Impact of Pristine Graphene on Intestinal Microbiota Assessed Using a

Research Article www.acsami.org

Cite This: ACS Appl. Mater. Interfaces 2019, 11, 25708−25719

Impact of Pristine Graphene on Intestinal Microbiota Assessed Using a Bioreactor-Rotary Cell Culture System Mohamed H. Lahiani, Kuppan Gokulan, Katherine Williams, and Sangeeta Khare* Division of Microbiology, National Center for Toxicological Research, U.S. Food and Drug Administration, 3900 NCTR Rd, Jefferson, Arkansas 72079, United States

Downloaded via UNIV OF SOUTHERN INDIANA on July 26, 2019 at 03:04:58 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: The increased use of graphene in consumer products such as food contact materials requires a thorough understanding of its effects on the gastrointestinal commensal bacterial population. During the first phase of study, three representative commensal bacterial species (L. acidophilus, B. longum, and E. coli) were exposed to different concentrations (1, 10, and 100 μg/mL) of pristine graphene for 3, 6, and 24 h in the Bioreactor Rotary Cell Culture System (BRCCS) which allowed a continuous interaction of intestinal microbiota with the pristine graphene without precipitation of test material. The results showed that pristine graphene had dosedependent effects on the growth of selective bacteria. To study the interaction of graphene with more diverse consortia of intestinal microbiota, fresh fecal samples from laboratory rats were used. Rat fecal slurry (3%) was maintained in an anaerobic environment and treated with different concentrations (1, 10, and 100 μg/mL) of pristine graphene for 3, 6, and 24 h. Counts of viable aerobic and anaerobic bacteria were assessed and fecal slurries were also collected for microbial population shift analysis using quantitative real-time PCR, as well as 16s rRNA sequencing. The results showed a significant two-fold increase in both aerobic and anaerobic bacterial counts (expressed as colony forming unit; CFU) during the first 3 h of exposure to all pristine graphene concentrations. However, 24 h of continuous exposure resulted in a 120% decrease in the CFU of aerobic bacteria at the highest concentration and the anaerobic bacteria CFU remained unchanged. Multivariate analysis of the q-PCR data showed that the exposure time, as well as the graphene concentrations, impacted the bacterial population abundance. Community analysis of graphene-treated fecal samples by 16S sequencing revealed significant alteration of 15 taxonomic groups, including 9 species. The increased abundance of butyrateproducing bacteria (Clostridium f imetarium, Clostridium hylemona, and Sutterella wadsworthensis) was correlated with an increase of the short-chain fatty acid, butyric acid after exposure to graphene. These results clearly indicate that graphene may cause adverse effects on the intestinal microbiome at the doses equal to 100 μg/mL. Further experiments using ex vivo intestinal explants (nonanimal model) could reveal the mechanisms by which graphene could perturb the microbe-host intestinal mucosa homeostasis. KEYWORDS: graphene, carbon-based nanomaterial, intestinal bacteria, microbiome, toxicity, next-generation sequencing, short chain fatty acids (SCFA)

1. INTRODUCTION The use of nanoparticles (NPs) in commercial products, agriculture, food safety, biomedical advances, and pharmaceutical applications has become the subject of extensive research in recent years.1−4 Carbon-based nanomaterials, including graphene, are a unique type of NPs with high strength, relative biocompatibility, and modifiable functionality.5−7 Graphene, a one-atom-thick monolayer composed of Sp2-hybridized carbon atoms and its functionally altered derivatives, have emerged as a new class of nanomaterials with exceptional abilities. For example, graphene has demonstrated potential for use in horticulture as a fertilizer and enhancer of seed germination.8−12 Moreover, due to its food-protection properties, graphene-doped polymers have been proposed for use in food packaging. For example, a nanocomposite of polystyrene© 2019 American Chemical Society

graphene synthesized by Compton et al. was shown to decrease oxygen permeation and light transmission.13 Similarly, other graphene-derived materials exhibited high thermal stability,14 moisture protection, protection against UV light,14 and heat resistance.15 All of these properties, including antimicrobial effects,16 make graphene a plausible candidate for food packaging applications to increase the shelf life of perishable food. However, there is a potential for the migration of these NPs into the food matrices which poses regulatory concerns.17 Received: May 1, 2019 Accepted: July 1, 2019 Published: July 1, 2019 25708

DOI: 10.1021/acsami.9b07635 ACS Appl. Mater. Interfaces 2019, 11, 25708−25719

Research Article

ACS Applied Materials & Interfaces

Figure 1. Schematic of the experimental design for the test of pristine graphene impact on pure bacterial cultures and rat fecal samples using a dynamic approach.

functional similarity to the human counterpart in terms of activation of biological pathways that are known to be dependent on intestinal homeostasis due to microbiome-host interaction.30 The study design is shown in Figure 1.

The introduction of graphene into consumer products and its impact on human health remains of substantial concern to the regulatory and research communities.18−20 Human exposure to graphene can occur directly via inhalation, intentional ingestion or injection, and skin contact or indirectly via food contaminated with NPs or as released from food packaging. Indeed, all of these routes can lead to the introduction of graphene into the digestive system21,22 where it can interact with the resident bacterial population. Any alteration of the bacterial microbiota can impact gastrointestinal homeostasis by altering the efficacy of the digestive system, triggering the immune system, and/or decreasing the ability to fight pathogens.23,24 Despite the early work on understanding graphene-bacteria interaction using in vitro and in vivo models, toxicological studies have shown contradictory results regarding the safety of graphene.25 While some groups have reported that graphene is bactericidal to numerous organisms,26−28 others have shown that the conditions of the cell culture and the type of starting graphene materials are the leading factors in the effects observed.25 In fact, some studies have shown that the precipitation of the graphene into the culture media may affect the antimicrobial effect and lead to bacterial proliferation and biofilm formation.27−29 Hence, it is important to develop and utilize an in vitro model that mimics the continuous interaction of graphene with bacteria that would exist in the human gastrointestinal system. The aim of this research was to study the interaction of pristine graphene with bacteria in a dynamic system using in vitro models. In particular, the first objective was to study the survival of three model commensal bacteria growing in a rotary culture system after exposure to pristine graphene. The second objective was to investigate the shifts of intestinal bacterial populations after exposure to pristine graphene using three different methodologies: Real-time PCR analysis of major phyla, families, and genera; V3−V4 based 16S sequencing of the rat gut microbial population; and quantification of short chain fatty acids (SCFA) using HPLC. In the second study, fresh fecal samples from laboratory rats were used considering

2. RESULTS AND DISCUSSION 2.1. Growth and Survival of Pure Bacterial Culture after Exposure to Pristine Graphene. The initial assessment of pristine graphene impact on bacterial cultures was performed under static conditions (in growth culture medium without shaking), in which a bacterial culture was spiked with different concentrations of graphene suspensions in culture media and optical density was measured at various time intervals. The results indicated that L. acidophilus growth was promoted at certain graphene concentrations (Figure 2A). For example, while the L. acidophilus bacterial control reached log phase at 14 h, the bacteria treated with 10 μg/mL of pristine graphene reached the same point in only 10 h. Moreover, the same bacteria treated with 1 or 10 μg/mL graphene reached a higher optical density overall (stationary phase; Figure 2A). The other two studied bacteria (B. longum and E. coli) showed no significant differences between the control and graphenetreated bacteria at 1 or 10ug/mL. (Figure 2B,C). The exposure of B. longum to graphene at 100 μg/mL showed no signal of growth of the bacteria (Figure 2B). Moreover, the exposure of E. coli to graphene at the same concentration (100 μg/mL) delayed bacteria from reaching the stationary phase by over 14 h (Figure 2C). To further study whether a dynamic experimental system could change the response of bacteria toward pristine graphene, we repeated the experiment using a rotary culture system. We assessed the bacterial survival at three time points (2, 4, and 24 h). Figure 3 shows the colony forming units (CFU) expressed as a percentage change compared to control samples (considered as 100%). The in vitro dynamic system allowed a continuous interaction of the graphene with the bacterial cultures and no sign of graphene deposition was observed even at the 25709


Research Article


Figure 2. Growth curves of L. acidophilus (A), B. longum (B), and E. coli (C) after exposure to graphene (Gr) at 1, 10, and 100 μg/mL. Control samples were spiked with ultrapure water only. Error bars showing standard error values. ¥, p < 0.05 of graphene 1 μg/mL compared to control. *, p < 0.05 graphene 10 μg/mL compared to control, and t, p < 0.05 graphene 100 μg/mL compared to control.

Figure 3. Live bacterial count of L. acidophilus (A), B. longum (B), and E. coli (C) after exposure to graphene (Gr) at 1, 10, and 100 μg/mL at 3 time points (2, 4, and 24 h) grown in rotary culture system. Error bars represent standard error values (n = 4). *, p < 0.05 compared to control at 3 h. ¥, p < 0.05 compared to control at 6 h. t, p < 0.05 compared to control at 24 h.

highest concentration of 100 μg/mL (Figure S1). This interaction resulted in a different pattern of growth/survival for the three model bacteria. The CFUs of L. acidophilus increased 5 times in the 1 and 10 μg/mL graphene-treated samples compared to the control after 24 h of exposure (Figure 3A). At the highest concentration, the level of CFUs was increased two times as compared to the control (Figure 3A). These results were in correlation with the previous bacterial growth data generated under static conditions, in which a significant increase of L. acidophilus in the 1 and 10 μg/mL treated samples were also found. B. longum showed a different pattern of growth after exposure to graphene as compared to the graphene exposures in the stationary cultures. Indeed, a significant increase in CFUs was noted at all concentrations during the first 4 h of exposure. At the 24-h time point, the number of B. longum CFUs in the graphene-treated samples was similar to the control at all concentrations (Figure 3B). These results were different from the growth curves obtained during the early analysis and no correlation was observed. The E. coli culture showed no significant impacts of exposure to graphene at any concentration (Figure 3C). Moreover, no clear correlation was found between the results obtained using the growth curves study (static condition) and the live bacterial culture subjected to continuous interaction between bacteria and graphene. 2.2. Effect of Pristine Graphene on the Survival of Rat Fecal Bacteria. The results showed that 1 μg/mL graphene had minimal to no effect on the survival of total anaerobic bacterial counts at all time points (Figure 4A). While graphene exposure at 10 and 100 μg/mL concentrations exhibited a 1.5fold increase of total anaerobic bacteria, for the 10 μg/mL concentration, this level returned to the control level by the 6 h time point and remained at that level at the 24 h sampling. Whereas, for the 100 μg/mL graphene exposure, the decrease in CFU observed at 6 h was transient. However, rat fecal

samples treated with graphene at all concentrations showed a significant increase (p = 0.035) in the total aerobic bacteria counts during the first 6 h of exposure. After 6 h, total aerobic counts of graphene-treated samples were equal to or below the untreated control (Figure 4B). Two selective bacterial culture mediums (LMRS and BSA) were used to study the survival of Lactobacillus and Bif idobacterium, respectively (Figure 4C, 4D). Exposure to graphene at all concentrations resulted in the decrease of Bif idobacterium CFUs during the first 6 h postexposure (Figure 4C). After 24 h, no significant changes in CFUs of Bif idobacterium were noted between the graphene-treated or nontreated samples. These results at 24 h were similar to the previous results obtained after exposing a pure culture of B. longum to graphene at all concentrations. However, direct graphene exposure to B. longum resulted in an increase of CFUs during the first 6 h (Figure 3B), while the exposure of graphene to the rat fecal samples resulted in a decrease of total Bif idobacterium CFUs during the same period of time (Figure 4C). Similarly, the exposure of graphene to rat fecal samples resulted in a significant increase of Lactobacillus CFUs at the earliest time point (3 h; Figure 4D). In contrast, graphenetreated fecal samples showed lower CFUs as compared to the controls at 6 and 24 h time points (except at the lowest concentration). After 24 h exposure of the fecal samples, we did not observe the significant increase of Lactobacillus CFUs (Figure 4D) that was previously observed in the graphenetreated pure culture of L. acidophilus (Figure 3A). The selective media used to isolate Lactobacillus and Bif idobacterium genera from the fecal samples has the potential to select for a wide range of bacterial species in the same genera and does not necessarily represent a specific species. Due to the complexity of identifying individual bacterial colonies on selective plates, we also isolated DNA from treated and nontreated fecal samples and examined the impact of 25710


Research Article


Figure 4. Total bacterial counts after exposure to graphene at 1, 10, and 100 μg/mL and at 3 different time points (3, 6, and 24 h). Brucella blood agar (A), tryptic soy blood agar (B), Bif idobacterium selective agar (C), Lactobacillus-MRS agar (D). Error bars show standard error values. All CFU values are expressed as a percentage relative to the control values. *, p < 0.05 compared to control at 3 h. ¥, p < 0.05 compared to control at 6 h. t, p < 0.05 compared to control at 24 h.

effect on the observed results. While the deviation between samples is much larger at the 3 h graphene exposure time, this deviation is much less pronounced following 24 h graphene exposure (Figure S2A). Moreover, PCA analysis in Figure S2B showed that the clusters of graphene-treated microbiomes are forming concentric circles where the center is formed by the control samples. As we move away from the center, the dose of graphene increase and the deviation between the samples increases simultaneously (Figure S2B). Taking all this data into consideration, we can see that both the time of exposure and the concentration of graphene were important factors in the observed responses. Indeed, it was clear that the major changes seen in the relative abundance of rat intestinal bacteria occurred at the 3 h time point. Following 24 h incubation, the intestinal microbiome population ratio in all graphene-treated samples resembled the control untreated samples. To further understand the major players in these bacterial microbiome shifts, we relied on next-generation sequencing analysis to obtain a detailed perspective of the different bacteria groups affected by the treatment of pristine graphene. 2.4. Shifts in Bacterial Populations after Exposure to Graphene Evaluated by Next-Generation Sequencing. The relative abundance of phylotypes was summarized at the phylum and other lower taxonomic levels. A total of 13 predominant bacterial phyla were found in all samples (graphene-treated and control), with Bacteroidetes and Firmicutes being the most abundant phyla in all treatment groups (Figure 6A). Moreover, we identified the other

graphene on the relative abundance of major phyla, families, and genera in rat fecal samples. 2.3. Impact of Graphene on the Relative Abundance of Major Bacterial Phyla, Families, and Genera in Rat Fecal Samples. Figure 5 shows the relative abundance of major phyla, families, and genera at 3 and 24 h after treatment with graphene. The results of this analysis showed that the ratio of Firmicutes and Bacteroidetes phyla changed significantly after exposure to the highest concentration of graphene (100 μg/mL). At the same concentration, the abundance of the Enterobacteriaceae family (representing several members of pathogenic and nonpathogenic bacteria) was found to increase in comparison to the untreated control. Moreover, in comparison to the untreated control, we observed significant dose-dependent changes at the genus level. At the low concentration (1 μg/mL), the abundance of Bif idobacterium, Lactobacillus, and Bacteroides remained unchanged at 3 and 24 h postexposure to graphene. At the highest concentration (100 μg/mL), Bif idobacterium abundance increased, accompanied by a significant decrease of Lactobacillus at 3 and 24 h postexposure (Figure 5). To further analyze the results obtained from the real-time PCR work, we utilized a multivariate analysis method (Principal component analysis) to identify the major factors underlying the effects observed upon exposure of rat fecal samples to graphene. Figure S2 presents the principal component analysis of the qPCR data based on (A) time and (B) groups studied. The results show that time of interaction between the graphene and bacteria has a major 25711


Research Article


Figure 5. Relative abundance of rat fecal bacteria after exposure to graphene at 0 (control), 1, 10, and 100 μg/mL as analyzed by real-time PCR. Bacteria are grouped based on (A) phylum, (B) genus, and (C) family.

with the highest concentration of graphene (100 μg/mL). At the species level, 9 organisms were found to have significant changes in abundance between the control and graphenetreated groups (Figure 7B). These 9 species were as follows: Parasutterella excrementihominis, Alistipes putredinis, Citrobacter spp., Sutterella wadsworthensis, Sutterella stercoricanis, Lachnoclostridium clostridium symbiosum, Lachnoclostridium clostridium f imetarium, Lachnoclostridium clostridium hylemona, and Ruminococcus champanellensis. The abundance changes observed under different concentrations of graphene and exposure times can be categorized into 2 different patterns: (1) a significant increase in bacterial abundance during the first 6 h of exposure, followed a decrease to near control levels after 24 h of exposure. For instance, Clostridium hylemona, Clostridium symbiosum, and Clostridium f imetarium, which belong to the class of Clostridia, had significant increases in abundance at all graphene concentrations during the first 6 h of exposure with a decrease and recovery after 24 h treatment; (2) a significant decrease of bacterial abundance during the first 6 h of exposure, followed by a recovery after 24 h of exposure. The species that exhibited such a pattern included Sutterella wadsworthensis, Sutterella stercoricanis, and Parasutterella excrementihominis (Betaproteobacteria), Ruminococcus champanellensis (Clostridia), and Alistipes putredinis (Bacter-

subclasses as follows: 13 classes (Figure S3), 47 orders (Figure S4), 98 families (Figure S5), 196 genera (Figure S6), and 485 species (Figure 6B). The abundance of both predominant phyla (Firmicutes and Bacteroidetes) was not significantly affected by the graphene treatment at any concentration (Figure 6A). To clearly detect the significant changes that occurred among the taxonomic classes in the rat intestinal microbiota, we used a statistical approach (one way ANOVA with posthoc Tukey test) to evaluate which taxonomic classes were significantly altered (p < 0.05) in abundance after exposure to graphene (Figure S7). Figure 7 presents a summary of all taxonomic groups (from phylum to species) that showed differences among the treatment groups studied. Among the phyla, abundance of Proteobacteria and Verrucomicrobia exhibited a significant increase between the 6 and 24 h time points in both control and graphene-treated groups (Figure 7). At the 100 μg/mL concentration of graphene, both of these phyla showed a significant decrease in relative abundance at 24 h (Figure 7). This relative decrease of abundance was also observed at the class level with betaproteobacteria, at the order level with burkholderiales, at the family level with sutterellaceae, and at the genus level with Parasutterella. Among all of these subclasses, the abundance of bacteria was comparatively lower in rat fecal samples treated 25712


Research Article


Figure 6. (A) Relative abundance of the phylotypes at the phylum level. (B) sPLS-DA based on phylotypes. The green dashed line circle represents control sample following 6 h of graphene treatment. The blue dashed line circle represents rat fecal samples treated with graphene following 6 graphene treatment. The red dashed line circle represents all graphene-treated and control samples.

iodia). Another species group which belongs to the class of Gammaproteobacteria, Citrobacter spp., showed no significant change during the first 6 h of exposure to graphene, followed by a significant increase of bacterial abundance after 24 h. Taking into consideration the data collected for all the bacterial taxonomic classes detected in rat fecal samples after graphene treatment at all concentrations and during both exposure time points (6 and 24 h), we used a sPLSDiscriminant Analysis (Figure 6B) approach to determine the main contributing factors. The sPLS-Discriminant Analysis was found to be an efficient variable selection methodology with clear graphical outputs.31 The analysis of our results revealed an interesting pattern of population grouping. For instance, we found that the data can be grouped into three different population groups (Figure 6B). The first group is composed of control rat fecal samples at 6 h. The second group is composed of graphene-treated samples at 6 h. The last group is composed of all graphene-treated and nontreated samples at 24 h postexposure. This analysis shows clearly that the two major factors behind the population abundance changes observed in this study are graphene dose and treatment time. As the most pronounced factor, graphene

treatment dose had clearly affected the microbiota of rat fecal samples during the initial 3 h exposure time. As the secondary treatment time factor was concerned, it took 24 h for the bacterial population abundance changes in the graphenetreated fecal samples to stabilize and return to near the untreated control sample levels. In an effort to further understand the changes that occurred during and after the exposure of rat fecal samples to graphene, we quantified SCFA levels using HPLC. The SCFAs are key players in the interplay between diet and the gut microbiota.32 Therefore, any change to the microbiota could impact the production of SCFAs profiles found in the gut. 2.5. Impact of Pristine Graphene Exposure on the Production of SCFAs. SCFA analysis of the fecal samples showed that the SCFA responses can be categorized into two distinct groups. The first group included SCFAs levels that were decreased or unchanged in comparison to the initial untreated fecal samples (before treatment). This group encompasses succinic acid, acetic acid, propionic acid, hexanoic acid, valeric acid, and isovaleric acid. Among these SCFAs, no differences were observed between the control and graphene treated samples. The second group includes SCFAs 25713


Research Article


Figure 7. Different taxonomic groups showing significant changes in relative abundance after exposure to graphene as compared with controls determined by one-way ANOVA analysis. (A) Phylum is shown in green bars, class in black bars, order in purple bars, family in light blue bars, genus in red bars and (B) species in yellow bars. C stands for control, E1 stands for 1 μg/mL graphene, E2 stands for 10 μg/mL graphene and E3 stands for 100 μg/mL graphene. ¥, p < 0.05 compared to control at 3 h. *, p < 0.05 compared to control at 24 h. Error bars represent standard error values.

with increased levels at the 3 and 24 h time points and includes butyric acid (Figure 8). The levels of butyric acid increased significantly by 1.8 (p = 0.04961) and 2.1 (p = 0.04909) fold for the graphene-treated samples 10 μg/mL and 100 μg/mL, respectively, and during the 3 h of exposure. The levels of butyric acids in control samples remained unchanged after 24 h. However, the butyric acid levels in fecal samples increased at the 24 h time point for all graphene concentrations (Figure 8). The purpose of this study was to assess the effect different doses of graphene treatment had on the rat gut microbiota using a BRCCS. The use of this dynamic system is an innovative approach in the search for a model that can replicate the continuous interaction of the bacteria and graphene that will occur in the human gastrointestinal tract. Conducting experiments with pristine graphene nanomaterial presents many unique challenges, as the test material precipitates during incubation with the model cells or bacterial

population. The 24 h time point was chosen for the maximum exposure time, as this is relevant to digestive system physiology of human; where interaction time of residence of digesta in intestinal tract is less than 24 h.33,34 This innovative BRCCS provides a model in which there is a continuous interaction of test materials with target bacteria without any signs of precipitation. To our knowledge, this is the first study to use such a system to mimic the continuous interaction of the NPs with the gut microbes. The chemical nature of pristine graphene makes it prone to agglomeration and low dispersion in the growth media.35 Therefore, many studies have relied on the modification of graphene surface chemistry using covalent and noncovalent bonding, an approach that can change the characteristics of the NPs.36 Bacterial toxicity assays in this context remain controversial due to the inconsistency of experimental design.35 Here, we have shown that pristine graphene can sediment in bacterial growth media in static 25714


Research Article


Figure 8. SCFA quantification in rat fecal samples after exposure to graphene at 3 and 24 h. (A) succinic acid, (B) propionic acid, (C) valeric acid, (D) acetic acid, (E) hexanoic acid, (F) isovaleic acid, and (G) butyric acid. Data expressed in fold change in comparison to the initial untreated fecal sample used. Error bars showing standard error values. ¥, p < 0.05 compared to control at 3 h. *, p < 0.05 compared to control at 24 h. Error bars represent standard error values (n = 4).

conditions, while a homogeneous distribution of graphene is observed in dynamic conditions (Figure S1). Previously, it was found that in static settings graphene precipitation can act as a scaffold for bacterial attachment, proliferation and biofilm formation.28,37 Our results clearly provide evidence that such observation is specific to certain bacteria. For instance, an early proliferation of L. acidophilus was observed during exposure to all concentrations of graphene. However, there was no similar indication of early proliferation for B. longum or E. coli. In dynamic settings, a different pattern of bacterial growth was observed, depending on the species studied. First, no obvious changes were seen in the survival of E. coli treated with graphene at any concentrations. Second, B. longum survival after exposure to graphene was concentration dependent. We noticed an early proliferation of B. longum bacteria at the lowest concentration of graphene (1 μg/mL) exposure and no changes during the exposure to 10 and 100 μg/mL concentrations. Third, L. acidophilus showed a higher bacterial count after exposure to graphene at all concentrations. Taken together, these results clearly demonstrate that the in vitro bacterial behavior and response to graphene can change dramatically between static to dynamic systems. Since the dynamic system minimizes the bias of NPs precipitation as well as aggregation, this system was chosen to further analyze the impact of graphene on rat intestinal microbiota. Moreover, the 3% fecal slurry also remained in a uniformly suspended form

that allowed a complete interaction of graphene with the fecal bacteria. In this study, the key factors related to the impact of graphene on the rat intestinal microbiota were duration of exposure and concentration of the nanomaterial. In fact, both the qPCR data and the next generation sequencing data demonstrated that during the first 3 h, graphene exposure led to significant shifts in bacterial communities in comparison to the untreated control. However, after 24 h, the shifts in bacterial communities became less pronounced in comparison to the control. Moreover, the highest graphene concentrations had the largest impact on shifts in bacterial populations. For instance, high graphene concentrations were shown to impact the relative abundance of two major phyla (Proteobacteria and Verrucomicrobia; Figure 7A). A similar impact of graphene on Proteobacteria and Verrucomicrobia phyla was also reported by Nguyen et al. 38 after adding graphene to microbial communities from sludge.39 In fact, in the Nguyen study, 10 mg/L of graphene was sufficient to significantly decrease Proteobacteria and Verrucomicrobia abundance and change the microbial communities of the sludge. However, such impact was considered by the investigators to be a transient response and was observed only during the first few hrs of the study.38,39 Our study further supports the concept that bacterial communities of the intestinal tract are resilient and able to recover from the short-term impact of pristine graphene exposure. Furthermore, the initial impact of graphene on the 25715


Research Article


under aerobic or anaerobic conditions in our laboratory.36 The same strategy was used to assess the effects of pristine graphene. The vessels were custom-made to accommodate the requirements of an anaerobic environment for bacterial culture. Pure bacterial cultures or fecal materials were prepared in separate tubes and then fed into the vessels using a 10 mL syringe. The same syringes were also used to obtain samples at specific time points. 4.3. Bacterial Cultivation and Preparation. Three model bacteria were used for the analysis of graphene-bacteria interaction. These cultures included Lactobacillus acidophilus (ATCC4356), Bif idobacterium longum (ATCC35183), and Escherichia coli (ATCC10798) and were originally purchased from the American Type Culture Collection (ATCC; Manassas, VA). Prior to their use, the bacterial strains were thawed, and the glycerol was removed by washing the bacterial pellet with each species’ respective nutrient media: Man, Rogosa, and Sharpe (MRS) broth for L. acidophilus; chopped meat broth (CM) for B. longum; and tryptic soy agar (TSA) for E. coli. B. longum was cultured under anaerobic conditions and at 37 °C. L. acidophilus was kept at 37 °C and in a 5% CO2 environment. E. coli was grown under aerobic conditions at 37 °C. 4.4. Measurement of Bacterial Growth and Survival. 4.4.1. Growth Measurement. Bacterial growth after exposure to graphene (0, 1, 10, and 100 μg/mL) was measured in real time by optical density measurement at OD600 nm using a Cytation3 Cell Imaging Multimode Reader (Biotek, Winooski, VT). Bacterial cells were allowed to reach mid log-phase growth each time before inoculation. The growth curve of E. coli was initiated by spiking ∼1 × 104 CFU/ml into TSA medium (pH 7.4) containing graphene at concentrations of 1, 10, and 100 μg/mL, after which the media was incubated at 37 °C with real time OD recording for 22 h every 30 min. Growth curves for L. acidophilus were set up in a similar way, however the organisms were spiked into diluted MRS media containing the different treatments and incubated at 37 °C with 5% CO2. B. longum was spiked into CM media and incubated anaerobically at 37 °C. 4.4.2. Survival Measurement. Bacteria survival after treatment with graphene in the rotary culture system was evaluated at three time points (2, 4, and 24 h). An aqueous suspension of triple autoclaved graphene was prepared at 1, 10, and 100 μg/mL as outlined above. The suspensions were then gently mixed with the corresponding growth media and spiked with ∼1 × 106 CFU/ml of each bacterial culture. Control samples were treated with ultrapure water only. After incubation, serial dilutions were prepared and 100 μL of diluted samples were inoculated onto four different bacterial growth agar plates brucella blood agar (BRU), bifidobacterium selective agar (BSA), Lactobacillus-MRS agar (LMRS), and tryptic soy blood agar (TSBA). Plates were incubated either anaerobically (BRU, BSA, and LMRS) or aerobically (TSBA) at 37 °C for 24 h, and colonies were counted using Scan1200 Colony Counter (Interscience, Woburn, MA). 4.5. Rat Fecal Sample Preparation and Exposure. Fecal samples were obtained from the colon of three healthy male Sprague− Dawley rats aged 4−6 months that were not previously exposed to graphene. The culture conditions of fecal microbiota used in this experiment were previously described.33 Briefly, the fecal samples were weighed and transferred to anaerobic media, Maximum Recovery Diluent (MRD, LbM IDG, Bury, U.K.) buffer, to a final concentration of 25% (w/v). Subsequently, fecal samples were cultured in low-concentration carbohydrate medium (LCM) media (10 mL) at a final 3% inoculum concentration. Graphene suspensions were added to the fecal samples at 1, 10, and 100 μg/mL, and a similar volume of water was added to the control sample. Before the exposure of different treatments to the microbiota, an initial sample was extracted to serve as a baseline control. All samples were incubated anaerobically at 37 °C and 1500 μL of each sample was removed using an 18-gauge syringe at 3, 6, and 24 h postexposure. At the same time, an equal amount of suspension was injected in the sample vessel via syringe to maintain the total volume. All samples were used immediately (live bacteria analysis) or saved at −80 °C for

bacterial communities was in synchrony with an increase of butyric acid production (Figure 8). This finding further correlated with the increase in abundance of 3 species belonging to the genus Clostridium. These bacteria include Clostridium hylemonae, Clostridium fimetarium, and Clostridium symbiosum (Figure 7). These organisms are not well characterized; however, they were classified under the genus Clostridium due to their production of butyric acid as a fermentation product.40 Since butyric acid is known to positively modulate the immune system,41 it becomes clear that graphene perturbation to the gut microbiome could have sequential effect on the host immune system. Moreover, butyrate is known to induce epithelial cell proliferation and can affect paracellular tight junction.42,43 Therefore, the butyrate production could be observed as a biomarker for such responses. Due to the high drug loading capacity of graphene, it is considered as a great potential vector for delivery.2 However, recent reports suggest that long-term accumulation of such carriers in the epithelial cells could be cytotoxic and lead to prevention of cell proliferation and induction of cell apoptosis.2,44−46 The results in this study show that the gastrointestinal microbiota is a resilient organ that can absorb the toxic effects of graphene. The ability of the microbiota to recover from a short perturbation after graphene exposure is led by the interactions of different microbial communities with each other to regulate the abundance of different microbes. Nevertheless, pristine graphene nanoparticles have low degradability and can be released into the environment through fecal discharge or waste. Consequently, pristine graphene can affect a wider range of microbes in other aquatic or soil system and can change ecosystems.1 Risks assessment studies will be required to shed light on the environment hazard after both acute and chronic exposures.

3. CONCLUSION The results presented here clearly demonstrate that the dynamic in vitro model is a suitable platform to study bacteria-nanoparticles interactions. Indeed, insoluble nanoparticles such as pristine graphene can stay in suspension and interact with the bacteria throughout the time of study. Our study further demonstrates that the toxicity of pristine graphene to the intestinal microbiome are time and dose dependent. The intestinal microbiota is a key player in the equilibrium between the host and the functional and protective aspects of the GI tract.47 The disruption of the microbiome during the early time exposure to graphene is transient and not dose dependent. Our finding could advance the knowledge about nanoparticles-microbiome interaction and further help regulatory agencies to understand the risks associated with the use of graphene in consumer use products. 4. MATERIALS AND METHODS 4.1. Graphene Characterization. Graphene was characterized at the Nanocore facility, National Center for Toxicological Research, Arkansas, as was previously described;48 graphene was autoclaved three times to remove any endotoxins contamination and suspended in 0.5% BSA. 4.2. Bioreactor-Rotary Cell Culture System Set Up. The bioreactor-rotary cell culture system (BRCCS) was used to allow interaction of graphene and bacteria. This system contained four separate bioreactor vessels revolving at a similar speed [Synthecon (Houston, TX)]. The BRCCS is commonly used for the culture of eukaryotic cells. This system was recently adapted to culture bacteria 25716


Research Article

ACS Applied Materials & Interfaces future use [(DNA extraction and Short Chain Fatty Acid (SCFAs) analysis)]. 4.6. Live Bacteria Analysis. The intestinal fecal bacteria survival after treatment with graphene in the rotary culture system was evaluated at three time points (3, 6, and 24 h). After incubation, serial dilutions were prepared and 100 μL of diluted samples were inoculated into four different bacterial growth agar plates: Brucella Blood Agar (BRU), Bifidobacterium Selective Agar (BSA), Lactobacillus-MRS Agar (LMRS) and Tryptic Soy Blood Agar (TSBA). Plates were incubated either anaerobically (BRU, BSA, and LMRS) or aerobically (TSBA) at 37 °C for 24 h and colonies were counted using a Scan1200 Colony Counter (Interscience, Woburn, MA). 4.7. Fecal Bacterial DNA Isolation and Real-Time PCR of Bacterial Groups. Bacterial DNA was extracted from 500 μL of fecal samples using QIAamp DNA Stool Mini Kit (Qiagen, Valencia, CA), as described earlier.37 Briefly, fecal samples were incubated with lysis buffer at 95 °C for 5 min and then centrifuged at 15,000 g for 1 min to pellet stool particles. The supernatant was incubated with one InhibitEX Tablet, followed by incubation with proteinase K at 65 °C for 10 min. After washing, DNA was eluted with DNase-free water. DNA quality and quantity were checked using Qubit Fluorometer (Thermo Fisher Scientific, Waltham, MA) and Cytation3 Cell Imaging Multimode Reader (Biotek, Winooski, VT). The identification of different bacterial groups was performed as described previously.49 Briefly, real-time PCR (qPCR) was used to amplify DNA fragments of the predominant phyla and representative genera and species of bacteria present in the fecal samples. Also, 3 μL of DNA template was used in all experiments with 10 μL of SYBR Green master mix, 2 μL of primer, and 5 μL of DNase-free water. The amplification was carried out with the following steps: 95 °C for 10 min, followed by denaturation at 95 °C for 15 s and annealing and extension at 60 °C for 1 min. The qPCR reaction was run in an ABI 7500 machine for 45 cycles and followed by a complete denaturation of the PCR product to obtain melting curves. Melting curve data was obtained from 60 to 95 °C at a temperature stepdown rate of 0.5 °C s-1 to allow the confirmation of specificity of the amplicon product. Real-time PCR data was normalized with 18S gene expression in rat genomic DNA as described previously (Petnicki-Ocwieja et al.). 4.8. V3−V4 Based 16S Next-Generation Sequencing. The same bacterial DNA (extracted for real-time PCR analysis) was further used for next-generation sequencing analysis. The quality of DNA was assessed using a Qubit Fluorometer (Thermo Fisher Scientific, Waltham, MA), and a similar amount of DNA was used for the 16S sequencing by MiSeq as described earlier.50,51 The 16S rRNA gene V4 variable region PCR primers 515/806 with barcode on the forward primer were used in a 28 cycle PCR (5 cycle used on PCR products) using the HotStarTaq Plus Master Mix Kit (Qiagen, U.S.A.) under the following conditions: 94 °C for 3 min, followed by 28 cycles of 94 °C for 30 s, 53 °C for 40 s and 72 °C for 1 min, after which a final elongation step at 72 °C for 5 min was performed. After amplification, PCR products were checked on a 2% agarose gel to determine the success of the amplification and the relative intensity of bands. Multiple samples were pooled together in equal proportions based on their molecular weight and DNA concentrations. Pooled samples were purified using calibrated Ampure XP beads. Then the pooled and purified PCR products were used to prepare the Illumina DNA library. Sequencing was performed on a MiSeq following the manufacturer’s guidelines. Sequence data was processed by joining the sequences, depletion of barcodes, removal of sequences

Impact of Pristine Graphene on Intestinal Microbiota Assessed Using a

Recommend Documents