Prevalence and Therapeutic Targeting of High-Level ERBB2 Amplification in NSCLC.

INTRODUCTION: ERBB2 amplification in lung cancer remains poorly characterized. HER2 (encoded by ERBB2) is a transmembrane tyrosine kinase capable of ligand-independent dimerization and signaling when overexpressed, and a common cause of HER2 overexpression is ERBB2 amplification. Here, we evaluated the clinicopathologic and genomic characteristics of ERBB2-amplified NSCLC and explored a HER2 antibody-drug conjugate (ADC) therapeutic strategy. METHODS: Our institutional next-generation DNA sequencing data (OncoPanel) from 5769 NSCLC samples (5075 patients) were queried for cases having high-level ERBB2 amplification (≥6 copies). Clinical and demographic characteristics were extracted from the electronic medical records. Efficacy of the pan-ERBB inhibitor afatinib or HER2 ADCs (trastuzumab deruxtecan and trastuzumab emtansine) was evaluated in NSCLC preclinical models and patients with ERBB2 amplification. RESULTS: High-level ERBB2 amplification was identified in 0.9% of lung adenocarcinomas and reliably predicted overexpression of HER2. ERBB2 amplification events are detected in two distinct clinicopathologic and genomic subsets of NSCLC: as the sole mitogenic driver in tumors arising in patients with a smoking history or as a concomitant alteration with other mitogenic drivers in patients with a light or never smoking history. We further reveal that trastuzumab deruxtecan is effective therapy in in vitro and in vivo preclinical models of NSCLC harboring ERBB2 amplification and report two cases of clinical activity of an anti-HER2 ADC in patients who acquired ERBB2 amplification after previous targeted therapy. CONCLUSIONS: High-level ERBB2 amplification reliably predicts HER2 overexpression in patients with NSCLC, and HER2 ADC is effective therapy in this population.

Machine-learning-based Structural Analysis of Interactions between Antibodies and Antigens.

Computational analysis of paratope-epitope interactions between antibodies and their corresponding antigens can facilitate our understanding of the molecular mechanism underlying humoral immunity and boost the design of new therapeutics for many diseases. The recent breakthrough in artificial intelligence has made it possible to predict protein-protein interactions and model their structures. Unfortunately, detecting antigen-binding sites associated with a specific antibody is still a challenging problem. To tackle this challenge, we implemented a deep learning model to characterize interaction patterns between antibodies and their corresponding antigens. With high accuracy, our model can distinguish between antibody-antigen complexes and other types of protein-protein complexes. More intriguingly, we can identify antigens from other common protein binding regions with an accuracy of higher than 70% even if we only have the epitope information. This indicates that antigens have distinct features on their surface that antibodies can recognize. Additionally, our model was unable to predict the partnerships between antibodies and their particular antigens. This result suggests that one antigen may be targeted by more than one antibody and that antibodies may bind to previously unidentified proteins. Taken together, our results support the precision of antibody-antigen interactions while also suggesting positive future progress in the prediction of specific pairing.

Mechanical vibration patterns elicit behavioral transitions and habituation in crawling Drosophila larvae.

How animals respond to repeatedly applied stimuli, and how animals respond to mechanical stimuli in particular, are important questions in behavioral neuroscience. We study adaptation to repeated mechanical agitation using the Drosophila larva. Vertical vibration stimuli elicit a discrete set of responses in crawling larvae: continuation, pause, turn, and reversal. Through high-throughput larva tracking, we characterize how the likelihood of each response depends on vibration intensity and on the timing of repeated vibration pulses. By examining transitions between behavioral states at the population and individual levels, we investigate how the animals habituate to the stimulus patterns. We identify time constants associated with desensitization to prolonged vibration, with re-sensitization during removal of a stimulus, and additional layers of habituation that operate in the overall response. Known memory-deficient mutants exhibit distinct behavior profiles and habituation time constants. An analogous simple electrical circuit suggests possible neural and molecular processes behind adaptive behavior.

Vepafestinib is a pharmacologically advanced RET-selective inhibitor with high CNS penetration and inhibitory activity against RET solvent front mutations.

RET receptor tyrosine kinase is activated in various cancers (lung, thyroid, colon and pancreatic, among others) through oncogenic fusions or gain-of-function single-nucleotide variants. Small-molecule RET kinase inhibitors became standard-of-care therapy for advanced malignancies driven by RET. The therapeutic benefit of RET inhibitors is limited, however, by acquired mutations in the drug target as well as brain metastasis, presumably due to inadequate brain penetration. Here, we perform preclinical characterization of vepafestinib (TAS0953/HM06), a next-generation RET inhibitor with a unique binding mode. We demonstrate that vepafestinib has best-in-class selectivity against RET, while exerting activity against commonly reported on-target resistance mutations (variants in RETL730, RETV804 and RETG810), and shows superior pharmacokinetic properties in the brain when compared to currently approved RET drugs. We further show that these properties translate into improved tumor control in an intracranial model of RET-driven cancer. Our results underscore the clinical potential of vepafestinib in treating RET-driven cancers.

Triple-negative Breast Carcinoma With Apocrine and Histiocytoid Features: A Clinicopathologic and Molecular Study of 18 Cases.

Triple-negative breast cancer (TNBC) is a heterogenous group of tumors. Most TNBCs are high-grade aggressive tumors, but a minority of TNBCs are not high grade, with relatively indolent behavior and specific morphologic and molecular features. We performed a clinicopathologic and molecular assessment of 18 non-high-grade TNBCs with apocrine and/or histiocytoid features. All were grade I or II with low Ki-67 (≤20%). Thirteen (72%) showed apocrine features, and 5 (28%) showed histiocytoid and lobular features. In all, 17/18 expressed the androgen receptor, and 13/13 expressed gross cystic disease fluid protein 15. Four (22.2%) patients were treated with neoadjuvant chemotherapy, but none achieved a pathologic complete response. In all, 2/18 patients (11%) had lymph node metastasis at the time of surgery. None of the cases had a recurrence or disease-specific death, with an average follow-up time of 38 months. Thirteen cases were profiled by targeted capture-based next-generation DNA sequencing. Genomic alterations (GAs) were most significant for PI3K-PKB/Akt pathway (69%) genes, including PIK3R1 (23%), PIK3CA (38%), and PTEN (23%), and RTK-RAS pathway (62%) including FGFR4 (46%) and ERBB2 (15%). TP53 GA was seen in only 31% of patients. Our findings support those on high-grade TNBCs with apocrine and/or histiocytoid features as a clinicopathologic and genetically distinct subgroup of TNBC. They can be defined by features including tubule formation, rare mitosis, low Ki-67 (≤20%), triple-negative status, expression of androgen receptor and/or gross cystic disease fluid protein 15, and GA in the PI3K-PKB/Akt and/or RTK-RAS pathway. These tumors are not sensitive to chemotherapy but have favorable clinical behavior. Tumor subtype definitions are the first step to implementing future trial designs to select these patients.

Estimating random close packing in polydisperse and bidisperse hard spheres via an equilibrium model of crowding.

We show that an analogy between crowding in fluid and jammed phases of hard spheres captures the density dependence of the kissing number for a family of numerically generated jammed states. We extend this analogy to jams of mixtures of hard spheres in d = 3 dimensions and, thus, obtain an estimate of the random close packing volume fraction, ϕRCP, as a function of size polydispersity. We first consider mixtures of particle sizes with discrete distributions. For binary systems, we show agreement between our predictions and simulations using both our own results and results reported in previous studies, as well as agreement with recent experiments from the literature. We then apply our approach to systems with continuous polydispersity using three different particle size distributions, namely, the log-normal, Gamma, and truncated power-law distributions. In all cases, we observe agreement between our theoretical findings and numerical results up to rather large polydispersities for all particle size distributions when using as reference our own simulations and results from the literature. In particular, we find ϕRCP to increase monotonically with the relative standard deviation, sσ, of the distribution and to saturate at a value that always remains below 1. A perturbative expansion yields a closed-form expression for ϕRCP that quantitatively captures a distribution-independent regime for sσ < 0.5. Beyond that regime, we show that the gradual loss in agreement is tied to the growth of the skewness of size distributions.

Shadow enhancers modulate distinct transcriptional parameters that differentially effect downstream patterning events.

Transcription in the early Drosophila blastoderm is coordinated by the collective action of hundreds of enhancers. Many genes are controlled by so-called 'shadow enhancers', which provide resilience to environment or genetic insult, allowing the embryo to robustly generate a precise transcriptional pattern. Emerging evidence suggests that many shadow enhancer pairs do not drive identical expression patterns, but the biological significance of this remains unclear. In this study, we characterize the shadow enhancer pair controlling the gene short gastrulation (sog). We removed either the intronic proximal enhancer or the upstream distal enhancer and monitored sog transcriptional kinetics. Notably, each enhancer differs in sog spatial expression, timing of activation and RNA Polymerase II loading rates. In addition, modeling of individual enhancer activities demonstrates that these enhancers integrate activation and repression signals differently. Whereas activation is due to the sum of the two enhancer activities, repression appears to depend on synergistic effects between enhancers. Finally, we examined the downstream signaling consequences resulting from the loss of either enhancer, and found changes in tissue patterning that can be explained by the differences in transcriptional kinetics measured.

Robust identification of temporal biomarkers in longitudinal omics studies.

MOTIVATION: Longitudinal studies increasingly collect rich 'omics' data sampled frequently over time and across large cohorts to capture dynamic health fluctuations and disease transitions. However, the generation of longitudinal omics data has preceded the development of analysis tools that can efficiently extract insights from such data. In particular, there is a need for statistical frameworks that can identify not only which omics features are differentially regulated between groups but also over what time intervals. Additionally, longitudinal omics data may have inconsistencies, including non-uniform sampling intervals, missing data points, subject dropout and differing numbers of samples per subject. RESULTS: In this work, we developed OmicsLonDA, a statistical method that provides robust identification of time intervals of temporal omics biomarkers. OmicsLonDA is based on a semi-parametric approach, in which we use smoothing splines to model longitudinal data and infer significant time intervals of omics features based on an empirical distribution constructed through a permutation procedure. We benchmarked OmicsLonDA on five simulated datasets with diverse temporal patterns, and the method showed specificity greater than 0.99 and sensitivity greater than 0.87. Applying OmicsLonDA to the iPOP cohort revealed temporal patterns of genes, proteins, metabolites and microbes that are differentially regulated in male versus female subjects following a respiratory infection. In addition, we applied OmicsLonDA to a longitudinal multi-omics dataset of pregnant women with and without preeclampsia, and OmicsLonDA identified potential lipid markers that are temporally significantly different between the two groups. AVAILABILITY AND IMPLEMENTATION: We provide an open-source R package (https://bioconductor.org/packages/OmicsLonDA), to enable widespread use. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Tracking Changes in Students' Online Self-Regulated Learning Behaviors and Achievement Goals Using Trace Clustering and Process Mining.

Success in online and blended courses requires engaging in self-regulated learning (SRL), especially for challenging STEM disciplines, such as physics. This involves students planning how they will navigate course assignments and activities, setting goals for completion, monitoring their progress and content understanding, and reflecting on how they completed each assignment. Based on Winne & Hadwin's COPES model, SRL is a series of events that temporally unfold during learning, impacted by changing internal and external factors, such as goal orientation and content difficulty. Thus, as goal orientation and content difficulty change throughout a course, so might students' use of SRL processes. This paper studies how students' SRL behavior and achievement goal orientation change over time in a large (N = 250) college introductory level physics course taught online. Students' achievement goal orientation was measured by repeated administration of the achievement goals questionnaire-revised (AGQ-R). Students' SRL behavior was measured by analyzing their clickstream event traces interacting with online learning modules via a combination of trace clustering and process mining. Event traces were first divided into groups similar in nature using agglomerative clustering, with similarity between traces determined based on a set of derived characteristics most reflective of students' SRL processes. We then generated causal nets for each cluster of traces via process mining and interpreted the underlying behavior and strategy of each causal net according to the COPES SRL framework. We then measured the frequency at which students adopted each causal net and assessed whether the adoption of different causal nets was associated with responses to the AGQ-R. By repeating the analysis for three sets of online learning modules assigned at the beginning, middle, and end of the semester, we examined how the frequency of each causal net changed over time, and how the change correlated with changes to the AGQ-R responses. Results have implications for measuring the temporal nature of SRL during online learning, as well as the factors impacting the use of SRL processes in an online physics course. Results also provide guidance for developing online instructional materials that foster effective SRL for students with different motivational profiles.