Method and device for the selection of processing parameters for nanomaterial compositions

Abstract

A method for the selection of processing parameters for nanomaterial compositions is performed by preparing a baseline model wherein a nanomaterial composition interacts with a disease model. It is then acquired a baseline interaction between the nanomaterial composition and the disease model. A perturbation is applied iteratively by changing at each iteration at least one parameter of the perturbation and acquiring at each iteration a perturbation parameter describing a protein corona formation on said first nanomaterial composition under the effect of the perturbation. Finally, it is identified and selected among the perturbation parameters at least one processing parameter minimizing the formation of protein corona.

Description

[0001] DESCRIPTION “Method and device for the selection of processing parameters for nanomaterial compositions” ★★★★★★★

[0002] Technical field

[0003] The present invention pertains to the technical field of the methods and devices for the selection of processing parameters for the production of nanomaterials compositions.

[0004] Advantageously, said nanomaterial compositions may then be implemented in the medical field to help / promote the delivery of pharmaceutical products to an individual.

[0005] Thus the present invention helps in particular in the identification of parameters that are optimal to improve the pharmacological efficiency of a nanomaterial composition. art

[0006] Nanomaterials offer excellent and promising properties for developing new diagnostic and therapeutic approaches. Among nanomaterials, metalbased nanoparticles (NPs) such as silver, gold, copper, iron, zinc, platinum, etc. have attracted much attention in the medical field. Noble metal nanoparticles (Ag, Au, Pt) have been used for various biomedical applications such as anticancer, radiotherapy enhancement, drug delivery, thermal ablation, antibacterial, diagnostic assays, antifungal, gene delivery and many others. In fact, metal nanoparticles have some unique chemical and physical properties that make them valuable and can be functionalized with a variety of functional groups such as peptides, antibodies, RNA and DNA to recognize different types of cells, including cancer cells. The surface of nanomaterials is usually coated with bio- recognizable molecules, which allow for improved biocompatibility and selectivity for molecular targets. However, in many cases, these aspects depend on the surface adsorption of biomolecules present in the biological environment, particularly plasma proteins, which results in the formation of the so-called "protein corona" (indicated for conciseness in the following as PC).

[0007] This is of relevance because PC can change the very identity of nanomaterials, altering their interaction with cells, with important repercussions on pharmacokinetics and therapeutic efficacy, limiting, for example, the penetration of transported drugs. Previous studies have shown that mechanical forces are involved in PC formation.

[0008] Therefore, it exists a strong felt need for the development of new systems and methods capable of rapidly simulating different conditions / aspects, in order to better understand the PC formation process so as to allow the production of novel and more efficient pharmaceutical products.

[0009] Examples of experiment carried out in this technical field can be found in the following scientific articles: D’Hollander Antoine et al. “Limiting the protein corona: a successful strategy for in vivo active targeting of anti- HER2 nanobody-functionalized nanostars”, Findlay Matthew et al. “Machine learning provides predictive analysis into silver nanoparticles protein corona formation from physicochemical properties”, Anuoluwa Bamidele Emmanuel et al. “Discovery and prediction capabilities in metalbased nanomaterials: an overview of the application of machine learning techniques and some recent advances”, Sagar Dhoble et al. “Decoding nanomaterial-biosystem interactions through machine learning”.

[0010] In this context, the technical purpose which forms the basis of the present application is to provide a method and a device which overcome the above-mentioned drawbacks of the available prior art.

[0011] In particular, when nanoparticles come into contact with a biological liquid, a layer of proteins called protein corona, PC, (hard-PC and soft-PC) can form and can change the identity of the nanoparticles themselves, including their distribution and therapeutic efficacy (ability to recognize and interact with their targets). The dynamic nature of the phenomenon can be best described by using the “hard” and “soft” corona terms. The 'hard corona' (hard-PC) is the inner layer of the PC that contains the proteins strongly adsorbed onto the surface of the nanomaterial. The 'soft corona' (soft-PC) is the outer layer of PC that contains the proteins that associate with the hard-PC through weak protein-protein interactions.

[0012] However, the process of protein corona formation could be altered, and the present invention concerns the selection of optimized processing parameters for the identification of condition that influence the PC formation so as to allow the production of nanomaterial compositions that can be efficiently used in the pharmaceutical field.

[0013] The indicated technical purpose and the specified aims are substantially achieved by method and a device comprising the technical features described in one or more of the appended claims.

[0014] More in detail, the invention relates to a method for the selection of processing parameters for nanomaterial compositions.

[0015] The method is performed by preparing a baseline model.

[0016] In the baseline model the nanomaterial composition interacts with a disease model.

[0017] It is acquired a baseline interaction between the nanomaterial composition and the disease model.

[0018] The baseline interaction is defined by at least one baseline parameter describing a protein corona formation on said nanomaterial composition.

[0019] It is then applied a perturbation to the baseline model.

[0020] The perturbation comprises at least a condition of modified gravity.

[0021] Said condition of modified gravity is selected among at least one of the following: microgravity, simulated microgravity, mechanical unloading, altered gravity, and hypogravity.

[0022] It is then acquired a perturbated interaction between the nanomaterial composition and the disease model.

[0023] The perturbated interaction is defined by at least one perturbation parameter describing a protein corona formation on said first nanomaterial composition under the effect of the perturbation.

[0024] The above steps are repeated iteratively varying at least on parameter of the perturbation.

[0025] The iteration is carried on until it is identified and selected among the parameters of the perturbation at least one processing parameter minimizing the formation of protein corona.

[0026] Further features and advantages of this invention are more apparent in the detailed description below, with reference to preferred, non-limiting, embodiments of a method and device of which the accompanying figures show experimental data and specifically:

[0027] - figure 1 shows a block diagram listing the main steps of the claimed method.

[0028] - figure 2 shows the total hydrodynamic diameter of gold nanoparticles, GNP (A), and gold nano stars, GNS (B), after incubation with human plasma wherein perturbation of controlled gravity (GC) and simulated microgravity (SMG) are applied;

[0029] - figure 3 shows the zeta potential of the same sample of figure 1 ;

[0030] - figure 4 shows PC quantification for the incubation of the same amount of GNP (A) or GNS (B) with human plasma at 37°C for different times in GC or SMG;

[0031] - figure 5 shows PC pattern measured via electrophoresis of protein corona formed around GNP (A-B) and GNS (C-D) after the incubation with human plasma at 37°C for different times (2, 4, 8 and 24 h, respectively) in GC or SMG. A set of reference bands (M) associated with molecular weights are displayed on the left of each data set. B, D) Lane intensity profiles as a function of molecular weight. Asterisks indicate where there are differences in the electrophoretic bands. Detailed description of preferred embodiments

[0032] The present description discloses preferred embodiments of a method and corresponding system for the selection of processing parameters for nanomaterial compositions.

[0033] In particular, the present method allows to individuate and then select the parameters that can be implemented so as to minimize, ideally eliminate, the formation of a protein corona, PC, on the nanomaterial.

[0034] Said PC can alter the behavior of the nanomaterial inside a biological system and as such it is desirable to reduce its formation.

[0035] The method is carried out by preparing a baseline model wherein a nanomaterial composition interacts with a disease model.

[0036] In other words, it is prepared a disease model, which can be for example a cellular culture used as in vitro model of a specific disease, and then a predefined nanomaterial composition is made to interact with said disease model, e.g. a specific nanomaterial composition in injected / inserted into the disease model, and then the baseline model is defined based on how / if the nanomaterial composition interacts with the disease model and how said interaction is affected by PC formation.

[0037] For example, the interaction, and thus the baseline model, may be defined in terms of quantity of nanomaterial actively interacting with the disease model and / or causing an alteration in the disease model.

[0038] In particular, the nanomaterial composition comprises gold and can further comprises other metal-based nanomaterial that can be preferably selected among: silver, gold, copper, iron, zinc, platinum.

[0039] More in detail, the method comprises a step of acquiring a baseline interaction between the nanomaterial composition and the disease model, by identifying, measuring, calculating at least one baseline parameter describing a protein corona formation on said nanomaterial composition.

[0040] Thus, the above-described steps provide a starting point for the present method by identifying the behavior and efficacy of a certain nanomaterial composition in interacting with a specific disease by measuring the entity of formation of the PC which is in turn responsible for an unwanted alteration in the capability of the nanomaterial composition to interact with the disease model.

[0041] In order to proceed with the identification of the optimal parameters that could avoid (or at least reduce to a minimum) the PC formation process, it is applied a perturbation to the baseline model.

[0042] Specifically, the perturbation is a condition of mechanical stimulation of the baseline model and can comprise at least one condition selected in the following list: microgravity, simulated microgravity, mechanical unloading, altered gravity, and hypogravity. The above conditions can be identified as follows:

[0043] - Microgravity: condition of an experiment concluded in space, subjected to actual weightlessness, e.g. inside a bioreactor experimental cube in satellite or space station orbit, or drop tower or other free fall mechanism.

[0044] - Simulated microgravity: gravitational vector averaged out by the rotation of a device equipped with 2 or more axis, such as clinostat, random positioning machine, 3D bioreactor, rotating wall vessels, other equipment capable of rotating a sample circularly or within a sphere like path.

[0045] - Mechanical unloading: analogous condition as simulated microgravity, or in case of an organism, limbs or other body parts are fully or partially lifted for reduced biomechnical stress.

[0046] -Altered gravity: a condition where the sample experience an acceleration between 0g-0.99g.

[0047] - Hypogravity: a condition where the sample experience an acceleration above 1 .01 g, e.g. with the use of centrifuges.

[0048] In other words, the perturbation is performed by varying the subjective gravity to which the model is subject.

[0049] It is then acquired a perturbated interaction between the nanomaterial composition and the disease model, wherein the perturbated interaction is defined by at least one perturbation parameter describing a protein corona formation on the nanomaterial composition under the effect of the perturbation.

[0050] In this scenario, the perturbation parameters allow to detect and understand how the PC formation changes when the disease model is made to interact with the nanomaterial composition in a perturbated state, wherein the perturbation is a condition of modified gravity.

[0051] The process is iteratively repeated modifying each time at least one parameter of the perturbation (by changing the kind of perturbation applied, or my modifying one or more numerical value defining the perturbation).

[0052] The perturbation may be applied by implementing a variety of devices that will be discussed more in detail in the following.

[0053] The above process leads to the acquisition of a plurality of different perturbated interaction each associated with a specific response and alteration of the baseline model.

[0054] It is then possible to identify and select among the perturbation parameters at least one processing parameter minimizing the formation of protein corona.

[0055] In other words, the method is carried out by iteratively subjecting the disease model to different condition (i.e. different perturbations) and then selecting the perturbation that minimizes the PC formation.

[0056] In particular, the iteration is carried one until a parameter or combination of parameter of the perturbation is found that causes the complete elimination of the PC or until the amount of PC generated is below a predetermined threshold or until a predetermined number of iterations have been carried out.

[0057] Further to the above, the described process allows to evaluate in a quick and efficient way different nanomaterial composition to assess their behavior and determine which specific composition may be processed in such a way that reduces and ideally eliminate the PC formation process. Advantageously, further to allowing the production of optimized nanomaterial composition, it is also possible to collect a considerable amount of information that may be used in further processes and to deepen the understanding of the interaction between specific nanomaterial compositions and diseases.

[0058] In particular, this is made possible by preparing a database, preferably a distributed database and memorizing on the database a dataset comprising at least:

[0059] - first data identifying the baseline model;

[0060] - second data identifying the nanomaterial composition;

[0061] - third data identifying the disease model;

[0062] - fourth data identifying the perturbation;

[0063] - fifth data identifying the at least one perturbation parameter;

[0064] - sixth data identifying the processing parameters.

[0065] All the collected data may then be made available over a predefined network to be consulted and / or used for further implementations.

[0066] In particular, the database may be used in an Artificial Intelligence environments as a training tool for a machine learning model.

[0067] In this context, a further step and implementation of the present method may provide for the training of an Artificial Intelligence so as to allow the creation of digital model of diseases and their interactions with nanomaterial compositions.

[0068] More in detail, the database can be used for the training of a machine learning model that comprises at least one input node, at least one processing node and at least one output node.

[0069] Said structure, and more in general the machine learning model, implements a reinforcement learning algorithm.

[0070] The reinforcement learning algorithm comprises a reward function configured to provide a positive feedback based on the perturbation parameter. In other words, whenever in response to a specific perturbation the protein corona achieves specific properties (such as specific binding affinities or structural arrangements) a positive feedback is provided to the model.

[0071] More in detail, the input node includes a memory location for storing the dataset.

[0072] The processing node is connected to the input node and is configured to calculate, based on the dataset, a probability parameter identifying a probability of the perturbation to minimize the formation of protein corona.

[0073] The output node is connected to the processing node and includes memory location for storing the probability parameter.

[0074] In this context, the iterative variation of the perturbation parameters can be performed at least in part based on the content of the output node.

[0075] In other words, the claimed method can benefit from the implementation of an Artificial Intelligence algorithm that is trained with the data collected on various disease models and the corresponding perturbations to guide and optimize the execution of further run of the method itself by identifying parameters that are statistically more likely to influence the formation of PC, so that said parameters can be investigated first or in general be given more weight in the identification of the desired optimal processing parameters for the production of the nanomaterial composition.

[0076] In general, the perturbated gravity condition is a highly scalable, highly configurable, and multifunctional approach.

[0077] For example, using high-throughput systems, of which examples will be provided below, hundreds of conditions can be run thus allowing the optimization of nanomaterials conjugation with different substances to control the formation of PCs exhibiting maximum drug activity.

[0078] The above also proves to be advantageous in the identification of Space experiments with a high probability of therapeutic potential.

[0079] In fact, it is possible to explore and study a wide variety of different experimental setting identifying those that may lead to more ideal or complete results if performed in a condition of absence of gravity (e.g. the above mentioned Space environment) so that a Space mission may be prepared a priori to focus only on the most promising experiments.

[0080] The present invention further pertains an apparatus which is specifically adapted to carry out the claimed method.

[0081] In particular, said data processing apparatus is configured to execute the steps described above.

[0082] From a structural point of view, the apparatus comprises: a baseline module, a gravity alteration module, an assessment module, a computational module and a selection module.

[0083] Each of the above listed modules may be a part of a single electronic device provided with appropriate computational power (e.g. a computer) or be defined by a different devices that are connected (in a cabled or wireless manner) within themselves.

[0084] In particular, the baseline module is configured to evaluate a baseline interaction between the nanomaterial composition and the disease model. Specifically, the baseline module comprises at least one baseline sensor configured to acquire baseline parameters and at least one processor connected to the sensor to receive the baseline parameter and generate the baseline interaction.

[0085] Said baseline sensor may be or comprise for example an optical sensor, an optoelectronic sensor or any analogous known sensor configured to acquire information that can define / identify the status of the disease model and its interaction with the nanomaterial, composition.

[0086] The gravity alteration module is instead configured to apply a perturbation to the baseline model.

[0087] For example, the gravity alteration module may be a rotary cell culture system or a rotating wall vessel bioreactor that through rotating walls creates a centrifugal force that keeps samples (the disease model and / or the nanomaterial composition) away from the walls.

[0088] In this context the rotary cell culture system is particularly adapted to operate with disease models that envision the use of 2D cell culture, while the rotating wall vessel bioreactor is more performant to operate with 3D cell cultures.

[0089] Alternatively, the gravity alteration module may also be implemented through a 2D or a 3D clinostat, which are devices that with one or two frames respectively that rotates at constant speed are able to average out gravitational vectors.

[0090] Furthermore, the gravity alteration module may be implemented with a random positioning machine that uses two independent frames rotating with randomized path averaging the gravitational vector and works optimally with disease models comprising cells that do not have a preferred growth direction.

[0091] Alternatively, a 3D bioreactor optimized for microgravity simulation may be implemented, which using two independent frames rotating with predefined and predetermined paths works also optimally when applying the method to assess cells that do not present a preferential growth direction.

[0092] Alternatively, a diamagnetic levitation device may be used (or analogous magnetic levitation devices), which levitate the baseline model between magnets able to equalize the gravitational vector.

[0093] The assessment module is instead configured to evaluate a perturbated interaction between the nanomaterial composition and the disease model. More in detail, the assessment module comprises at least one assessment sensor configured to acquire perturbation parameters and at least one assessment processor connected to the assessment sensor to receive the baseline parameter and generate the perturbated interaction.

[0094] The assessment sensor may be the same sensor as the baseline sensor or a different sensor.

[0095] In other words, the apparatus may comprise just one sensor (or one set of sensors) that operates as both the baseline and the assessment sensor.

[0096] In general, the assessment sensor may be or comprise any of the kind of sensors that are apt to form up the baseline sensor. The computational module is coupled to the gravity alteration module and is configured to iteratively modify the applied perturbation.

[0097] In other words, the computational module may be a control unit (e.g. a processor with computational power) connected to the gravity alteration module and able to set / modify its operating parameter so as to modify the characteristics of the perturbation to which the baseline model is exposed. Finally, the selection module comprises at least one selection processor configured to identify and select among the perturbation parameters at least one processing parameter minimizing the formation of protein corona.

[0098] Specifically, the selection processor receives the data produces by the apparatus or provided to the apparatus as an input and process them so as to identify the perturbation parameter (or parameters) that produces a reduction, ideally the elimination, of the PC.

[0099] Preferably, the apparatus further comprises a database, for example a distributed database like a computing cloud, configured to memorize a dataset, specifically a training dataset, comprising all the data generated by each module of the apparatus.

[0100] Furthermore the computational module and / or the selection module may further be configured to memorize and execute a machine learning model, which comprises at least one input node, at least one processing node and at least one output node.

[0101] As discussed above, the input node includes a memory location for storing at least the dataset, the at least one processing node is connected to the input node and configured to calculate, based on the dataset, a probability parameter identifying a probability of the perturbation to minimize the formation of protein corona and the output node (which is connected to the processing node) includes memory location for storing said probability parameter. said computation and / or selection modules are in this context configured to modify the applied perturbation based at least on the content of the output node.

[0102] Furthermore, in this context, the machine learning model defines a recurrent neural network configured to identify temporal dependencies between the application of a perturbation and the corresponding perturbation parameter.

[0103] In other words, the machine learning model is able to identify temporal patterns by processing sequences of actions (the perturbations) and their outcomes (the ensuing perturbation parameters).

[0104] In this way, the apparatus can iteratively refine the working parameters of the gravity alteration module to align the end result with temporal patterns observed in previously successful corona formations experiments.

Claims

CLAIMS1. Method for the selection of processing parameters for nanomaterial compositions comprising the steps of: a) preparing a baseline model wherein a nanomaterial composition comprising gold nanoparticles interacts with a disease model; b) acquiring a baseline interaction between the nanomaterial composition and the disease model, said baseline interaction being defined by at least one baseline parameter describing a protein corona formation on said nanomaterial composition; c) applying a perturbation to the baseline model, said perturbation comprising at least one condition selected among: microgravity, simulated microgravity, mechanical unloading, altered gravity, and hypogravity; d) acquiring a perturbated interaction between the nanomaterial composition and the disease model, said perturbated interaction being defined by at least one perturbation parameter describing a protein corona formation on said first nanomaterial composition under the effect of the perturbation; e) iteratively repeating steps c) and d) varying at least one parameter of the perturbation; f) identifying and selecting among the parameters of the perturbation at least one processing parameter minimizing the formation of protein corona.

2. Method according to claim 1 , wherein the nanomaterial composition also comprises metal-based nanoparticles other than gold nanoparticles.

3. Method according to claim 2, wherein said nanoparticles comprise at least one selected among: silver, copper, iron, zinc, platinum.

4. Method according to any of the preceding claims, comprising the stepsof:- preparing a database, preferably a distributed database;- memorizing on said database a dataset comprising at least:- first data identifying the baseline model;- second data identifying the nanomaterial composition;- third data identifying the disease model;- fourth data identifying the perturbation;- fifth data identifying the at least one perturbation parameter;- sixth data identifying the processing parameters.

5. Method according to claim 4, comprising the step of training a machine learning model with the dataset contained in said database, said machine learning model comprising:- at least one input node including a memory location for storing at least the dataset;- at least one processing node connected to the input node and configured to calculate, based on the dataset, a probability parameter identifying a probability of the perturbation to minimize the formation of protein corona;- at least one output node connected to the processing node and including memory location for storing said probability parameter; and step e) is performed based at least in part on the content of the output node.

6. Method according to claim 5, wherein the machine learning model implements a reinforcement learning algorithm comprising a reward function configured to provide a positive feedback based on said perturbation parameter.

7. A data processing apparatus configured for carrying out a method according to any of the preceding claims, said data processing apparatus comprising:- a baseline module configured to evaluate a baseline interaction between the nanomaterial composition and the disease model, said baseline module comprising at least one baseline sensor configured to acquire baseline parameters and at least one processor connected to the sensor to receive the baseline parameter and generate the baseline interaction;- a gravity alteration module configured to apply a perturbation to the baseline model;- an assessment module configured to evaluate a perturbated interaction between the nanomaterial composition and the disease model, said assessment module comprising at least one assessment sensor configured to acquire perturbation parameters and at least one assessment processor connected to the assessment sensor to receive the baseline parameter and generate the perturbated interaction;- a computational module coupled to the gravity alteration module configured to iteratively modify the applied perturbation;- a selection module comprising at least one selection processor configured to identify and select among the perturbation parameters at least one processing parameter minimizing the formation of protein corona8. Apparatus according to claim 7 comprising a database, preferably a distributed database, configured to memorize a dataset comprising all the data generated by each module of the apparatus.

9. Apparatus according to claim 8, wherein the computational module memorizes and executes a machine learning model, said machine learning model comprising:- at least one input node including a memory location for storing at least the dataset;- at least one processing node connected to the input node and configured to calculate, based on the dataset, a probability parameter identifying a probability of the perturbation to minimize the formation of protein corona;- at least one output node connected to the processing node and including memory location for storing said probability parameter; said computation module being configured to modify the applied perturbation based at least on the content of the output node.

10. Apparatus according to claim 9, wherein said machine learning model defines a recurrent neural network configured to identify temporal dependencies between the application of a perturbation and the corresponding perturbation parameter.

Images