Longitudinal characterization of primary osteosarcoma and derived subcutaneous and orthotopic relapsed patient-derived xenograft models.

Osteosarcoma is a rare bone cancer in adolescents and young adults with a dismal prognosis because of metastatic disease and chemoresistance. Despite multiple clinical trials, no improvement in outcome has occurred in decades. There is an urgent need to better understand resistant and metastatic disease and to generate in vivo models from relapsed tumors. We developed eight new patient-derived xenograft (PDX) subcutaneous and orthotopic/paratibial models derived from patients with recurrent osteosarcoma and compared the genetic and transcriptomic landscapes of the disease progression at diagnosis and relapse with the matching PDX. Whole exome sequencing showed that driver and copy-number alterations are conserved from diagnosis to relapse, with the emergence of somatic alterations of genes mostly involved in DNA repair, cell cycle checkpoints, and chromosome organization. All PDX patients conserve most of the genetic alterations identified at relapse. At the transcriptomic level, tumor cells maintain their ossification, chondrocytic, and trans-differentiation programs during progression and implantation in PDX models, as identified at the radiological and histological levels. A more complex phenotype, like the interaction with immune cells and osteoclasts or cancer testis antigen expression, seemed conserved and was hardly identifiable by histology. Despite NSG mouse immunodeficiency, four of the PDX models partially reconstructed the vascular and immune-microenvironment observed in patients, among which the macrophagic TREM2/TYROBP axis expression, recently linked to immunosuppression. Our multimodal analysis of osteosarcoma progression and PDX models is a valuable resource to understand resistance and metastatic spread mechanisms, as well as for the exploration of novel therapeutic strategies for advanced osteosarcoma.

Clinical outcome of pediatric medulloblastoma patients with Li-Fraumeni syndrome.

BACKGROUND: The prognosis for Li-Fraumeni syndrome (LFS) patients with medulloblastoma (MB) is poor. Comprehensive clinical data for this patient group is lacking, challenging the development of novel therapeutic strategies. Here, we present clinical and molecular data on a retrospective cohort of pediatric LFS MB patients. METHODS: In this multinational, multicenter retrospective cohort study, LFS patients under 21 years with MB and class 5 or class 4 constitutional TP53 variants were included. TP53 mutation status, methylation subgroup, treatment, progression free- (PFS) and overall survival (OS), recurrence patterns, and incidence of subsequent neoplasms were evaluated. RESULTS: The study evaluated 47 LFS individuals diagnosed with MB, mainly classified as DNA methylation subgroup "SHH_3" (86%). The majority (74%) of constitutional TP53 variants represented missense variants. The 2- and 5-year (y-) PFS were 36% and 20%, and 2- and 5y-OS were 53% and 23%, respectively. Patients who received postoperative radiotherapy (RT) (2y-PFS: 44%, 2y-OS: 60%) or chemotherapy before RT (2y-PFS: 32%, 2y-OS: 48%) had significantly better clinical outcome then patients who were not treated with RT (2y-PFS: 0%, 2y-OS: 25%). Patients treated according to protocols including high-intensity chemotherapy and patients who received only maintenance-type chemotherapy showed similar outcomes (2y-PFS: 42% and 35%, 2y-OS: 68% and 53%, respectively). CONCLUSIONS: LFS MB patients have a dismal prognosis. In the presented cohort use of RT significantly increased survival rates, whereas chemotherapy intensity did not influence their clinical outcome. Prospective collection of clinical data and development of novel treatments are required to improve the outcome of LFS MB patients.

Oncogenic chimeric transcription factors drive tumor-specific transcription, processing, and translation of silent genomic regions.

Many cancers are characterized by gene fusions encoding oncogenic chimeric transcription factors (TFs) such as EWS::FLI1 in Ewing sarcoma (EwS). Here, we find that EWS::FLI1 induces the robust expression of a specific set of novel spliced and polyadenylated transcripts within otherwise transcriptionally silent regions of the genome. These neogenes (NGs) are virtually undetectable in large collections of normal tissues or non-EwS tumors and can be silenced by CRISPR interference at regulatory EWS::FLI1-bound microsatellites. Ribosome profiling and proteomics further show that some NGs are translated into highly EwS-specific peptides. More generally, we show that hundreds of NGs can be detected in diverse cancers characterized by chimeric TFs. Altogether, this study identifies the transcription, processing, and translation of novel, specific, highly expressed multi-exonic transcripts from otherwise silent regions of the genome as a new activity of aberrant TFs in cancer.

Forecasting the outcome of spintronic experiments with Neural Ordinary Differential Equations.

Deep learning has an increasing impact to assist research, allowing, for example, the discovery of novel materials. Until now, however, these artificial intelligence techniques have fallen short of discovering the full differential equation of an experimental physical system. Here we show that a dynamical neural network, trained on a minimal amount of data, can predict the behavior of spintronic devices with high accuracy and an extremely efficient simulation time, compared to the micromagnetic simulations that are usually employed to model them. For this purpose, we re-frame the formalism of Neural Ordinary Differential Equations to the constraints of spintronics: few measured outputs, multiple inputs and internal parameters. We demonstrate with Neural Ordinary Differential Equations an acceleration factor over 200 compared to micromagnetic simulations for a complex problem - the simulation of a reservoir computer made of magnetic skyrmions (20 minutes compared to three days). In a second realization, we show that we can predict the noisy response of experimental spintronic nano-oscillators to varying inputs after training Neural Ordinary Differential Equations on five milliseconds of their measured response to a different set of inputs. Neural Ordinary Differential Equations can therefore constitute a disruptive tool for developing spintronic applications in complement to micromagnetic simulations, which are time-consuming and cannot fit experiments when noise or imperfections are present. Our approach can also be generalized to other electronic devices involving dynamics.

Clinical and molecular analysis of smoothened inhibitors in Sonic Hedgehog medulloblastoma.

BACKGROUND: Smoothened inhibitors (SMOi) have shown activity in Sonic Hedgehog (SHH) medulloblastoma, however this therapeutic class was not developed in children due to severe effects reported on growth. We hereby report long-term follow-up of young patients treated with SMOi for recurrent medulloblastoma. METHODS: Clinical data on response and toxicity from patients treated with vismodegib or sonidegib from 2011 to 2019 for a SHH medulloblastoma were retrospectively reviewed. Methylation analysis and whole exome sequencing were performed whenever possible. RESULTS: All patients with a somatic PTCH1 mutation responded to SMOi (6/8), including 2 prolonged complete responses. One patient was free of disease 8.2 years after treatment. SMOi was challenged again for 3 patients. Two of them had a response, one with SMOi alone, the other one in combination with temozolomide despite previous progression under monotherapy. SMO resistance mutations were found in patients from biopsy at relapse. Combination with temozolomide or surgery plus radiotherapy was associated with very long disease control in 2 patients. The most severe adverse events were myalgia and growth plate fusion with metaphyseal sclerosis. Normal growth velocity was recovered for 1 patient although her final height was below estimated target height. CONCLUSIONS: Targeting SMO in mutated PTCH1 is an interesting strategy for long-term responses. Combination of SMOi with chemotherapy or surgery and local radiotherapy is an appealing strategy to prevent early resistance and diminish SMOi exposure, especially in young patients. Inhibition of SHH pathway causes growth and development impairment but partial recovery of the growth velocity is possible.

Magnetoresistive Sensor in Two-Dimension on a 25 µm Thick Silicon Substrate for In Vivo Neuronal Measurements.

Neuronal electrical activity is widely studied in vivo, and the ability to measure its magnetic equivalent to obtain an undisturbed signal with both amplitude and direction information leading to neuronal signal mapping would be a promising tool for neuroscience. To provide such a tool, a probe with spin-electronics-based magnetic sensors with orthogonal axes of sensitivity for two directions of measurement is realized, thanks to a local magnetization re-orientation technique induced by Joule heating. This probe is tested under in vivo measurement conditions in the brain of an anesthetized rat. To be as close as possible to neurons and to create minimal damage during the probe's insertion, the tip thickness has been drastically decreased using a silicon-on-insulator substrate. Our probes provide the ability to perform in vivo magnetic measurements on two orthogonal axes on a 25 µm thick silicon tip with a sensitivity of 1.7%/mT along one axis and 0.9%/mT along the perpendicular axis in the sensor plane, for a limit of detection at 1 kHz of 1.0 and 1.3 nT, respectively. These probes have been tested through a phantom study and during an in vivo experiment. The robustness and stability over one year are demonstrated.

Role of non-linear data processing on speech recognition task in the framework of reservoir computing.

The reservoir computing neural network architecture is widely used to test hardware systems for neuromorphic computing. One of the preferred tasks for bench-marking such devices is automatic speech recognition. This task requires acoustic transformations from sound waveforms with varying amplitudes to frequency domain maps that can be seen as feature extraction techniques. Depending on the conversion method, these transformations sometimes obscure the contribution of the neuromorphic hardware to the overall speech recognition performance. Here, we quantify and separate the contributions of the acoustic transformations and the neuromorphic hardware to the speech recognition success rate. We show that the non-linearity in the acoustic transformation plays a critical role in feature extraction. We compute the gain in word success rate provided by a reservoir computing device compared to the acoustic transformation only, and show that it is an appropriate bench-mark for comparing different hardware. Finally, we experimentally and numerically quantify the impact of the different acoustic transformations for neuromorphic hardware based on magnetic nano-oscillators.

Neuromorphic computing with nanoscale spintronic oscillators.

Neurons in the brain behave as nonlinear oscillators, which develop rhythmic activity and interact to process information. Taking inspiration from this behaviour to realize high-density, low-power neuromorphic computing will require very large numbers of nanoscale nonlinear oscillators. A simple estimation indicates that to fit 108 oscillators organized in a two-dimensional array inside a chip the size of a thumb, the lateral dimension of each oscillator must be smaller than one micrometre. However, nanoscale devices tend to be noisy and to lack the stability that is required to process data in a reliable way. For this reason, despite multiple theoretical proposals and several candidates, including memristive and superconducting oscillators, a proof of concept of neuromorphic computing using nanoscale oscillators has yet to be demonstrated. Here we show experimentally that a nanoscale spintronic oscillator (a magnetic tunnel junction) can be used to achieve spoken-digit recognition with an accuracy similar to that of state-of-the-art neural networks. We also determine the regime of magnetization dynamics that leads to the greatest performance. These results, combined with the ability of the spintronic oscillators to interact with each other, and their long lifetime and low energy consumption, open up a path to fast, parallel, on-chip computation based on networks of oscillators.

Tunable inertia of chiral magnetic domain walls.

The time it takes to accelerate an object from zero to a given velocity depends on the applied force and the environment. If the force ceases, it takes exactly the same time to completely decelerate. A magnetic domain wall is a topological object that has been observed to follow this behaviour. Here we show that acceleration and deceleration times of chiral Neel walls driven by current are different in a system with low damping and moderate Dzyaloshinskii-Moriya exchange constant. The time needed to accelerate a domain wall with current via the spin Hall torque is much faster than the time it needs to decelerate once the current is turned off. The deceleration time is defined by the Dzyaloshinskii-Moriya exchange constant whereas the acceleration time depends on the spin Hall torque, enabling tunable inertia of chiral domain walls. Such unique feature of chiral domain walls can be utilized to move and position domain walls with lower current, key to the development of storage class memory devices.

Interface control of the magnetic chirality in CoFeB/MgO heterostructures with heavy-metal underlayers.

Recent advances in the understanding of spin orbital effects in ultrathin magnetic heterostructures have opened new paradigms to control magnetic moments electrically. The Dzyaloshinskii-Moriya interaction (DMI) is said to play a key role in forming a Néel-type domain wall that can be driven by the spin Hall torque. Here we show that the strength and sign of the DMI can be changed by modifying the adjacent heavy-metal underlayer (X) in perpendicularly magnetized X/CoFeB/MgO heterostructures. The sense of rotation of a domain wall spiral is reversed when the underlayer is changed from Hf, Ta to W and the strength of DMI varies as the filling of 5d orbitals, or the electronegativity, of the heavy-metal layer changes. The DMI can even be tuned by adding nitrogen to the underlayer, thus allowing interface engineering of the magnetic texture in ultrathin magnetic heterostructures.