Open Access

An unresolved issue in contemporary biomedicine is the overwhelming number and diversity of complex images that require annotation, analysis and interpretation. Recent advances in Deep Learning have revolutionized the field of computer vision, creating algorithms that compete with human experts in image segmentation tasks. However, these frameworks require large human-annotated datasets for training and the resulting “black box” models are difficult to interpret. In this study, we introduce Kartezio, a modular Cartesian Genetic Programming-based computational strategy that generates fully transparent and easily interpretable image processing pipelines by iteratively assembling and parameterizing computer vision functions. The pipelines thus generated exhibit comparable precision to state-of-the-art Deep Learning approaches on instance segmentation tasks, while requiring drastically smaller training datasets. This Few-Shot Learning method confers tremendous flexibility, speed, and functionality to this approach. We then deploy Kartezio to solve a series of semantic and instance segmentation problems, and demonstrate its utility across diverse images ranging from multiplexed tissue histo-pathology images to high resolution microscopy images. While the flexibility, robustness and practical utility of Kartezio make this fully explicable evolutionary designer a potential game-changer in the field of biomedical image processing, Kartezio remains complementary and potentially auxiliary to mainstream Deep Learning approaches.

Background: Besides the promising application potential of nanotechnologies in engineering, the use of nanomaterials in medicine is growing. New therapies employing innovative nanocarrier systems to increase specificity and efficacy of drug deliveryschemes to reach non-operable structures are already in clinical trials. However the influence of the nanoparticles (NPs) themselves is still unknown in medical applications, especially for complex interactions in least investigatable neural systems. The aim of this study was to investigate in vitro effects of coated silvernanoparticles (cAg-NPs) on the excitability of single neuronal cells and to integrate those findings into an in silico model to predict possible effects from single cells up to neuronal circuits and finally toneural field potentials generated by those circuits. Methods: First, patch-clamp measurements were performed to investigate the effects of nano-sized silver particles, surrounded by an organic coating, on excitability of single cells. Second, it was determined which parameters were altered by exposure to those nanoparticles using the Hodgkin-Huxley model of the sodium current. As a third step those findings were integrated into a well defined neuronal circuit of thalamocortical interactions to predict possible changes in network signaling due to the applied cAg-NPs, in silico. Fourthly, the model was extended to observe neural fields originating from Hodgkin-Huxley type neurons. Therefore it was investigated how the neural field potentials influence the spike generation in neurons that are physically located within these fields, if this feedback causes relevant changes in the underlying neuronal signaling within the circuit, and most important if the cAg-NPs effects on single neurons of the network are strong enough to cause observable changes in the generated field potentials themselves. Results: A rapid suppression of sodium currents was observed after exposure to cAg-NPs in the in vitro recordings. In numerical simulations of sodium currents the parameters most likely affected by cAg-NPs were identified. The effects of such changes on the activity of networks were then examined. In silico network modeling indicated effects of local cAg-NP application on firing patterns in all neurons in the circuit. It has been shown that field potentials have strong effects on the action potential generation of neurons that are exposed to those fields. Furthermore, it was also shown that this is also affecting the underlying neuronal signaling. The assumed cAg-NPs presence in the circuit’s thalamic cells were finally found to also have distinctive effects on the emerging neural field potentials. Conclusion: The sodium current measurements and simulations show that suppression of sodium currents by cAg-NPs results primarily in a reduction in the current amplitude right after a few seconds of particle addition. The network simulations on larger scale show that locally cAg-NPs induced changes result in diversification of activity in the entire circuit. This was also found for the field potential simulations on a more larger scale. The results indicated that local application of cAg-NPs may influence the activity throughout the network and may cause distortions in cortical field potentials in vivo. This multiscale model may subserve as basic approach to estimate the NPs affected spatiotemporal dynamics of cortical field potentials on a very small cortical patch. The electrophysiological detection of this simulated effect by utilizing the voltage sensitive dyes technique is part of the future work that will be carried out by the group ”Systems Neuroscience and Neurotechnology Unit”

The restoration of neuronal activity after injury or during aging requires neuron repopulation at the site of injury, directional regeneration of new nerves and oriented generation of new synapses. The ECM protein Laminin is abundant in neuronal microenvironment and is known to be involved in directing neuronal migration, differentiation and neurite development. In this thesis, a strategy for in vitro directional neurite growth in soft hydrogels is presented. It is based on the spatiotemporal control of the availability of Laminin adhesive motifs within synthetic hydrogels using light as an external guiding trigger. Different variants of Laminin mimetic peptides containing the IKVAV were selected as ligands to mediate control over axonal growth on biomaterials. The photo- cleavable groups 3-(4,5-dimethoxy-2-nitrophenyl)-2-butanol (DMNPB), 6-nitroveratryl alcohol (NVOC) and 2,2'-((3'-(1-hydroxypropan-2-yl)-4'-nitro-[1,1'-biphenyl] 4-yl)azanediyl)bis(ethan-1-ol) (HANBP) were inserted at the K rest of the peptide to temporally block IKVAV bioactivity. Poly(acrylamide) (PAAm) hydrogel films with varied stiffness from 0.2-70 kPa were used as 2D substrates to study IKVAV-guided directional growth of axons. Two novel acryl monomers carrying methylsulfone (MS) side chains were developed to tune specific coupling of thiol terminated IKVAV to the PAAm gel at physiological conditions. The ability of the photoactivatable IKVAV-containing peptides to trigger and support neurite outgrowth was studied and compared in different cell biology experiments using neural progenitor cells from mouse embryo. The ability of the photoactivatable IKVAV-containing peptides to trigger and support spatial organization of neurons was demonstrated by using masked irradiation. The in-situ light exposure of IK(HANBP)VAV by scanning lasers allowed spatially directed neurite development in 2D cell cultures. In the last part of the Thesis, an attempt to extend the photoactivation strategy to 3D environments was made by using two-photon activatable chromophores. The p-methoxynitrobiphenyl (PMNB) photoremovable group was introduced at aspartic acid residue of RGD sequence, a common adhesive motif used for cell attachment to biomaterials. Degradable hydrogels modified with RGD(PMNB)fC peptide were developed and 3D resolved spatial photoactivation inside the gel using two-photon laser guided migration of fibroblasts L929 within the 3D network was established. These results demonstrate that photoactivatable adhesive peptides can be used for spatiotemporal activation of attachment, migration and directional growth of cells in 2D and 3D cultures and provide a tool to control and pattern cell processes in relevant biomedical applications

Cancer immunotherapy has transformed cancer treatment, with chimeric antigen receptor (CAR)-T cell therapy being one of the most promising approaches. In the previous issue of Molecular Therapy – Oncolytics, Lainšček et al.1 outline a novel strategy for controlling CD19 CAR-T cell activity to address limitations currently hampering clinical practice. To better and dynamically align CAR-T cell activity with clinical needs, the authors developed a method for external control based on engineered endogenous transcription factors acting downstream of the CAR signaling pathway.

There are no targeted therapies for patients with triple-negative breast cancer (TNBC). TNBC is enriched in breast cancer stem cells (BCSC), which play a key role in metastasis, chemoresistance, relapse, and mortality. γδ T cells hold great potential in immunotherapy against cancer and might provide an approach to therapeutically target TNBC. γδ T cells are commonly observed to infiltrate solid tumors and have an extensive repertoire of tumor-sensing mechanisms, recognizing stress-induced molecules and phosphoantigens (pAgs) on transformed cells. Herein, we show that patient-derived triple-negative BCSCs are efficiently recognized and killed by ex vivo expanded γδ T cells from healthy donors. Orthotopically xenografted BCSCs, however, were refractory to γδ T-cell immunotherapy. We unraveled concerted differentiation and immune escape mechanisms: xenografted BCSCs lost stemness, expression of γδ T-cell ligands, adhesion molecules, and pAgs, thereby evading immune recognition by γδ T cells. Indeed, neither promigratory engineered γδ T cells, nor anti–PD-1 checkpoint blockade, significantly prolonged overall survival of tumor-bearing mice. BCSC immune escape was independent of the immune pressure exerted by the γδ T cells and could be pharmacologically reverted by zoledronate or IFNα treatment. These results pave the way for novel combinatorial immunotherapies for TNBC.