Quantifying the Efficacy of Magnetic Nanoparticles for MRI and Hyperthermia Applications via Machine Learning Methods

Magnetic nanoparticles are a prospective class of materials for use in biomedicine as agents for magnetic resonance imagining (MRI) and hyperthermia treatment. However, synthesis of nanoparticles with high efficacy is resource-intensive experimental work. In turn, the use of machine learning (ML) methods is becoming useful in materials design and serves as a great approach to designing nanomagnets for biomedicine. In this work, for the first time, an ML-based approach is developed for the prediction of main parameters of material efficacy, i.e., specific absorption rate (SAR) for hyperthermia and r1 /r2 relaxivities in MRI, with parameters of nanoparticles as well as experimental conditions as descriptors. For that, a unique database with more than 980 magnetic nanoparticles collected from scientific articles is assembled. Using this data, several tree-based ensemble models are trained to predict SAR, r1 and r2 relaxivity. After hyperparameter optimization, models reach performances of R2 = 0.86, R2 = 0.78, and R2 = 0.75, respectively. Testing the models on samples unseen during the training shows no performance drops. Finally, DiMag, an open access resource created to guide synthesis of novel nanosized magnets for MRI and hyperthermia treatment with machine learning and boost development of new biomedical agents, is developed.

Quantitative Prediction of Inorganic Nanomaterial Cellular Toxicity via Machine Learning

Organic chemistry has seen colossal progress due to machine learning (ML). However, the translation of artificial intelligence (AI) into materials science is challenging, where biological behavior prediction becomes even more complicated. Nanotoxicity is a critical parameter that describes their interaction with the living organisms screened in every bio-related research. To prevent excessive experiments, such properties have to be pre-evaluated. Several existing ML models partially fulfill the gap by predicting whether a nanomaterial is toxic or not. Yet, this binary categorization neglects the concentration dependencies crucial for experimental scientists. Here, an ML-based approach is proposed to the quantitative prediction of inorganic nanomaterial cytotoxicity achieving the precision expressed by 10-fold cross-validation (CV) Q²  = 0.86 with the root mean squared error (RMSE) of 12.2% obtained by the correlation-based feature selection and grid search-based model hyperparameters optimization. To provide further model flexibility, quantitative atom property-based nanomaterial descriptors are introduced allowing the model to extrapolate on unseen samples. Feature importance is calculated to find an interpretable model with optimal decision-making. These findings allow experimental scientists to perform primary in silico candidate screening and minimize the number of excessive, labor-intensive experiments enabling the rapid development of nanomaterials for medicinal purposes.

Inverse Material Search and Synthesis Verification by Hand Drawings via Transfer Learning and Contour Detection

Nano- and micromaterials of various morphologies and compositions have extensive use in many different areas. However, the search for procedures giving custom nanomaterials with the desired structure, shape, and size remains a challenge and is often implemented by manual article screening. Here, for the first time, scanning and transmission electron microscopy inverse image search and hand drawing-based search via transfer learning are developed, namely, VGG16 convolution neural network repurposing for image features extraction and image similarity determination. Moreover, the case use of this platform is demonstrated on the calcium carbonate system, where the data are acquired by random high throughput experimental synthesis, and on Au nanoparticles data extracted from the articles. This approach can be used for advanced nanomaterials search, synthesis procedure verification, and can be further combined with machine learning solutions to provide data-driven nanomaterials discovery.

Artificial intelligence to bring nanomedicine to life

The technology of drug delivery systems (DDSs) has demonstrated an outstanding performance and effectiveness in production of pharmaceuticals, as it is proved by many FDA-approved nanomedicines that have an enhanced selectivity, manageable drug release kinetics and synergistic therapeutic actions. Nonetheless, to date, the rational design and high-throughput development of nanomaterial-based DDSs for specific purposes is far from a routine practice and is still in its infancy, mainly due to the limitations in scientists' capabilities to effectively acquire, analyze, manage, and comprehend complex and ever-growing sets of experimental data, which is vital to develop DDSs with a set of desired functionalities. At the same time, this task is feasible for the data-driven approaches, high throughput experimentation techniques, process automatization, artificial intelligence (AI) technology, and machine learning (ML) approaches, which is referred to as The Fourth Paradigm of scientific research. Therefore, an integration of these approaches with nanomedicine and nanotechnology can potentially accelerate the rational design and high-throughput development of highly efficient nanoformulated drugs and smart materials with pre-defined functionalities. In this Review, we survey the important results and milestones achieved to date in the application of data science, high throughput, as well as automatization approaches, combined with AI and ML to design and optimize DDSs and related nanomaterials. This manuscript mission is not only to reflect the state-of-art in data-driven nanomedicine, but also show how recent findings in the related fields can transform the nanomedicine's image. We discuss how all these results can be used to boost nanomedicine translation to the clinic, as well as highlight the future directions for the development, data-driven, high throughput experimentation-, and AI-assisted design, as well as the production of nanoformulated drugs and smart materials with pre-defined properties and behavior. This Review will be of high interest to the chemists involved in materials science, nanotechnology, and DDSs development for biomedical applications, although the general nature of the presented approaches enables knowledge translation to many other fields of science.

DiZyme: Open-Access Expandable Resource for Quantitative Prediction of Nanozyme Catalytic Activity

Enzymes suffer from high cost, complex purification, and low stability. Development of low-cost artificial enzymes of comparative or higher effectiveness is desired. Given its complexity, it is desired to presume their activities prior to experiments. While computational approaches demonstrate success in modeling nanozyme activities, they require assumptions about the system to be made. Machine learning (ML) is an alternative approach towards data-driven material property prediction achieving high performance even on multicomponent complex systems. Despite the growing demand for customized nanozymes, there is no open access nanozyme database. Here, a user-friendly expandable database of >300 existing inorganic nanozymes is developed by data collection from >100 articles. Data analysis is performed to reveal the features responsible for catalytic activities of nanozymes, and new descriptors are proposed for its ML-assisted prediction. A random forest regression (RFR) model for evaluation of nanozyme peroxidase activity is developed and optimized by correlation-based feature selection and hyperparameter tuning, achieving performance up to R²  = 0.796 for K_cat and R²  = 0.627 for K_m . Experiment-confirmed unknown nanozyme activity prediction is also demonstrated. Moreover, the DiZyme expandable, open-access resource containing the database, predictive algorithm, and visualization tool is developed to boost novel nanozyme discovery worldwide (https://dizyme.net).

Magnetically Controlled Carbonate Nanocomposite with Ciprofloxacin for Biofilm Eradication

Biofilms are the reason for a vast majority of chronic inflammation cases and most acute inflammation. The treatment of biofilms still is a complicated task due to the low efficiency of drug delivery and high resistivity of the involved bacteria to harmful factors. Here we describe a magnetically controlled nanocomposite with a stimuli-responsive release profile based on calcium carbonate and magnetite with an encapsulated antibiotic (ciprofloxacin) that can be used to solve this problem. The material magnetic properties allowed targeted delivery, accumulation, and penetration of the composite in the biofilm, as well as the rapid triggered release of the entrapped antibiotic. Under the influence of an RF magnetic field with a frequency of 210 kHz, the composite underwent a phase transition from vaterite into calcite and promoted the release of ciprofloxacin. The effectiveness of the composite was tested against formed biofilms of E. coli and S. aureus and showed a 71% reduction in E. coli biofilm biomass and an 85% reduction in S. aureus biofilms. The efficiency of the composite with entrapped ciprofloxacin was higher than for the free antibiotic in the same concentration, up to 72%. The developed composite is a promising material for the treatment of biofilm-associated inflammations.

MO934COVID-19 IN RENAL TRANSPLANT RECIPIENTS (RTR)

Background and Aims

SARS-CoV-2 infection has a severe course in immunocompromised (RTR) patients. The aim is to study the clinical course and risk factors for adverse outcomes and results of COVID-19 treatment in RTR.

Method

At the beginning of the study there were 2580 RTR observed at Moscow Nephrology Center, by the end of it there were 2776 RTR. A retrospective uncontrolled observational study included 279 RTR (M: 172/F: 107, aged 49.9±10.9 yrs.), infected with SARS-CoV-2 from April 1 to November 30, 2020. The period after kidney transplantation before the onset of the disease was 54,0 months (14.0;108.0). After confirmation of COVID-19 by PCR and chest СТ MMF/Аza were canceled, CNI dose was minimized (target blood level was CyA 30-50 ng/ml, Тac 1,5-3 ng/ml), a CS dose was increased to 10-15 ng/day. Observation endpoints: discharge/recovery or death.

Results

The number of RTR infected with SARS-CoV-2 from April 1 to May 31, 2020 was 108; there were 42 RTR from June 1 to August 31, 2020; and 129 RTR - from September 1 to November 30, 2020. 59 RTR (21,1%) had a mild course of COVID-19. Patients with moderate and severe course (220/78.9%) were treated in the hospital. The period from the onset of the disease to the hospitalization was 7.1 ± 5.1 days. Severe lung damage (> 50%) occurred in 97 of 220 (44.1%); decrease in SpO2 <95% was seen in 128 of 220 (58.2%); 31 patients died. Thus, hospital mortality was 14.1%, overall mortality was 11.1%. Scr during the course increased from 160.9 ± 68.2 μmol/l to 185.4 ± 130.9 μmol/l (p <0.01) with no signs of acute rejection; and after the recovery, it decreased to 158.1 ± 63.2 μmol/l (p <0.01). Risk factors associated with fatal outcome were analyzed among the survivors (group 1; n-189) and the deceased (group 2; n-31). Groups 1 and 2 differed in the frequency of severe lung damage (69/36.9% vs 24/77.4%, p <0.001); the Charlson comorbidity index (4.4 ± 1.7 vs 6.1 ± 2.5, p <0.001); the frequency of IMV use (0 vs 23, p <0.0001), Scr upon admission (160.3 ± 67.1 µmol/l vs 208.9 ± 99.4 µmol/l, p <0.03), Hb levels (116.3 ± 21.8 g/l vs 91.7 ± 23.9 g/l, p <0.001), white blood cell сount (11.1 ± 4.8 × 109/L vs 18.1 ± 7.5 × 109/L, p <0.001), lymphocyte count (0.7 ± 0.4 × 109/l vs 0.4 ± 0.4 × 109/L, p <0.02), albumin (32.4 ± 4.1 g/l vs 25.8 ± 2.8 g/l, p <0.001), glucose (6.1 ± 1.9 mmol/l vs 7.8 ± 2.8 mmol/l, p <0.001), LDG (305.6 ± 135.6 U/l vs 800.8 ± 313.8 U/l, p <0.0001), CRP (74.1 ± 68.4 mg/L vs 160.7 ± 74.4 mg/L, p <0.0001), D-dimer (967.3 ± 949.0 μg/L vs 2810.1 ± 1807.7 μg/L, p <0.0001) and the frequency of procalcitonin increase (29.5 vs 86.4%, p <0.001). The independent factors of adverse outcome (Cox model) were high levels of comorbidity index (p <0.006) and procalcitonin (p <0.006), as well as the IMV use (p <0.0001). It was not possible to establish differences in Groups 1 and 2 depending on the use of individual drugs (Corticosteroids, Baricitinib, Monoclonal Ab IL-6/IL-17/IL-1β, antiCOVID plasma) as well as their combinations.

Conclusion

The frequency of SARS-CoV-2 infection in RTR was more than 2 times lower in summer compared to spring and autumn, which suggests a seasonal nature of this infection. The course of the disease was characterized by high hospital and general mortality. High values of the comorbidity index, procalcitonin and the IMV use were independent predictors of the fatal outcome.

One-pot synthesis of template-free hollow anisotropic CaCO3 structures: towards inorganic shape-mimicking drug delivery systems

A major obstacle in the introduction of nanoformulated drugs has been the fact that the shape of the drug delivery systems (DDSs) - the most important parameter driven by the nature of viruses and bacteria - remains almost out-of-scope in artificial systems. Here we propose a potential solution for this problem by developing a template-free approach for the formulation of hollow bacteria-like CaCO₃-based pH-sensitive DDSs with controllable anisotropy and click-release behavior.

Magnetically controlled protein nanocontainers as a drug depot for the hemostatic agent

PURPOSE

Currently, there is a number of successfully implemented local hemostatic agents for external bleedings in forms of wound dressings and other topical materials. However, little has been done in the field of intravenous hemostatic agents. Here, we propose a new procedure to fabricate biocompatible protein nanocontainers (NCs) for intravenous injection allowing magneto-controllable delivery and short-term release of the hemostatic agent ε-aminocaproic acid (EACA).

METHODS

The nanocontainers were synthesized by the desolvation method from bovine serum albumin (BSA) using methanol without any further crosslinking. Polyethylene glycol (PEG) was used both as a stabilization agent and for size control. Characterization of nanocontainers was performed by the transmission and scanning electron microscopy, dynamic light scattering, X-ray diffraction, and FTIR spectroscopy. Cytotoxicity was estimated using MTT assay. The dopant release from nanocontainers was measured spectrophotometrically using rhodamine B as a model molecule. The specific hemostatic activity was assessed by analyzing clot lysis and formation curve (CloFAL). Moreover, the ability for magneto targeting was estimated using the original flow setup made of a syringe pump and silicon contours.

RESULTS

Fabricated nanocontainers had an average size of 186±24 nm and were constructed from building blocks-nanoparticles with average size ranged from 10 to 20 nm. PEG shell was also observed around nanocontainers with thickness 5-10 nm. NCs were proved to be completely non-cytotoxic even at concentrations up to 8 mg BSA/mL. Uptake capacity was near 36% while release within the first day was 17%. The analysis of the CloFAL curve showed the ability of NCs to inhibit the clot lysis successfully, and the ability of magneto targeting was confirmed under flow conditions.

CONCLUSION

The ability of synthesized NCs to deliver and release the therapeutic drug, as well as to accumulate at the desired site under the action of the magnetic field was proved experimentally.

Acetohydroxyacid synthase: a proposed structure for regulatory subunits supported by evidence from mutagenesis

Valine inhibition of acetohydroxyacid synthase (AHAS) plays an important role in regulation of biosynthesis of branched-chain amino acids in bacteria. Bacterial AHASs are composed of separate catalytic and regulatory subunits; while the catalytic subunits appear to be homologous with several other thiamin diphosphate-dependent enzymes, there has been no model for the structure of the small, regulatory subunits (SSUs). AHAS III is one of three isozymes in Escherichia coli. Its large subunit (encoded by ilvI) by itself has 3-5 % activity of the holoenzyme and is not sensitive to inhibition by valine. The SSU (encoded by ilvH) associates with the large subunit and is required for full catalytic activity and valine sensitivity. The isolated SSU binds valine. The properties of several mutant SSUs shed light on the relation between their structure and regulatory function. Three mutant SSUs were obtained from spontaneous Val(R) bacterial mutants and three more were designed on the basis of an alignment of SSU sequences from valine-sensitive and resistant isozymes, or consideration of the molecular model developed here. Mutant SSUs N11A, G14D, N29H and A36V, when reconstituted with wild-type large subunit, lead to a holoenzyme with drastically reduced valine sensitivity, but with a specific activity similar to that of the wild-type. The isolated G14D and N29H subunits do not bind valine. Mutant Q59L leads to a valine-sensitive holoenzyme and isolated Q59L binds valine. T34I has an intermediate valine sensitivity. The effects of mutations on the affinity of the large subunits for SSUs also vary. D. Fischer's hybrid fold prediction method suggested a fold similarity between the N terminus of the ilvH product and the C-terminal regulatory domain of 3-phosphoglycerate dehydrogenase. On the basis of this prediction, together with the properties of the mutants, a model for the structure of the AHAS SSUs and the location of the valine-binding sites can be proposed.