T lymphocytes utilize amoeboid migration to navigate effectively within complex microenvironments. The precise rearrangement of the actin cytoskeleton required for cellular forward propulsion is mediated by actin regulators, including the actin‐related protein 2/3 (Arp2/3) complex, a macromolecular machine that nucleates branched actin filaments at the leading edge. The consequences of modulating Arp2/3 activity on the biophysical properties of the actomyosin cortex and downstream T cell function are incompletely understood. We report that even a moderate decrease of Arp3 levels in T cells profoundly affects actin cortex integrity. Reduction in total F‐actin content leads to reduced cortical tension and disrupted lamellipodia formation. Instead, in Arp3‐knockdown cells, the motility mode is dominated by blebbing migration characterized by transient, balloon‐like protrusions at the leading edge. Although this migration mode seems to be compatible with interstitial migration in three‐dimensional environments, diminished locomotion kinetics and impaired cytotoxicity interfere with optimal T cell function. These findings define the importance of finely tuned, Arp2/3‐dependent mechanophysical membrane integrity in cytotoxic effector T lymphocyte activities.

Mechanical stimuli such as tension, compression, and shear stress play critical roles in the physiological functions of red blood cells (RBCs) and their homeostasis, ATP release, and rheological properties. Intracellular calcium (Ca²⁺) mobilization reflects RBC mechanosensing as they transverse the complex vasculature. Emerging studies have demonstrated the presence of mechanosensitive Ca²⁺ permeable ion channels and their function has been implicated in the regulation of RBC volume and deformability. However, how these mechanoreceptors trigger Ca²⁺ influx and subsequent cellular responses are still unclear. Here, we introduce a fluorescence-coupled micropipette aspiration assay to examine RBC mechanosensing at the single-cell level. To achieve a wide range of cell aspirations, we implemented and compared two negative pressure adjusting apparatuses: a homemade water manometer (− 2.94 to 0 mmH₂O) and a pneumatic high-speed pressure clamp (− 25 to 0 mmHg). To visualize Ca²⁺ influx, RBCs were pre-loaded with an intensiometric probe Cal-520 AM, then imaged under a confocal microscope with concurrent bright-field and fluorescent imaging at acquisition rates of 10 frames per second. Remarkably, we observed the related changes in intracellular Ca²⁺ levels immediately after aspirating individual RBCs in a pressure-dependent manner. The RBC aspirated by the water manometer only displayed 1.1-fold increase in fluorescence intensity, whereas the RBC aspirated by the pneumatic clamp showed up to threefold increase. These results demonstrated the water manometer as a gentle tool for cell manipulation with minimal pre-activation, while the high-speed pneumatic clamp as a much stronger pressure actuator to examine cell mechanosensing directly. Together, this multimodal platform enables us to precisely control aspiration and membrane tension, and subsequently correlate this with intracellular calcium concentration dynamics in a robust and reproducible manner.

This paper investigates the adaptability of four state-of-the-art artificial intelligence (AI) models to the Australian mammographic context through transfer learning, explores the impact of image enhancement on model performance and analyses the relationship between AI outputs and histopathological features for clinical relevance and accuracy assessment. A total of 1712 screening mammograms (n = 856 cancer cases and n = 856 matched normal cases) were used in this study. The 856 cases with cancer lesions were annotated by two expert radiologists and the level of concordance between their annotations was used to establish two sets: a ‘high-concordances subset’ with 99% agreement of cancer location and an ‘entire dataset’ with all cases included. The area under the receiver operating characteristic curve (AUC) was used to evaluate the performance of Globally aware Multiple Instance Classifier (GMIC), Global-Local Activation Maps (GLAM), I&H and End2End AI models, both in the pretrained and transfer learning modes, with and without applying the Contrast Limited Adaptive Histogram Equalization (CLAHE) algorithm. The four AI models with and without transfer learning in the high-concordance subset outperformed those in the entire dataset. Applying the CLAHE algorithm to mammograms improved the performance of the AI models. In the high-concordance subset with the transfer learning and CLAHE algorithm applied, the AUC of the GMIC model was highest (0.912), followed by the GLAM model (0.909), I&H (0.893) and End2End (0.875). There were significant differences (p < 0.05) in the performances of the four AI models between the high-concordance subset and the entire dataset. The AI models demonstrated significant differences in malignancy probability concerning different tumour size categories in mammograms. The performance of AI models was affected by several factors such as concordance classification, image enhancement and transfer learning. Mammograms with a strong concordance with radiologists’ annotations, applying image enhancement and transfer learning could enhance the accuracy of AI models.

Tropomyosins (Tpms) are rod-shaped proteins that interact head-to-tail to form a continuous polymer along both sides of most cellular actin filaments. Head-to-tail interaction between adjacent Tpm molecules and the formation of an overlap complex between them leads to the assembly of actin filaments with one type of Tpm isoform in time and space. Variations in the affinity of tropomyosin isoforms for different actin structures are proposed as a potential sorting mechanism. However, the detailed mechanisms of the spatio-temporal sorting of Tpms remain elusive. In this study, we investigated the early intermediates during actin–tropomyosin filament assembly, using a skeletal/cardiac Tpm isoform (Tpm1.1) and a cytoskeletal isoform (Tpm1.6) that differ only in the last 27 amino acids. We investigated how the muscle isoform Tpm1.1 and the cytoskeletal isoform Tpm1.6 nucleate domains on the actin filament, and tested whether (1) recruitment is affected by the actin isoform (muscle vs. cytoskeletal) and (2) whether there is specificity in recruiting the same isoform to a domain at these early stages. To address these questions, actin filaments were exposed to low concentrations of fluorescent tropomyosins in solution. The filaments were immobilized onto glass coverslips and the pattern of decoration was visualized by TIRF microscopy. We show that at the early assembly stage, tropomyosins formed multiple distinct fluorescent domains (here termed “cluster”) on the actin filaments. An automated image analysis algorithm was developed and validated to identify clusters and estimate the number of tropomyosins in each cluster. The analysis showed that tropomyosin isoform sorting onto an actin filament is unlikely to be driven by a preference for nucleating on the corresponding muscle or cytoskeletal actin isoforms, but rather is facilitated by a higher probability of incorporating the same tropomyosin isoforms into an early assembly intermediate. We showed that the 27 amino acids at the end of each tropomyosin seem to provide enough molecular information for the attachment of the same tropomyosin isoforms adjacent to each other on an actin filament. This results in the formation of homogeneous clusters composed of the same isoform rather than clusters with mixed isoforms.

Tropomyosins (Tpm) are rod-shaped proteins that interact head-to-tail to form a continuous polymer along both sides of most cellular actin filaments. Head-to-tail interaction between adjacent Tpm molecules and the formation of an overlap complex between them leads to the assembly of actin filaments with one type of Tpm isoform in time and space. Variations in the affinity of tropomyosin isoforms for different actin structures are proposed as a potential sorting mechanism. However, the detailed mechanisms of spatio-temporal sorting of Tpms remain elusive. In this study, we investigated the early intermediates during actin-tropomyosin filament assembly, using skeletal/cardiac Tpm isoform (Tpm1.1) and a cytoskeletal isoform (Tpm1.6) that differ only in the last 27 amino acids. We investigated how the muscle isoform Tpm1.1 and the cytoskeletal isoform Tpm1.6 nucleate domains on the actin filament and tested whether (1) recruitment is affected by the actin isoform (muscle vs cytoskeletal) and (2) whether there is specificity in recruiting the same isoform to a domain at these early stages. To address these questions, actin filaments were exposed to low concentrations of fluorescent tropomyosins in solution. The filaments were immobilized onto glass coverslips and the pattern of decoration was visualized by TIRF microscopy. We show that at the early assembly stage, tropomyosins formed multiple distinct fluorescent domains (here termed “cluster“) on the actin filaments. An automated image analysis algorithm was developed and validated to identify clusters and estimate the number of tropomyosins in each cluster. The analysis showed that tropomyosin isoform sorting onto an actin filament is unlikely to be driven by a preference for nucleating on the corresponding muscle or cytoskeletal actin isoforms but rather is facilitated by a higher probability of incorporating the same tropomyosin isoforms into an early assembly intermediate. We showed that the 27 amino acids at the end of each tropomyosin seem to provide enough molecular information for attachment of the same tropomyosin isoforms adjacent to each other on an actin filament. This results in the formation of homogeneous clusters composed of the same isoform rather than clusters with mixed isoforms.