Cooperative regulation of coupled oncoprotein synthesis and stability in triple-negative breast cancer by EGFR and CDK12/13.

Evidence has long suggested that epidermal growth factor receptor (EGFR) may play a prominent role in triple-negative breast cancer (TNBC) pathogenesis, but clinical trials of EGFR inhibitors have yielded disappointing results. Using a candidate drug screen, we identified that inhibition of cyclin-dependent kinases 12 and 13 (CDK12/13) dramatically sensitizes diverse models of TNBC to EGFR blockade. This combination therapy drives cell death through the 4E-BP1-dependent suppression of the translation and translation-linked turnover of driver oncoproteins, including MYC. A genome-wide CRISPR/Cas9 screen identified the CCR4-NOT complex as a major determinant of sensitivity to the combination therapy whose loss renders 4E-BP1 unresponsive to drug-induced dephosphorylation, thereby rescuing MYC translational suppression and promoting MYC stability. The central roles of CCR4-NOT and 4E-BP1 in response to the combination therapy were further underscored by the observation of CNOT1 loss and rescue of 4E-BP1 phosphorylation in TNBC cells that naturally evolved therapy resistance. Thus, pharmacological inhibition of CDK12/13 reveals a long-proposed EGFR dependence in TNBC that functions through the cooperative regulation of translation-coupled oncoprotein stability.

Backdoor Adjustment of Confounding by Provenance for Robust Text Classification of Multi-institutional Clinical Notes.

Natural Language Processing (NLP) methods have been broadly applied to clinical tasks. Machine learning and deep learning approaches have been used to improve the performance of clinical NLP. However, these approaches require sufficiently large datasets for training, and trained models have been shown to transfer poorly across sites. These issues have led to the promotion of data collection and integration across different institutions for accurate and portable models. However, this can introduce a form of bias called confounding by provenance. When source-specific data distributions differ at deployment, this may harm model performance. To address this issue, we evaluate the utility of backdoor adjustment for text classification in a multi-site dataset of clinical notes annotated for mentions of substance abuse. Using an evaluation framework devised to measure robustness to distributional shifts, we assess the utility of backdoor adjustment. Our results indicate that backdoor adjustment can effectively mitigate for confounding shift.

Leveraging patient derived models of FGFR2 fusion positive intrahepatic cholangiocarcinoma to identify synergistic therapies.

Intrahepatic cholangiocarcinoma (ICC) remains a deadly malignancy lacking systemic therapies for advanced disease. Recent advancements include selective FGFR1-3 inhibitors for the 15% of ICC patients harboring fusions, although survival is limited by poor response and resistance. Herein we report generation of a patient-derived FGFR2 fusion-positive ICC model system consisting of a cell line, organoid, and xenograft, which have undergone complete histologic, genomic, and phenotypic characterization, including testing standard-of-care systemic therapies. Using these FGFR2 fusion-positive ICC models, we conducted an unbiased high-throughput small molecule screen to prioritize combination strategies with FGFR inhibition, from which HDAC inhibition together with pemigatinib was validated in vitro and in vivo as a synergistic therapy for ICC. Additionally, we demonstrate broad utility of the FGFR/HDAC combination for other FGFR fusion-positive solid tumors. These data are directly translatable and justify early phase trials to establish dosing, safety, and therapeutic efficacy of this synergistic combination.

P2RY2-AKT activation is a therapeutically actionable consequence of XPO1 inhibition in acute myeloid leukemia.

Selinexor is a first-in-class inhibitor of the nuclear exportin XPO1 that was recently approved by the US Food and Drug Administration for the treatment of multiple myeloma and diffuse large B-cell lymphoma. In relapsed/refractory acute myeloid leukemia (AML), selinexor has shown promising activity, suggesting that selinexor-based combination therapies may have clinical potential. Here, motivated by the hypothesis that selinexor's nuclear sequestration of diverse substrates imposes pleiotropic fitness effects on AML cells, we systematically catalog the pro- and anti-fitness consequences of selinexor treatment. We discover that selinexor activates PI3Kγ-dependent AKT signaling in AML by upregulating the purinergic receptor P2RY2. Inhibiting this axis potentiates the anti-leukemic effects of selinexor in AML cell lines, patient-derived primary cultures and multiple mouse models of AML. In a syngeneic, MLL-AF9-driven mouse model of AML, treatment with selinexor and ipatasertib outperforms both standard-of-care chemotherapy and chemotherapy with selinexor. Together, these findings establish drug-induced P2RY2-AKT signaling as an actionable consequence of XPO1 inhibition in AML.

Re-Examination of the Exacerbating Effect of Inflammasome Components during Radiation Injury.

Radiation can be applied for therapeutic benefit against cancer or may result in devastating harm due to accidental or intentional release of nuclear energy. In all cases, radiation exposure causes molecular and cellular damage, resulting in the production of inflammatory factors and danger signals. Several classes of innate immune receptors sense the released damage associated molecules and activate cellular response pathways, including the induction of inflammasome signaling that impacts IL-1ß/IL-18 maturation and cell death. A previous report indicated inflammasomes aggravate acute radiation syndrome. In contrast, here we find that inflammasome components do not exacerbate gamma-radiation-induced injury by examining heterozygous and gene-deletion littermate controls in addition to wild-type mice. Absence of some inflammasome genes, such as caspase-1/11 and Nlrp3, enhance susceptibility of treated mice to acute radiation injury, indicating importance of the inflammasome pathway in radioprotection. Surprisingly, we discover that the survival outcome may be sex-dependent as more inflammasome-deficient male mice are susceptible to radiation-induced injury. We discuss parameters that may influence the role of inflammasomes as radioprotective or radioexacerbating factors in recovery from radiation injury including the use of littermate controls, the sex of the animals, differences in microbiota within the colonies and other experimental conditions. Under the conditions tested, inflammasome components do not exacerbate radiation injury, but rather provide protective benefit.

Adaptation and selection shape clonal evolution of tumors during residual disease and recurrence.

The survival and recurrence of residual tumor cells following therapy constitutes one of the biggest obstacles to obtaining cures in breast cancer, but it remains unclear how the clonal composition of tumors changes during relapse. We use cellular barcoding to monitor clonal dynamics during tumor recurrence in vivo. We find that clonal diversity decreases during tumor regression, residual disease, and recurrence. The recurrence of dormant residual cells follows several distinct routes. Approximately half of the recurrent tumors exhibit clonal dominance with a small number of subclones comprising the vast majority of the tumor; these clonal recurrences are frequently dependent upon Met gene amplification. A second group of recurrent tumors comprises thousands of subclones, has a clonal architecture similar to primary tumors, and is dependent upon the Jak/Stat pathway. Thus the regrowth of dormant tumors proceeds via multiple routes, producing recurrent tumors with distinct clonal composition, genetic alterations, and drug sensitivities.