Translating the regulatory landscape of medical devices to create fit-for-purpose artificial intelligence (AI) cytometry solutions.

The implementation of medical software and artificial intelligence (AI) algorithms into routine clinical cytometry diagnostic practice requires a thorough understanding of regulatory requirements and challenges throughout the cytometry software product lifecycle. To provide cytometry software developers, computational scientists, researchers, industry professionals, and diagnostic physicians/pathologists with an introduction to European Union (EU) and United States (US) regulatory frameworks. Informed by community feedback and needs assessment established during two international cytometry workshops, this article provides an overview of regulatory landscapes as they pertain to the application of AI, AI-enabled medical devices, and Software as a Medical Device in diagnostic flow cytometry. Evolving regulatory frameworks are discussed, and specific examples regarding cytometry instruments, analysis software and clinical flow cytometry in-vitro diagnostic assays are provided. An important consideration for cytometry software development is the modular approach. As such, modules can be segregated and treated as independent components based on the medical purpose and risk and become subjected to a range of context-dependent compliance and regulatory requirements throughout their life cycle. Knowledge of regulatory and compliance requirements enhances the communication and collaboration between developers, researchers, end-users and regulators. This connection is essential to translate scientific innovation into diagnostic practice and to continue to shape the development and revision of new policies, standards, and approaches.

Recommendations for using artificial intelligence in clinical flow cytometry.

Flow cytometry is a key clinical tool in the diagnosis of many hematologic malignancies and traditionally requires close inspection of digital data by hematopathologists with expert domain knowledge. Advances in artificial intelligence (AI) are transferable to flow cytometry and have the potential to improve efficiency and prioritization of cases, reduce errors, and highlight fundamental, previously unrecognized associations with underlying biological processes. As a multidisciplinary group of stakeholders, we review a range of critical considerations for appropriately applying AI to clinical flow cytometry, including use case identification, low and high risk use cases, validation, revalidation, computational considerations, and the present regulatory frameworks surrounding AI in clinical medicine. In particular, we provide practical guidance for the development, implementation, and suggestions for potential regulation of AI-based methods in the clinical flow cytometry laboratory. We expect these recommendations to be a helpful initial framework of reference, which will also require additional updates as the field matures.

Recommendations for the development and validation of flow cytometry-based receptor occupancy assays.

Receptor occupancy measurements demonstrate the binding of a biotherapeutic agent to its extra-cellular target and represent an integral component of the pharmacodynamic (PD) portfolio utilized to advance the development and commercialization of a therapeutic agent. Coupled with traditional pharmacokinetic (PK) assessments derived from serum drug concentration, receptor occupancy data can be used to model PK/PD relationships and validate dose selection decisions throughout the drug development lifecycle. Receptor occupancy assays can be even more challenging to develop than other flow cytometric methods (e.g. surface immunophenotyping). In addition to typical considerations regarding stability of the cell type of interest, stability of the target-bound therapeutic agent and stability of the target receptor must be taken into account. Reagent selection is also challenging as reagents need to be evaluated for the potential to compete with the therapeutic agent and bind with comparable affinity. This article provides technical guidance for the development and validation of cytometry-based receptor occupancy assays.

Role of receptor occupancy assays by flow cytometry in drug development.

The measurement of the binding of a biotherapeutic to its cellular target, receptor occupancy (RO), is increasingly important in development of biologically-based therapeutic agents. Receptor occupancy (RO) assays by flow cytometry describe the qualitative and/or quantitative assessment of the binding of a therapeutic agent to its cell surface target. Such RO assays can be as simple as measuring the number of cell surface receptors bound by an antireceptor therapeutic agent or can be designed to address more complicated scenarios such as internalization or shedding events once a receptor engages the administered therapeutic agent. Data generated from RO assays can also be used to model whether given doses of an experimental therapeutic agent and their administration schedules lead to predicted levels of receptor occupancy and whether the receptor is modulated (up or down) on cells engaged by the therapeutic agent. There are a variety of approaches that can be used when undertaking RO assays and with the ability to measure distinct subsets in heterogeneous populations, flow cytometry is ideally suited to RO measurements. This article highlights the importance of RO assays on the flow cytometric platform in the development of biotherapeutic agents.

Cheating by type 3 secretion system-negative Pseudomonas aeruginosa during pulmonary infection.

The opportunistic pathogen Pseudomonas aeruginosa expresses a type 3 secretion system (T3SS) strongly associated with bacterial virulence in murine models and human patients. T3SS effectors target host innate immune mechanisms, and T3SS-defective mutants are cleared more efficiently than T3SS-positive bacteria by an immunocompetent host. Nonetheless, T3SS-negative isolates are recovered from many patients with documented P. aeruginosa infections, leading us to test whether T3SS-negative strains could have a selective advantage during in vivo infection. Mice were infected with mixtures of T3SS-positive WT P. aeruginosa plus isogenic T3SS-OFF or constitutively T3SS-ON mutants. Relative fitness of bacteria in this acute pneumonia model was reflected by the competitive index of mutants relative to WT. T3SS-OFF strains outcompeted WT PA103 in vivo, whereas a T3SS-ON mutant showed decreased fitness compared with WT. In vitro growth rates of WT and T3SS-OFF bacteria were determined under T3SS-inducing conditions and did not differ significantly. Increased fitness of T3SS-OFF bacteria was no longer observed at high ratios of T3SS-OFF to WT, a feature characteristic of bacterial cheaters. Cheating by T3SS-OFF bacteria occurred only when T3SS-positive bacteria expressed the phospholipase A2 effector Exotoxin U (ExoU). T3SS-OFF bacteria showed no fitness advantage in competition experiments carried out in immunodeficient MyD88-knockout mice or in neutrophil-depleted animals. Our findings indicate that T3SS-negative isolates benefit from the public good provided by ExoU-mediated killing of recruited innate immune cells. Whether this transient increase in fitness observed for T3SS-negative strains in mice contributes to the observed persistence of T3SS-negative isolates in humans is of ongoing interest.

Examining chemical compound biodegradation at low concentrations through bacterial cell proliferation.

We show proof of principle for assessing compound biodegradation at 1-2 mg C per L by measuring microbial community growth over time with direct cell counting by flow cytometry. The concept is based on the assumption that the microbial community will increase in cell number through incorporation of carbon from the added test compound into new cells in the absence of (as much as possible) other assimilable carbon. We show on pure cultures of the bacterium Pseudomonas azelaica that specific population growth can be measured with as low as 0.1 mg 2-hydroxybiphenyl per L, whereas in mixed community 1 mg 2-hydroxybiphenyl per L still supported growth. Growth was also detected with a set of fragrance compounds dosed at 1-2 mg C per L into diluted activated sludge and freshwater lake communities at starting densities of 10(4) cells per ml. Yield approximations from the observed community growth was to some extent in agreement with standard OECD biodegradation test results for all, except one of the examined compounds.

Reversible and irreversible pollutant-induced bacterial cellular stress effects measured by ethidium bromide uptake and efflux.

Chemical pollution is known to affect microbial community composition but it is poorly understood how toxic compounds influence physiology of single cells that may lay at the basis of loss of reproductive fitness. Here we analyze physiological disturbances of a variety of chemical pollutants at single cell level using the bacterium Pseudomonas fluorescens in an oligotrophic growth assay. As a proxy for physiological disturbance we measured changes in geometric mean ethidium bromide (EB) fluorescence intensities in subpopulations of live and dividing cells exposed or not exposed to different dosages of tetradecane, 4-chlorophenol, 2-chlorobiphenyl, naphthalene, benzene, mercury chloride, or water-dissolved oil fractions. Because ethidium bromide efflux is an energy-dependent process any disturbance in cellular energy generation is visible as an increased cytoplasmic fluorescence. Interestingly, all pollutants even at the lowest dosage of 1 nmol/mL culture produced significantly increased ethidium bromide fluorescence compared to nonexposed controls. Ethidium bromide fluorescence intensities increased upon pollutant exposure dosage up to a saturation level, and were weakly (r(2) = 0.3905) inversely correlated to the proportion of live cells at that time point in culture. Temporal increase in EB fluorescence of growing cells is indicative for toxic but reversible effects. Cells displaying high continued EB fluorescence levels experience constant and permanent damage, and no longer contribute to population growth. The procedure developed here using bacterial ethidium bromide efflux pump activity may be a useful complement to screen sublethal toxicity effects of chemicals.

A flow cytometry based oligotrophic pollutant exposure test to detect bacterial growth inhibition and cell injury.

Toxicity of chemical pollutants in aquatic environments is often addressed by assays that inquire reproductive inhibition of test microorganisms, such as algae or bacteria. Those tests, however, assess growth of populations as a whole via macroscopic methods such as culture turbidity or colony-forming units. Here we use flow cytometry to interrogate the fate of individual cells in low-density populations of the bacterium Pseudomonas fluorescens SV3 exposed or not under oligotrophic conditions to a number of common pollutants, some of which derive from oil contamination. Cells were stained at regular time intervals during the exposure assay with fluorescent dyes that detect membrane injury (i.e., live-dead assay). Reduction of population growth rates was observed upon toxicant insult and depended on the type of toxicant. Modeling and cell staining indicate that population growth rate decrease is a combined effect of an increased number of injured cells that may or may not multiply, and live cells dividing at normal growth rates. The oligotrophic assay concept presented here could be a useful complement for existing biomarker assays in compliance with new regulations on chemical effect studies or, more specifically, for judging recovery after exposure to fluctuating toxicant conditions.

Development of a multistrain bacterial bioreporter platform for the monitoring of hydrocarbon contaminants in marine environments.

Petroleum hydrocarbons are common contaminants in marine and freshwater aquatic habitats, often occurring as a result of oil spillage. Rapid and reliable on-site tools for measuring the bioavailable hydrocarbon fractions, i.e., those that are most likely to cause toxic effects or are available for biodegradation, would assist in assessing potential ecological damage and following the progress of cleanup operations. Here we examined the suitability of a set of different rapid bioassays (2-3 h) using bacteria expressing the LuxAB luciferase to measure the presence of short-chain linear alkanes, monoaromatic and polyaromatic compounds, biphenyls, and DNA-damaging agents in seawater after a laboratory-scale oil spill. Five independent spills of 20 mL of NSO-1 crude oil with 2 L of seawater (North Sea or Mediterranean Sea) were carried out in 5 L glass flasks for periods of up to 10 days. Bioassays readily detected ephemeral concentrations of short-chain alkanes and BTEX (i.e., benzene, toluene, ethylbenzene, and xylenes) in the seawater within minutes to hours after the spill, increasing to a maximum of up to 80 muM within 6-24 h, after which they decreased to low or undetectable levels. The strong decrease in short-chain alkanes and BTEX may have been due to their volatilization or biodegradation, which was supported by changes in the microbial community composition. Two- and three-ring PAHs appeared in the seawater phase after 24 h with a concentration up to 1 muM naphthalene equivalents and remained above 0.5 muM for the duration of the experiment. DNA-damage-sensitive bioreporters did not produce any signal with the oil-spilled aqueous-phase samples, whereas bioassays for (hydroxy)biphenyls showed occasional responses. Chemical analysis for alkanes and PAHs in contaminated seawater samples supported the bioassay data, but did not show the typical ephemeral peaks observed with the bioassays. We conclude that bacterium-based bioassays can be a suitable alternative for rapid on-site quantitative measurement of hydrocarbons in seawater.

Intracellular excision and reintegration dynamics of the ICEclc genomic island of Pseudomonas knackmussii sp. strain B13.

Genomic islands are DNA elements acquired by horizontal gene transfer that are common to a large number of bacterial genomes, which can contribute specific adaptive functions, e.g. virulence, metabolic capacities or antibiotic resistances. Some genomic islands are still self-transferable and display an intricate life-style, reminiscent of both bacteriophages and conjugative plasmids. Here we studied the dynamical process of genomic island excision and intracellular reintegration using the integrative and conjugative element ICEclc from Pseudomonas knackmussii B13 as model. By using self-transfer of ICEclc from strain B13 to Pseudomonas putida and Cupriavidus necator as recipients, we show that ICEclc can target a number of different tRNA(Gly) genes in a bacterial genome, but only those which carry the GCC anticodon. Two conditional traps were designed for ICEclc based on the attR sequence, and we could show that ICEclc will insert with different frequencies in such traps producing brightly fluorescent cells. Starting from clonal primary transconjugants we demonstrate that ICEclc is excising and reintegrating at detectable frequencies, even in the absence of recipient. Recombination site analysis provided evidence to explain the characteristics of a larger number of genomic island insertions observed in a variety of strains, including Bordetella petri, Pseudomonas aeruginosa and Burkholderia.

Toxicity of triclosan, penconazole and metalaxyl on Caulobacter crescentus and a freshwater microbial community as assessed by flow cytometry.

Biocides are widely used for domestic hygiene, agricultural and industrial applications. Their widespread use has resulted in their introduction into the environment and raised concerns about potential deleterious effects on aquatic ecosystems. In this study, the toxicity of the biocides triclosan, penconazole and metalaxyl were evaluated with the freshwater bacterium Caulobacter crescentus and with a freshwater microbial community using a combination of single- and double-stain flow cytometric assays. Growth of C. crescentus and the freshwater community were repressed by triclosan but not by penconazole or metalaxyl at concentrations up to 250 µM. The repressive effect of triclosan was dependent on culture conditions. Caulobacter crescentus was more sensitive to triclosan when grown with high glucose at high cell density than when grown directly in sterilized lake water at low cell density. This suggests that the use of conventional growth conditions may overestimate biocide toxicity. Additional experiments showed that the freshwater community was more sensitive to triclosan than C. crescentus, with 10 nM of triclosan being sufficient to repress growth and change the phylogenetic composition of the community. These results demonstrate that isolate-based assays may underestimate biocide toxicity and highlight the importance of assessing toxicity directly on natural microbial communities. Because 10 nM of triclosan is within the range of concentrations observed in freshwater systems, these results also raise concerns about the risk of introducing triclosan into the environment.

Use of flow cytometric methods for single-cell analysis in environmental microbiology.

Flow cytometry (FCM) is emerging as an important tool in environmental microbiology. Although flow cytometry applications have to date largely been restricted to certain specialized fields of microbiology, such as the bacterial cell cycle and marine phytoplankton communities, technical advances in instrumentation and methodology are leading to its increased popularity and extending its range of applications. Here we will focus on a number of recent flow cytometry developments important for addressing questions in environmental microbiology. These include (i) the study of microbial physiology under environmentally relevant conditions, (ii) new methods to identify active microbial populations and to isolate previously uncultured microorganisms, and (iii) the development of high-throughput autofluorescence bioreporter assays.

Methylotrophic extremophilic yeast Trichosporon sp.: a soil-derived isolate with potential applications in environmental biotechnology.

A yeast isolate revealing unique enzymatic activities and substrate-dependent polymorphism was obtained from autochthonous microflora of soil heavily polluted with oily slurries. By means of standard yeast identification procedures the strain was identified as Trichosporon cutaneum. Further molecular PCR product analyses of ribosomal DNA confirmed the identity of the isolate with the genus Trichosporon. As it grew on methanol as a sole carbon source, the strain appeared to be methylotrophic. Furthermore, it was also able to utilize formaldehyde. A multi-substrate growth potential was shown with several other carbon sources: glucose, glycerol, ethanol as well as petroleum derivatives and phenol. Optimum growth temperature was determined at 25 degrees C, and strong inhibition of growth at 37 degrees C together with the original soil habitat indicated lack of pathogenicity in warm-blooded animals and humans. The unusually high tolerance to xenobiotics such as diesel oil (>30 g/l), methanol (50 g/l), phenol (2 g/l) and formaldehyde (7.5 g/l) proved that the isolate was an extremophilic organism. With high-density cultures, formaldehyde was totally removed at initial concentrations up to 7.5 g/l within 24 h, which is the highest biodegradation capability ever reported. Partial biodegradation of methanol (13 g/l) and diesel fuel (20 g/l) was also observed. Enzymatic studies revealed atypical methylotrophic pathway reactions, lacking alcohol oxidase, as compared with the conventional methylotroph Hansenula polymorpha. However, the activities of glutathione-dependent formaldehyde dehydrogenase, formaldehyde reductase, formate dehydrogenase and unspecific aldehyde dehydrogenase(s) were present. An additional glutathione-dependent aldehyde dehydrogenase activity was also detected. Metabolic and biochemical characteristics of the isolated yeast open up new possibilities for environmental biotechnology. Some potential applications in soil bioremediation and wastewater decontamination are discussed.