Citations

Abstract Background Alzheimer’s disease (AD) is a progressive neurological disorder with no specific curative medications. Sophisticated clinical skills are crucial to optimize treatment regimens given the multiple coexisting comorbidities in the patient population. Objective Here, we propose a study to leverage reinforcement learning (RL) to learn the clinicians’ decisions for AD patients based on the longitude data from electronic health records. Methods In this study, we selected 1736 patients from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. We focused on the two most frequent concomitant diseases—depression, and hypertension, thus creating 5 data cohorts (ie, Whole Data, AD, AD-Hypertension, AD-Depression, and AD-Depression-Hypertension). We modeled the treatment learning into an RL problem by defining states, actions, and rewards. We built a regression model and decision tree to generate multiple states, used six combinations of medications (ie, cholinesterase inhibitors, memantine, memantine-cholinesterase inhibitors, hypertension drugs, supplements, or no drugs) as actions, and Mini-Mental State Exam (MMSE) scores as rewards. Results Given the proper dataset, the RL model can generate an optimal policy (regimen plan) that outperforms the clinician’s treatment regimen. Optimal policies (ie, policy iteration and Q-learning) had lower rewards than the clinician’s policy (mean −3.03 and −2.93 vs. −2.93, respectively) for smaller datasets but had higher rewards for larger datasets (mean −4.68 and −2.82 vs. −4.57, respectively). Conclusions Our results highlight the potential of using RL to generate the optimal treatment based on the patients’ longitude records. Our work can lead the path towards developing RL-based decision support systems that could help manage AD with comorbidities.

VEXAS is caused by somatic mutations in UBA1 (UBA1mut) and characterized by heterogenous systemic auto-inflammation and progressive hematologic manifestations, meeting criteria for myelodysplastic syndrome (MDS) and plasma cell dyscrasias. The landscape of myeloid-related gene mutations leading to typical clonal hematopoiesis (CH) in these patients is unknown. Retrospectively, we screened 80 VEXAS patients for CH in their peripheral blood (PB) and correlated findings with clinical outcomes in 77. UBA1mutwere most common at hotspot p.M41 (median variant allele frequency/VAF = 75%). Typical CH mutations co-occurred with UBA1mut in 60% of patients, mostly in DNMT3A and TET2, and were not associated with inflammatory or hematologic manifestations. In prospective single-cell proteogenomic sequencing (scDNA), UBA1mutwas the dominant clone, present mostly in branched clonal trajectories. Based on integrated bulk and scDNA analyses, clonality in VEXAS followed two major patterns: with either typical CH preceding UBA1mutselection in a clone (Pattern 1), or occurring as an UBA1mutsubclone or in independent clones (Pattern 2). VAF in PB differed markedly between DNMT3A and TET2 clones (median VAF of 25% vs 1%). DNMT3A and TET2 clones associated with hierarchies representing patterns 1 and 2, respectively. Overall survival for all patients was 60% at 10 years. Transfusion-dependent anemia, moderate thrombocytopenia, and typical CH mutations, each correlated with poor outcome. In VEXAS, UBA1mut cells are the primary cause of systemic inflammation and marrow failure, being a new molecularly defined somatic entity associated with MDS. VEXAS-associated MDS is distinct from classical MDS in its presentation and clinical course.

Abstract Recurrent chromosomal rearrangements are a hallmark of hematologic malignancies and play critical roles in pathogenesis. The TP53 analog TP63 is rearranged in 5-10% of diverse subtypes of both aggressive T- and B-cell lymphomas. Patients with TP63-rearranged lymphomas have dismal outcomes, with 5-year overall survival rates between 0-17%, depending on cohorts. The function and mechanisms of TP63 rearrangements and TP63 fusion proteins in tumorigenesis are poorly understood. As a result, attempts to treat these patients to date have been largely empiric. Thus, there is an urgent need to understand how TP63 fusions contribute to tumorigenesis and to translate the findings into novel therapeutic options for these patients. Here, we demonstrated that TP63 fusions are essential for the propagation of T-cell lymphomas (TCLs). Knockdown of TP63 fusions with specific shRNAs in TCL cell lines harboring TP63 fusions suppressed both cell growth in vitro and tumor growth in vivo. Retroviral expression of TBL1XR1-TP63, the most common TP63 fusion, conferred cytokine independence in Ba/F3 cells, consistent with its role as an oncogene. To investigate the role of TP63 fusions in T- and B-cell lymphomagenesis, we engineered a CAG-Loxp-Stop-Loxp-TBL1XR1-TP63 conditional knock-in mouse model and crossed with hCD2-Cre mice. This results in expression beginning during early lymphoid development. As observed in patients, transgenic mice developed multiple subtypes of both T- and B-cell lymphoma. To define the effects and mechanisms of TP63 fusions within T cells, we performed CRISPR scanning, transcriptomic, epigenomic, and proteomic analyses. Our data showed that domains within both the N-terminal TBL1XR1 and C-terminal TP63 portions contribute to the function of this fusion. We found that the N-terminal component of TP63 fusions interacts with components of the NCOR/SMRT complex. At the same time, the C-terminal portion of TP63 (which recapitulates the deltaN-p63 isoform expressed in some carcinomas) interacts with the enhancer modifier KMT2D and its complex members. TBL1XR1-TP63 binds to a novel distal enhancer to drive MYC expression, and thus upregulates the expression of the histone H3K27 methylase EZH2. Finally, we assessed whether EZH2 is a vulnerability of TP63-rearranged lymphomas. We found that knockdown of EZH2 in TP63-rearranged lines significantly impaired cell growth, as did treatment with the EZH2 and 1 dual inhibitor valemetostat. Valemetostat, which is now being tested in patients with lymphoma, counteracted the oncogenic effects of TP63 fusions in multiple preclinical models in vivo. Together, our results identify the TP63 fusion as a highly unique oncogenic driver in lymphomagenesis capable of recruiting multiple epigenetic modifier complexes and inducing a targetable dependence on EZH2. Citation Format: Gongwei Wu, Noriaki Yoshida, Jihe Liu, Xiaoyang Zhang, Yuan Xiong, Tayla Heavican-Foral, Huiyun Liu, Geoffrey Nelson, Lu Yang, Renee Chen, Katherine Donovan, Marcus Jones, Mikhail Roshal, Yanming Zhang, Ran Xu, Ajit Nirmal, Salvia Jain, Catharine Leahy, Kristen Jones, Kristen Stevenson, Natasha Galasso, Nivetha Ganesan, Tiffany Chang, Wen-Chao Wu, Abner Louissaint, Lydie Debaize, Hojong Yoon, Paola Dal Cin, Wing Chan Chan, Shannan Ho Sui Ho Sui, Samuel Ng, Andrew Feldman, Steven M. Horwitz, Mathew Meyerson, Karen Adelman, Eric Fischer, Chun-Wei Chen, David Weinstock, Myles Brown. TP63 fusions drive enhancer rewiring, lymphomagenesis, and dependence on EZH2. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5755.