity, and pharmacokinetic profile of the drug. To increase safety and efficacy, a physiologically based pharmacokinetic modeling (PBPK) approach was used to predict the optimal dose for treatment initiation. Methods: Clinical data were used in combination with measured physicochemical and blood-binding data to inform the model. Hence a middle-out approach was used to build the parent model for a population of adult cancer patients. While the default Sim-Cancer population available within the Simcyp Simulator (v21) was used, the unbound fraction in plasma for the healthy population was incorporated into the PBPK model. By using the reverse translational tool and a retrograde approach, the irinotecan metabolism was described by incorporating CYP3A (Km 14.4±1.6 μM, Vmax 14.7± 0.6 pmol/min/mg protein) and an additional clearance. This additional clearance was then explained by scaled esterases (CES1 and CES2) enzyme kinetic from in vitro data. Besides the clinical study that was used to build the model, other independent intravenously administered irinotecan studies were used for verification. Results: The model predicted the plasma concentration of the parent drug and its metabolite (SN-38) in good agreement with observed data from the clinical study that was used to build the model, as well as from several clinical studies that were not used to build the initial model, covering a dose range of 50-750 mg/m2. Conclusion: We developed a PBPK model that accurately predicts the pharmacokinetic behavior of irinotecan and its metabolite (SN-38) in an adult cancer population. Future directions involve the adaption of the model to the pediatric cancer population to predict the optimal dose for initiation of first-in-human Phase I studies and to identify potential drug interactions and unanticipated toxicities when this “backbone” regimen is combined with other investigated anticancer agents. This will then lead to an increase in individualized treatment. Citation Format: Kristina Zoran Denic, Sibylle Neuhoff, Joel Reid, Rachel Kudget. A physiologically based pharmacokinetic modeling approach for predicting the exposure of irinotecan and its active metabolite (SN-38) in cancer patients [abstract]. In: Proceedings of the AACR-NCI-EORTC Virtual International Conference on Molecular Targets and Cancer Therapeutics; 2023 Oct 11-15; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2023;22(12 Suppl):Abstract nr A140.