Periodic Reporting for period 1 - nanoPASS (Bridging the gaps in nanosafety for animal-free prediction of adverse outcomes)
Reporting period: 2023-01-01 to 2024-06-30
nanoPASS develops cost-effective, animal-free predictions of long-term adverse outcomes (AOs) of inhaled nanomaterials that can be demonstrated to work in regulatory and industrial settings.
To this end, we are shifting the focus of nanosafety testing from late endpoints to early key events (KEs) leading to adverse outcomes (AOs). As such tests can only be based on mechanistic understanding, we need to close the knowledge gaps and match KEs in vitro and in vivo. The key bridging methods provided by nanoPASS consortium are intravital in vivo microscopy, quantitative time-lapse in vitro microscopies, and automated identification of the modes of action (i.e. KE relationships) with proprietary in silico algorithms, supported by datamining of the worlds’ largest in vivo database and single-cell omics data, and computational modelling of structure-function relationships.
We are implementing this approach in the context of 4 AOs related to inhalation of nanomaterial: reduced lung function (AOP302), cardiovascular disease (AOP237), chronic inflammation (KE in AOP173, fibrosis; renumbered AOP33 after recent endorsement), and neurodegeneration (no AOP proposed), to which we have added cancer (AOP451) beyond the proposal.
The developed in vitro/in silico prediction models are being calibrated against in vivo data with 40+ benchmark materials, and validated with real-life materials at different stages of the life cycle from 5 diverse industrial cases: construction, electronics waste, advanced materials for catalysis, nanoplastics, and medical materials.
To this end, we are shifting the focus of nanosafety testing from late endpoints to early key events (KEs) leading to adverse outcomes (AOs). As such tests can only be based on mechanistic understanding, we need to close the knowledge gaps and match KEs in vitro and in vivo. The key bridging methods provided by nanoPASS consortium are intravital in vivo microscopy, quantitative time-lapse in vitro microscopies, and automated identification of the modes of action (i.e. KE relationships) with proprietary in silico algorithms, supported by datamining of the worlds’ largest in vivo database and single-cell omics data, and computational modelling of structure-function relationships.
We are implementing this approach in the context of 4 AOs related to inhalation of nanomaterial: reduced lung function (AOP302), cardiovascular disease (AOP237), chronic inflammation (KE in AOP173, fibrosis; renumbered AOP33 after recent endorsement), and neurodegeneration (no AOP proposed), to which we have added cancer (AOP451) beyond the proposal.
The developed in vitro/in silico prediction models are being calibrated against in vivo data with 40+ benchmark materials, and validated with real-life materials at different stages of the life cycle from 5 diverse industrial cases: construction, electronics waste, advanced materials for catalysis, nanoplastics, and medical materials.
With the combination of bottom-up (from in vitro observations) and top-down approach (from in vivo data), we have compiled more than 40 KE candidates for the addressed AOs, among which we next aim to identify bridging events to finalise the corresponding in vitro models.
To track the dynamics of in vitro events, we have consolidated our high-throughput high-resolution time-lapse microscopy with label-free detection of nanomaterial, which is now available at two partner locations for inter-laboratory comparison. For the 40 KE candidates identified above, we tuned the algorithms for quantitative image analysis to extract time-dependent observables for in silico prediction of AOs. A state-of-the-art Stimulated Raman Scattering microscope has been set up for in vitro chemical fingerprinting to characterise materials’ modifications in complex environments in the next stage.
We have incorporated new events into our model for in vitro/in silico translation for quantitative AO prediction. Simultaneous detection of many early events and identification of relations between those events allows us to predict material-specific triggering of material-agnostic outcomes, as different materials can cause the same symptoms via different early MoAs. We have identified new descriptors for in silico characterisation of complex materials, advanced algorithms for prediction of interaction between nanomaterials and biomolecules, and developed a methodology to predict gene expression patterns for exposure to silica, a well characterised benchmark material.
To calibrate our in vitro/in silico prediction models, we have compiled in vivo datasets available from previous projects for 35 benchmark materials (including several from JRC) together with their characterisation.
To showcase industrial application of the prediction tests, we collected samples for 5 diverse families of industrial materials for different applications and at different stages of their life cycle:
· construction: cement, together with intermediate production steps
· electronics: waste from electric and electronic equipment
· catalysis: advanced bimetallic nanoparticles
· plastic consumer products: degraded 3D-printed plastics
· medical: dental filling material
At one of the industrial sites, we have performed personal exposure measurements for risk assessment, based on which we will develop a risk-mitigation plan and estimate the potential global economic benefits of in vitro/in silico tests for product development as well as workforce protection.
To track the dynamics of in vitro events, we have consolidated our high-throughput high-resolution time-lapse microscopy with label-free detection of nanomaterial, which is now available at two partner locations for inter-laboratory comparison. For the 40 KE candidates identified above, we tuned the algorithms for quantitative image analysis to extract time-dependent observables for in silico prediction of AOs. A state-of-the-art Stimulated Raman Scattering microscope has been set up for in vitro chemical fingerprinting to characterise materials’ modifications in complex environments in the next stage.
We have incorporated new events into our model for in vitro/in silico translation for quantitative AO prediction. Simultaneous detection of many early events and identification of relations between those events allows us to predict material-specific triggering of material-agnostic outcomes, as different materials can cause the same symptoms via different early MoAs. We have identified new descriptors for in silico characterisation of complex materials, advanced algorithms for prediction of interaction between nanomaterials and biomolecules, and developed a methodology to predict gene expression patterns for exposure to silica, a well characterised benchmark material.
To calibrate our in vitro/in silico prediction models, we have compiled in vivo datasets available from previous projects for 35 benchmark materials (including several from JRC) together with their characterisation.
To showcase industrial application of the prediction tests, we collected samples for 5 diverse families of industrial materials for different applications and at different stages of their life cycle:
· construction: cement, together with intermediate production steps
· electronics: waste from electric and electronic equipment
· catalysis: advanced bimetallic nanoparticles
· plastic consumer products: degraded 3D-printed plastics
· medical: dental filling material
At one of the industrial sites, we have performed personal exposure measurements for risk assessment, based on which we will develop a risk-mitigation plan and estimate the potential global economic benefits of in vitro/in silico tests for product development as well as workforce protection.
Most importantly, we have successfully fine-tuned the calibration of the first generation of In finite prediction of chronic lung inflammation to cover various early responses. With delivery in less than a week, it clearly demonstrates the potential to address the industrial need for time- and cost-effective long-term hazard assessment of different nanomaterials.
The prediction comprises of monitoring 17 biologically-relevant functional early KEs, relatable to the earliest in vivo KEs and with the AOs throughout the well-defined AOPs, in a live lung-epithelial-response-mimicking in vitro model with automated high-resolution imaging, employing automated in-vitro-to-in-silico translation providing insight into early Mode-of-Action (MoA). This fine-tuned In finite platform is available for industrial testing as a prototype technology, announced at EuroTox and SOT conferences/fairs and through industrial networks. In the next period, nanoPASS aims to validate the aforementioned prediction platform via industrial cases and evolve the second generation of the prediction platforms.
Besides, we would like to stress the following achievements:
• Reduced lung function (AOP302): We have nearly completed the development of our acellular model and prepared a detailed SOP for ECVAM application.
• Chronic inflammation (KE in fibrosis AOP173, renumbered into AOP33 after endorsement): We have contributed to endorsement of AOP33, the first nano-relevant AOP.
• Cardiovascular disease (AOP237): We submitted a proposal for a quantitative AOP to OECD via AOPwiki, which is currently under review for compliance.
• Neurodegeneration (no AOP proposed): We have developed the first environmentally (not genetically!) triggered in vitro model expressing in vivo-relevant events (neurite dystrophy, formation of amyloid-β and tau-containing plaques) (Nanotoxicology 2024). As a proof of principle, we tested it as a Alzheimer’s disease model for prediction of drug efficiency.
• Cancer (AOP451) – beyond the plan: We have identified multiple KE candidates related to abnormal cell division (bioRxiv 2024, medRxiv & Nature Communications 2024) for prediction of materials’ carcinogenicity.
The prediction comprises of monitoring 17 biologically-relevant functional early KEs, relatable to the earliest in vivo KEs and with the AOs throughout the well-defined AOPs, in a live lung-epithelial-response-mimicking in vitro model with automated high-resolution imaging, employing automated in-vitro-to-in-silico translation providing insight into early Mode-of-Action (MoA). This fine-tuned In finite platform is available for industrial testing as a prototype technology, announced at EuroTox and SOT conferences/fairs and through industrial networks. In the next period, nanoPASS aims to validate the aforementioned prediction platform via industrial cases and evolve the second generation of the prediction platforms.
Besides, we would like to stress the following achievements:
• Reduced lung function (AOP302): We have nearly completed the development of our acellular model and prepared a detailed SOP for ECVAM application.
• Chronic inflammation (KE in fibrosis AOP173, renumbered into AOP33 after endorsement): We have contributed to endorsement of AOP33, the first nano-relevant AOP.
• Cardiovascular disease (AOP237): We submitted a proposal for a quantitative AOP to OECD via AOPwiki, which is currently under review for compliance.
• Neurodegeneration (no AOP proposed): We have developed the first environmentally (not genetically!) triggered in vitro model expressing in vivo-relevant events (neurite dystrophy, formation of amyloid-β and tau-containing plaques) (Nanotoxicology 2024). As a proof of principle, we tested it as a Alzheimer’s disease model for prediction of drug efficiency.
• Cancer (AOP451) – beyond the plan: We have identified multiple KE candidates related to abnormal cell division (bioRxiv 2024, medRxiv & Nature Communications 2024) for prediction of materials’ carcinogenicity.