Neuroprotective effect of mealworm protein hydrolysate-derived bioactive peptides in human microglial cells

Journal of Insects as Food and Feed (2024) DOI:10.1163/23524588-00001313 brill.com/jiff Research article Neuroprotective effect of mealworm protein hydrolysate-derived bioactive peptides in human microglial cells T. Gonzalez-de la Rosa1,2 , M.A. del Valle-Alonso1, S. Montserrat de la Paz1,2* and F. Rivero-Pino1,2,3 1Department of Medical Biochemistry, Molecular Biology, and Immunology, School of Medicine, University of Seville, Av. Sanchez Pizjuan s/n, 41009 Seville, Spain; 2Instituto de Biomedicina de Sevilla, IBiS, Hospital Universitario Virgen del Rocio/CSIC/University of Seville, 41013 Seville, Spain; 3European Food Safety Authority, Nutrition and Food Innovation Unit, Novel Foods Team, 43126 Parma, Italy; *[email protected] Received 6 July 2024 | Accepted 18 November 2024 | Published online 7 December 2024 Abstract The larva of Tenebrio molitor, better known as mealworm, is one of the most studied insects today, since it was recently classified as safe for human consumption. Specifically, its high protein content makes it a splendid candidate for the search for bioactive peptides, with properties ranging from anti-inflammatory to neuroprotective. In this study, the effect of a digested hydrolysate on microglia cells (which previously demonstrated high antiinflammatory activity in intestinal CACO-2 cells) was analysed, specifically on their levels of gene expression of various cytokines, such as IL-6, IL-10, or IL-1β, among others. In addition, four chemically synthesised peptides, selected from this digested hydrolysate based on in silico analysis, including the estimation of the bioactive properties of these peptides, their physicochemical properties and their toxicokinetic, were evaluated in the microglia cells as well. In addition, molecular docking assays of the peptides were performed with two receptors relevant to inflammation. Relevant changes in the expression of IL-6, IL-1β, and BDNF were observed due to the action of the peptides, especially in the case of IYVDAVIN and SLPSLPEPV. These results provide an insight on how specific peptides found in Tenebrio molitor could be used as therapeutics agent in the immunomodulation of brain cells. Keywords bioactive peptides – microglia – neuroprotection – protein hydrolysates 1 Introduction Faced with the prospect of constant growth in the world’s population, and the environmental problems arising from the excessive production of animal products, it is imperative to find new sources of food that are healthy and nutritious, but at the same time sustainable and respectful of the environment (Galanakis, 2024). In this context, there are many strategies that can be followed from a biotechnological approach to try to guarantee safe foodfor the population, while seeking to minimise the impact generated on the environment derived from its production (Haessner et al., 2024). In this context, entomophagy, defined as the practice of including insects in the human diet, has gained the greatest relevance in recent years. This is due both to the simplicity of insect farming compared to traditional livestock (Oonincx and De Boer, 2012), and to its reduced ecologi- Published with license by Koninklijke Brill BV | DOI: 10.1163/23524588-00001313 © T. Gonzalez-de la Rosa et al., 2024 | ISSN: 2352-4588 (online) This is an open access article distributed under the terms of the CC BY 4.0 license. 2T. Gonzalez-de la Rosa et al. cal impact compared to other livestock farms. In the use of insects as food for humans, it is relevant to mention their interesting nutritional composition, given their high content of chitin, unsaturated fatty acids and protein content (Stull and Weir, 2023), as well as vitamins such as those of group B (Finke and Oonincx, 2014), or minerals such as iron or zinc. In addition to this, the peptide sequences within the structure of the protein itself are being studied. It has been shown that the cleavage of the native protein generates a set of peptides (i.e. protein hydrolysate), among which some can exert some type of biological activity, being called bioactive peptides. It should be noted that once these peptides are absorbed in the intestinal lumen, most are hydrolyzed by peptidases in the cytosol of the enterocyte itself, and that only a few will be resistant to these and then transported through the basolateral membrane to the bloodstream (Amigo and Hernández-Ledesma, 2020), potentially exerting activity on specific targets. In this way, the diet, especially the action of bioactive peptides, can have a certain modulating effect on the immune system, by regulating the release of pro-inflammatory cytokines carried out by enterocytes (Tourkochristou et al., 2021; del Hierro et al., 2024). Recently, our research group reported the full peptidome of different T. molitor hydrolysates and their respective digested and evaluated the samples in relation to the immunomodulatory potential in CACO- 2 cells (Gonzalez-de la Rosa et al., 2024), but no chemically synthesised peptides were evaluated. A protein hydrolysate usually has between 1000 and 6000 sequences, so synthesising all the sequences to validate its activity is an expensive and laborious job that can hardly be carried out (Rivero-Pino et al., 2023). To address this challenge, there are currently numerous in silico tools that allow to determine many of the properties of a peptide through rapid computer analysis, to then select those of interest (Rivero-Pino et al., 2023; Fatoki et al., 2022). Precisely in the case of the nervous system, some peptides from the diet have shown promising neuroprotective results (Patel and Mani, 2024), reducing both inflammatory and oxidative stress levels in nervous tissue. Neuroprotection refers to the set of molecular, biochemical, and physiological mechanisms that favour the proper development of the nervous system and prevent or delay the deterioration or cell death of neurons (Bohlson and Tenner, 2023). This ability is of even greater interest in the field of neurodegenerative diseases, such as Alzheimer’s disease (Albadrani et al., 2024) or Parkinson’s disease (Tosefsky et al., 2024), among others, in which it would be crucial to increase the neuronal survival rate significantly. Finding bioactive compounds in the diet that have significant neuroprotective properties would be decisive in improving the manifestation and symptomatology of neurodegenerative diseases, whose incidence is increasingly high among the population, given the longer life expectancy and the complexity of their treatment at the pharmacological level compared to pathologies associated with other organs less complex than the brain (Skibska et al., 2024). In this context, the role of the blood-brain barrier (BBB), a highly selective and specific exchange system with the endothelium that protects the brain, is of interest, and it must be determined whether bioactive peptides are permeable, so that they can exert their neuroprotective function once in the nervous tissue (Badaut et al., 2024; Galland et al., 2024). The aim of this study was therefore to evaluate the anti-inflammatory activity in cells of the central nervous system that function as elements of the immune system (microglia) by analysing the regulation in the gene expression of pro and anti-inflammatory cytokines, which can be exerted by selected peptides, and the digested hydrolysate from which they come.The digested hydrolysate previously defined as highly antiinflammatory in CACO-2 cells (called dTMH10A) was administered to microglia cells previously stimulated with inflammatory stimuli to observe if it exerts such an immunomodulatory effect. From the characterized peptidome of the hydrolysate and its digest obtained by the INFOGEST protocol, 4 peptides were selected based on an exhaustive in silico analysis, which elucidates which may be the candidate peptides that may have bioactive capacity. These peptides were also evaluated in terms of their possible interactions with the membrane receptors (TLR4/MD2 and TREM2) of the cells on which they can exert regulation through molecular docking. 2 Materials and methods Samples, chemicals and cells Tenebrio molitor flour (TMF) was donated by Insekt Label Biotech (Vizcaya, Spain). The enzyme Alcalase (2.4 L) was donated by Novozymes (Bagsvaerd, Denmark). All chemicals (reagents and solvents) used were of analytical grade and provided by Sigma Chemical Co., Bachem AG (Bubendorf, CH, USA), and Gibco (Waltham, MA, USA). HMC3 (CRL-3304) human microglia cells were purchased from ATCC. Journal of Insects as Food and Feed 0 (2024) 1–17 Neuroprotective effect of mealworm protein hydrolysate 3 Hydrolysis of Tenebrio molitor flour TMF hydrolysis was performed in a jacketed reactor with agitation, at a constant temperature (50 °C) and pH (8), and monitored with the pH-stat method (Camacho et al., 2001) that allows determining the degree of hydrolysis (DH). To do this, TMF was diluted to a concentration of 28 g protein/L in distilled water, brought to the indicated temperature and pH conditions, and then, Alcalase was added at an enzyme/substrate ratio of 0.375% and the protease was deactivated to stop the reaction when the DH reached 10%, by freezing at −80 °C. The sample was therefore named TMH10A. This sample was exerting high anti-inflammatory activity in CACO-2 cells compared to protein hydrolysates with higher DH (Gonzalez-de la Rosa et al., 2024). The sample was then freeze-dried and stored until analysis. Chemical characterization of samples A proximate analysis encompassing protein, fat, and fibre content was evaluated by elemental microanalysis, gravimetry and the acidic fibre detergent method, respectively, as described in Gonzalez-de la Rosa et al. (2024). The amino acids content was measured after chemical hydrolysis of the sample, using a Thermo- Scientific liquid chromatography system consisting of a binary Dionex Ultimate 3000 RS UHPLC, connected to a quadrupole-Orbitrap QExactive mass spectrometer, whereas the lipid profile was analysed using an Agilent 6890N gas chromatograph equipped with a flame ionization detector. The total phenolic content was quantified in each sample according to the Folin– Ciocalteu spectrophotometric method, and the mineral content using a plasma-paired inductive optical emission spectroscope after a microwave-assisted digestion. Full details of this methodology can be found elsewhere (Gonzalez-de la Rosa et al., 2024). INFOGEST, standard in vitro digestion model Simulated gastrointestinal digestion (SGID) was performed following the INFOGEST protocol (Brodkorb et al., 2019) with slight modifications. The digestive enzymes included were α-amylase (EC 3.2.1.1), pepsin (EC 3.4.23.1), pancreatic lipase (EC 3.1.1.3), pancreatin and bile salts, diluted in simulated salivary, gastric, and intestinal fluids, respectively. To carry it out, 2 g of the sample (TMH10A) were diluted 1:2 and subjected to the different phases: oral, gastric, and intestinal for 0.5, 2, and 2 h, respectively. At the end of the SGID, enzymatic deactivation was carried out by freezing each sample at −80 °C and then freeze-drying and storing it until subsequent analyses, now calling it the digested hydrolysate as dTMH10A. Extraction, purification and identification of peptide sequences The samples (TMH10A and dTMH10A) were acidified with 0.5% trifluoracetic acid (v/v).The desalination and concentration step were performed with ZipTip C18 (Millipore, Billerica, MA, USA) and the samples were dried in a vacuum concentrator. The LC-TIMS-MS/MS was performed using a nanoElute for nanoflow with ultra-high pressure with the LC system (Bruker Daltonics, Bremen, Germany) paired with a timsTOF Pro2 mass spectrometer, equipped with a CaptiveSpray nanoelectrospray as an ion source (Bruker Daltonics) according to the process described in the P202230873 patent. The complete identification protocol can be found in González-de la Rosa et al. (2024). Bioactivity prediction by in silico analysis Gastro-resistant peptides (those identified in both samples) with a molecular weight (MW) of less than 1000 Da were subjected to in silico analysis: (1) Toxin- Pred software, was used to determine hydrophobicity, amphipaticity, steric hindrance, charge, and MW(Gupta et al., 2013); (2) PepCalc software, was used to measure solubility, number of residues, and isoelectric point (Lear et al., 2016); (3) The anti-inflammatory activity of peptides was estimated using the PreAIP software (Khatun et al., 2019); (4) PeptideRanker, was used to classify sequences according to their predicted bioactive probability (Mooney et al., 2012); (5) AnOxPePred-1.0, was used to predict antioxidant activity (Olsen et al., 2020); (6) PASTA2.0, was used to estimate the secondary structure (Walsh et al., 2014); (7) The blood half-life was predicted using PlifePred (Mathur et al., 2018); (8) The SCMB3PP tool was used to predict the ability of peptides to penetrate the BBB (Charoenkwan et al., 2022); and (9) AllerTOP v.2.0, to predict sequence allergenicity (Dimitrov et al., 2014). Based on these results, four peptides were selected and chemically synthesized by bioNova Científica (Madrid, Spain). The synthesised peptides were Peptide 1: GGGGGLGGGGGL; Peptide 2: GLIGAPIAAPI; Peptide 3: IYVDAVIN; and Peptide 4: SLPSLPEPV, with 95% purity. Cell viability HMC3 microglial cells were incubated with different concentrations of lipopolysaccharide (LPS), β-amyloid (Aβ), as inflammation stimuli, dTMH10A, and the 4 selected peptides in 96-well plates (105 cells/well) for Journal of Insects as Food and Feed 0 (2024) 1–17 4T. Gonzalez-de la Rosa et al. 24 h. Increasing concentrations ranging from 10 μg/mL to 1 mg/mL were used. The 3-(4,5-dimethylthiazole-2- yl)-2,5-diphenyltetrazolium (MTT, Sigma Aldrich) bromide solution was then added at 0.5 mg/mL and incubated for 3 h until the purple precipitate was visible.The MTT crystals were solubilized with dimethyl sulfoxide and then measured with a plate reader at 570 nm corrected for 650 nm. HMC3 culture and treatments with dTMH10A and peptides HMC3 microglial cells were maintained in a highglucose Dulbecco’s Modified Eagle’s (DMEM) cellular medium, supplemented with 10% heat-inactivated fetal bovine serum and 1% penicillin/streptomycin. They were incubated in an atmosphere with 5% CO2 at 37 °C and medium renewal every 2-3 days. The experiments were conducted in 12-well plates, culturing the HMC3 cells at a density of 1 × 106 cells per well. Previously, HMC3 cells were pre-stimulated with a proinflammatory component, either LPS at 0.5 μg/mL or Aβ peptide (500 mM). After1hofstimulation, the cells were treated with dTMH10A and the four peptides for 24 h. For the hydrolysate, treatment was carried out at two concentrations, both 50 μg/mL and 100 μg/mL; while for peptides only one concentration was used, 50 μg/mL. Analyses were carried out in triplicate. RNA and RT-qPCR isolation The total RNA of HMC3 cells stimulated with LPS and Aβ was extracted using gTPzol (CViral, Seville, Spain). The A260/A280 ratio of a NanoDrop ND-2000 spectrophotometer (Thermo Scientific, Madrid, Spain) was used to test the quality of the RNA extract. RNA (1 μg) was then reverse-transcripted (iScript, Bio-Rad, Madrid, Spain). The resulting cDNA was used as a template for real-time PCR amplifications at a concentration of 10 ng/μL. The levels of gene-specific mRNAs were determined with the CFX96 system (Bio-Rad) the pairs of primers for each gene and also for hypoxanthine phosphoribosyltransferase-1 (HPRT) as a housekeeping gene as constitutive, all of them detailed in Supplementary Table S1. All amplification reactions were performed quadruplicate with the mean number of cycles (Ct) for mRNA expression in the target genes, quantified with the standard method 2−(ΔΔCt). All data were standardized with respect to the endogenous reference gene (HPRT) and expressed as a relative value of expression with respect to the value of expression in the samples of control cells. Statistical analysis All values are presented as means ± standard deviations (SD). The data were evaluated using GraphPad Prism version 9.1.2(San Diego, CA, USA).The statistical significance of any differences between group parameters was determined using a one-variable analysis of variances (ANOVA), followed by a multiple comparison Dunnet test (for RT-qPCRs). Pvalues less than 0.05 were considered significant. Molecular docking Molecular docking was performed to determine the binding affinity energy of the selected peptides with TLR4/MD2 and TREM2 receptors. The crystallized X-ray structure of both the TLR4/MD2 (PDB: 3FXI) complex containing LPS and the TREM2 (PDB: 5UD7) were obtained from the RCSB PDB (Protein Data Bank, http://www.rcsb.org/) database. The ligands and all water molecules were removed from the PDB file of the receiver, while the polar hydrogen atoms were added using the UCSF Chimera software. The 3D structures of the peptides were obtained and minimized by the USCF Chimera. The molecular structures of TLR4/MD2, TREM2 and the four peptides were then converted to PDBQT format with AutoDock Tools. The AGFR program was used to calculate the positions and sizes of the specific docking boxes for each peptide in each of the two receptors. Next, AutoDock Crank Pep was used to perform the docking analysis. Finally, the dockings with the best potential score were selected and visualized via Biovia Discovery StudioVisualizer, as well as the 2D two-dimensional structure of the surface of both the ligand and the protein. 3 Results and discussion Chemical characterization of hydrolysate The composition of the hydrolysate encompassed protein, fat, neutral fibre, and phenol content in percentage of 40.17 ± 1.55, 33.57 ± 3.94, 10.62 ± 0.22, and 2.00 ± 0.15, respectively. The protein and fat content of the hydrolysate was in line with previous literature on T. molitor hydrolysates (Rivero-Pino et al., 2024a,b; Precup et al., 2022; van Huis et al., 2021). On the other hand, in the analysis of fibre, the acid detergent fibre was 5.12, thus, it could be estimated a chitin content of 5.5 (Hahn et al., 2018). In addition to that, the phenol content of insects has been scarcely reported, but in this study, it was noted a considerable amount which could be of interest in future studies. Journal of Insects as Food and Feed 0 (2024) 1–17 Neuroprotective effect of mealworm protein hydrolysate 5 Table 1 Summary of the outcomes revealed by the in silico tools evaluated for the 129 gastrorresistant <1000 Da peptides identified PreAIP score PlifePred (seconds) SCMB3PP (threshold = 749) MAX 0.529 1069.91 868.75 MIN 0.253 724.91 696.29 Average 0.355 837.49 789.25 High confidence AIP 6 (4.65%) 116 B3PP (89.9%) Medium confidence AIP 23 (17.8%) 13 Non-B3PP (10.1%) Low confidence AIP 43 (33.3%) Negative AIP 57 (44.1%) Pre-AIP = http://kurata14.bio.kyutech.ac.jp/PreAIP/. It classifies peptides for their confidence to be anti-inflammatory, as negative, low, medium, high; Plifepred = https://webs.iiitd.edu.in//raghava/plifepred/. It estimated the half-life of peptides in blood (in seconds); SCMB3PP = http://pmlabstack.pythonanywhere.com/SCMB3PP It determines the likelihood to cross the BBB (threshold of 749 established by the authors). In relation to the amino acids content, the presence of all 9 essential amino acids was detected, in quantities relevant to daily nutritional requirements. Other amino acids, such as phenylalanine (aromatic) or valine (hydrophobic), may be of interest as they could enhance the antioxidant activity in peptides that contain them. On the other hand, the most abundant fatty acids were oleic acid (37.57%), linoleic acid (27.90%) and palmitic acid (19.12%), whereas for the minerals, the ones found in higher concentrations were sodium, phosphorus and potassium, with >10 g/kg (Supplementary Table S2). In silico peptide analysis Physicochemical characterization The peptidome of the TMH10A sample contained 14998 sequences, while the dTMH10A contained 1941 sequences. This reduction in identified sequences may be due to the fact that the technique used does not detect sequences smaller than 6 amino acids in length, which are those that can be generated after digestion. Generally, in the field of bioactive peptides, less than 1000 Da gastro-resistant peptides are those relevant at the level of bioactivity (Peighambardoust et al., 2021; Picariello et al., 2021), and were therefore the focus of the study. A total of 129 sequences met these criteria and were subjected to in silico tests. A summary of the results obtained for these 129 sequences in relation to the PreAIP, PlifePred, and SCMB3PP tools is shown in Table 1, where it can be observed than 90% of the peptides would cross the BBB and their half-life in blood is comparable to other hormones able to act in the human body (Gonzalez-de la Rosa et al., 2024). In relation to the predicted anti-inflammatory activity, the values were comparable to recent literature identifying anti-inflammatory peptides (Gopinatth et al., 2021). From these 129, the 20 most active were selected according to the PreAIP tool. In SupplementaryTable S3, for these 20 peptides, their predicted physicochemical properties are reported. These parameters allow to elucidate how peptides will behave and their bioactive potential, for example, a peptide with a high amphipathic character, as well as a low steric impairment, is likely to have a better bioactive character, since it will be easier to stabilize the conformation of the enzymes on which it could have a modulating effect (Mora et al., 2020). However, this in silico approach is not enough to state the bioactivity of any compound. Toxicokinetics and bioactivity Table 2 shows the prediction results of the 20 selected sequences, regarding their anti-inflammatory, antioxidant and allergenicity, as well as the general bioactivity (PeptideRanker), the half-life of the peptides in the blood plasma, and their permeability with respect to the BBB. PeptideRanker provides numerical values for each sequence, where a value greater than 0.5 (or ideally 0.8) provides great reliability to determine that the peptide is bioactive, according to the developers of the web tool themselves (Mooney et al., 2012). The most promising peptides according to this tool would be GGGGGLGGGGGL and GGGSGGAGGGGL. In relation to the anti-inflammatory activity (evaluated by Pre- AIP), a value greater than 0.47 would allow us to affirm with high confidence that this peptide sequence is antiinflammatory. Of the 20 peptides, 6 of them obtained this high-confidence result, and the other 14 a value between 0.47 and 0.39; having a medium confidence, comparable to peptides from other sources, as mentioned previously. In addition, all 20 peptides obtained high results in the AnOxPePred analysis tool, which identifies the capacity to scavenge free radicals (FRS) Journal of Insects as Food and Feed 0 (2024) 1–17 6T. Gonzalez-de la Rosa et al. Table 2 Predicted toxicokinetic and bioactivity of the 20 gastro-resistant peptides with MW <1000 Da that obtained the highest values in the Pre-AIP analysis Peptides Peptide Ranker a PreAIP b Prediction with PreAIP AnOXPred c Allergenicity prediction d Halflife(s) e SCMB3PP f FRS CHEL GLIGAPIAAPI 0.542384 0.529 High confidence 0.41735241 0.23994091 PNA 839.11 814.10 IYVDAVIN 0.0980126 0.525 High confidence 0.28964424 0.17704557 PA 824.11 784.00 SLPSLPEPV 0.485013 0.521 High confidence 0.43029708 0.27932003 PNA 834.11 796.62 GGGGGLGGGGGL 0.929482 0.507 High confidence 0.50041258 0.19938439 PNA 834.81 830.73 GLIGAPIAAPL 0.650559 0.476 High confidence 0.40337378 0.25729752 PNA 843.91 816.00 GHPVETPE 0.138711 0.472 High confidence 0.43283486 0.24516445 PNA 821.71 748.57 WIPDINE 0.442267 0.460 Medium confidence 0.384121 0.25140926 PA 837.51 773.50 GGGSGGAGGGGL 0.865732 0.444 Medium confidence 0.45576787 0.19912483 PNA 834,81 821.27 IATPALGAL 0.460566 0.435 Medium confidence 0.35637504 0.24583387 PNA 862.31 730.62 QVHPDTGIS 0.161728 0.431 Medium confidence 0.39366096 0.23587096 PNA 938.41 791.50 EFQFGVSL 0.406814 0.430 Medium confidence 0.35691726 0.22165833 PA 724.91 780.43 GSGLIGGGLY 0.593491 0.428 Medium confidence 0.47397298 0.23423967 PNA 826.71 798.67 IAAPIAPIA 0.45534 0.425 Medium confidence 0.33304119 0.23399849 PA 834.81 804.62 IVGGDEAAE 0.0702873 0.424 Medium confidence 0.35231122 0.20415857 PA 872.01 756.00 GYEPPVQE 0.151726 0.422 Medium confidence 0.48824069 0.25539362 PA 821.21 800.43 SVVPSPK 0.214148 0.416 Medium confidence 0.3808786 0.25845125 PA 835.81 777.50 FTGPSLGPK 0.514775 0.415 Medium confidence 0.43408331 0.24583943 PA 913.71 785.38 SWGPSHPE 0.599721 0.408 Medium confidence 0.54359734 0.28828046 PNA 868.41 813.86 VHPDTGIS 0.161388 0.406 Medium confidence 0.37468761 0.23069441 PNA 847.31 790.43 GGIGGTGGQ 0.241236 0.402 Medium confidence 0.46653453 0.23650835 PNA 834.81 809.75 a The probability that peptides have bioactivity was evaluated with PeptideRanker (http://bioware.ucd.ie/~compass/biowareweb), a server that classifies them with scores between 0 (least active) and 1 (most active). b The Pre-AIP tool; (http://kurata14.bio.kyutech.ac.jp/PreAIP/) was used to predict the anti-inflammatory potential of peptides, classifying them from low, (<0.39) to high, (>0.47); N = negative; M = medium. c AnOxPePred Tool (http://services.bioinformatics.dtu.dk/service.php?AnOxPePred-1.0) predicts the ability to entrain the free radical and chelating ion in peptide scores. d Allergenicity prediction was made with AllerTOPv.20. (https://www.ddg-pharmfac.net/AllerTOP/). PA = Probably allergen; PNA = Probably non-allergenic. e The blood half-life (in seconds) was estimated using PlifePred, accessible from https://webs.iiitd.edu.in//raghava/plifepred/. f SCMB3PP (http://pmlabstack.pythonanywhere.com/SCMB3PP) estimates the probability of the peptide crossing the blood-brain barrier (threshold of 749). All tools were used in May 2024. Journal of Insects as Food and Feed 0 (2024) 1–17 Neuroprotective effect of mealworm protein hydrolysate 7 and chelating ion (CHEL); and therefore determine the antioxidant character of the peptides, with results between 0.35 and 0.57; and 0.18 and 0.30, respectively, which are in line with peptides that have shown antioxidant activity as chemically synthesised compounds (Yesiltas et al., 2022). After all the relevant analyses, the four peptides selected were those with a higher score in Pre-AIP that reflects high confidence of being anti-inflammatory, ensuring that they maintained adequate antioxidant values in the AnOxPePred tool, a blood half-life sufficient to exceed the average threshold, considered so that they can be captured from the bloodstream, and finally a value in the tool SCMB3PP above the threshold, which implies that they are capable of crossing the BBB, so that their bioactive and neuroimmunomodulatory character can have an effect on cells of the nervous system. These results were taken into account to select the peptides to be synthesised chemically, which were subject of cell-based analyses as described in the following sections. These peptides were, subsequently: (1) reported experimentally to resist the gastrointestinal digestion; (2) their MW suggests their ability to cross the intestinal barrier; (3) the in silico predictions suggest that their half-life is enough to achieve different targets; (4) the in silico prediction suggest that these sequences are able to cross the BBB. Then, their anti-inflammatory activity in microglia cells is the purpose of the study (i.e. to determine the effect on the gene expression of pro and anti-inflammatory cytokines in the cells of the nervous system, exerted by the digested hydrolysate and the selected peptides). Anti-inflammatory activity The relative expression of different cytokines is observed in Figures 1 and 2, for dTMH10A and the synthetic peptides, respectively. In Figure 1, the relative expressions of IL-6 (Figure 1A), TNF-α (Figure 1B) and IL-1β (Figure 1C), three genes of a proinflammatory nature are depicted. In these cases, in conditions of inflammation (stimuli with LPS), the gene expression increased. In the samples treated with the dTMH10A, a tendency towards a decreased gene expression is observed, although not statistically significant. On the other hand, IL-10 (Figure 1D) is an anti-inflammatory gene, so as can be seen its expression levels fall drastically in the positive controls, while when the cells are treated with dTMH10A, it reflects a certain increase in expression, although not reaching the initial levels of the positive control. Although not statistically significant, it does indicate that the presence of hydrolysate reduces to some degree the level of inflammation produced by the previous stimuli. BDNF (Figure 1E), a gene of the neurotrophin family that acts as a nerve growth factor, and for which its increased expression in controls can be explained by the need for this factor for the survival of nerve cells, and that in the presence of a negative inflammatory stimulus its expression is increased in order to maintain cell survival. Then, a drastic increase in expression in treatments was observed, where the digested hydrolysate seems to favour expression. Figure 2 shows the results obtained with the synthetic peptides. In the case of the three proinflammatory genes, the results reflect a drastic decrease in their expression in the peptide treatments, to such an extent that, in the case of both IL-6 (Figure 2A) and IL-1 β (Figure 2C), the four treatments were significantly different from the control in the case of the Aβ stimulus. In fact, except for the treatment with GGGGGLGGGGGL (P1) on Aβ in IL-6, the results indicate changes statistically relevant. In the case of TNF-α (Figure 2B), however, although it is true that a reduction in expression is observed in the treatments, it is not significant and also remains at higher levels of expression than those of the negative control. For the IL-10 gene (Figure 2D), a drastic decrease in its expression is observed in positive controls due to its anti-inflammatory nature, but later it is observed that the treatments successfully counteract this loss of expression, being especially interesting the case of treatment with SLPSLPEPV (P4), which results in a particularly high expression in both Aβ-stimulated and LPS- stimulated cells, obtaining in the latter case statistical significance. Finally, the BDNF gene (Figure 2E) also shows an increase in its expression when exposed to inflammatory stimuli, although it behaves very differently against both, against LPS its expression increases markedly in the positive control, and the presence of the 4 peptides in the treatment does not achieve a very high increase; it is in the case of Aβ where a much more drastic increase in expression is observed, especially with IYVDAVIN (P3) and SLPSLPEPV (P4), both presenting statistical significance, for P3 a p-value of 0.01 and for P4ap-value of 0.05. The regulation of gene expression is given by the interaction of the stimulus, LPS or Aβ with the cell membrane receptors, which generate signaling cascades that affect the epigenetic balance of the genome, thus modifying the expression levels of certain genes (Iyer et al., 2010). In accordance with this, the results of the relative expression of the TLR4 receptor (Figure 3) reflect this interaction, since in conditions that only had Journal of Insects as Food and Feed 0 (2024) 1–17 8T. Gonzalez-de la Rosa et al. Figure 1 Relative transcriptional activity of different genes (A) IL-6; (B) TNF-α; (C) IL-1β; (D) IL-10, and (E) BDNF. In each of them, 7 conditions are reflected, one negative control, two positive controls (cells pre-stimulated with LPS and Aβ, respectively, but without treatment) and four treatments with dTMH10A (two for each inflammatory stimulus, one of them at 50 μg/mL and the other at 100 μg/mL). The values are represented as normalized values with respect to the standard, for which the control was taken, and presented as a mean ± standard deviation (SD with an n= 4). The level of statistical significance was set at a p-value less than 0.05, which was not obtained in these graphs. inflammatory stimulus, the expression of the receptor was clearly increased compared to the control, while when the treatments were applied, the expression was drastically reduced, being that the peptides interact directly or indirectly with the signaling pathway that controls the expression of TLR4. Therefore, when the expression of this is reduced, so does the number of interactions it generates with the inflammatory stimulus, and therefore the level of activation of the signaling cascades that would cause, for example, the increase in the expression of proinflammatory genes is also reduced. It is in this succession of signalling that bioactive peptides would exert their modulation, binding to both receptors and enzymes, modifying their conformations, active centers, and binding sites (Takahashi et al., 2009). In the case of the two receptors studied, TLR4/MD2 and TREM2 (Figure 4), molecular docking also showed that another point in the regulation of gene expression affected by peptides could be the interaction of the receptor itself with the inflammatory stimulus, since the peptides proved to be able to bind to the receptor, with a large number of molecular interactions and an affinity that favors their binding. being able in the process to alter the receptor’s recognition domain for these molecules. Finally, comparing the results of both experiments, although both correspond to the hypothesis, it is notable that the anti-inflammatory effect of the treatment with each of the peptides separately was more effective than the treatment with the complete hydrolysate, generating much more statistical significance. Given that the Journal of Insects as Food and Feed 0 (2024) 1–17 Neuroprotective effect of mealworm protein hydrolysate 9 Figure 2 Relative transcriptional activity of different genes (A) IL-6; (B) TNF-α; (C) IL-1β; (D) IL-10, and (E) BDNF. In each of them, 9 conditions are reflected, one negative control, two positive controls (cells pre-stimulated with LPS and Aβ, respectively, but without treatment) and four treatments with the 4 selected peptides (50 μg/mL). P1 = GGGGGLGGGGGL; P2 = GLIGAPIAAPI; P3 = IYVDAVIN; and P4 = SLPSLPEPV. The values are represented as normalized values with respect to the standard, for which the control was taken, and presented as a mean ± standard deviation (SD with an n= 4). The level of statistical significance was established at a p-value below 0.05, which is reflected in those conditions that meet it. peptides were selected based on the prediction of the in silico tool for their inflammatory activity, this seems to corroborate it, more than in the hydrolysate, where although they will also be present, they will not be the majority or the only peptides interacting, and other sequences may interfere or interact with them, reducing their activity in this aspect. Molecular docking The four candidate peptides were subjected to the molecular docking study, and their interactions with the receptors of the two genes were observed, for which the difference in transcriptional activity in the different treatments, TLR4 and TREM2, was also quantified, and whose relative expression plots are represented here in Figures 3 and 4, respectively. For each receptor-peptide interaction, negative affinity values were obtained in Kcal/mol, in the case of TLR4 they vary between −1.6 and −10.9 kcal/mol; and in the case of TREM2 between −7.4 and −12 kcal/mol. The fact that the eight affinity values of the four peptides with the two genes result in negative affinities means that the binding between peptide and receptor is energetically favoured, so it can occur spontaneously when both structures come into contact (Kastritis and Bonvin, 2013), and therefore these peptides will be able to exert their immunomodulatory effect. Figures 3 and 4 also represent the 3D structures of the receptors, the peptides, their binding site and the interactions that are established between them when the interaction occurs. In Figures 5 and 6, the 2D structure of each of the four peptides is represented, and Journal of Insects as Food and Feed 0 (2024) 1–17 16 T. Gonzalez-de la Rosa et al. Haessner, P., Haessner, J. and McMurtrey, M., 2024. Trends and challenges in the food supply chain. Journal of Strategic Innovation and Sustainability 19: 115. https://doi.org/10 .33423/jsis.v19i1.6868 Hahn, T., Roth, A., Febel, E. and Zibek, S., 2018. New methods for high-accuracy insect chitin measurement. Journal of the Science of Food and Agriculture 98: 5069-5073. https://doi.org/10.1002/jsfa.9044 Hubbard, R.E. and Haider, M.K., 2010. Hydrogen bonds in proteins: role and strength. eLS. https://doi.org/10.1002 /9780470015902.a0003011.pub2 Iyer, S.S., Ghaffari, A.A. and Cheng, G., 2010. Lipopolysaccharide-mediated IL-10 transcriptional regulation requires sequential induction of type I IFNs and IL-27 in macrophages.The Journal of Immunology 185: 6599-6607. https://doi .org/10.4049/jimmunol.1002041 Kastritis, P.L. and Bonvin, A.M., 2013. On the binding affinity of macromolecular interactions: daring to ask why proteins interact. Journal of The Royal Society Interface 10: 20120835. https://doi.org/10.1098/rsif.2012.0835 Khatun, M.S., Hasan, M.M. and Kurata, H., 2019. PreAIP: computational prediction of anti-inflammatory peptides by integrating multiple complementary features. Frontiers in Genetics 10. https://doi.org/10.3389/fgene.2019.00129 Lear, S. and Cobb, S.L., 2016. Pep-Calc. com: a set of web utilities for the calculation of peptide and peptoid properties and automatic mass spectral peak assignment. Journal of Computer-Aided Molecular Design 30: 271-277. Mathur, D., Singh, S., Mehta, A., Agrawal, P. and Raghava, G.P., 2018. In silico approaches for predicting the half-life of natural and modified peptides in blood. PLOS ONE 13: e0196829. https://doi.org/10.1371/journal.pone.0196829 Mooney, C., Haslam, N.J., Pollastri, G. and Shields, D.C., 2012. Towards the improved discovery and design of functional peptides: common features of diverse classes permit generalized prediction of bioactivity. PLOS ONE 7: e45012. https://doi.org/10.1371/journal.pone.0045012 Mora, L., Gonzalez-Rogel, D., Heres, A. and Toldrà, F., 2020. Iberian dry-cured ham as a potential source of α-glucosi- dase-inhibitory peptides. Journal of Functional Foods 67: 103840. https://doi.org/10.1016/j.jff.2020.103840 Olsen, T.H., Yesiltas, B., Marin, F.I., Pertseva, M., García- Moreno, P.J., Gregersen, S., Overgaard, M.T., Jacobsen, C., Lund, O., Hansen, E.B. and Marcatili, P., 2020. AnOx- PePred: using deep learning for the prediction of antioxidative properties of peptides. Scientific Reports 10: 21471. https://doi.org/10.1038/s41598-020-78319-w Oonincx, D.G. and De Boer, I.J., 2012. Environmental impact of the production of mealworms as a protein source for humans–a life cycle assessment. PLOS ONE 7: e51145. https://doi.org/10.1371/journal.pone.0051145 Patel, K. and Mani, A., 2024. Food-derived peptides as promising neuroprotective agents: mechanism and therapeutic potential. Current Topics in Medicinal Chemistry 24: 1212-1229. https://doi.org/10.2174/01156802662892482403 22061723 Peighambardoust, S.H., Karami, Z., Pateiro, M. and Lorenzo, J.M., 2021. A review on health-promoting, biological, and functional aspects of bioactive peptides in food applications. Biomolecules 11: 631. https://doi.org/10.3390 /biom11050631 Picariello, G., Siano, F., Di Stasio, L., Mamone, G., Addeo, F. and Ferranti, P., 2023. Structural properties of food proteins underlying stability or susceptibility to human gastrointestinal digestion. Current Opinion in Food Science 50: 100992. https://doi.org/10.1016/j.cofs.2023.100992 Precup, G., Ververis, E., Azzollini, D., Rivero-Pino, F., Zakidou, P. and Germini, A., 2022. The safety assessment of insects and products thereof as novel foods in the European Union. In: Novel Foods and Edible Insects in the European Union, p. 123. https://doi.org/10.1007/978-3-031 -13494-4_7 Rivero-Pino, F., Millan-Linares, M.C. and Montserrat-de-la- Paz, S., 2023. Strengths and limitations of in silico tools to assess physicochemical properties, bioactivity, and bioavailability of food-derived peptides. Trends in Food Science and Technology 138: 433-440. https://doi.org/10 .1016/j.tifs.2023.06.023 Rivero-Pino, F., Gonzalez-de la Rosa, T. and Montserrat-de la Paz, S., 2024a. Edible insects as a source of biopeptides and their role in immunonutrition. Food and Function 5: 2789-2798. https://doi.org/10.1039/D3FO03901C Rivero-Pino, F., Ververis, E., Pérez-Gálvez, R. and Naska, A., 2024b. Biological properties and safety aspects of edible insects. In: García-Vaquero, M. and Carlos Álvarez García, C. (eds.) Insects as Food and Food Ingredients. Academic Press, Cambridge, MA, USA, pp. 173-190. https://doi.org/10 .1016/B978-0-323-95594-2.00006-9 Skibska, A. and Perlikowska, R., 2024. Natural plant materials as a source of neuroprotective peptides. Current Medicinal Chemistry 31: 5027-5045. https://doi.org/10.2174 /0929867331666230703145043 Stull, V.J. and Weir, T.L., 2023. Chitin and omega-3 fatty acids in edible insects have underexplored benefits for the gut microbiome and human health. Nature Food 4: 283-287. https://doi.org/10.1038/s43016-023-00728-7 Takahashi, K., Sugi, Y., Hosono, A. and Kaminogawa, S., 2009. Epigenetic regulation of TLR4 gene expression in intestinal epithelial cells for the maintenance of intestinal homeostasis. The Journal of Immunology 183: 6522-6529. https://doi.org/10.4049/jimmunol.0901271 Journal of Insects as Food and Feed 0 (2024) 1–17 Neuroprotective effect of mealworm protein hydrolysate 17 Tosefsky, K.N., Zhu, J., Wang, Y.N., Lam, J.S., Cammalleri, A. and Appel-Cresswell, S., 2024. The role of diet in Parkinson’s disease. Journal of Parkinson’s Disease 14: S21-S34. https://doi.org/10.3233/JPD-230264 Tourkochristou, E., Triantos, C. and Mouzaki, A., 2021. The influence of nutritional factors on immunological outcomes. Frontiers in Immunology 12: 665968. https://doi .org/10.3389/fimmu.2021.665968 Van Huis, A., Rumpold, B., Maya, C. and Roos, N., 2021. Nutritional qualities and enhancement of edible insects. Annual Review of Nutrition 41: 551-576. https://doi.org/10 .1146/annurev-nutr-041520-010856 Walsh, I., Seno, F., Tosatto, S.C. and Trovato, A., 2014. PASTA 2.0: an improved server for protein aggregation prediction. Nucleic Acids Research 42: W301-W307. https://doi.org/10 .1093/nar/gku399 Yesiltas, B., García-Moreno, P.J., Gregersen, S., Olsen, T.H., Jones, N.C., Hoffmann, S.V., Marcatili, P., Overgaard, M.T., Hansen, E.B. and Jacobsen, C., 2022. Antioxidant peptides derived from potato, seaweed, microbial and spinach proteins: Oxidative stability of 5% fish oil-in-water emulsions. Food Chemistry 385: 132699. https://doi.org/10.1016 /j.foodchem.2022.132699 Journal of Insects as Food and Feed 0 (2024) 1–17