Compositions and methods for therapeutic treatment

Abstract

The invention provides AI systems that output proposed combinations of agents such as active pharmaceutical ingredients and other agents useful for a therapeutic objectives. Systems of the invention propose cocktails of those agents and optionally simulate or model effects of those agents to evaluate the potential effects of those agents in accomplishing the therapeutic objective. Specific embodiments provide a skin care composition from the AI system. The composition includes (i) a mixture of agents that includes epidermal growth factor (EGF), fibroblast growth factor (FGF), and insulin-like growth factor (IGF-1) and one or more of: a collagen-stimulating peptide, a booster of nicotinamide phospho-ribosyl-transferase (NAMPT) expression, coenzyme Q10, a nucleic acid repair enzyme, and a ceramide; (ii) one or more liposomes encapsulating at least a portion of the mixture of agents; and (iii) a moisturizer carrying the liposomes and the mixture of agents.

Description

TITLE

[0001] COMPOSITIONS AND METHODS FOR THERAPEUTIC TREATMENTBACKGROUND

[0002] Conventional treatments for skin, e.g., to remedy damage or to promote skin health, have unfortunate drawbacks. Many such treatments are only temporary, ineffective, or come with side effects. The loss of collagen and elastic proteins present in the dermal layers causes a breakdown of resiliency and skin thickness over time, which are important to healthy skin. Some conventional surgical interventions for skin treatment include plastic surgery, laser surgery, skin peels, and injection of microbial toxins. However, those approaches may result in detrimental complications, are often painful, and must be repeated with time. Non-invasive remedies include topical formulations consisting of moisturizers, retinoic acids, vitamins, and other agents. However, none of those methods provide all the health benefits that many people desire, and are often very complicated and expensive. Some topical formulations may act as irritants to the skin, only weakly elicit wound healing responses, and fail to successfully replenish the thinning skin with adequate agents for treatment and / or prevention of defects.SUMMARY

[0003] The invention provides Al-based systems that automatically research provided therapeutic objectives and output proposed combinations of agents such as active pharmaceutical ingredients (APIs), growth factors, signaling mediators, co-factors, or other bio-active agents, shown in the literature or in medical data to be potentially useful for the therapeutic objectives. Systems of the invention propose cocktails of those agents and optionally simulate or model effects of those agents in vivo in an in silico predictive model (e.g., using artificial intelligence (Al) or machine learning (ML) modules) to evaluate the potential effects of those agents in accomplishing the therapeutic objective. The systems are suitable for therapeutic objectives where intracellular signaling, cellular development or activity, paracrine signaling, cellular or extracellular environment, and / or age, environment, or metabolic stressors may be important. For example, systems of the invention may be usedto proposes therapeutic mixtures for the promotion of healthy outcomes with regard to skin health, osteoarthritis, inflammation, or other biological phenomenon.

[0004] Preferred embodiments use an Al-based system that includes components such as a research agent, a cocktail agent, and a validation agent that — respectively — research literature and medical data to identify relevant biomechanisms that underpin the therapeutic objective; identify cocktails of ingredients that address or exploit those mechanisms; and perform the Al-based in silico modeling or simulation to score each candidate cocktail for its effectiveness for the therapeutic objective. A system user may then formulate one or a plurality of those cocktails and test them (e.g., in an animal or clinical model). In fact, in certain embodiments, systems of the invention are operably linked to microfluidic platforms that formulate the cocktails, including by the preparation of biphasic or multiphasic delivery vehicles.

[0005] Noting that preferred embodiments involve intracellular signaling, cellular development or activity, paracrine signaling, cellular or extracellular environment, and other conditions where cells and their environments, and signaling among the cells, and a composition of intracellular components are involved, compositions of the invention may be best delivered by biphasic or multiphasic delivery systems such as those that use liposomes, oleosomes, micelles, exosomes, lipid encapsulation, artificial cells, hydrogels, or other delivery systems that can deliver agents to both intracellular and extracellular environments and recapitulate intercellular signaling. For those purposes, systems herein may include a carrier material such as a gel, hydrogel, moisturizer or suspension and active agents of the cocktail / composition may be encapsulated within, e.g., liposomes carried by the carrier material. Other active agents (e.g., signaling factors or matrix-remodeling factors) may be suspended in the carrier material. Some agents may be carried in part within the liposomes and in part within the carrier material. Delivery systems may be biphasic in that a lipophilic phase may be stably combined with an aqueous phase, such as liposomes encapsulating polar proteins carried in a lipophilic or oily gel or moisturizer carrier. In some embodiments, systems are multiphasic. For example, certain embodiments use proteins that are integrated into oleosomes as ingredients within a multi-ingredient composition, wherein the composition is, in-turn, encapsulated in liposomes. Such a recursive, multiphasic packing may be one of the outputs arrived at by the validation agents using the in silico modeling.

[0006] Embodiments herein have been used to design a skin care composition and a composition for the treatment of osteoarthritis. Thus the invention provides Al-based systemsand methods for the discovery and design of therapeutic compositions, therapeutic compositions originating from such systems, and methods of treating conditions using such compositions.

[0007] In certain aspects, the invention provides a method for producing a composition. The method includes: querying a research agent of a machine learning system with a therapeutic objective. The research agent finds and ranks document from a library by relevance to the objective, summarizes documents for which the relevance is ranked above a threshold, and passes summaries of the summarized documents to a cocktail agent of the machine learning system. The method includes selecting, by the cocktail agent, a plurality of ingredients and simulating, using a validation agent of the machine learning system, effects of the ingredients and concentrations of the ingredients. From those, the method includes outputting an identified set of ingredients predicted to accomplish the therapeutic objective. In some embodiments, at least one output of the validation agent are passed back to the cocktail agent, whereby the cocktail agent creates a second plurality of ingredients based on the at least one output of the validation agent. The validation agent and the cocktail agent operate in a feedback loop in which cocktail agent selects different ingredients until the validation agent outputs an acceptable effectiveness score. Methods may include outputting the identified set of ingredients with an effectiveness score that includes a measure of an estimated likelihood of accomplishing the therapeutic objective. The effectiveness score may include a prediction of adverse effects.

[0008] The machine learning system may access medical knowledge input comprising biomedical data, clinical data, and research literature. In some embodiments, the research agent queries the library with the therapeutic objective to get a list of results, filters results by availability, retrieves available results, and summarizes the retrieved, available results. The research agent and / or the cocktail agent may score the plurality of ingredients for one or more of: efficacy, cartilage regeneration, and inflammation reduction. The cocktail agent may include a large language model (LLM) module and / or the validation agent may include a graph neural network (GNN) module. The validation agent may predict release kinetics of active agents in liposome compositions, optimize membrane stability and / or targeting ligands for specific tissues; and / or simulate absorption, biodistribution, and bioavailability of the plurality of ingredients.

[0009] In certain hardware embodiments, the machine learning system is coupled to a microfluidics system that is operable to combine elements of the plurality of ingredients;form a biphasic delivery system; and encapsulate at least a portion of the plurality of ingredients in the biphasic delivery system.

[0010] In certain skin care embodiments, the therapeutic objective comprises skin health. In skin health embodiments, the plurality of ingredients may include one or a combination of epidermal growth factor (EGF), fibroblast growth factor (FGF), insulin-like growth factor (IGF-1) a collagen-stimulating peptide, a booster of nicotinamide adenine dinucleotide (NAD+), coenzyme Q10, a nucleic acid repair enzyme, and a ceramide

[0011] In other osteoarthritis embodiments, the therapeutic objective comprises osteoarthritis. The plurality of ingredients may include growth factors, cytokines, non-coding RNAs, extracellular vesicles, exosomes, proteins, enzymes, ECM components, hormones, and antioxidants, wherein the growth factors include one or more of TGF- P, IGF-1, BMP-7, and IL-Ra.

[0012] Aspects of the disclosure provide a skin care composition. The composition includes (i) a mixture of agents that includes epidermal growth factor (EGF), fibroblast growth factor (FGF), and insulin-like growth factor (IGF-1) and one or more of: a collagen-stimulating peptide, a booster of nicotinamide adenine dinucleotide (NAD+), coenzyme Q10, a nucleic acid repair enzyme, and a ceramide; (ii) one or more liposomes encapsulating at least a portion of the mixture of agents; and (iii) a carrier carrying the liposomes and the mixture of agents. The FGF may be FGF-2. The EGF and / or the FGF-2 are fused with an oleosome within the mixture. The mixture preferably includes all of the EGF; the FGF; the IGF-1; at least two distinct collagen-stimulating peptides; the booster of NAD+; the coenzyme Q10 formulated with vitamin E; photolyase as the nucleic acid repair enzyme, and the ceramide.

[0013] The carrier may be moisturizer that includes one or more of a humectant, an occlusive, a barrier-repair agent, and an emollient. In some embodiments, the carrier is a serum and a moisturizer that includes glycerin, propanediol, squalane, pentylene glycol, polyglyceryl-2 stearate, glyceryl stearate, stearyl alcohol, hyaluronic acid, hydrogenated lecithin, phytosterols, caprylyl glycol, polyacrylate crosspolymer- 11, and 1,2-hexanediol. In certain embodiments, the carrier is a moisturizer comprising glycerin, propanediol, squalane, pentylene glycol, polyglyceryl-2 stearate, glyceryl stearate, stearyl alcohol, and hyaluronic acid, and one or more of hydrogenated lecithin, phytosterols, caprylyl glycol, polyacrylate crosspolymer-11, and 1,2-hexanediol. The carrier may consist essentially of glycerin, propanediol, squalane, pentylene glycol, polyglyceryl-2 stearate, glyceryl stearate, stearylalcohol, hyaluronic acid, hydrogenated lecithin, phytosterols, caprylyl glycol, polyacrylate crosspolymer-11, and 1,2-hexanediol.

[0014] In various embodiments: the collagen-stimulating peptide comprises a mixture of at least palmitoyl tripeptide-5 and palmitoyl pentapeptide-4; the Coenzyme Q10 may be provided in a formulation with vitamin E; the nucleic acid repair enzyme may be a photolyase (e.g., optionally within its own lipid nanoparticle delivery system); and / or the booster of NAD+ may be an extract from sunflower sprouts that increases nicotinamide phospho-ribosyl transferase (NAMPT) expression.

[0015] In certain embodiments, the mixture consists essentially of: the EGF within an EGF- oleosome fusions; the FGF as FGF-2 with FGF-oleosome fusions; the IGF- produced in plant cells; palmitoyl tripeptide-5 and palmitoyl pentapeptide-4 as collagen-stimulating peptides; sunflower sprout extract as the booster of NAMPT expression; the coenzyme Q10 formulated with vitamin E; photolyase as the nucleic acid repair enzyme, and the ceramide. The composition may specifically consist essentially of (w / w): 0.5-1 % the EGF-oleosome fusions; 0.5-1 % the FGF-oleosome fusions; 0.5-1.5 % the IGF-1; 0.1- 5 % palmitoyl tripeptide-5; 0.1- 5 % palmitoyl pentapeptide-4 as collagen-stimulating peptides; 1-3 % a compound comprising the coenzyme Q10 formulated with the vitamin E; 0.5-2 % sunflower sprout extract as the booster of NAMPT expression; 0.1- 5 % photolyase as the nucleic acid repair enzyme; 0.1-5 % the ceramide; the liposomes; and balance % the carrier.

[0016] In some embodiments, the coenzyme Q10 and the ceramide are distributed among the moisturizer. The one or more liposomes may encapsulate growth factors and water-soluble components of the mixture. In some embodiments, the EGF and FGF are present in oleosome fusions. The nucleic acid repair enzyme may be an algal photolyase introduced into the composition in photolyase liposomes. The booster of NAD+ may be a sunflower sprout extract that increases NAMPT expression. The IGF-1 may be a product derived from transgenic plants. In certain embodiments, the EGF and / or the FGF are obtained from transgenic plants of the family Camelina (flax), by a process that includes suspecting seeds from the transgenic plants in suspension, centrifuging and washing the suspension, and collecting a supernatant layer comprising oleosomes fused to the EGF and / or the FGF.

[0017] The composition may be provided in a container or package for topical, e.g., such as a foil squeeze-tube with screw cap.

[0018] Related aspects provide a method of treating skin to promote skin health. The method includes applying, to skin of a subject, a composition includes (i) a mixture of agents thatincludes epidermal growth factor (EGF), fibroblast growth factor (FGF), and insulin-like growth factor (IGF-1) and one or more of: a collagen-stimulating peptide, a booster of NAD+, coenzyme Q10, a nucleic acid repair enzyme, and a ceramide; (ii) one or more liposomes encapsulating at least a portion of the mixture of agents; and (iii) a moisturizer carrying the liposomes and the mixture of agents.

[0019] In related aspects, the invention provides methods and compositions that use microparticle-encapsulated therapeutics that collectively function like engineered synthetic cells. These compositions are designed to replicate and harness the biological models of paracrine signaling provided by stem cells as well as the proteins and biological factors that are secreted by such cells into the extracellular space, or the "secretome". Compositions of the invention include a cocktail of therapeutic agents such as paracrine signaling mediators that recapitulates a therapeutic secretome and associated biological aspects of cell-to-cell interaction. Compositions and methods of the invention provide a fabricated synthetic cell niche that is optimized to promote tissue regeneration and targeted therapeutic applications.

[0020] Systems and methods of the invention use machine learning models to identify, model, and mimic the mechanisms of action of living stem cells. A suitable machine learning system such as a large language model (LLM) may be trained on data such as scientific literature, clinical trial data, and proprietary stem cell information. By profiling and analyzing the data, the machine learning system proposes or identifies compositions and dosages for targeting specific pathologies. The machine learning system may select, for example, specific therapeutic agents such as proteins, cytokines, or other bioactive molecules based on their known or predicted effects and concentration levels in natural secretomes and the associations between those therapeutic agents and beneficial clinical outcomes.

[0021] Once the machine learning system has analyzed the data, the system generates an optimal composition for treating a condition, which composition may include or be described as a recombinant protein cocktail (RPC). A protein cocktail or other composition of the invention may generally include one or more proteins as well as other therapeutic agents such as small molecules, steroid hormones, biological macromolecules, chemicals, metabolites, or other agents. For example, a composition for the treatment of osteoarthiritis may include a combination of several or all of transforming growth factor beta (TGF- P ), insulin-like growth factor- 1 (IGF-1 ), bone morphogenetic protein-7 (BMP-7), and interleukin- 1 receptor antagonist (IL-IRa ).

[0022] The selected therapeutic agents in such a composition are preferably encapsulated within microparticles, such as microparticles made of a suitable polymer such as poly(lactic- co-glycolic) acid (PLGA). The microparticles may be delivered in a suitable delivery system such a hydrogel that may include hyaluronic acid (HA). The composition is delivered to targeted tissue sites, and the delivery system preferably promotes a controlled and sustained release. Hyaluronic acid is one preferred ingredient enhance the delivery and retention of the composition at the targeted tissue site.

[0023] Certain preferred embodiments promote the retention and controlled release of the composition at the treatment site through the inclusion of agent or mechanism that provides for in situ gelation. In such embodiments, a microparticle-encapsulated therapeutic cocktail may be carried in a hydrogel designed to exhibit a transition from a liquid to a gel once it is administered at the target site within the body. By undergoing gelation in situ, the hydrogel retains the microparticles at the treatment site (inhibiting those particles from entering circulation or being eliminated), providing for a controlled release of the cocktail of therapeutic agents, e.g., paracrine signaling mediators, at the treatment site.

[0024] In certain aspects, the invention provides a therapeutic composition. The composition includes a mixture of a plurality of therapeutic agents (preferably at least some paracrine signaling agents), microparticles encapsulating the mixture, and a hydrogel suspending the microparticles. The hydrogel exhibits gelation in situ once delivered into a patient. The paracrine signaling agents may include one or more growth factors such as TGF-P, IGF-1, BMP-7, and / or IL-lRa. The microparticles may be composed of a biocompatible polymer such as poly(lactic-co-glycolic acid) (PLGA) and the hydrogel may include hyaluronic acid. In some cases, a crosslinking agent may be added to the hydrogel to facilitate gelling. In preferred embodiments, the plurality of therapeutic agents is selected by, and is the output of, a machine learning system such as a large language model (LLM). The machine learning system may be trained to, and operated to, determine therapeutically effective formulations and dosages for each of the components of the therapeutic composition.

[0025] In other aspects, the invention provides a method for producing a therapeutic The method includes operating a machine learning system to select therapeutic agents such as paracrine signaling mediators and amounts thereof. Using the cocktail of agents selected by the machine learning system, the method includes creating a primary emulsion comprising the selected agents emulsified in a polymer, emulsifying the primary emulsion into a solvent to create a double emulsion, evaporating the solvent away to provide microparticles thatencapsulate the agents, and mixing the microparticle-encapsulated agents into a hydrogel. The hydrogel preferably include hyaluronic acid and a reagent or mechanism (such as a solgel composition, ferrocene, or a cross-linking agent) that will provide for in situ gelation once the therapeutic composition is used. The machine learning system is preferably a large language model that has been trained on training data such as clinical trial results, medical literature, and in-house / in-silico cellular models and data. The machine learning system may be prompted with inputs such as medical history, allergies, current medications, or demographic information. Based on the results of the training and based on input inquiries, the machine learning system may be prompted to generate an output that describes a therapeutic composition for a particular pathology or use-case, including for a specific patient with his or her own medical history, demographic information, or other information. The therapeutic composition may comprise a cocktail of therapeutic agents (preferably including one or more paracrine signaling agents), a microparticle, and a hydrogel. The therapeutic agents may include growth factors such as TGF-P, IGF-1, BMP-7, and IL-lRa. The microparticles may be composed of PLGA. The method may include packaging the hydrogel- suspended microparticles, carrying the therapeutic cocktail, in a suitable container (e.g., vials or tubes) and storing the composition in a suitable manner, e.g., freezing (e.g., in a -80 degree C freezer) for subsequent use. In a preferred embodiment, the hydrogel may gel (e.g., transition from a liquid to a gel) in situ once administered into a treatment site in a body of a patient.

[0026] In related aspects, the invention provides a method for treating a patient. The method includes administering a therapeutic composition targeting a disease. The therapeutic composition is a mixture of paracrine signaling agents encapsulated by microparticles and suspended in a hydrogel. The hydrogel is capable of transitioning from a liquid to a gel in situ in the patient. The patient may have osteoarthritis. The hydrogel may be transitioned from a liquid to a gel through various mean such as CO2 induction, a mucoadhesive polymer, a ferrocene and hyaluronic acid combination, ultrasound, or adding crosslinking agents to the hydrogel. A machine learning system, such as a large language model, may determine a therapeutically effective dose for the patient. The patient may suffer from osteoarthritis.BRIEF DESCRIPTION OF DRAWINGS

[0027] FIG. 1 diagrams a method for producing a composition.

[0028] FIG. 2 diagrams a workflow of an Al system of the invention.

[0029] FIG. 3 is a screen capture of an input into the Al system.

[0030] FIG. 4 shows output of a cocktail agent of the Al system.

[0031] FIG. 5 is output from a validation agent of the Al system.

[0032] FIG. 6 shows a skin care composition according to certain embodiments.DETAILED DESCRIPTION

[0033] FIG. 1 diagrams a method 101 for producing a composition. The method 101 includes querying a research agent of a machine learning system with a therapeutic objective. The research agent finds and ranks document from a library by relevance to the objective, summarizes documents for which the relevance is ranked above a threshold, and passes summaries of the summarized documents to a cocktail agent of the machine learning system. The method 101 includes selecting 121, by the cocktail agent, a plurality of ingredients; simulating 131, using a validation agent of the machine learning system, effects of the ingredients and concentrations of the ingredients. Finally, the method 101 includes outputting 141 an identified set of ingredients predicted to accomplish the therapeutic objective.

[0034] FIG. 2 diagrams a workflow of the system. The research agent 203 gathers information from research papers, online sources, and extract data from these relevant articles. The cocktail agent 207 (e.g., an LLM Assistant) trained on given data sources predicts a "cocktail" (presumably of therapeutic compounds or treatments). The A CNN / validation agent 213 (e.g., LLM-based) that attempts to "work out in the dish" (likely simulating or predicting biological interactions). The workflow includes a feedback loop that continues until the solution works effectively in the simulated dish environment. The system then produces a final output.

[0035] In the disclosed workflow, the research agent 203 queries available medical knowledge input. The medical knowledge input may include biomedical data, clinical data, and research literature (e.g., PubMed). For example, the input may be a therapeutic objective that the research agent 203 passes to a library such as PubMed. The research agent queries the library with the therapeutic objective to get a list of results, filters results by availability, retrieves available results, and summarizes the retrieved, available results, and passes the summaries to the cocktail agent 207. The cocktail agent 207 may output a list of ingredients, e.g., a treatment composition and may further include dosage estimation and / or side effect analysis. Those outputs are preferable passed to the validation agent 213.

[0036] The validation agent 213 may be implemented using a graphical neural network or a LangGraph. The validation agent 213 may execute operations for treatment reasoning, e.g., domain analysis, interaction reasoning, and validity checking. In preferred embodiments, outputs of the validation agent 213 include one or more of an extracellular matrix impact (e.g., regeneration) score, an inflammation level score, a tissue balance status, or a treatment success probability estimate.

[0037] FIG. 3 is a screen capture of an input into an initial prototype example of the system, specifically showing inputs (text on right) to the research agent 203.

[0038] FIG. 4 shows a portion of an initial output of a cocktail agent 207. The final cocktail for this query is discussed more fully below. The cocktail agent preferably includes a large language model (LLM) module and / or the validation agent includes a graph neural network (GNN) module.

[0039] FIG. 5 is an exemplary output from a validation agent 213 of an early prototype of the system. Compared to the cocktail agent, the validation agent 213 preferably provides ready - to-make formulae, with specific amounts of ingredients and predictive scores for efficacy and side-effects. Such a production-ready output may be arrived at by an iterative feedback loop with the cocktail agent 207. In feedback loop embodiments, at least one output of the validation agent are passed back to the cocktail agent, whereby the cocktail agent creates a second plurality of ingredients based on the at least one output of the validation agent. As shown, the validation agent and the cocktail agent operate in a feedback loop in which cocktail agent selects different ingredients until the validation agent outputs an acceptable effectiveness score. The validation agent 213 may give the ingredients with an effectiveness score, wherein the effectiveness score includes a measure of an estimated likelihood of accomplishing the therapeutic objective. The effectiveness score may address potential adverse effects or unwanted side effects. The validation agent 213 may further predict release kinetics of active agents in liposome compositions, optimize membrane stability and / or targeting ligands for specific tissues; and / or simulates= absorption, biodistribution, and bioavailability of the plurality of ingredients.

[0040] In some embodiments, systems of the invention are coupled to a microfluidics system that is operable to combine elements of the plurality of ingredients; form a biphasic delivery system; and encapsulate at least a portion of the plurality of ingredients in the biphasic delivery system.

[0041] The disclosed system was used to design skin care compositions. In such embodiments, the therapeutic objective used as input included skin health.

[0042] FIG. 6 shows a skin care composition 601 according to certain embodiments. The composition 601 includes a mixture of agents that includes epidermal growth factor (EGF) 605, fibroblast growth factor (FGF) 609, and insulin-like growth factor (IGF-1) 613 and one or more of: a collagen-stimulating peptide 613, a booster of nicotinamide phospho-ribosyl- transf erase (NAMPT) expression 623, coenzyme Q10 627, a nucleic acid repair enzyme 627, and a ceramide 635, one or more liposomes 617 encapsulating at least a portion of the mixture of agents; and a carrier 655 carrying the liposomes 617 and the mixture of agents. The FGF 609 is preferably FGF-2. The EGF 605 and / or the FGF-2 may be fused with an oleosome 675 within the mixture. Specifically, in the composition 601, the mixture preferably includes the EGF 605; the FGF 609; the IGF-1 613; at least two distinct collagen- stimulating peptides (e.g., preferably palmitoyl tripeptide-5 and palmitoyl pentapeptide-4 ); the booster of NAMPT expression 623 is preferably provided as an extract from sunflower sprouts; the coenzyme Q10 627 is preferably formulated with vitamin E; the nucleic acid repair enzyme 627 is preferably photolyase (e.g., from algae, optionally in its own biphasic delivery vehicle independently of or encapsulated within the liposomes 617), and the ceramide 635 may be suspended in the carrier 655 (or encapsulated within the liposomes 617).

[0043] The carrier 655 may include one or more of a humectant, an occlusive, a barrier-repair agent, and an emollient. The carrier may be a moisturizer that includes glycerin, propanediol, squalane, pentylene glycol, polyglyceryl-2 stearate, glyceryl stearate, stearyl alcohol, and hyaluronic acid, and one or more of hydrogenated lecithin, phytosterols, caprylyl glycol, polyacrylate crosspolymer- 11, and 1,2-hexanediol.

[0044] In certain embodiments, the mixture of agents consists essentially of: the EGF 605 within an EGF-oleosome 675 fusion; the FGF 609 as FGF-2 with FGF-oleosome 675 fusions; the IGF-1 613 produced from plant cells; palmitoyl tripeptide-5 and palmitoyl pentapeptide-4 as collagen-stimulating peptides 613; sunflower sprout extract as the booster of NAMPT expression 623; the coenzyme Q10 627 formulated with vitamin E; photolyase as the nucleic acid repair enzyme 627, and the ceramide 635. The specific amounts of the mixture, in w / w may be:

[0045] 0.5-1 % the EGF-oleosome fusions;

[0046] 0.5-1 % the FGF-oleosome fusions;

[0047] 0.5-1.5 % the IGF-l;

[0048] 0.1- 5 % palmitoyl tripeptide-5;

[0049] 0.1- 5 % palmitoyl pentapeptide-4 as collagen-stimulating peptides;

[0050] 1-3 % a compound comprising the coenzyme Q10 formulated with the vitamin E;

[0051] 0.5-2 % sunflower sprout extract as the booster of NAMPT expression;

[0052] 0.1- 5 % photolyase as the nucleic acid repair enzyme;

[0053] 0.1-5 % the ceramide;

[0054] the liposomes; and

[0055] balance % the carrier. The carrier may be a moisturizer that includes a combination of glycerin, propanediol, squalane, pentylene glycol, polyglyceryl-2 stearate, glyceryl stearate, stearyl alcohol, hyaluronic acid, hydrogenated lecithin, phytosterols, caprylyl glycol, polyacrylate crosspolymer- 11, an dl,2-hexanediol. The carrier preferably includes glycerin, propanediol, squalane, pentylene glycol, polyglyceryl-2 stearate, glyceryl stearate, stearyl alcohol, and hyaluronic acid, and one or more of hydrogenated lecithin, phytosterols, caprylyl glycol, polyacrylate crosspolymer- 11, and 1,2-hexanediol. The carrier may consist essentially of glycerin, propanediol, squalane, pentylene glycol, polyglyceryl-2 stearate, glyceryl stearate, stearyl alcohol, hyaluronic acid, hydrogenated lecithin, phytosterols, caprylyl glycol, polyacrylate crosspolymer- 11, and 1,2-hexanediol.

[0056] As noted, the EGF 605 and FGF 609 may be present in oleosome fusions. The EGF and / or the FGF may be obtained from transgenic plants of the family Camelina, by a process that includes suspecting seeds from the transgenic plants in suspension, centrifuging and washing the suspension, and collecting a supernatant layer comprising oleosomes fused to the EGF and / or the FGF. This produces the oleosomes 675, a naturally occurring lipid-based delivery system fused with the growth factor. The nucleic acid repair enzyme 627 is preferably an algal photolyase introduced into the composition in photolyase liposomes. The booster of NAMPT expression 623 may be introduced into the mixture as a sunflower sprout extract. The IGF-1 613 may be derived from transgenic plants.

[0057] The composition may include (e.g., within the mixture of agents) any of a copper peptide GHK-cu peptide, tri -peptide 1, hexapeptide-9, palmitoyl tripeptide-7, matrikine peptides, photolyse, plankton extract, DNA repair enzyme oxoguanine glycosylase (OGGI), nicotinamide mononucleotide (NMN), NAD+, nicotinamide riboside, epigallocatechin gallate (EGCG), CoQlO liposome blend, and a prepared component ingredient that consists essentially of water, glycerin, steareth-20, palmitoyl pentapeptide-4, palmitoyl tetrapeptide-7,acetyl hexapeptide-8, Ginkgo Biloba leaf extract, caprylyl glycol, and ethylhexylglycerin (which prepared component ingredient is sold under the name SPECPLEX GPS) in any combination.

[0058] Embodiments herein may include any of EGF, FGT, IGF-1, a collagen-stimulating peptide, a booster of NAD+, coenzyme Q10, a nucleic acid repair enzyme, a ceramide, a copper peptide GHK-cu peptide, tri-peptide 1, hexapeptide-9, palmitoyl tripeptide-7, matrikine peptides, photolyse, plankton extract, DNA repair enzyme oxoguanine glycosylase (OGGI), nicotinamide mononucleotide (NMN), NAD+, nicotinamide riboside, epigallocatechin gallate (EGCG), CoQlO liposome blend, water, glycerin, steareth-20, palmitoyl pentapeptide-4, palmitoyl tetrapeptide-7, acetyl hexapeptide-8, Ginkgo Biloba leaf extract, caprylyl glycol, ethylhexylglycerin, a prepared component ingredient that consists essentially of water, glycerin, steareth-20, palmitoyl pentapeptide-4, palmitoyl tetrapeptide-7, acetyl hexapeptide-8, Ginkgo Biloba leaf extract, caprylyl glycol, and ethylhexylglycerin (which prepared component ingredient is sold under the name SPECPLEX GPS), the EGF within an EGF-oleosome fusions; the FGF as FGF-2 with FGF-oleosome fusions; the IGF-1 produced in plant cells; palmitoyl tripeptide-5 and palmitoyl pentapeptide-4 as collagen- stimulating peptides; sunflower sprout extract as the booster of NAD+; the coenzyme Q10 formulated with vitamin E; and photolyase optionally within a photosome as the nucleic acid repair enzyme— in any combination such as any of those shown herein.

[0059] The mixture of agents may consist essentially of any plurality of the EGF-oleosome fusions; the FGF-oleosome fusions; the IGF-1; palmitoyl tripeptide-5; palmitoyl pentapeptide-4 as collagen-stimulating peptides; a compound comprising the coenzyme Q10 formulated with the vitamin E (e.g., within an independent CoQlO liposome blend); sunflower sprout extract as the booster of NAD+; photolyase as the nucleic acid repair enzyme; the ceramide; the liposomes; glycerin; steareth-20; palmitoyl pentapeptide-4; palmitoyl tetrapeptide-7; acetyl hexapeptide-8; Ginkgo Biloba leaf extract; caprylyl glycol; and ethylhexylglycerin in any combination.

[0060] In embodiments, the mixture of agents is preferably encapsulated at least in part, or substantially, or entirely in liposomes or a biphasic or multiphasic delivery packages. For example, one or more of the growth factors may be directly carried by oleosomes and / or the the photoloyase may be carried in photosomes and / or the CoQlO maybe carried in a CoQlO liposome blend, those oleosomes and / or photosomes and / or CoQlO liposomes being presentamong the mixture of agents which is encapsulated (in part, substantially, or entirely) in liposomes.

[0061] The composition is provided in a container or package for topical delivery.

[0062] In related aspects, the invention provides a method of treating skin to promote skin health. The method includes treating skin of a subject with a composition 601. Such methods employ engineered combinations of cytokine research + synergistic molecule research to achieve simulated youthful cell signaling and liposomal delivery of active ingredients for skin anti-aging and aging prevention purposes. The composition 601 is an output of the disclosed Al platform which uses a multi-agentic architecture to search through existing research and datasets to find the best combinations of cytokines, growth factors, proteins, and peptides and optimal concentration ranges for topical application in skin aging, injection in osteoarthritis, topical in wound healing, and other regenerative use cases.

[0063] As used in the method, the composition 601 contains plant-derived or human recombinant cytokine / growth factor molecules encapsulated in liposome or plant-derived liposomes: including EGF, FGF, IGF-1, Palmitoyl Tripeptide-5, Palmitoyl pentapeptide-4, NAD precursor or booster, antioxidant such as CoQlO, and encapsulated photolyase. The composition 601 is incorporated into a moisturizer 655 to provide a single-product serum for skin health and longevity. The composition 601 is engineered to address all major hallmarks of skin aging and delay visible aging in skin.

[0064] As used in the method and the composition 601, EGF, FGF, and IGF can create a powerful synergy where EGF promotes skin cell turnover, FGF supports deeper skin structure and collagen synthesis, and IGF enhances overall cellular function and hydration. This multifaceted approach targets various pathways involved in skin aging and repair. EGF stimulates skin cell proliferation, promotes wound healing, and enhances skin hydration and elasticity. EGF is valuable for epidermal regeneration.

[0065] FGF promotes angiogenesis, collagen synthesis, and wound healing. FGF is vital for fibroblast proliferation, which contributes to skin firmness and elasticity. When EGF 605 and FGF 609 are encapsulated in a liposome 617 and / or plant-derived oleosome 675, penetration effects increase and can reach into the epidermis to stimulate the above process and remain stable in formulation.

[0066] EGF, FGF, or other agents herein may be provided in fusion with oleosomes. Oleosomes are natural oil droplets, abundant in plants and more specifically in seeds, composing 20-50 wt% of their mass. The structure of oleosomes is used by seeds to safelystore energy in the form of triacylglycerols and use it during germination. For this, the phospholipid / protein membrane that covers and protects the triacylglycerols give oleosomes very good stability against physical and chemical stresses. The remarkable propertyperformance relationships of oleosomes is used here as a biphasic delivery system for EGF, FGF, or other agents to take advantage of the sophisticated membrane of the oleosome. For detail, see Nikiforidis, 2019, Structure and function of oleosomes, Adv Colloid Interf Sci 274: 102039, incorporated by reference.

[0067] IGF-1 supports cell growth and proliferation, enhances collagen production, and improves skin elasticity. IGF-1 is crucial for overall skin rejuvenation. Embodiment herein use a plant-based synthetic human Insulin-like Growth Factor-1 (plant-IGFl) obtained from Nicotiana benthamiana by transient expression in a non-GMO system which is able to restore the activity of the skin basal layers, as well as to activate the proteasome and decrease carbonylation, for an effective clearance of damaged proteins. The IGF-1 protects, detoxifies and involved in the clearance of UV damaged proteins.

[0068] Palmitoyl Tripeptide-5 mimics a natural signaling peptide that activates TGF-P and starts a signaling cascade to direct fibroblasts to produce more collagen I and II and reduce enzymatic activity that breaks down collagen. Palmitoyl pentapeptide-4 mimics matrikines, the naturally occurring peptides produced when collagen breaks down. These signal the skin to “repair itself’ by producing more structural proteins. It's a bio-mimetic strategy that reactivates skin’s own regenerative functions.

[0069] Palmitoyl Tripeptide-5 and Palmitoyl pentapeptide-4 work together to boost TGF-b (used as a replacement for TGF-b since whole TGF-b can be irritating), enhance collagen synthesis, and complement the cytokines to mimic natural cell signaling found in youthful skin to regenerate and repair damage.

[0070] Coenzyme Q10 is a fat-soluble vitamin-like substance present in every cell of the human body. It serves as a coenzyme vital to production of energy within cells. It is also a very powerful antioxidant or free radical scavenger. Coenzyme Q10 when formulated with vitamin E, when topically applied, can penetrate the cell layers of the skin and may lessen the depth of wrinkles due to photo aging, as well as epithelial cell turnover.

[0071] A Helianthus annuus (Sunflower) Sprout Extract used to activate and upregulate NAMPT expression, which in turn simulates NAD+ production and upregulates NAD+ expression in vitro. Nicotinamide phosphoribosyltransferase (NAMPT) converts nicotinamide to nicotinamide mononucleotide (NMN), a key NAD intermediate.

[0072] The method and the composition 601 use a liposome encapsulated photolyase 627, used as a DNA repair enzyme post sun-exposure to repair and prevent further DNA damage. The repair enzyme 627 may be obtained from commercial vendors and may be obtained packaged in its own lipid nanoparticle or liposome with the implication that the composition 601 potentially includes multiple lipid nanoparticles and / or liposomes 617 (see also oleosome 675), some of those being nested within one another, as shown.

[0073] The ceramide complex is preferably a multi-ceramide blend mimicking the natural skin lipid barrier to provide durable barrier reinforcement, hydration retention, and stratum comeum repair. Ceramides are waxy lipid molecules with sphingosine and a fatty acid joined by an amide bond. Ceramides are found in high concentrations within the cell membrane of eukaryotic cells, since they are component lipids that make up sphingomyelin, one of the major lipids in the lipid bilayer. Ceramides can participate in a variety of cellular signaling: examples include regulating differentiation, proliferation, and programmed cell death (PCD) of cells.

[0074] Any suitable moisturizer 655 may be included. Suitable moisturizers for use with a serum contains moisturizing ingredients like humectants (hyaluronic acid, glycerin), emollients (ceramides, shea butter), and occlusives (petroleum jelly), along with antioxidants, and plant extracts like aloe vera and green tea to soothe and protect the skin.

[0075] The disclosed composition 601 when used together at optimal concentration (e.g., the w / w % as shown) support epidermal regeneration, accelerate renewal, promote cell survival and collagen synthesis, and counteract senescence-associated changes. The composition 601 should repair DNA damage caused by UV rays and ROS associated with urban pollutants. The ingredients should prevent melanin synthesis in dark spots and fade existing ones. It should also increase collagen I and II synthesis and reduce enzymatic collagen degradation. It should help enhance intracellular NAD+levels to improve mitochondrial metabolism, DNA repair, and cellular resilience against stress-induced senescence. The formulation mimics the natural skin lipid barrier to increase hydration, decrease appearance of fine lines and wrinkles, reduce TEWL (trans-epidermal water loss), increase elasticity, and brightness.

[0076] Other embodiments and applications are within the scope of the disclosure.

[0077] In some embodiments of the disclosed method 101, the therapeutic objective comprises osteoarthritis. The method may provide compositions that deliver paracrine signaling agents or other therapeutic agents as a plurality of ingredients for the treatment of osteoarthritis. The plurality of ingredients comprises growth factors, cytokines, non-codingRNAs, extracellular vesicles, exosomes, proteins, enzymes, ECM components, hormones, and antioxidants, wherein the growth factors include one or more of TGF- P, IGF-1, BMP-7, and IL-Ra.

[0078] The present disclosure relates to compositions and methods for making and using said compositions that deliver paracrine signaling agents or other therapeutic agents to patients in need thereof. Embodiments of the present invention include therapeutic compositions and methods of making and using said compositions. The compositions may include paracrine signaling agents encapsulated in microparticles. As discussed herein, the microparticles may be suspended in a hydrogel that forms a gel once inside the patient. A large language model may be used when formulating the therapeutic composition and to determine appropriate dosages for a patient.

[0079] The therapeutic composition may include a mixture comprising a plurality of paracrine signaling agents. Paracrine signaling is a type of cellular communication in which a cell produces a signal to induce changes in nearby cells, altering the behavior of those cells. Signaling molecules known as paracrine signaling agents diffuse over a relatively short distance. Cells that produce paracrine signaling agents secrete them into the immediate extracellular environment. The signaling agents then travel to nearby cells in which the gradient of agents received determines the outcome.

[0080] Most paracrine signaling agents employ a relatively streamlined set of receptors and pathways. In fact, different organs in the body - even between different species - are known to utilize similar sets of paracrine signaling agents in differential development. The conserved receptors and pathways have been organized into four major families based on similar structures: TGF-P superfamily, fibroblast growth factor (FGF) family, Hedgehog family, and Wnt family. Binding of a paracrine signaling agents to its respective receptor initiates signal transduction cascades, eliciting different responses.

[0081] Transforming growth factor (TGF) is a family of proteins that includes 33 members that encode dimeric, secreted polypeptides that regulate development. Many developmental processes are under the control of TGF's including gastrulation, axis symmetry of the body, organ morphogenesis, and tissue homeostasis in adults. This family includes TGF-pi, TGF- P2, TGF-P3, and TGF-P5. They are involved in positively and negatively regulation of cell division, the formation of the extracellular matrix between cells, apoptosis, and embryogenesis. Due to those biological roles, TGF proteins are likely candidates for inclusion in therapeutic compositions of the invention.

[0082] In certain embodiments, TGF-pi is used because that protein stimulates the synthesis of collagen and fibronectin and inhibits the degradation of the extracellular matrix. Ultimately, TGF-pi increases the production of extracellular matrix by epithelial cells. TGF- P proteins regulate epithelia by controlling where and when they branch to form kidney, lung, and salivary gland ducts. TGF-P also plays a crucial role in stem cell differentiation as well as T-cell regulation and differentiation.

[0083] The FGF family of paracrine factors has a broad range of functions, but primarily stimulate proliferation and differentiation. To fulfill many diverse functions, FGFs can be alternatively spliced or even have different initiation codons to create hundreds of different FGF isoforms. One of the most important functions of the FGF receptors (FGFR) is in limb development. This signaling involves nine different alternatively spliced isoforms of the receptor. Fgf8 and FgflO are two of the critical players in limb development. Paracrine signaling through fibroblast growth factors and its respective receptors utilizes the receptor tyrosine pathway. Binding of FGF to FGFR phosphorylates the idle kinase and activates the receptor tyrosine kinase (RTK) pathway. The RTK pathway begins at the cell membrane surface, where a ligand binds to its specific receptor. Ligands that bind to RTKs include fibroblast growth factors, epidermal growth factors, platelet-derived growth factors, and stem cell factor. Due to the role of FGF factors in cell health, proliferation, and differentiation, compositions of the disclosure may preferably include one or more FGF agent.

[0084] Embodiments of the disclosure may include one or more signaling factors of the hedgehog protein family. The hedgehog protein family is involved in induction of cell types and the creation of tissue boundaries and patterning. Sonic hedgehog (SHH) has various roles in vertebrae development, mediating signaling and regulating the organization of central nervous system, limb, and somite polarity. Desert hedgehog (DHH) is expressed in the Sertoli cells involved in spermatogenesis. Indian hedgehog (IHH) is expressed in the gut and cartilage, important in postnatal bone growth.

[0085] The Wnt protein family includes a large number of cysteine-rich glycoproteins. The Wnt proteins activate signal transduction cascades via three different pathways, the canonical Wnt pathway, the noncanonical planar cell polarity (PCP) pathway, and the noncanonical Wnt / Ca2+ pathway. Wnt proteins control a wide range of developmental processes and have been seen as necessary for control of spindle orientation, cell polarity, cadherin mediated adhesion, and early development of embryos in many different organisms.

[0086] Another preferred embodiment uses insulin-like growth factor 1 (IGF-1), also called somatomedin C. IGF-1 is a hormone similar in molecular structure to insulin which plays an important role in childhood growth and has anabolic effects in adults. IGF-1 is a protein that in humans is encoded by the IGF1 gene. IGF-1 is produced primarily by the liver. Production is stimulated by growth hormone (GH). Most of IGF-1 is bound to one of 6 binding proteins (IGF-BP). IGFBP-1 is regulated by insulin. A synthetic analog of IGF-1, mecasermin, is used for the treatment of growth failure in children with severe IGF-1 deficiency. Cyclic glycineproline (cGP) is a metabolite of hormone insulin-like growth factor-1 (IGF-1). cGP has a cyclic structure, lipophilic nature, and is enzymatically stable which makes cGP a more favourable candidate for manipulating the binding-release process between IGF-1 and its binding protein, thereby normalising IGF-1 function. IGF-1 is a key mediator of anabolic activities in numerous tissues and cells, such as growth hormone-stimulated growth, metabolism and protein translation.

[0087] Due to IGF-l's participation in the GH-IGF-1 axis, IGF-1 contributes to the maintenance of muscle strength, muscle mass, and development of the skeleton, brain, eye and lung during fetal development, which makes IGF-1 a good candidate for inclusion in a therapeutic composition of the invention. IGF-1 is a primary mediator of the effects of growth hormone (GH). Growth hormone is made in the anterior pituitary gland, released into the bloodstream, and then stimulates the liver to produce IGF-1. IGF-1 then stimulates systemic body growth and has growth-promoting effects on almost every cell in the body, especially skeletal muscle, cartilage, bone, liver, kidney, nerve, skin, hematopoietic, and lung cells. In addition to the insulin-like effects, IGF-1 can also regulate cellular DNA synthesis.

[0088] IGF-1 binds to at least two cell surface receptor tyrosine kinases: the IGF-1 receptor (IGF1R), and the insulin receptor. Its primary action is mediated by binding to its specific receptor, IGF1R, which is present on the surface of many cell types in many tissues. Binding to the IGF1R initiates intracellular signaling. IGF-1 is one of the most potent natural activators of the Akt signaling pathway, a stimulator of cell growth and proliferation, and a potent inhibitor of programmed cell death. The IGF-1 receptor and insulin receptor are two closely related members of a transmembrane tetrameric tyrosine kinase receptor family. They control vital brain functions, such as survival, growth, energy metabolism, longevity, neuroprotection and neuroregeneration.

[0089] As a major growth factor, IGF-1 is responsible for stimulating growth of all cell types and causing significant metabolic effects. One important metabolic effect of IGF-1 issignaling cells that sufficient nutrients are available for them to undergo hypertrophy and cell division. Its effects also include inhibiting cell apoptosis and increasing the production of cellular proteins. IGF-1 receptors are ubiquitous, which allows for metabolic changes caused by IGF-1 to occur in all cell types. IGF-l's metabolic effects are far-reaching and can coordinate protein, carbohydrate, and fat metabolism in a variety of different cell types. The regulation of IGF-l's metabolic effects on target tissues is also coordinated with other hormones such as growth hormone and insulin.

[0090] Similarly to IGF-1, IGF-2 is mainly produced in the liver and after it is released into circulation, it stimulates growth and cell proliferation. IGF-2 is thought to be a fetal growth factor, as it is essential for a normal embryonic development and is highly expressed in embryonic and neonatal tissues.

[0091] Another preferred paracrine signaling agent is bone morphogenetic protein 7 or BMP7 (also known as osteogenic protein-1 or OP-1). BMP7 is a protein that is encoded by the BMP7 gene. BMP7 plays a key role in the transformation of mesenchymal cells into bone and cartilage. BMP7 is inhibited by noggin and a similar protein, chordin, which are expressed in the Spemann-Mangold Organizer. BMP7 may be involved in bone homeostasis. It is expressed in the brain, kidneys and bladder. In fact, for therapeutic applications such as osteoarthritis and acute tissue injury (e.g., rotator cuff tears or similar joint injuries), due to their role in cartilage development one or more of the bone morphogenic proteins, such as BMP7, may preferably be included in a composition of the invention.

[0092] BMP7 induces the phosphorylation of SMAD1 and SMAD5, which in turn induce transcription of numerous osteogenic genes. It has been demonstrated that BMP7 treatment is sufficient to induce all of the genetic markers of osteoblast differentiation in many cell types. Endogenous BMP-7 is an inhibitor of the TGF-P signaling cascade that induces fibrosis.

[0093] Certain embodiments of the invention also use the interleukin- 1 receptor antagonist (IL-IRA) that is a protein encoded by the IL1RN gene. IL-IRA is an agent that binds non- productively to the cell surface interleukin-1 receptor (IL-1R), the same receptor that binds interleukin 1 family (IL-1), preventing IL-1 's from sending a signal to that cell. IL- IRA is a member of the interleukin 1 cytokine family. IL-IRA is secreted by various types of cells including immune cells, epithelial cells, and adipocytes, and is a natural inhibitor of the pro- inflammatory effect of ILip. IL-IRA inhibits the activities of interleukin 1, alpha (ILIA) and interleukin 1, beta (IL1B), and modulates a variety of interleukin 1 related immune and inflammatory responses. IL- IRA mRNA reduces pain and joint inflammation by blockinginflammatory cascade signals that lead to osteoarthritis progression. Any one or more of interleukin proteins or interleukin receptor antagonists such as IL-IRa may be included in a composition of the invention, particularly for conditions associated with pain or inflammation such as tumor or osteoarthritis treatment.

[0094] In addition to or in place of the proteins detailed above, other embodiments of the invention may comprise therapeutic agents such as cytokines, non-coding RNAs, extracellular vesicles (EVs), exosomes, proteins, enzymes, extracellular matrix (ECM) components, hormones, small molecules, or antioxidants.

[0095] Some embodiments of the invention use recombinant proteins. A recombinant protein is a protein that results from the expression of recombinant DNA. Proteins of this type can be purchased or formed by laboratory methods.

[0096] In some embodiments the therapeutic agents may include one or more antioxidants. Antioxidants are compounds that inhibit oxidation (usually occurring as autoxi dati on), a chemical reaction that can produce free radicals. Antioxidants such as glutathione, mycothiol, or bacillithiol, and enzyme systems like superoxide dismutase, can prevent damage from oxidative stress and may be used in the invention.

[0097] Certain embodiments of the invention include one or more hormones including potentially steroid hormones as one or more of the signaling agents. A hormone is a class of signaling molecules in multicellular organisms that are sent to distant organs or tissues. Numerous kinds of molecules can be classified as hormones such as eicosanoids (e.g. prostaglandins and thromboxanes), steroids (e.g. oestrogen and brassinosteroid), amino acid derivatives (e.g. epinephrine and auxin), protein or peptides (e.g. insulin and CLE peptides), and gases (e.g. ethylene and nitric oxide). Hormones affect distant cells by binding to specific receptor proteins in the target cell, resulting in a change in cell function. When a hormone binds to the receptor, it results in the activation of a signal transduction pathway that typically activates gene transcription, resulting in increased expression of target proteins. Hormones can also act in non-genomic pathways that synergize with genomic effects. Exemplary hormones that are contemplated by the invention are estrogen, progesterone, corticosteroids, and growth hormones.

[0098] In preferred embodiments of the invention, a mixture of the various therapeutic agents are encapsulated in microparticles. The microparticles may comprise a polymer such as collagen, chitosan, cellulose or poly(lactic-co-gly colic acid) (PLGA). In preferredembodiments, the polymer is PLGA. PLGA is a copolymer made of successive monomeric units of glycolic or lactic acid that are linked together by ester linkages.

[0099] The microparticle-encapsulated paracrine signaling agents may be suspended in a delivery vehicle such as a gel, cream, sol-gel, or liquid suspension. The delivery vehicle may be a hydrogel that may undergo gelation in situ in a patient. A hydrogel is a biphasic material; a mixture of porous, permeable solids and typically at least about 10% by weight or volume of interstitial fluid composed completely or mainly by water. In hydrogels, the porous permeable solid may comprise a water insoluble three-dimensional network of natural or synthetic polymers and a fluid, having absorbed a large amount of water or biological fluids. The crosslinks which bond the polymers of a hydrogel fall under two general categories: physical hydrogels and chemical hydrogels.

[0100] Chemical hydrogels have covalent cross-linking bonds, whereas physical hydrogels have non-covalent bonds. Chemical hydrogels can result in strong reversible or irreversible gels due to the covalent bonding. Chemical hydrogels that contain reversible covalent crosslinking bonds, such as hydrogels of thiomers being cross-linked via disulfide bonds, are nontoxic and are used in numerous medicinal products. Physical hydrogels usually have high biocompatibility, are not toxic, and are also easily reversible by simply changing an external stimulus such as pH, ion concentration, or temperature.

[0101] Physical crosslinks consist of hydrogen bonds, hydrophobic interactions, and chain entanglements. A hydrogel generated through the use of physical crosslinks is sometimes called a reversible hydrogel. Chemical crosslinks consist of covalent bonds between polymer strands. Hydrogels generated in this manner are sometimes called permanent hydrogels.

[0102] Hydrogels may be prepared using a variety of polymeric materials, which can be divided broadly into two categories according to their origin: natural or synthetic polymers. Natural polymers for hydrogel preparation include hyaluronic acid, chitosan, heparin, alginate, gelatin and fibrin. Common synthetic polymers include polyvinyl alcohol, polyethylene glycol, sodium polyacrylate, acrylate polymers, polyethylene glycol (PEG), polyacrylate, polyvinylpyrrolidone (PVP), and copolymers thereof. In preferred embodiments, the hydrogel is comprised of hyaluronic acid.

[0103] In addition to one of the above listed polymers, the hydrogel may also comprise a crosslinking agent. The crosslinking allows a hydrogel to transition from a liquid to a gel. Examples of cross-linking agents that may be used include tetraethylmethylenediamine (TEMED), disulfides, diallyl hydrogen phosphate, bis-(2-methylacryloyloxyethyl)phosphate,STMT-OOl / OIWO 30704.0003 divinyl carbonate, diallyl carbonate, bis(2-methylallyl)carbonate, diallyl urea, di vinyl sulfate, diallyl sulfate, diethylene glycol divinyl ether, piperazine diacrylamide, N,N'- ethylenebisacrylamide, ethyleneglycoldiallyl ether, ethylene glycol dimethacrylate, trimethyl oylpropane trimethacrylate, diethyleneglycol dimethacrylate, bis(glycerol dimethacrylate) phosphate, glycerol dimethylacrylate, and divinyl ketone. In preferred embodiments, the transition of the hydrogel from a liquid to a gel occurs in situ once the therapeutic composition has been administered to a patient.

[0104] The invention further provides methods for producing a therapeutic composition. In preferred embodiments, a machine learning system such as a large language model may be used to select therapeutic agents such as paracrine signaling mediators and determine the amounts that should be used in the composition. A machine learning (ML) system is a type of artificial intelligence concerned with the development and study of statistical algorithms that can learn from data and generalize to unseen data, and thus perform tasks without explicit instructions. To use an ML system, a user would need to train the system by inputting relevant data. In the present invention, relevant data may be clinical trial data, scientific articles about various drugs and diseases, patient data, and any other available data that a physician may use when diagnosing a patient. From that, the ML system may learn what combinations of therapies or treatments were most effective for a patient with a specific disease. The ML system will then generate an output of combinations and quantities of therapeutic agents that predicted to be effective, which a user may follow when creating the therapeutic composition of the invention.

[0105] Many different ML systems may be used such as neural networks, random forests, decision trees, inference algorithms, and Bayesian networks. In a preferred embodiment, a large language model (LLM) is used. An LLM is a computational model capable of language generation or other natural language processing tasks. As language models, LLMs acquire these abilities by learning statistical relationships from vast amounts of text during a selfsupervised and semi-supervised training process. Because machine learning algorithms process numbers rather than text, the text must be converted to numbers. In the first step, a vocabulary may be decided upon, then integer indices may be arbitrarily but uniquely assigned to each vocabulary entry, and finally, an embedding is associated to the integer index. Algorithms include byte-pair encoding (BPE) and WordPiece. In the context of training LLMs, datasets are typically cleaned by removing toxic passages from the dataset, discarding low-quality data, and de-duplication. Cleaned datasets can increase trainingefficiency and lead to improved downstream performance. A trained LLM may be used to clean datasets for training another LLM.

[0106] The LLM has preferably been trained on comprehensive datasets, including scientific literature, clinical trial data, and proprietary stem cell information. A patient's information may also be input into the LLM. Patient information may include medical history, allergies, current medications, or demographic information. The LLM profiles and analyzes the patient information against existing data, then outputs the components of the composition (i.e. what specific paracrine signaling agents to use and amounts thereof) that would be most effective at treating the patient. An example may be a selection of one or more of TGF-P, IGF-1, BMP- 7, and IL-lRa.

[0107] Having selected the therapeutic agents by a machine learning system to include in a composition of the invention, the invention includes laboratory methods and techniques for preparing a composition for delivery to a patient.

[0108] In a preferred embodiment, the selected therapeutic agents are obtained, mixed, and dissolved in an aqueous phase to create a solution. The solution may then be emulsified in a polymer such as PLGA. PLGA may also first be dissolved in an organic solvent (e.g., di chloromethane) to create a primary emulsion. The primary emulsion may then be emulsified into a larger aqueous phase that contains a surfactant (e.g., polyvinyl alcohol) to form a double emulsion. The solvent may then be evaporated to form microparticles that encapsulate the paracrine signaling agents, which are then collected and dried.

[0109] The collected and dried microparticles are then mixed into a hydrogel solution to ensure even distribution. The hydrogel solution may be made by dissolving or combining a hydrogel material into a buffer solution and optionally adding a crosslinking agent. Possible hydrogel compounds and crosslinking agents have been discussed above.

[0110] Embodiments of the invention also provide a method for treating a patient with a disease (e.g., osteoarthritis). Osteoarthritis is a type of degenerative joint disease that results from breakdown of joint cartilage and underlying bone. Here, a patient may be injected with the therapeutic composition described throughout the application directly at the site of broken-down joint cartilage.[OHl] Once the therapeutic composition has been administered to the patient, it may form a gel. In situ gelation of the composition may occur in a number of ways. For example, ultrasound stimulation as discussed in Mater. Horiz., 2023,10, 3507-22 may cause a hydrogelthat is composed of fibrinogen and thrombin to gel. The ultrasound waves may also rupture the microparticles containing the paracrine signaling agents, allowing them to disperse.

[0112] Another example is through CO2 induction using methods disclosed in Preibisch et al., 2020, In Situ Measurement Methods for the CO2-Induced Gelation of Biopolymer Systems. Gels. Here, the hydrogel may be formulated with the various polymers discussed throughout the application, and then pressurized CO2 is applied to cause gelation.

[0113] Yet another catalyst for gelation may be the introduction of sandwich ferrocene into a hydrogel that is composed of hyaluronic acid, as discussed in Zhang et al., 2022, In situ gelation strategy based on ferrocene-hyaluronic acid organic copolymer biomaterial for exudate management and multi-modal wound healing, Acta Biomaterialia, 154: 180-93.

[0114] A preferred embodiment uses cross-linking agents to transition the hydrogel from a liquid into a gel. The cross-linking agent may be added to the hydrogel before or after administration to the patient.

[0115] In certain embodiments, a large language model is used to determine the most therapeutically effective dose for the patient. The LLM may use information about the patient such as medical records, demographic information, and allergies. The LLM analyzes the patient's information against scientific literature, clinical trial data, and proprietary stem cell information. Based on the results of the analysis, the LLM will recommend a formulation and dose of the therapeutic composition.

[0116] The invention provides Al systems that output proposed combinations of agents such as active pharmaceutical ingredients and other agents useful for a therapeutic objectives. Systems of the invention propose cocktails of those agents and optionally simulate or model effects of those agents to evaluate the potential effects of those agents in accomplishing the therapeutic objective. Specific embodiments provide a skin care composition from the Al system. The composition includes (i) a mixture of agents that includes epidermal growth factor (EGF), fibroblast growth factor (FGF), and insulin-like growth factor (IGF-1) and one or more of: a collagen-stimulating peptide, a booster of nicotinamide phospho-ribosyl- transf erase (NAMPT) expression, coenzyme Q10, a nucleic acid repair enzyme, and a ceramide; (ii) one or more liposomes encapsulating at least a portion of the mixture of agents; and (iii) a moisturizer carrying the liposomes and the mixture of agents.

Claims

STMT-OOl / OIWO 30704.0003WHAT IS CLAIMED IS1. A skin care composition comprising: a mixture of agents that includes epidermal growth factor (EGF), fibroblast growth factor (FGF), and insulin-like growth factor 1 (IGF-1) and one or more of: a collagen-stimulating peptide, a booster of nicotinamide adenine dinucleotide (NAD+), coenzyme Q10, a nucleic acid repair enzyme, and a ceramide; one or more liposomes encapsulating at least a portion of the mixture of agents; and a carrier carrying the liposomes and the mixture of agents.

2. The composition of claim 1, wherein the FGF comprises FGF-2, and further wherein the EGF and / or the FGF-2 are fused with an oleosome within the mixture.

3. The composition of claim 1, wherein the mixture comprises: the EGF the FGF the IGF-1 at least two distinct collagen-stimulating peptides the booster of NAD+ the coenzyme Q10 formulated with vitamin E photolyase as the nucleic acid repair enzyme, and the ceramide.

4. The composition of claim 1, wherein the carrier is a moisturizer comprising one or more of a humectant, an occlusive, a barrier-repair agent, and an emollient.

5. The composition of claim 1, wherein the collagen-stimulating peptide comprises a mixture of at least palmitoyl tripeptide-5 and palmitoyl pentapeptide-4.

6. The composition of claim 1, further comprising vitamin E.

7. The composition of claim 1, wherein the nucleic acid repair enzyme comprises photolyase8. The composition of claim 1, wherein the booster of NAD+ is provided as an extract from sunflower sprouts and includes an agent that increase expression of nicotinamide phospho-ribosyl transferase (NAMPT) enzyme, resulting in a boost of NAD+.

9. The composition of claim 1, wherein the mixture consists essentially of: the EGF within an EGF-oleosome fusions the FGF as FGF-2 with FGF-oleosome fusions the IGF-1 produced in plant cells palmitoyl tripeptide-5 and palmitoyl pentapeptide-4 as collagen-stimulating peptides sunflower sprout extract to increase NAMPT expressionand boost NAD+ the coenzyme Q10 formulated with vitamin E photolyase as the nucleic acid repair enzyme, and the ceramide.

10. The composition of claim 9, consisting essentially of (w / w): 0.5-1 % the EGF- oleosome fusions 0.5-1 % the FGF-oleosome fusions 0.5-1.5 % the IGF-1 0.1- 5 % palmitoyl tripeptide-5 0.1- 5 % palmitoyl pentapeptide-4 as collagen-stimulating peptides 1-3 % a compound comprising the coenzyme Q10 formulated with the vitamin E 0.5-2 % sunflower sprout extract as the booster of NAMPT expression 0.1- 5 % photolyase as the nucleic acid repair enzyme; 0.1-5 % the ceramide the liposomes; and balance the carrier, wherein the carrier is a moisturizer comprising glycerin, propanediol, squalane, pentylene glycol, polyglyceryl-2 stearate, glyceryl stearate, stearyl alcohol, and hyaluronic acid, and one or more of hydrogenated lecithin, phytosterols, caprylyl glycol, polyacrylate crosspolymer- 11, 1,2-hexanediol.

11. The composition of claim 1, wherein the coenzyme Q10 and the ceramide are distributed among the carrier.

12. The composition of claim 1, wherein the one or more liposomes encapsulate growth factors and water-soluble components of the mixture.

13. The composition of claim 1, wherein the EGF and FGF are present in oleosome fusions; wherein the nucleic acid repair enzyme comprises an algal photolyase introduced into the composition in photolyase liposomes; the booster of NAD+ is introduced into the mixture as a sunflower sprout extract; and the IGF-1 is derived transgenic plants.

14. The composition of claim 1, wherein the EGF and / or the FGF are obtained from transgenic plants of the family Camelina, by a process that includes suspecting seeds from the transgenic plants in suspension, centrifuging and washing the suspension, and collecting a supernatant layer comprising oleosomes fused to the EGF and / or the FGF.

15. The composition of claim 1, wherein the composition is provided in a container or package for topical application.

16. A method for producing a composition, the method comprising: querying a research agent of a machine learning system with a therapeutic objective, wherein the research agent finds and ranks document from a library by relevance to the objective,STMT-OOl / OIWO 30704.0003 summarizes documents for which the relevance is ranked above a threshold, and passes summaries of the summarized documents to a cocktail agent of the machine learning system selecting, by the cocktail agent, a plurality of ingredients simulating, using a validation agent of the machine learning system, effects of the ingredients and concentrations of the ingredients; and outputting an identified set of ingredients predicted to accomplish the therapeutic objective.

17. The method of claim 16, wherein at least one output of the validation agent are passed back to the cocktail agent, whereby the cocktail agent creates a second plurality of ingredients based on the at least one output of the validation agent.

18. The method of claim 16, wherein the validation agent and the cocktail agent operate in a feedback loop in which cocktail agent selects different ingredients until the validation agent outputs an acceptable effectiveness score.

19. The method of claim 16, further comprising outputting the identified set of ingredients with an effectiveness score, wherein the effectiveness score includes a measure of an estimated likelihood of accomplishing the therapeutic objective.

20. The method of claim 19, wherein the effectiveness score includes predicting of adverse effects.

21. The method of claim 16, wherein the machine learning system access medical knowledge input comprising biomedical data, clinical data, and research literature.

22. The method of claim 16, wherein the research agent queries the library with the therapeutic objective to get a list of results, filters results by availability, retrieves available results, and summarizes the retrieved, available results.

23. The method of claim 16, wherein the therapeutic objective comprises skin health.

24. The method of claim 23, wherein the plurality of ingredients comprises epidermal growth factor (EGF), fibroblast growth factor (FGF), insulin-like growth factor 1 (IGF-1) a collagen-stimulating peptide, a booster of NAD+, coenzyme Q10, a nucleic acid repair enzyme, and a ceramide25. The method of claim 24, wherein the research agent and / or the cocktail agent scores the plurality of ingredients for one or more of: efficacy, cartilage regeneration, and inflammation reduction.STMT-OOl / OIWO 30704.000326. The method of claim 16, wherein the therapeutic objective comprises osteoarthritis.

27. The method of claim 26, wherein the plurality of ingredients comprises growth factors, cytokines, non-coding RNAs, extracellular vesicles, exosomes, proteins, enzymes, ECM components, hormones, and antioxidants, wherein the growth factors include one or more of TGF- P, IGF-1, BMP-7, and IL-lRa.

28. The method of claim 16, wherein the validation agent: predicts release kinetics of active agents in liposome compositions, optimizes membrane stability and / or targeting ligands for specific tissues; and / or simulates absorption, biodistribution, and bioavailability of the plurality of ingredients.

29. The method of claim 16, wherein the machine learning system is coupled to a microfluidics system that: combines elements of the plurality of ingredients; forms a biphasic delivery system; and encapsulates at least a portion of the plurality of ingredients in the biphasic delivery system.

30. The method of claim 16, wherein the cocktail agent includes a large language model (LLM) module and / or the validation agent includes a graph neural network (GNN) module.

31. A therapeutic composition comprising: a mixture comprising a plurality of paracrine signaling agents; a plurality of microparticles encapsulating the agents of the mixture; and a hydrogel suspending the microparticles, wherein the hydrogel exhibits in situ gelation once delivered into a patient.

32. The composition of claim 31, wherein the paracrine signaling agents include one or more of TGF-P, IGF-1, BMP-7, and IL-lRa.

33. The composition of claim 31, where the microparticles are composed of poly(lactic- co-glycolic acid) (PLGA).

34. The composition of claim 31, wherein the hydrogel is composed of hyaluronic acid.

35. The composition of claim 31, wherein a crosslinking agent is added to the hydrogel.

36. The composition of claim 31, wherein a large language model is used to determine therapeutically effective formulations and dosages of the therapeutic composition.

37. A method for producing a therapeutic composition, the method comprising: using a machine learning system to select paracrine agents, and amounts thereof; creating aSTMT-OOl / OIWO 30704.0003 primary emulsion comprising the selected agents emulsified in a polymer; emulsifying the primary emulsion into a solvent to create a double emulsion; evaporating the solvent away to provide microparticles that encapsulate the paracrine agents; and mixing the microparticle-encapsulated paracrine agents into a hydrogel solution.

38. The method of claim 37, wherein the machine learning system selects paracrine agents and amounts thereof after a user inputs patient information.

39. The method of claim 38, wherein the patient information consists of a patient's medical history, allergies, current medications, or demographic information.

40. The method of claim 37, wherein the therapeutic composition comprises paracrine agents, microparticles, and a hydrogel.

41. The method of claim 37, wherein the paracrine agents are growth factors selected from one or more of TGF-P, IGF-1, BMP-7, and IL-lRa.

42. The method of claim 37, where the microparticles are composed of poly(lactic-co- glycolic acid) (PLGA).

43. The method of claim 37, wherein the hydrogel, once administered to a patient, gels.

44. A method for treating a patient, the method comprising: administering to a patient a therapeutic composition comprising a mixture comprising a plurality of paracrine signaling agents, a plurality of microparticles encapsulating the therapeutic agents of the mixture, and a hydrogel suspending the microparticles; and transitioning the therapeutic composition from a liquid to a gel in situ in the patient.

45. The method of claim 44, wherein a machine learning system determines a therapeutically effective dose for the patient.

46. The method of claim 45, wherein the machine learning system is a large language model.

47. The method of claim 44, wherein the patient has osteoarthritis.

48. The method of claim 44, wherein the transitioning step is done by CO2 induction.

49. The method of claim 44, wherein the transitioning step is done by combining ferrocene and hyaluronic acid.

50. The method of claim 44, wherein the transitioning step is done using ultrasound.STMT-OOl / OIWO 30704.000351. The method of claim 44, wherein the transitioning step is done using crosslinking agents.

52. A method for replicating and delivering paracrine signaling models of stem cells for therapeutic applications, the method comprising: using a large language model (LLM) trained on scientific, clinical, and cell data to identify and optimize a composition and dosage of a recombinant protein cocktail (RPC), encapsulating the RPC in microparticles, and delivering the RPC in conjunction with hyaluronic acid to targeted tissue sites in a patient.

53. The method of claim 52, wherein the encapsulated RPC is tailored to treat osteoarthritis.

54. The method of claim 53, wherein the encapsulated RPC includes at least one protein selected from the list of TGF- p, IGF-1, BMP-7, and IL-lRa.

55. The method of claim 52, wherein the delivery step is a controlled and sustained release of the therapeutic agents inside the RPC.

56. The method of claim 52, wherein the LLM is continuously refined based on new scientific data and clinical outcomes.

57. The method of claim 52, wherein the combination of therapeutic agents is one or more of growth factors, cytokines, non-coding RNAs, extracellular vesicles, exosomes, proteins, enzymes, ECM components, hormones, and antioxidants.

Images