Cytotoxicity of Insects (T. Molitor, G. Bimaculatis, S. Gregaria and A. Domesticus) in A Simulated Gastrointestinal Digestion/Human Intestinal Cell Model

Insects as an alternative nutritional source for protein and other nutrients has gained increasing attention in the last decade. On the other hand, insects produce several compounds toxic to human cells associated with e.g. antitumoral activity and as such has been proposed as tentative anti-cancer agents. In this work, we used a combined simulated gastrointestinal digestion and human absorptive cell model, Caco-2, common in nutritional studies but also in cancer research, to investigate cytotoxicity of common insect species in a simulated oral intake setting. Caco-2 cells were exposed to various concentrations of digested insects (Gryllus bimaculatus, Schistocerca gregaria, Tenebrio molitor, Acheta domesticus) which showed a signicant decrease in survival and viability in response to the digests. T. molitor affected the viability of Caco-2 cells the most, compared to the other species. Notable effects on protein expression of several calcium homeostasis and cell response related proteins, FOSB, BANK1 and EGFL7, a metabolic homeostasis oxidative stress response related protein, FOXO1, and an apoptosis related protein, H0-1/HM0X/HSP32 was identied. We conclude that these insect species were cytotoxic to the intestinal cells in a simulated oral intake setting.


Introduction
Human nutrition and insects. Global micronutrient de ciencies are among the most prevalent de ciencies within malnutrition challenges and seen disproportionately in women of reproductive age (WRA), pregnant women, and children in developing countries [1]. Suboptimal levels in any of the primary micronutrients: iron, zinc, vitamin A, folate, vitamin D, and iodine, are associated with an increased risk of the following consequences: maternal mortality, maternal morbidity, low birth weights, and subpar health and cognitive development of children [2]. Globally, upwards of 40% of pregnant women and more than 290 million children under the age of 5 are anaemic [3][4][5][6]. Insects have been suggested as a novel solution to anaemia. Edible insects have been documented to contain high iron levels and have thus surfaced a potential alternative food choice [7,8]. Overall insects are good sources of energy, protein, fat, minerals, and vitamins within the human diet [9]. The energy provided by insects has been established to be on par with other fresh meat sources per fresh weight [9]. Mean estimates show energy levels to be around 400-500 kcal per 100 g of dry insect matter, making it a competitive food option when presented next to other protein sources. From a micronutrient perspective, levels of the key micronutrients: iron, zinc, vitamin A, folate, vitamin D, and iodine vary depending on insect species, however, folate, vitamin D, iodine are generally not present in notable amount. Iron, speci cally, has been shown to range from 18 to 1562 mg/100g dry insect matter across species, with the low end of the spectrum encompassed by ants, the mid-level being termites and the highest levels found in crickets [10]. For this reason, suggestions have been made to use insects documented to be high in iron as an alternative food source to help reduce the prevalence of iron de ciency anaemia.
Insects as potential cancer therapy. Insects produce a large number of species-dependent bioactive substances needed for survival, and for this, insects have attracted interest from the pharmacology research eld and several investigations have shown harmful effects to human cancer cells of different origin e.g. anti-proliferative compounds extracted from the Texas grasshopper were shown to inhibit growth of human leukemic leucocytes [11] and insect (Hydrillodes repugnalis) tea was shown to inhibit viability of human hepatoma (HepG2) cells [12]. Yet another study, looking at both HepG2 and Caco-2 (undifferentiated cells, 4 days post-seeding) observed decreased viability after incubation with T.molitor (mealworm larvae) [13]. Also, in vivo studies in mice have shown that insect alleferons have antitumoral activity [14]. These are only a few examples from a large body of studies on insects and insect-derived compounds for the potential use in cancer therapy.
The present study. The aim of the present study was to examine toxicity of insects to human intestinal cells in a simulated gastrointestinal setting using a standard model for nutrient absorption. As such, the insects went through a simulated gastrointestinal digestion procedure before they were further digested and absorbed by human intestinal Caco-2 cells. The digests were incubated with undifferentiated cells (colonic characteristics) as most cytotoxicity studies are done in undifferentiated cells, but also with fully differentiated Caco-2 epithelia (duodenal characteristics) as appropriate for in vitro studies on intestinal uptake/absorption of nutrients to also cover the effects of gastrointestinal processing of the nutrient before it meets the absorptive cells. We investigated four popular insect species Gryllus bimaculatus (black cricket), Tenebrio molitor (yellow mealworm), Schistocerca gregaria (desert locust), and Acheta domesticus (house cricket). We also used two species from an alternative source (in the Netherlands) to verify that the observed cytotoxicity was not connected to the primary source (in the UK). We undertook viability and cell survival experiments at different concentrations of the insect digests complemented with catalase activity testing and protein arrays to gain mechanistic insight.

Materials And Methods
Materials. Unless otherwise stated, all reagents and chemicals were purchased from Sigma-Aldrich. αamylase (A3176, ≥ 10 units/mg solid), porcine pepsin (P-7125, ≥ 400 units/mg protein), porcine pancreatin (P-1750) and porcine bile extract (B-8631) were stored as directed until use. A Milli-Q plus system (Merck Millipore, Darmstadt, Germany) was used to purify water to a minimum resistance of 18.2 MΩ cm. All other chemicals, such as solvents used, were of analytical grade. Solutions of enzymes were prepared fresh before each use. The rst round of insects was sourced from a farm in the United Kingdom. Insects were received fresh, immediately frozen to kill, then oven dried at 80℃ for 36 hours, prior to grinding in a coffee grinder. All were stored at -20℃ when not in use. The second round of insects were purchased from the online platform Bug Bazaar.
Simulated gastrointestinal digestion. 1 g of each insect sample was weighed out for digestion. To each sample, 5ml of α-amylase solution (75 units/ml in MQ) was added and sample agitated on a rotating plate (200 RPM) for two minutes. Following this 5ml of MQ water was added and pH adjusted to 2 using HCl (6M). During peptic digestion, 0.3ml of pepsin solution (0.15 g/ml) was added and incubated at 37℃ for 60 mins on rotating plate (70 RPM), followed by pH to 7 with NaOH (5M) to stop peptic digestion. Afterward, 1.7 ml of pancreatin-bile solution (24 mg/ml bile extract and 4 mg/ml pancreatin) was added to start pancreatin-bile digestion and samples incubated at 37℃ for 60 mins on rotating plate (70 RPM).
Following digestion, sample volumes were brought up to 15ml using MQ water and stored at -20℃ until use. Samples were not heat inactivated to avoid degradation of the insect material which would not be true to physiological reality i.e. digestion does not involve a point at which material reaches above body temperature. Trials using only non-heat-inactivated in vitro digestion uids at varying concentrations showed these had no signi cant impact on cell survival or viability over 24 hours at levels utilised in this trial (supplemental data).
Cell survival measures. Cells were plated in 24-well plates (90,000 cells/well, Corning CellBIND) and cultured for 24 hours or 14 days, dependent on if the experiment was with differentiated or nondifferentiated cells. Cell medium was changed 24 hours prior to the start of the trial. The cells were incubated with digest dilutions for 24 hours, following which digests and medium were aspirated and cells washed twice with DPBS. Cells were lysed in 100 µl cold RIPA buffer with added PPI, collected and stored at -80℃ until analysis. Digest dilutions with which cells were treated, were made by mixing digests with cell medium already present on the cells. In order to approximate actual insect concentrations represented by these dilutions (mg/ml) on the cells, it was assumed that 100% of the insect was digested in the 15 ml of digestion uid. Given this, it was assumed that base digest prior to dilution contained 66.67 mg of insect per ml of digestion uid. It is acknowledged that 100% digestion of the insect samples is an overestimate, as evidenced by visible precipitate in the digests. Total cell density (surface attached) as a proxy for cell survival after treatment with the various insect digests and controls were estimated by measuring the total protein content per well by using the Pierce BCA Protein Assay Kit [33].
Cell viability estimated by the MTT assay. The cells were cultured, maintained, and seeded as stated above in 2.3. and 2.5; cells were seeded in 96-well plates (10,000 cells/well, Corning CellBIND) and cultured for 24 hours. Cell viability was measured based on metabolic reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide as measured by MTT Cell Growth Kit (MTT, Merck Millipore, CT02). Dimethyl sulfoxide (DMSO) was used to dissolve MTT crystals in place of acidi ed isopropanol, as it was ineffective.
Expression of stress-related protein markers. The protein levels of apoptosis-related proteins were identi ed using the Proteome Pro ler™ antibody arrays (Human Apoptosis Array Kit; R&D Systems, MN, USA), which estimates the protein levels of 35 different apoptosis-related human proteins. Pooled samples of cell lysates were loaded on the arrays at 15 µg protein/membrane. Membranes with controls, digest controls, and insect digests were processed and developed simultaneously in the same run for each trial. The chemiluminescent signal was developed using a Chemidoc XRS+ (Bio-Rad Laboratories Inc., Hercules, USA). Samples from cells exposed to tolerable and cytotoxic levels of T. molitor and G. bimaculatus were additionally sent to OLink Proteomics (Uppsala, Sweden) for analysis with their organdamage panel which analyses 92 protein biomarkers relevant to processes involved in biological response to organ damage.
Veri cation trials. Veri cation trials were done to account for potential impact of processing or source of insects utilised. The rst veri cation utilised insect samples heated to 105℃ for 60 minutes in noncon uent cells. The second veri cation examined freeze dried insects, T. molitor and Locusta migratoria, which were explicitly marketed as human food from a farm in the Netherlands in non-con uent cells.
Based on continually updating research these samples were run through a slightly modi ed simulated gastrointestinal digestion as follows. 1 g of each insect was weighed out. To each sample 4 ml of salt solution (140 mM NaCl and 5 mM KCl) pre-heated to 37°C was added, following which the pH was adjusted to 7 using 1 M NaHCO 3 . Sample volume adjusted to 6.5 ml and 500 µl of the amylase stock solution (1050 U/ml; Sigma-Aldrich A3176-1MU, 10 U/mg) was added to each sample and then incubated at 37°C with shaking (165 rpm) for 4 mins. The pH was immediately lowered to 2 using 1 M HCl, and volume of each sample adjusted to 9.5 ml with salt solution. Then 500 µl of the pepsin stock solution (40000 U/ml; Sigma-Aldrich P7012-5G, 2529 U/mg) was added to each sample and they were incubated at 37°C with shaking (165 rpm) for 60 mins. Subsequently the pH was raised to 5.5 using 1 M NaHCO 3 and 2 ml of bile solution (10.635 mg/ml; Sigma-Aldrich B8631-100G, 1.68 U/mg) and 0.5 ml of pancreatin solution (7 mg/ml; Sigma-Aldrich P7545, 2.6U/mg) were added. The pH was adjusted to 7 using 1 M NaHCO 3 , total sample volume adjusted to 15 ml using salt solution and incubated at 37°C with shaking (165 rpm) for 60 mins. Digests were stored at -20°C until use. Cells were treated with concentrations previously identi ed as tolerable or cytotoxic and dilution was done in blank digestion uids to ensure all cells were equally exposed to digestion enzymes. with Student's t-test, equal variances not assumed, two-tailed and signi cant differences were considered at p < 0.05.

Results
Cell survival decreased at increasing concentrations of insects. Cell survival, based on total protein from live cells, was signi cantly reduced by all insect digests. The dilution point at which cell survival was signi cantly impacted varied between species and cells treated with A. domesticus digests showed signi cantly reduced survival at all concentrations as presented in Figure 1, while cells treated with G. bimaculatus and S. gregaria digests all had signi cantly reduced survival rates only at dilutions at and below 0.167 mg/ml. The cells were most resilient to T. molitor digests with signi cant reductions only seen at dilutions levels at and below 0.667 mg/ml.

Decreasing cell viability with increasing concentration of insect digests. Cell viability tests demonstrated
that all insect digests signi cantly reduced cell viability. Average cell viability as measured by the MTT assay, is presented as percentage of control (untreated cells in medium) in Figure 2. Cells treated with T. molitor digests showed signi cantly reduced viability at and below dilutions of 0.667 mg/ml, consistent with the cell survival data in Figure 1. However, the decrease in cell survival of the other three species was evident at slightly lower concentrations of insect digests than the measurable effect on cellular viability. A. domesticus digests only signi cantly reduce cell viability at and below dilutions of 1.333 mg/ml, while G. bimaculatus and S. gregaria digests only have a signi cant reduction at and below dilutions of 3.333 mg/ml.
Stress-related protein markers. Changes in protein expression of one apoptosis related protein with notable presence were identi ed. HO-1/HMOX-1/HSP32 protein levels consistently showed marked change when analysed with Proteome Pro lerÔ arrays (Human apoptosis array kit, R&D systems). Array results are plotted in graphs to illustrate this in Figure 3a and 3b, but due to limited sample availability (n=1, pooled triplicate), the array results could not be used for statistical analysis. The data suggested that there were no marked changes in the expression of common apoptosis-related proteins such as procaspase-3, cleaved caspase-3, bad, and bax (data not shown). HO-1/HMOX-1/HSP32 or heme oxygenase 1, also known as heat shock protein 32, is a stress-related cytoprotective molecule vital in the protection against oxidative injury. Notable changes in several proteins related to cellular calcium and metabolic homeostasis were identi ed in a protein panel analysis (n=3; Olink proteomics, Uppsala, Sweden), Figure   3c and 3d. Cells displayed the largest change in FOSB expression for both insect species and also followed the same trends of an initial increase in TNNI3, followed by a decline. FOSB and TNNI3 are responsible for cellular responses to calcium ions and calcium ion homeostasis, respectively. Similarly, EGFL7 controls calcium binding and BANK1 is involved in calcium mobilisation from intracellular storage. While FOXO1 regulates metabolic homeostasis in response to oxidative stress; it is the main regulator of redox balance [15].
Catalase activity in Caco-2 cells was increased. Catalase activity was measured as a secondary marker for oxidative stress response in the cells [16,17]. The catalase activity was analysed in cells exposed to insect samples at the dilution point, 1:20 (3.335 mg/ml at which cell viability or survival was signi cantly reduced across all tested insect species. The catalase activity in cells exposed to insect digests was elevated, with some variability between species. G. bimaculatus and S. gregaria caused the highest increase (116% increase (p = .002) and 177% respectively, (p =.000)) and T. molitor and A. domesticus a slightly smaller increase (43% (p = .001) and 63% (p =.000) respectively) compared to digestion uid control, Figure 4. This distinct increase in catalase activity further suggests that the cells were responding to an oxidative stressor [18][19][20].
Veri cation trials in another insect source -cell survival. Veri cation trials utilising insect samples from an alternative source (the Netherlands) con rmed previous ndings that cell survival, based on total protein from live cells, was signi cantly reduced by all insect digests at the 3.333 mg/ml dilution level as presented in Figure 5. For T. molitor signi cant reductions were again seen at dilutions levels at and below 0.667 mg/ml. Trials utilising insects which were heated to 105℃ for 60 minutes, Figure 6, prior to digestion and treating the cells con rmed similar trends, although more variation in response was observed suggesting processing temperature may play a role in toxicity of the insects to cells.

Discussion
Oxidative stress as a possible cause for cell cytotoxicity. While these analyses do not elucidate the exact mechanism by which the insect digests are killing the cells, they do point towards an oxidative stress reaction. The protein panel analysis supports the hypothesis of oxidative stress related apoptosis. Additionally, based on the increased expression of the calcium and metabolic homeostasis related proteins it is plausible that the cells are exhibiting calcium-mediated mitochondrial apoptosis induced through oxidative stress. Past evidence has demonstrated oxidative stress, particularly reactive oxygen species, act upstream in this apoptotic pathway [21,22]. Increased expression of FOSB has been observed in cancer cells exposed to an accumulation of reactive oxygen species with pro-apoptotic outcomes [23].
Upregulation of EGFL7 has also been documented to have a cytoprotective effect in cells when faced with oxidative stress [24]; thus the increase seen here is potentially indicative of the cells responding to oxidative stress.
Based on previous literature, a possible explanation might be linked to catechol moieties in the insects.
Catechol moieties cross-link chitin to protein in insect exoskeletons and are vital to cuticle sclerotization during development [25]. These catechol moieties can also act as precursors for certain quinonoids, particularly o-quinone, which are derived via the action of phenoloxidases, key components of insect immune systems and responses [25,26]. Quinones are cytotoxic through several mechanisms, one of which is the induction of oxidative stress through reduction to a semiquinone radical which reduces oxygen to superoxide radicals which then reform the quinone -a futile redox cycle which increases levels of hydrogen peroxide [27]. The suggested increase of HSP32 protein levels in the present study could be an indicator of increased hydrogen peroxide levels in the cells. Quinones are documented within insect cellular immune responses, in addition to hydrogen peroxide already being present as a defence mechanism produced by insect phagocytes [28]. Further, research has demonstrated that some quinones, depending on molecular structure, decompose above 100℃ which may account for the more varied response seen in Figure 6, when cells were treated with insects heated above 100℃ [29].
Although there is an increasing number of studies on insects and insect-derived molecules in cancer cell and mice models, very little is known about the effects of insects and their defence proteins on healthy human cells or healthy mice. A literature search on insects and primary human cells did not reveal any such studies. Since there is a thin line between cancerous and non-cancerous cells, we need to also be investigating how the cytotoxic effects observed in cancer cells relate to healthy cells. Figure 1 Cell survival, estimated by total protein from live cells, expressed as a percent of untreated controls. The cells were treated with varying dilutions of insect digests (data are means ±SE, n=4    Stress-related protein expression of Caco-2 cells exposed to insect digests at varying concentrations. (a)

Figure 4
Catalase activity (U/ml) normalised to total protein (mg/ml) in human intestinal Caco-2 cells treated for 24 hours with insect digests at dilution 3.335 mg/ml (means ±SE, n = 2). An asterisk (*) indicates a signi cant difference (p<0.05) from the digest control. There was no signi cant difference between untreated cells and cells treated with blank digestion uids (p = 0.85).

Figure 5
Cell survival, estimated by total protein from live cells, expressed as a percent of the untreated control of insects from the Netherlands. The cells were treated with varying dilutions of insect digests (data are means ± max/min, n=2). T-tests done using technical cell well replicates, n=6. φ indicates signi cant difference of T. molitor from control, * indicates signi cant difference from control for all insects (p < 0.05).

Figure 6
Cell survival, estimated by total protein from live cells, expressed as a percent of the untreated control of insects heated to 105℃ for 60 minutes. The cells were treated with varying dilutions of insect digests (data are means ± max/min, n=2). T-tests done using technical cell well replicates, n=6. φ indicates signi cant difference of T. molitor from control, Ω indicates signi cant difference of G. bimaculatus from control, ¥ indicates signi cant difference of S. gregaria from control, * indicates signi cant difference from control for all insects (p < 0.05).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.