The mechanistic effects of human digestion on magnesium oxide nanoparticles: implications for probiotics Lacticaseibacillus rhamnosus GG and Bifidobacterium bifidum VPI 1124

Abstract

The effects of nanoparticles (NPs) on the human gut microbiota are of high interest due to the link between the gut homeostasis and overall human health. The human intake of metal oxide NPs has increased due to its use in the food industry as food additives. Specifically, magnesium oxide nanoparticles (MgO-NPs) have been described as antimicrobial and antibiofilm. Therefore, in this work we investigated the effects of the food additive MgO-NPs, on the probiotic and commensal Gram-positive Lactobacillus rhamnosus GG and Bifidobacterium bifidum VPI 1124. The physicochemical characterization showed that food additive MgO is formed by nanoparticles (MgO-NPs) and after a simulated digestion, MgO-NPs partially dissociate into Mg²⁺. Moreover, nanoparticulate structures containing magnesium were found embedded in organic material. Exposures to MgO-NPs for 4 and 24 hours increased the bacterial viability of both L. rhamnosus and B. bifidum when in biofilms but not when as planktonic cells. High doses of MgO-NPs significantly stimulated the biofilm development of L. rhamnosus, but not B. bifidum. It is likely that the effects are primarily due to the presence of ionic Mg²⁺. Evidence from the NPs characterization indicate that interactions bacteria/NPs are unfavorable as both structures are negatively charged, which would create repulsive forces.

Graphical Abstract

Ingested dietary MgO-NPs could form different magnesium aggregates, that would reach the small intestine and interact with gut microbiota.

1. Introduction

Magnesium oxide (MgO) is an approved food additive (E 530) used in the preparation of several food products, including milk and dairy powders, canned peas, frozen desserts, and cacao products. MgO is also widely used, in the food industry, as an anti-caking agent and flow enhancer, and its concentration is not limited except for chocolate products, which may contain up to 7% MgO relative to its dry weight (1,2). Due to its basic properties, MgO is particularly effective in acid neutralization and has been used as an antiacid for relief of heartburn and dyspepsia (3). Moreover, chemical grade MgO is commonly used as a short-term osmotic laxative. Orally administered MgO reacts with the gastrointestinal fluids as follows: 2HCl + MgO → MgCl₂ + H₂O in the stomach, followed by MgCl₂ + 2NaHCO₃ → 2NaCl + Mg(HCO₃)₂ in the intestinal tract. Mg(HCO₃)₂ increases the osmotic pressure within the intestines, stimulating water exudation that softens and expands the contents of the intestines, stimulating defecation (4).

Previous studies have shown that food additives contain nano-sized components, specifically metal oxides such as titanium dioxide (TiO₂) and silica dioxide (SiO₂), E 171 and E 551, respectively (5,6). The consumption of food and products containing metal oxide nanoparticles (NPs) increases the opportunities for direct ingestion of NPs and their interaction with the gastrointestinal tract (GIT). Oral intake of NPs has been correlated to multiple detrimental effects in the human body, including inflammatory response, DNA damage, irritable bowel disease, and bacteria dysbiosis (7–9). This new relevant scientific evidence of detrimental effects has concerned governmental agencies (e.g. European Food and Safety Authority) and triggered re-evaluations on the safety assessment of TiO₂ and SiO₂ as food additives (10,11).

To date, the effects of magnesium oxide nanoparticles (MgO-NPs) on the human GIT are poorly understood. Under biological conditions, different sized MgO complexes are formed, some within the nanometer range (12). In vivo, oral intake of MgO-NPs resulted in significant DNA damage, biochemical alterations, and oxidative stress in female Wistar rats whilst also accumulating in the liver and kidney tissues apart from urine and feces (13) ¹³. Ghobadian and colleagues (2015) reported cellular apoptosis, reactive oxygen species and malformation of zebrafish embryos exposed to MgO-NPs (14). Gelli et al., (2015) highlighted the pulmonary toxicity that MgO-NPs produced to rats, and Mekky and coworkers (2020) revealed elevated serum alanine aminotransferase and aspartate aminotransferase levels suggesting potential toxic effects of nano Mg on the liver tissue samples of a rat convulsion model (15,16). In vitro, contradictory results have been reported about the toxicity of MgO-NPs. While some authors described that exposure of various mammalian cell lines to MgO-NPs led to cytotoxic effects, such as reduced cell viability, mitochondrial and lysosomal malfunctions, DNA damage, oxidative stress, and apoptosis; Mittag et al., (2019) indicated that MgO-NPs had no cytotoxic or genotoxic effects in intestinal HT29 cells and did not induce apoptotis, cell cycle changes, or oxidative stress (12,17–19). Recent studies investigated the potential applications of MgO-NPs as antimicrobial, antifungal, antibiofilm, and as biocidal for medical and agricultural purposes (20–22). MgO-NPs had been demonstrated to inhibit Gram-positive, Gram-negative, and Gram-positive endospore-forming bacteria (23–25). Moreover, Mg-ONPs (> 0.5 mg mL⁻¹) can diminish the adhesion and biofilm formation of infectious yeast (e.g. Candida albicans and Candida glabrata) and bacterial pathogens including Escherichia coli, Pseudomonas aeruginosa, and Staphylococcus aureus (22). This is of particular concern, as MgO-NPs in the human GIT derived from the consumption of food supplements (e.g. MgO tablets), or food-containing E 530 could have detrimental effects on the intestinal microbiota, leading to gut dysbiosis - altering microbial diversity, abundance and functionality - which has been associated with functional gastrointestinal disorders such as irritable bowel syndrome (IBS) (26).

In the present study, we investigated the physicochemical biotransformation of the food additive MgO in a simulated in vitro digestion model and its effects on two common commensal bacteria present in the human small intestine: Lactobacillus rhamnosus and Bifidobacterium bifidum. We extrapolated the human daily consumption of MgO to an average dose of 4.3×10⁻⁴ mg mL⁻¹ and considered 4.3×10⁻³ mg mL⁻¹ and 4.3×10⁻⁵ mg /mL⁻¹ as high and low MgO intake respectively. Planktonic and biofilm cultures of L. rhamnosus GG and B. bifidum VPI 1124 were exposed to in vitro digested MgO-NPs for a period of 4 and 24 hours and bacterial viability, adherence to surfaces and biofilm growth was quantified. Overall, our findings showed that when digested, the nanoparticulate food additive MgO partially dissociates into ionic Mg²⁺, diminishing the NPs size and concentrations. Although no detrimental effects were associated to MgO-NPs in general, different outcomes were detected based on the bacteria mode of growth (biofilms vs planktonic), bacterial species (Lactobacillus Vs Bifidobacterium), concentration, and exposure times.

It is worth to mention that little is known about the effects of food additive MgO-NPs on the human gastrointestinal tract and gut microbiota interface. The recently described roles that gut microbiota play in gastrointestinal function, and human health together with the increasing use of NPs in the food industry, makes this kind of studies of an urgent need. Here, we show for the first time that post in vitro digestion, the food additive MgO-NPs only partially dissociate in ions (51.2 % at intestine step), and that either (1) the size of MgO-NPs diminishes; or (2) dissociated magnesium re-precipitate forming nanoparticulate crystals of MgO or Mg(OH)₂, which could potentially arrive to the intestinal tract, bio-accumulate and affect the gut homeostasis. However, the evidence led us to believe that the beneficial effects of food additive MgO-NPs on bacterial growth and attachment are due to the presence of dissociated Mg²⁺ since the reduced presence of nanoparticulate structures and the negative charge of MgO-NPs diminishes the likelihood of bacteria/NPs interaction. Last but not least, we have demonstrated that the effects of in vitro digested food additive MgO-NPs may vary depending on the concentration, the exposure time, the bacterial mode of growth (planktonic Vs biofilms) and on the bacterial specie. Therefore, these findings highlight the importance of using physiologically relevant conditions when performing in vitro nanotoxicology as the physicochemical properties of the NPs can drastically change as well as their beneficial or detrimental effects.

2. Experimental

2.1. Nanoparticle Dose Calculation and Dispersion

The recommended dietary allowance is currently set at 420 mg/day for adult men and 320 mg/day for adult women (27). However, data from the National Health and Nutrition Examination Survey (NHANES) of 2013-1016 found that 48% of Americans of all ages ingest less magnesium from food and beverages than the estimated average required (EARs). Food additive MgO doses were calculated using the average MgO intake among users of dietary supplements (267 mg for women) (28). A total of 1.34×10⁻⁴ mg cm⁻² was established as the mean daily intake of MgO normalized to the surface area of the gastrointestinal tract (2x10⁶ cm²) (29). This value was shifted one order of magnitude higher and lower to mimic MgO exposure in low or deficient (1.34x10⁻⁵ mg cm⁻²), medium or standard (1.34x10⁻⁴ mg cm⁻²), and high or overdosed (1.34x10⁻³ mg cm⁻²) doses. According to our equipment dimensions, the low, medium, and high MgO doses were estimated to be 4.3×10⁻⁵ mg/mL, 4.3×10⁻⁴ mg mL⁻¹, and 4.3×10⁻³ mg mL⁻¹, respectively.

Food additive grade MgO-NPs were purchased from Spectrum Chemical MFG (CORP., Gardena, CA) and were prepared and dispersed using a protocol based on those established by the Organization for Economic Co-operation and Development (OECD) and the National Institute of Standards and Technology (NIST) (30,31). Briefly, in a 20 mL scintillation vial, 10 mg of the desired NP powder was dissolved in 10 mL of sterile 18 MΩ DI water and sonicated for 2 min at 10% amplitude and continuous mode. The probe sonicator equipped with a disruptor horn of ½” diameter (BRANSON Sonifier^® SFX550, Emerson Electric Co) was fully immersed in the NP suspension without touching the scintillation vial. To avoid heating the samples, the scintillation vial containing 1 mg mL⁻¹ of each NP was placed in ice bath while sonicating. The disruptor horn was sterilized before and after NP preparation by sonicating a solution of 50% ethanol for 5 min.

2.2. In Vitro Digestion of Nanoparticles

Nanoparticles (NPs) were submitted to an in vitro digestion closely mimicking the physiology of a human digestion process. This protocol was adapted from Glahn et al., (1998) and Moreno-Olivas et al., (2019) (32,33). Briefly, after sonication MgO-NPs were mixed with 10 mL of NaCl (140 Mm) + KCl (5 mM) solution (pH=2) and adjusted to pH 2 mimicking the stomach environment. Following, 0.5 mL of pepsin solution (0.57 mM) was added to the sample. The digestion bolus was then incubated on a rocker for 1 hour at 37°C. After incubation, the pH was adjusted to 5.5-6 using NaHCO₃ (1 M). Then, pancreatin (13.3 mM )- bile (8.27 mM) solution was prepared using NaHCO₃ (0.1 M) and 2.5 mL were added to the mix adjusting the pH 7. The final volume of the digestion bolus containing MgO-NPs was adjusted to 40 mL with NaCl (140 mM)+ KCl (5mM) (pH=6.7). The in vitro digestion was performed with sterile reagents under sterile conditions. All reagents are from Sigma Aldrich (Sant Louis, MO, USA).

2.3. Characterization of food additive MgO

In order to determine physicochemical changes of the food additive MgO during the sample preparation, the MgO NPs were characterized when dissolved in: (i) 18 MΩ sterile DI water; (ii) BHI medium containing 0.5% of Dextrose, 0.05% of L-Cysteine and 0.1% bacteriological agar; (iii) in vitro digestion bolus-1 (Stomach step at pH 2); and (iv) in vitro digestion bolus-2 (Intestine step at pH 7). Transmission electron microscopy (TEM) was then used to measure the primary particle diameter and morphology of particulate MgO. Briefly, a drop (μL) of 0.1 mg mL⁻¹ of MgO suspension in 18 MΩ sterile DI water was dispensed on the top of a 400-mesh copper TEM grid (Ted Pella, Inc.) and allowed to dry. TEM images of random fields of view were acquired using a JEOL JEM-2100F (JEOL, Peabody, MA) and processed with Image J software to measure the diameter of approximately 100 particles. The hydrodynamic size was measured using both the Zetsizer Nano ZS90 using dynamic light scattering (DLS) and Nanosight (Malvern Panalytical Ltd) which uses the Nanoparticle Tracking Analysis software (NTA). Both DLS and NTA utilize the properties of both light scattering and Brownian motion to obtain the NP size distribution of samples in liquid suspension. The zeta potential was evaluated by laser doppler electrophoresis (LDE) using a Malvern Zetasizer Nano ZS90 (Malvern Panalytical Ltd). Sample (0.1 mg/mL) measurements were performed in Malvern disposable polycarbonate folded capillary cells with gold plated beryllium–copper electrodes (DTS1070), which were rinsed with 18 MΩ sterile DI water to clean any dust contamination before sample filling. The refractive index used for MgO was 1.73. The samples were equilibrated in the instrument chamber for 120 seconds and measured at 25°C. Three independent experiments were analyzed (n = 3).

2.4. Inductively Coupled Plasma Mass Spectrometry (ICP-MS)

ICP-MS was used to quantify the amount of ionic Mg²⁺ released into the medium from food additive MgO when diluted in: (i) 18 MΩ sterile DI water; (ii) BHI medium containing 0.5% of Dextrose, 0.05% of L-Cysteine and 0.1% of bacteriological agar; (iii) in vitro digestion bolus-1 (Stomach step at pH 2); and (iv) in vitro digestion bolus-2 (Intestine step at pH 7). For this, MgO was diluted as above described at the highest concentration used in this study (4.3×10⁻³ mg mL⁻¹). To separate the ions from the NPs, samples were centrifuged at 10,600 x g, for 10 minutes (Eppendorf Centrifuge 5417 R, with a rotor F-45-30-11, Brinkmann Instruments, Inc, Westbury, NY). The supernatant was then analyzed by ICP-MS.

Briefly, 100 μL of each sample was pre-digested in borosilicate glass tubes with 3 mL of a concentrated ultra-pure nitric acid and perchloric acid mixture (60:40 v/v) for 16 h at room temperature. Samples were then placed in a digestion block (Martin Machine, Ivesdale, IL, USA) and heated incrementally over 4 h to a temperature of 120°C with refluxing. This step is intended to completely dissolve analytes and to decompose solids while avoiding loss of the sample and contamination. After incubation at 120°C for 2 h, 5 mL of concentrated ultra-pure nitric acid was subsequently added to each sample before raising the digestion block temperature to 145°C for an additional 2 hours. The temperature of the digestion block was then raised to 190°C and maintained for at least 10 minutes before samples were allowed to cool at room temperature. In this case, the use of nitric acid is common for metal dissolution and stabilization. Digested samples were re-suspended in 20 mL of ultrapure water prior to analysis using ICP-MS (Agilent ICP-MS 7500 Series, Agilent Technologies, Santa Clara, Ca, USA) with quality control standards (High Purity Standards, Charleston, SC, USA) following every 10 samples. Yttrium purchased from High Purity Standards (10M67-1) was used as an internal standard. To ensure batch-to-batch accuracy and to correct for matrix inference, all samples were digested and measured with 0.5 μg/mL of Yttrium (final concentration).

2.5. Bacterial Strains and Growth conditions

Major bacterial species/genus that inhabit the human small intestine consist of Bifidobacterium, Bacteroides, Lactobacillus, Clostridium, Fusobacterium and Enterobacteria (34). In this study, the two commensal, well-characterized, human derived probiotic Gram-positive strains Lactobacillus rhamnosus GG and Bifidobacterium bifidum VPI 1124 were used. L. rhamnosus and B. bifidum overnight cultures were grown on test tubes containing complete brain heart infusion media (BHI), consisting of BHI (Becton, Dickinson and Company, Franklin Lakes, NJ) supplemented with 0.5% Dextrose, 0.05% L-Cysteine and 0.1% Bacteriological Agar (Becton, Dickinson and Company, Franklin Lakes, NJ) for 48 h. L. rhamnosus was cultured at 37 °C with 5% CO₂, and B. bifidum was cultures at 37 °C anaerobically using BD GasPack^™ EZ anaerobe pouch systems (Becton, Dickinson and Company, Franklin Lakes, NJ).

2.6. Bacterial Viability Quantification

Bacterial viability was quantified for planktonic and biofilm cultures. For planktonic cells, overnight cultures were diluted to 10³ CFU mL⁻¹ (previously established to estimate the concentration of bacteria in the upper small intestine) and placed on 24-well plates (Costar^® Corning Incorporated, USA) using BHI (Becton, Dickinson and Company, Franklin Lakes, NJ). Bacterial cells were subsequently exposed to low (4.3×10⁻⁵ mg mL⁻¹), medium (4.3×10⁻⁴ mg mL⁻¹) and high (4.3×10⁻³ mg mL⁻¹) concentrations of in vitro digested MgO, previously diluted in complete BHI, for 4 h and 24 h. The negative control was the in vitro digestion bolus with no NPs and diluted in complete BHI. Following exposure, samples were homogenized, collected, serially diluted, and drop plated on the appropriate agar medium. L. rhamnosus was plated onto MRS (Becton, Dickinson and Company, Franklin Lakes, NJ) and B. bifidum onto Bifidobacterium agar plates (HIMEDIA^® Laboratories Pvt. Ltd.). Colonies were allowed to develop for a period of 48 h at 37 °C in 5% CO₂ for L. rhamnosus and at 37 °C anaerobically for B. bifidum. For biofilms, both overnight cultures were diluted to 10³ CFU mL⁻¹, wells of 24-well plates were then inoculated. Biofilms were allowed to form for a period of 4 days before exposure to NPs. To achieve similar MgO/bacteria ratios to the planktonic experiments, 4 days biofilms were disrupted and re-inoculated at approximately 10³ CFU mL⁻¹. Then bacterial exposure and quantification was carried out as previously stated. At least two independent biological experiments were performed using triplicate technical replicates per NP concentration exposure.

2.7. Bacterial Attachment and Biofilm Formation Assay

To study the effects of food additive MgO on the ability of L. rhamnosus and B. bifidum to form biofilms, both bacterial attachment and biofilm development were assessed. For the initial attachment assay, both L. rhamnosus and B. bifidum were transferred to clear 96-well plates (Costar^® Corning Incorporated, USA) at a concentration of 10³ CFU mL⁻¹ and exposed to low (4.3×10⁻⁵ mg mL⁻¹), medium (4.3×10⁻⁴ mg mL⁻¹), and high (4.3×10⁻³ mg mL⁻¹) doses of MgO. The negative control was the in vitro digestion bolus with no NPs and diluted in complete BHI. Then, the attached bacteria were fixed with 4% paraformaldehyde (PFA) (Sigma Aldrich, Sant Louis, MO, USA) at room temperature for 20 min and detected and quantified after 4, 6, 12 and 24 hours by staining each well with 0.1 % of crystal violet (Sigma Aldrich, Sant Louis, MO, USA) for 15 min. The crystal violet was then removed, rinsed (4x) with deionized water, to eliminate the excess of crystal violet, and left to dry. Finally, 95% ethanol was added to each well to resuspend the attached biomass, and plates were incubated for 15 minutes on a Roto Mix (Type 508000, Thermolyne, DE) at 200 rpm for 15 min. A plate reader (Synergy2, BioTek^® Instruments, Inc) was used to measure the plates at 570 nm absorbance (35). Similarly, since our previous studies suggested that both L. rhamnosus and B. bifidum can form biofilms that reach a steady state within a couple of days, and can be maintained for up to 6 days, the overall biomass during development were quantified using 0.1% crystal violet but at time points of 48, 72, 96, and 120 hours. Biofilm development was quantified every 24 hours because of the low growth rate of both L. rhamnosus and B. bifidum.

2.8. Immunocytochemistry and Confocal Microscopy

The confocal laser scanning microscope (CLSM) (LSM 880, Carl Zeiss Microscopy, LLC) equipped with an oil immersion objective (PL APO 63x/1.40 oil DIC, Carl Zeiss Microscopy, LLC) was used to visualize changes in biofilm morphology and colocalize potential NPs/bacteria interactions. Since metal oxide NPs present reflectance properties, they are easily detected using the laser 488 nm and the reflectance option of the confocal microscope (marked in bright green color by the ZEN lite software, Carl Zeiss Microscopy). The DNA of L. rhamnosus and B. bifidum were stained with DRAQ5^™ (Thermo Fisher Scientific, USA), which allowed to visualize the bacteria in red. For further differentiation, L. rhamnosus was tagged in red, while B. bifidum was tagged in violet, artificially – using the color palette. For confocal imaging, only biofilms of L. rhamnosus and B. bifidum were exposed to low, medium, and high dose of food additive MgO (4.3x10⁻³ mg mL⁻¹) in 24-well black plates with transparent bottom. Biofilms were cultured for 4 days, scraped, and resuspended at an approximate concentration of 10³ CFU mL⁻¹ and then mixed with the NPs, which enabled standardization of cells and NPs. Following exposure to NPs, the 24-well plates were centrifuged at 4000 rpm for 5 min so that the cells clusters and the NPs would precipitate on the bottom, and to enable their visualization. Then, the samples were fixed with 4% PFA (Sigma Aldrich, Sant Louis, MO, USA) in 1x PBS for 20-30 min at room temperature (RT). After fixation, bacterial DNA was stained with DRAQ5^™ (1:1000). The staining solution was incubated in the dark for 20 min at RT. Finally, the 24-well plate was left to dry. Samples were cured overnight in the dark and imaged the next day. Z Stacks image series were taken with 1 μm intervals. Overall, 4 images were taken per sample for further analysis. Overall biofilm biomass was analyzed using the ImageJ plugin COMSTAT2 (www.comstat2.dk).

2.9. Ionic control calculations and experimental design

Magnesium chloride (MgCl₂) (Sigma Aldrich, Sant Louis, MO, USA) was used as an ionic control since is highly soluble in water, forming Mg²⁺. To apply equal or similar amounts of ionic Mg, the moles of Mg from MgO NP used in the above-mentioned experiments were extrapolated to the same moles of Mg in the ionic control MgCl₂. The calculations were as indicated in equation 1.

The lowest concentration of MgONPs is 4.3 x 10⁻⁴ mg mL⁻¹ or 4.3 x 10⁻⁸ g mL⁻¹;

$$\begin{gathered} \frac{4.3x{10}^{-8}gMgO}{1}x\frac{1molMgO}{40.3044gMgO}x\frac{1molMg}{1molMgO}=1.06x{10}^{-9}moleMg \\ \frac{95.211g{\mathit{MgCl}}_2}{1mol{\mathit{MgCl}}_2}x\frac{1.06x{10}^{-9}molMg}{1}x\frac{1mol{\mathit{MgCl}}_2}{1molMg}=\mathbf{1}.\mathbf{016}x{\mathbf{10}}^{\mathbf{-}\mathbf{7}}\mathit{g}{\mathit{MgCl}}_{\mathbf{2}} \end{gathered}$$ Equation 1

Therefore, high (1.016×10⁻² mg mL⁻¹), medium (1.016×10⁻³ mg mL⁻¹) and low (1.016×10⁻⁴ mg mL⁻¹) concentrations of MgCl₂ were used to investigate the effects of dissolved Mg on the biofilm attachment and formation. For experiments with MgCl₂, samples were processes exactly as the NPs in section 2.2, sonicated, digested and serial diluted into the respective concentrations, and the bacterial exposures were performed as indicated in section 2.7. The negative control was the in vitro digestion bolus with no MgCl₂ and diluted in complete BHI.

2.10. Statistical Analysis

All assays were performed in 3 independent experiments, with at least 3 replicates per treatment. Results are expressed as mean ± standard error. Before statistical analysis, data was submitted to the two most recommended normality tests to determine whether the values come from a Gaussian distribution. For that purpose, both (i) D’Agostino-Pearson omnibus normality test, and (ii) Shapiro-Wilk normality test was determined using GraphPad Prism version 9. After passing the normality test, a one-way ANOVA with Tukey’s or Dunnet’s post-test was used to compare differences between means. Data was analyzed with GraphPad Prism version 9.00 for Windows (GraphPad Software, San Diego California USA, http://graphpad.com). Differences between means were considered significant at P<0.05.

3. Results

3.1. Characterization of food additive MgO in different media suspensions

The food additive MgO was dispersed following OECD guidelines and subsequently submitted to an in vitro digestion - simulating the human digestion process with changes in the pH and addition of gastric enzymes. The digested MgO was then diluted to physiologically relevant concentrations in BHI before determining its effect on bacteria. Hence, food additive MgO was exposed to mechanical (peristaltic movements simulated with a rocket) and chemical (by the addition of salts, acids, and digestive enzymes) interactions that can drastically influence its physicochemical characteristics and reactivity, and consequently its interactions with live organisms, organs, tissues, and molecular structures. To understand those potential changes, we fully characterized MgO at every stage of the sample process before exposing the commensal bacteria. When analyzing the morphology of the MgO captured with TEM (Fig. 1A) we found that the food additive MgO formed crystalline structures in the nanoparticle (NP) range with no defined shape or size forming aggregates and/or agglomerates. Using ImageJ, the primary particles that could be distinguished from the aggregated clusters were measured and plotted in a frequency histogram (Fig. 1B). The primary particle size of MgO-NPs was found to follow a Gaussian distribution with sizes ranging from 20 to 130 nm. The average particle diameter of MgO-NPs was around 65 nm. However, when in aqueous suspension, MgO-NPs size tended to increase, mostly due to aggregation or agglomeration events (Fig. 1D). The hydrodynamic distribution of MgO-NPs in water also followed a narrowed Gaussian distribution ranging from 100 to 450 nm (Fig. 1C). Both the Nanosight and Nano Zetasizer were used to analyze the hydrodynamic size (d_H) of MgO-NPs in aqueous suspensions. As shown in Figure 1D, although both pieces of equipment use dynamic light scattering to detect and measure NPs, a large variation in MgO-NPs d_H were detected for all the treatments. While Nanosight detected aggregates and/or agglomerates of MgO-NPs of approximately 208.24, 328.60, 162.36, and 193.80 nm when suspended in water, BHI+0.1% Agar, stomach and intestine digesta respectively; the Nano Zetasizer detected higher values (d_H-Water = 985.14 nm; d_H-BHI = 12473.68 nm; d_H-Stomach = 3831 nm; and d_H-Intestine = 274.9 nm). The polydispersity index (PdI) was also measured with the Nano Zetasizer. PdI is a dimensionless value of the broadness of size distribution calculated from the cumulants analysis. Values range from 0 to 1, being <0.05 very monodisperse; <0.08 nearly monodisperse; 0.08-0.7 mid-range polydisperse; and >0.7 very polydisperse. Thus, our measurements indicate that the size range of MgO-NPs suspended in water (0.434), BHI+0.01% agar (0.56), stomach (0.993) and intestine (0.675) digesta is widespread and extremely heterogeneous, particularly in the stomach. Moreover, the NPs’ zeta potential (ζ) (mV) and electrophoretic mobility (μm·cm/V·s) were analyzed using laser doppler electrophoresis technique (LDE), where the stability and mobility of a particle suspended in liquid is evaluated under an applied electric field. MgO-NPs were mid-range stable in water (ζ = −17.36 mV) and highly stable in intestine digesta (ζ = −41.2 mV), considering ±30 mV the reference value for a good NP stability in aqueous systems. However, in stomach digesta MgO-NPs were poorly stable (ζ = −2.19 mV) most likely because of the low pH (pH=2). Zeta potential measurements in more complex media such as BHI+0.1% agar (ζ = −7881.1) will be always difficult to interpret because of the viscosity of the sample. Finally, MgO-NPs seems to maintain a negatively charged particle surface through all the process independently of the media complexity and pH (Fig. 1D).

Figure 1. MgO characterization.

(A) Transmission electron microscopy (TEM) representative image of MgO-NPs in its dry form (red asterisk indicates a lattice fringe effect from a MgO NP). (B) Frequency histogram of the primary particle diameter (nm) of MgO-NPs in its dry state. Measures were performed with ImageJ software from TEM images. (C) Hydrodynamic size (nm) distribution of hydrated MgO-NPs measured by NTA. (D) Table summarizing the average of the primary particle diameter (nm), hydrodynamic size (nm) measured with Nanosight and Nano Zetasizer, polydispersity index (PdI), zeta potential values (mV) and electrophoretic mobility (μm·cm V·s⁻¹) of MgO-NPs suspended in different media. Data consists of mean±SEM.

3.2. Effects of in vitro digestion on food additive MgO

MgO is conventionally considered insoluble in water and stable at high temperatures. However, metal oxide nanoparticles are known to dissociate into ions when immersed in different physiologically relevant solutions (17,36). Therefore, we aimed to deeply understand the physicochemical changes of the food additive MgO, its potential bioavailability, and its behavior along the simulated digestion process. First, after confirming the presence of nano-crystalline structures of MgO by TEM in Fig. 1, we measured the ionic release of Mg²⁺ from the food additive MgO-NPs in the various media suspensions. Suspensions of MgO-NPs at 4.3×10⁻³ mg mL⁻¹ in water, BHI, stomach and intestine digesta were ultracentrifuged (10,600 xg for 10 min) to separate the putative Mg²⁺ ions (supernatant) from MgO-NPs (pellet) and the supernatants were analyzed by ICP-MS. The obtained results (Fig. 2G) indicate that MgO-NPs partly dissociated into Mg²⁺ in all the suspensions analyzed, being significantly influenced by the stomach digestion step (65.8%), followed by BHI (59.9%), intestine digestion step (51.2%), and water (39.9%). This dissociation seemed to be reversible by the pH of the intestine digestion step (51.2%) since the content of Mg²⁺ in the intestinal supernatant was significantly lower than in the stomach digestion (Fig. 2G). Moreover, the concentration of particles found in the stomach (Fig. 2E) and intestine (Fig. 2F) digestion were lower than those quantified in water (Fig. 1C). Pellets of digested MgO were also analyzed by TEM to determine whether MgO-NPs were present (Fig. 2A and C). Despite the high complexity of both samples and the high content of organic material, crystalline structures (yellow arrows), potentially from nanoparticulate MgO, were detected by the lattice fringe effect (diffracted wave from a crystal). To confirm this, EDS analysis of the area was performed, and elemental Magnesium and Oxygen were detected in both the stomach (Fig. 2B) and the intestine digestion (Fig. 2D). EDS analysis did not detect Mg in controls of stomach digestion or intestine digestion lacking MgO-NPs (Fig. S2). The lattice fringe effect was also observed when using TEM with NP of food additive MgO (Fig. 1A -indicated with a red asterisk). These findings suggest that along the human digestion, food additive MgO could partially dissociate into ions, but the remaining NP diminish its size to ≤ 10 nm, further explaining the low size peaks (marked with a black asterisk) found below 100 nm when analyzing digested MgO with the Nanosight (Fig. 2E and F).

Figure 2. Effects of the in vitro digestion on MgO.

(A and A’) Transmission electron microscopy (TEM) images of stomach digested MgO (yellow arrows indicates a lattice fringe effect from MgO-NPs embedded in organic material). (C and C’) TEM images of fully digested MgO (yellow arrows indicates a lattice fringe effect from MgO-NPs embedded in inorganic material). (B and D) EDS spectra showing the presence of magnesium (Mg) and oxygen (O) in stomach and fully digested samples, respectively. (E and F) Hydrodynamic size (nm) distribution of digested MgO in stomach and intestine, respectively, measured by NTA. (G) Net content of ionic magnesium (Mg⁺) detected in the supernatant of food additive MgO samples analyzed by inductive coupled plasma – mass spectrometry. Data shown as mean±SEM.

Because metal oxide nanoparticles have the capability of refracting polarized light, we also analyzed the food additive MgO in different suspension media using confocal laser scanning microscopy (Fig. 3). To increase the likelihood to observe MgO-NPs once we found that the NP size diminishes (by TEM), 0.5 mL of food additive MgO at a concentration of 0.25 mg mL⁻¹ (stock solution) were dispensed onto a microscope slide and mounted with a coverslip. To visualize MgO-NPs we made use of a 63x objective and 488 laser. MgO-NPs suspended in water were detected (Fig. 3A.2) and observed (bright green) to form large clusters of particulate MgO aggregating and/or agglomerating (Fig. 3A.3). Blank samples (negative control) did not contain any refracting NPs (Fig. 3A.1). MgO-NPs were also detected in bacterial media BHI+0.1% agar (Fig. 3B.2 and B.3) forming NP aggregates, albeit the presence of background noise, in all the samples analyzed, most likely due to the presence of bacteriological agar (Fig. 3B.1). However, the intensity of MgO-NPs was above background noise and clearly distinguished. Similar results were obtained when analyzing the negative controls of stomach and intestine digesta (Fig. 3C.1 and D.1). The organic material from both stomach and intestine digesta was detected by the microscope as background noise. However, aggregates of MgO-NPs were once again detected in a higher signal (brighter green) (Fig. 3C.2 and D.2). In agreement with the TEM and ICP-MS data, aggregates of MgO-NPs were found to be smaller when immersed in stomach conditions, suggesting that low pH triggers the MgO-NPs dissociation (Fig. 3C.2 and C.3). Interestingly, fully digested MgO-NPs (intestine step) where detected (bright green) embedded into some background noise (red asterisk and red box), which could be the organic material present in digesta.

Figure 3. MgO Nanoparticles detected by confocal laser scanning microscopy.

Samples of MgO resuspended in water, bacteriological media (BHI + 0.1% agar), partial (stomach step), and fully digested (intestine step) were observed under confocal microscopy, as metallic NPs (seen in bright green) reflect the polarized light. (A.1, B.1, C.1 and D.1) Blanks of each medium. (A.2, B.2, C.2 and D.2) 2D images of MgO resuspended in water, BHI, and partial and fully digested. (A.3, B.3, C.3 and D.3) 3D images of MgO resuspended in water, BHI, and partial and fully digested. Red asterisks indicate MgO-NPs embedded in organic material and are zoomed in the red box. Scale bar represents 50 μm.

3.3. Effects of food additive MgO on bacterial viability

In this study, both L. rhamnosus and B. bifidum were exposed to digested food additive MgO for 4 hours to mimic the time for one meal to enter and exit the small intestine. It takes 3 to 5 hours from entry in the duodenum (first section of the small intestine) to exit from the ileum (last segment of the small intestine) (37). In addition, a longer exposure time (24 hours) was chosen to study the effects of putative bio-persistent MgO entrapped within the GI tract mucosa. Moreover, the effects of food additive MgO were evaluated for planktonic and biofilm derived cells. We found that exposures to low, medium, and high concentrations of in vitro digested food additive MgO did not affect the viability of planktonic cells of L. rhamnosus and B. bifidum after 4 hours (Fig. 4A and C) or 24 hours (Fig. 4B and D). However, when exposing biofilm derived cells of L. rhamnosus and B. bifidum to in vitro digested food additive MgO, significant differences in cell viability were quantified (Fig. 5). High and medium standard human intake concentrations of MgO significantly increased the viability of L. rhamnosus biofilm derived cells after 4 hours (Fig. 5A) and 24 hours (Fig. 5B). Similarly, viability of B. bifidum derived biofilm cells also increased significantly upon exposure to MgO at 4 hours (Fig. 5C) and 24 hours (Fig. 5D), albeit being in all concentrations used. Is worth to mention that (i) the initial cell density of planktonic and biofilm derived cells were standardized to 10³ CFU mL⁻¹ to enable an equal NP/bacteria ratio, and (ii) to ensure that the digestion solution by itself did not affect the cell viability (data not shown).

Figure 4. Viable counts of planktonic derived cells.

Planktonic cells of L. rhamnosus (A and B) and B. bifidum (C and D) were exposed to low (4.3×10⁻⁵ mg mL⁻¹), medium (4.3×10⁻⁴ mg mL⁻¹) and high (4.3×10⁻³ mg mL⁻¹) concentrations of MgO and viability was quantified after 4 (A and C) and 24 h (B and D). Data represented as mean±SEM and analyzed using one-way ANOVA with a Tukey’s post-hoc test. (*) p<0.05; (**) p<0.001; and (***) p<0.0001.

Figure 5. Viable counts of biofilm derived cells.

Biofilms cells of L. rhamnosus (A and B) and B. bifidum (C and D) were exposed to low (4.3×10⁻⁵ mg mL⁻¹), medium (4.3×10⁻⁴ mg mL⁻¹)and high (4.3×10⁻³ mg mL⁻¹) concentrations of MgO and viability was quantified after 4 (A and C) and 24 h (B and D). Data represented as mean±SEM and analyzed using one-way ANOVA with a Tukey post-hoc test. (*) p<0.05; (**) p<0.001; and (***) p<0.0001.

3.4. Effects of food additive MgO on bacterial initial attachment and biofilm development

Probiotic bacteria are known to permanently colonize the gut mucosa as biofilms, acting as a prevention barrier against pathogen colonization (38). Together the intestinal epithelial barrier, bacterial biofilms, and their secretions (polysaccharides, secreted proteins, and extracellular DNAs) provide protection against exogenous threats. Therefore, determining the bacterial attachment and biofilm development during exposure to food additive MgO, is of high importance to predict potential consequences in the health of the human GI tract.

Using the standardized crystal violet assay, we determined the attachment ability of both commensal bacteria as well as monitored the biofilm development for 5 days. The initial attachment for each species varied. L. rhamnosus showed attachment between 6 and 12 hours following the bacterial inoculum (Fig. 6A), while B. bifidum needed 24- and 48-hours following inoculation (Fig. 6B). Exposure of L. rhamnosus to the food additive MgO did not result in changes of bacterial attachment in the first 12 hours of biofilm development, however from 24 hours onwards high concentrations of MgO significantly stimulated biofilm development of L. rhamnosus (Fig. 6A). In contrast, B. bifidum biofilms (Fig. 6B) exposed to low and medium concentrations of the food additive MgO already showed a significant decrease of the overall biomass at 48 hours, compared to control. However, higher concentrations of food additive MgO significantly increased the biofilm biomass of B. bifidum after 48 hours. This trend shifted drastically after 72 hours onwards, where low and medium concentrations maintained similar levels of biofilm biomass than control, and high concentrations of food additive MgO significantly decreased the B. bifidum biomass (Fig. 6B).

Figure 6. Fold increase of the overall biomass attached after exposures to food additive MgO.

The initial attachment and biofilm development of (A) L. rhamnosus and (B) B. bifidum were determined using crystal violet after exposing both bacterial strains to low (4.3×10⁻⁵ mg mL⁻¹), medium (4.3×10⁻⁴ mg /mL⁻¹) and high (4.3×10⁻³ mg /mL⁻¹) concentrations of digested MgO up to 5 days. Data is shown as mean±SD and analyzed using Two-way ANOVA with Dunnett’s post-test. (*) p<0.05; (**) p<0.01; (***) p<0.001; (****) p<0.0001.

3.5. Evaluation of biofilm colonies by confocal microscopy

Since both L. rhamnosus and B. bifidum biofilm derived cells were more influenced by the food additive MgO than their planktonic counterparts, confocal microscopy was once again used to detect and characterize any morphological and structural changes, as well as, to visualize biofilms/MgO-NPs interactions. Given that this study worked with physiologically relevant doses of food additive MgO, which are relatively low compared with other NP-related studies, and that MgO partially dissociate into ions, the likelihood of encountering biofilms/MgO-NPs interacting is extremely low. However, we found some as examples in Fig. 7, image A.3 (pointed with a white asterisk), where an aggregate of MgO-NPs settled on a biofilm microcolony. Further examples of interactions biofilm/MgO-NPs can be observed in the supplemental information (Fig. S3). However, no differences were detected for biofilm/MgO-NPs interactions throughout the exposure time and the bacterial type.

Figure 7. Confocal images of biofilms exposed to food additive MgO.

To assess changes in biofilm morphology and interactions biofilms/MgO-NPs, bacterial DNA was stained with Draq5 and confocal laser scanning microscopy (Carl Zeiss LSM880) was used to record Z-stacks. Ortho images of L. rhamnosus biofilms (red) after 4 hours (A.1-A.4) and 24 hours (B.1-B.4), and B. bifidum biofilms (violet) after 4 hours (C.1-C.4) and 24 hours (D.1-D.4) of exposure to low (4.3×10⁻⁵ mg mL⁻¹), medium (4.3×10⁻⁴ mg mL⁻¹), and high (4.3×10⁻³ mg mL⁻¹) concentrations of MgO (green/yellow). White asterisks represent putative biofilm/MgO-NPs interactions. Scale bars represent 50 μm.

As anticipated, exposure to the food additive MgO resulted in an increase of the size of L. rhamnosus biofilm microcolonies (Fig. 7) from 4 hours to 24 hours, independently of the MgO concentration (Fig. 7A and 7B). Microcolonies of L. rhamnosus were composed of higher bacterial cell numbers, which were compacted in the center of the well, with long bacillus filaments at the edges however, no morphological differences were detected in any of the treatments. Comparing to L. rhamnosus, biofilm development of B. bifidum was slower with smaller and more compacted biofilm microcolonies (Fig. 7D), further confirming the viability data (Fig. 5). Exposure of B. bifidum to high concentrations of food additive MgO resulted in larger microcolonies (Fig. 7C4 and D4). Comstat analysis reinforced the trend observed in the confocal images, where the overall biomass of either L. rhamnosus and B. bifidum biofilms after 24 hours of medium and high MgO-NPs exposures was significantly higher than the control group (Fig. S4).

4. Discussion

Research from the past two decades suggest that the gastrointestinal (GI) microbiota may be linked to the homeostasis of the intestinal tract and the overall health of the human host (39). When impaired, common inflammatory and metabolic disorders including Crohn’s disease, ulcerative colitis; and malnutrition, type 2 diabetes, and obesity, respectively, are more likely to be developed (40,41). Environmental factors likely have a major impact on microbial dysbiosis; childbirth mode, air pollution, antibiotic use, diet, and urban environments have all previously been reported (42). Notably, the influence of diet and dietary or food-containing nanoparticles (NPs) on the intestinal microbiota and their advantageous or adverse effects are currently in the spotlight. Recent studies have reported gut microbial shifts and changes in species abundance after oral exposures to silver and titanium dioxide NPs, both used in a wide range of consumer products such as food packaging and industrial baked goods, respectively (43,44). Hence, due to the wide, and increasing use of NPs (organic and inorganic) in the food industry (food additives, food supplements or food packaging), a growing interest in studying the impacts of foodborne NPs on the intestinal microbiota has emerged. Moreover, chronic GI exposures to NPs including direct effects on microbiota and indirect effects due to NPs-mediated immune system dysfunctions, require further scrutiny. In this study, we aimed to understand the direct effects of nanoparticulate food additive MgO on the two well-studied human Gram-positive commensal bacteria L. rhamnosus GG and B. bifidum VPI 1124. Both are early colonizers of the GI tract on newborns and of high importance, as they play a crucial role in the human gut homeostasis by preventing and fighting pathogen infections (45,46).

The main route of entry of MgO to the human body is by oral ingestion where, once ingested, the MgO will be subjected to mechanical and chemical interactions along the human digestion process. MgO presents a crystalline cubic structure and is very stable at high temperatures, which has led to conventionally classify MgO as “stable” and “relatively inert” like other metal oxides (47). However, our previous study revealed that simulated human digestion has a strong impact on metal oxide NPs’ size and reactivity (36). Therefore, to realistically understand the impact of food additive MgO on commensal bacteria, we performed in vitro digestion of MgO and thereafter assessed its structural changes. Primary particles of the food additive MgO presented the conventional cubic structure with a mean size of 65 nm approximately, although MgO aggregated forming 200 nm particles when suspended in sterile MΩ water (Fig. 1C). This tendency is commonly seen when sonicating metal oxide suspension as their surface atoms get excited and react with neighbor NPs forming ionic bounds (36). Nevertheless, the ICP-MS results showed that MgO partially dissociated in ionic Mg²⁺ when immersed in water, bacterial medium (completed BHI+0.01% agar), and when partially (stomach) or fully (intestine) in vitro digested (Fig. 2G). These results are in agreement with those by Wetteland and co-workers (2018), who demonstrated that nano-sized MgO and Mg(OH)₂ dissociated significantly in biologically relevant substances (DMEM, HEPES and Simulated Body Fluid) (47). Other metal oxide NPs such as SiO₂ and ZnO have also been reported to partially (65.5 %) and fully (100 %) dissolve, respectively, after a simulated human digestion (48). Although biological fluids and environments are highly complex and dynamic, the impact of pH plays the most significant part in the dissociation of MgO. Other theories, like an actual nucleophilic substitution of O²⁻ or OH⁻ by Cl⁻ (from the digestion’s inorganic salts), are less contemplated because the energy of Mg=O bonds is greater than that of Mg–Cl, making nucleophilic substitution by chloride energetically unfavorable (47,49). When measuring the hydrodynamic size of MgO suspension by dynamic light scattering, different and misleading results were obtained using Nanosight and Nano Zetasizer. While Nano Zetasizer obtained measurements of particles above 1000 nm for water, BHI and stomach digesta, Nanosight detected particles ranging from 150 to 300 nm for all the treatments (Fig. 1D). It has been already discussed that DLS techniques are not well suited for complex media analysis as NPs cannot be distinguished from the other media-components such as macromolecules and proteins aggregation (SI Table 1). However, using TEM-EDX (Fig. 2) we were able to reveal the presence of crystal structures (emitting the lattice fringe effect) that contained Mg in both stomach (Fig. 2B) and intestine samples (Fig. 2D). These crystalline structures detected in the nanometer range are thought to be re-precipitates of MgO-NPs, precipitates of Mg(OH)₂, or a hybrid of MgO with Mg(OH)₂ at the particle surfaces, as magnesium could react with water (50). Contrarily, in a study where the effects of pH on MgO were studied in the absence of organic material such as gastric enzymes and inorganic salts, Schneider et al. (2021) did not detect any nanoparticulate structure by TEM after subjecting MgO-NPs to a simulated digestion (pH, transit time and temperature) (49). Although diverse dissolution-precipitation models for MgO have been proposed in pure water and validated at certain temperatures (JMAK model), there are several parameters that can affect the aquatic chemistry (e.g. media impurities, particle size and shape, internal porosity, water acidity, CO₂ content, presence of anions, etc.) and consequently interfere in the demonstration of an equilibrium model explaining the Mg re-precipitation (51). In our study, MgO samples were concentrated by ultracentrifugation to increase the chance of detecting NPs as sample volumes that can be analyzed by TEM are extremely small (10 to 20 μl). These findings were reinforced by confocal microscopy. As described in previous studies, this technique allows the detection of metal oxide NPs due to their capability to refract polarized light (using the 488 nm laser) being distinguished from other fluorescently stained structures (Fig. 3) (52). Interestingly, aggregates of NPs were seen embedded into big matrices of organic material (Fig. 3D), which could completely isolate NPs from the intestinal lumen-microbiota interface and nullify its potential reactivity.

The relationship between food additive MgO-NPs and intestinal microbiota has been scarcely studied. Most previous studies have focused on the antimicrobial and antibiofilm capacity of MgO-NPs. In general, MgO-NPs showed biocidal activity against Gram-positive and Gram-negative bacteria, bacterial spores, and viruses at relatively high particle concentration (0.1 to 1.5 mg mL⁻¹) (53). For example, MgO-NPs impaired biofilm formation of E. coli, Klebsiella pneumoniae and S. aureus at 0.25, 0.125 and 0.5 mg mL⁻¹, respectively (20). In this study, however, we have aimed to work with physiologically relevant concentrations of food additive MgO-NPs (4.3×10⁻⁴ mg mL⁻¹), which were extrapolated from the daily human consumption of magnesium (267 mg day⁻¹). Our results show that after short-term exposures (4 and 24 hours), in vitro digested food additive MgO-NPs generally do not affect the bacterial viability of both Gram-positives L. rhamnosus and B. bifidum when growing as planktonic cells (free floating cells) (Fig. 4). However, digested MgO-NPs significantly increased the viable counts of biofilm-forming L. rhamnosus and B. bifidum already at 4 h of exposure. These results suggest that; (i) food additive MgO-NPs could be distinctly assimilated based on the bacterial mode of growth (free floating cells Vs sessile biofilm communities); and/or (ii) food additive MgO-NPs could target biofilm-specific components such as the extracellular polymeric substances (EPS). These short-term benefits were also detected in the initial bacterial attachment (from 12 to 48 hours) of both bacterial strains (Fig. 6). Despite this, differences on biofilm development between strains started appearing after 72 hours of attached biomass monitoring. While significant increments on biofilm biomass were detected for L. rhamnosus, the biofilm development of B. bifidum seemed to be significantly compromised by high and medium food additive MgO-NPs after 72, 96 and 120 hours.

In general, three main antibacterial mechanisms of NPs have been discussed: (1) mechanical interaction and consequently damage to the bacterial cell wall because of electrostatic interactions, (2) oxidative stress derived from the presence of reactive oxygen species, and (3) malfunction of proteins and extracellular structures because of metal cations release (54). In the present work, we can clearly discard the first hypothesis as the food additive MgO-NPs were negatively charged in all the four media suspensions (Fig. 1D), which most likely would create electrostatic repulsions, as the bacterial cell wall and the biofilm EPS are structures also negatively charged. For the second hypothesis, Leung et al. (2014) previously demonstrated the absence of ROS production (measured by electron spin resonance), absence of oxidative stress, and absence of lipid peroxidation after exposing the Gram-negative bacteria E. coli to 1 and 0.1 mg mL⁻¹ MgO-NPs (positively charged). They found that the toxic effects of three different MgO-NPs towards E. coli resulted from cell membrane damage due to direct NP/cell wall interaction and attachment involving phosphate groups (50). Putative interaction between aggregates of food additive MgO-NPs and biofilms of L. rhamnosus and B. bifidum were detected by confocal microscopy (Fig. S3). Nevertheless, the use of physiologically relevant concentrations of MgO-NPs resulting in relatively low NPs concentration, when compared with other studies, diminishes the chance of mechanical interaction and consequently bacterial toxicity. On the other side, the release of metal cations (third hypothesis) such as Mg²⁺ from MgO NPs, could elucidate most of the results found along this investigation. Although metal cations such as Zn²⁺ and Cu²⁺ were seen interacting with sulfhydryl groups in enzymes, with amine and carboxyl groups on microbial cells, and causing mismetallation of proteins affecting cell metabolism ending in cell death, Mg²⁺ seemed to play a beneficial role when administrated in adequate amounts (55–58).

In bacteria, Mg²⁺ is the second-most abundant cation (59). The roles of Mg²⁺ in homeostasis, sensing and transport have been extensively investigated in Gram-negative bacteria, including Salmonella enterica and E. coli. Thus, Mg²⁺ was shown to act as cofactor in ATP-dependent phosphorylation in several enzymatic reactions, stabilize ribosomes and membranes, influence RNA folding and the nucleic acid-protein interactions among others (60). In Lactobacillus spp., divalent ions such as Mg²⁺ are also required as an essential cofactor for stimulating the activity of metalloenzymes such as aminopeptidases, dipeptidases, catalases, D-xylose isomerase, L-arabinose isomerase, and ribozymes. The addition of Mg²⁺ in growth media has been reported to increase growth of Lactobacillus bifermantans and L. rhamnosus FTCD 8313 (61). Similar effects were seen in cultures of B. bifidum sub-spp. pennsylvanicum, where Mg²⁺ deficiency markedly reduced bacterial growth and lowered the content of cellular lipid-galactose (62). These studies are in agreement with our bacterial viability results after short term exposures (Fig. 4 and 5).

In biofilm attachment and development, the role of divalent cations such as Mg²⁺ appear to be multi-faceted. The presence of magnesium is thought to assist in the bacterial initial attachment to biotic or abiotic surfaces by conditioning film formation, bridging between molecules, modifying cell surface adhesins and reducing the apparent surface charge of bacteria cell wall (58). In concordance, our results showed a significant increase on initial bacterial attachment when both commensal Gram-positives L. rhamnosus (12 to 24 hours) and B. bifidum (24 to 48 hours) where exposed to high concentrations of digested food additive MgO (Fig. 7). Although not statistically significant, a similar trend of increment on initial bacterial attachment was detected when exposing both L. rhamnosus and B. bifidum to ionic Mg²⁺ adding MgCl₂ to the cell culture medium (Fig. S5). Similarly, Hisano et al. (2014) described that exogenous Mg²⁺ (0-5 mM) resulted in a higher cell to cell aggregation of Aggregatibacter actinomycetemcomitans, which positively contributes to biofilm formation (63). Mangwani et al. (2014) also found that Ca²⁺ and Mg²⁺ (20 mM) significantly increased the average thickness, roughness coefficient and surface area of biomass of Pseudomonas mendocina biofilms (marine bacterium). Whereas Ca²⁺ was found to enhance the production of EPS, Mg²⁺ significantly increased the cell growth of P. mendocina in biofilms (64). Interestingly, exopolysaccharides extracted from Streptococcus mutants and L. rhamnosus cultures showed high binding affinity toward divalent cations (65). However, caution is needed when extrapolating the Mg²⁺ response of one species to another or one strain to the whole species, as it was observed in P. aeruginosa that 20 mM Mg²⁺ increased biofilm formation by strain PPF-1 but not by the other three strains tested (PAO1, LESB58 and Urg-7) (66). In our study, after 48 h the biofilm development of B. bifidum was significantly reduced when exposed to food additive MgO-NPs (Fig. 6), but not when exposed to MgCl₂ (Fig. S5B). This could be explained by a reduced bioavailability of ionic Mg²⁺ when exposing B. bifidum with the food additive MgO-NPs as some Mg might be re-precipitating after long exposures forming crystalline or nanoparticulate structures. Contrarily, the study of Larsen et al. (2007) on probiotic research, did not detect significant changes in the adhesion of Lactobacilli cells to IPEC-J2 cells in the presence of Mg²⁺ (67). Meanwhile, contradictory results were found in in vivo investigations. On the one hand, an in vivo study by García-Legorreta et al. (2020), where rats were fed with control (1000 mg kg⁻¹), low (60 mg kg⁻¹), and high (6000 mg kg⁻¹) magnesium content diets, showed that high dietary magnesium decreased the bacterial community diversity while low dietary magnesium did not modify diversity (68). On the other hand, Winther et al. (2015) concluded that diet deficient in magnesium modify bacterial diversity in mice contributing to the development of depressive-like behavior (69). In agreement, Pachikian et al. (2010) found that mice with magnesium deficient diets (70 mg kg⁻¹) had a lower gut bifidobacterial content (1.5 log reduction) as well as 36 to 50% lower mRNA content from components controlling the integrity of the gut barrier (zonula occludens-1, occluding, proglucagon) (70). These contradictory results reinforce the above-mentioned idea that extreme oscillations (excess or deficiency) on magnesium contents in the gut is critical for the gut homeostasis and microbiota populations, essentially commensal bacteria.

5. Conclusions

The effects of in vitro digested food additive MgO-NPs on two human-derived commensal bacteria were investigated. The NPs characterization showed that food additive MgO-NPs partially dissociate in ionic Mg due to the low pH in the stomach step of a simulated in vitro digestion and partially remained in nanoparticulate form. Although further experiments are needed to demonstrate the Mg re-precipitation or the formation of hybrid MgO-Mg(OH)₂ NPs (e.g. particulate sampling to sample by size fractions, ion chromatography to measure saturation rates, TEM or SEM coupled to EDX and RAMAN spectroscopy to provide information about composition, structure and crystallinity of the sample), we hypothesize that some magnesium could also re-precipitate forming crystalline structures (most likely as MgO and Mg(OH)₂) in the range of nm as detected by TEM and confocal microscopy. Exposures to in vitro digested food additive MgO-NPs increased the bacterial viability of both L. rhamnosus and B. bifidum when growing as biofilms community but not when in planktonic form. Exposures to food additive MgO-NPs affected differently the biofilm development of L. rhamnosus and B. bifidum. While high concentrations of in vitro digested food additive MgO-NPs significantly increased the biofilm development of L. rhamnosus up to 120 hours, the biofilm development of B. bifidum was seen benefited only after 48 hours of exposure. Following the electrostatic laws, we believe that the negative charge of the remaining nanoparticulate fraction would create electrostatic repulsion with the bacterial cell wall impeding their interaction and consequently impeding detrimental effects. Therefore, we suggest that the ionic fraction of the in vitro digested food additive MgO - formed by divalent cations of magnesium (Mg²⁺) -, plays the major role in the detected results of this study. However, further assays could help to reinforce this hypothesis such as the use of magnesium single-element ion cation and non-digested MgONPs in minimum bacterial growth media. Moreover, future investigations are needed to further investigate whether the nanoparticulate Mg detected by TEM and confocal after the in vitro digestion can bio-accumulate in the GIT, for example after long-term exposures, and consequently affect the gut homeostasis.

List of abbreviations

BHI

Brain Heart Infusion

DMEM

Dulbecco’s Modified Eagles Medium

DLS

Diffractive Light Scattering

EDS

Energy-dispersive X-ray Spectroscopy

EPS

Extracellular Polymeric Substances

GIT

Gastrointestinal Tract

HEPES

Hydroxyethyl Piperazineethanesulfonic Acid

ICP-MS

Inductively Coupled Plasma – Mass Spectrometry

LDE

Laser Doppler Electrophoresis

MgO

Magnesium Oxide

NPs

Nanoparticles

NTA

Nanoparticle Tracking Analysis

OECD

Organisation for Economic Co-operation and Development

TEM

Transmission Electron Microscopy

Availability of data and materials.

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.