Chronic Manganese Exposure and the Enteric Nervous System: An in Vitro and Mouse in Vivo Study

Abstract

Background:

Chronic environmental exposure to manganese (Mn) can cause debilitating damage to the central nervous system. However, its potential toxic effects on the enteric nervous system (ENS) have yet to be assessed.

Objective:

We examined the effect of Mn on the ENS using both cell and animal models.

Method:

Rat enteric glial cells (EGCs) and mouse primary enteric cultures were exposed to increasing concentrations of Mn and cell viability and mitochondrial health were assessed using various morphological and functional assays. C57BL/6 mice were exposed daily to a sublethal dose of Mn ($15\mathrm{mg}/\mathrm{kg}/\mathrm{d}$) for 30 d. Gut peristalsis, enteric inflammation, gut microbiome profile, and fecal metabolite composition were assessed at the end of exposure.

Results:

EGC mitochondria were highly susceptible to Mn neurotoxicity, as evidenced by lower mitochondrial mass, adenosine triphosphate–linked respiration, and aconitase activity as well as higher mitochondrial superoxide, upon Mn exposure. Minor differences were seen in the mouse model: specifically, longer intestinal transit times and higher levels of colonic inflammation.

Conclusion:

Based on our findings from this study, Mn preferentially induced mitochondrial dysfunction in a rat EGC line and in vivo resulted in inflammation in the ENS. https://doi.org/10.1289/EHP7877

Introduction

Manganese (Mn) is an essential trace metal for many enzymatic reactions, taking part in several metabolic as well as biological functions in the body (Bowman et al. 2011; Harischandra et al. 2019). Despite its nutritional benefits, prenatal (Zota et al. 2009) and early postnatal (Claus Henn et al. 2010) exposures to Mn were shown to be associated with low birth weight and lower mental development scores, respectively, implying the role of Mn as both an essential nutrient and a toxicant. Mn exposure occurs commonly in both occupational and environmental settings, and the metal is widely used in industrial and agricultural products. Moreover, chronic exposure to Mn via welding fumes, pesticides such as Maneb, or Mn-contaminated well water can lead to manganism, a neurological disorder sharing certain features with Parkinson’s disease (PD) (Ferraz et al. 1988; Frisbie et al. 2012; McMahon et al. 2019). Mn levels are primarily regulated by gastrointestinal (GI) absorption and hepatobiliary excretion in the biliary system (Claus Henn et al. 2010; Davis et al. 1993). Interestingly, mutations in the metal exporter SLC30A10 have been shown to cause the first known disease of inherited Mn excess that is characterized by broad neurological, hepatic, and hematological disturbances (Mercadante et al. 2019; Quadri et al. 2012; Taylor et al. 2019; Tuschl et al. 2016; Zaki et al. 2018).

In the general population, food and water are the main sources of Mn exposure (O’Neal and Zheng 2015). In 2001, the Food and Nutrition Board established the adequate intake for Mn to be between $1.8\text{\hspace{0.17em}and\hspace{0.17em}}2.3\mathrm{mg}/\mathrm{kg}$ for adults (Malpartida et al. 2021). Mn is naturally present in many surface water and groundwater sources and in soils that may erode into these waters. Although the World Health Organization has discontinued its guidelines for Mn levels in water (World Health Organization 2017), an earlier U.S. Environmental Protection Agency guideline suggested $300\mathrm{\mu }\mathrm{g}/\mathrm{L}$ Mn as an acceptable concentration in drinking water. Yet, a survey conducted by McMahon et al. (2019) found $>300\mathrm{\mu }\mathrm{g}/\mathrm{L}$ Mn concentrations in 43,334 wells in the United States, raising questions on whether these levels were harmful to the population that drank from these wells. Frisbie et al. (2012) documented several nations with water bodies having $>400\mathrm{\mu }\mathrm{g}/\mathrm{L}$ Mn concentrations. Accumulating evidence suggests that early-life exposure to Mn in drinking water contributes to cognitive and behavioral disorders (Bjørklund et al. 2017; Bouchard et al. 2007, 2011; Khan et al. 2011, 2012; Oulhote et al. 2014; Schullehner et al. 2020). A few case reports have highlighted a negative correlation between high Mn intake via parenteral nutrition or food consumption and a patient’s motor function and behavior (Bleich et al. 1999; Fell et al. 1996; Holzgraefe et al. 1986; Schuh 2016; Walter et al. 2016). Thus, the above-mentioned cases support the oral intake of Mn as a major concern in environmental metal toxicology. However, these cases mainly refer to dysfunctions in the central nervous system (CNS) because they only sporadically mention the presence of GI-related abnormalities. Although the reported incidences of GI-related Mn toxicity are few, gastritis was identified in a man with acute Mn toxicity (Huang and Lin 2004), and peptic ulcers were identified in guinea pigs orally administered $10\mathrm{mg}/\mathrm{kg}/\mathrm{d}$ Mn for 30 d (Chandra and Imam 1973). As we begin to get clarity on the gut-brain axis, several theories regarding the development and progression of neurodegenerative diseases have focused on whether these neurological disorders begin in the enteric nervous system (ENS) (Pellegrini et al. 2015). If this origin is indeed the case, then changes in GI physiology are likely to occur before any CNS-controlled symptoms occur.

The ENS comprises enteric neurons and glial cells present in two distinct layers in the GI tract and plays an important role in governing nutrient absorption and peristalsis, among other functions. In particular, enteric glial cells (EGCs) are present not only in close contact with enteric neurons but are also present in the muscle layers and lamina propria (Gulbransen and Sharkey 2012). Similar to astrocytes in the CNS, EGCs play an important role in maintaining tight junctions and communicate with other cells, including surveilling immune cells (Veiga-Fernandes and Pachnis 2017; von Boyen and Steinkamp 2010). Incidentally, astrocytes are particularly sensitive to Mn toxicity (Harischandra et al. 2019; Sarkar et al. 2018); however, whether EGCs follow a similar trend has not yet been studied.

With much research devoted to understanding the mechanisms behind Mn toxicity in the brain, scant attention has been paid to its toxic effects in the gut. Although some effort has been made to study the effects of Mn exposure on gut physiology and the microbiome (Chi et al. 2017), to our knowledge, no published data exist exploring Mn’s potential toxic role in the ENS. Thus, in the present study, we systematically characterized the adverse effects of Mn on the ENS using cell and animal models of Mn toxicity.

Methods

Chemicals and Reagents

Manganese (II) chloride tetrahydrate (${\text{MnCl}}_2$) and $\mathrm{\beta }\text{-actin}$ antibody were purchased from Sigma. Neurobasal medium and Dulbecco’s modified Eagle’s media (DMEM), B27 and N2 supplements, fetal bovine serum (FBS), Trypsin-EDTA (TE), L-glutamine, penicillin, and streptomycin were purchased from Invitrogen. Antibodies for PGP9.5 (Cat. No. AB1761-I) and glial fibrillary acidic protein (GFAP) (Cat. No. MAB3402) were purchased from Millipore while iNOS (Cat. No. sc-651) was purchased from Santa Cruz. Mouse recombinant GDNF (Cat. No. 450-44) was purchased from PeproTech.

Cell Cultures

The rat EGC line (Cat. No. CRL-2,690) was purchased from ATCC. Cells were grown in DMEM media supplemented with 10% v/v FBS, 1% v/v L-glutamine, penicillin ($100\text{\hspace{0.17em}units}/\mathrm{mL}$), and streptomycin ($100\text{\hspace{0.17em}units}/\mathrm{mL}$), and maintained at 37°C in a humidified atmosphere of 5% ${\mathrm{CO}}_2$.

To obtain a primary enteric mixed culture, intestines from a total of sixteen E15 C57BL/6 mice (given same diet and husbandry conditions as mice in our in vivo study; see Animal Studies) were obtained and minced in sterile, ice-cold DMEM media containing antibiotics. The finely chopped pieces were enzymatically digested in serum-free media containing collagenase ($0.2\mathrm{mg}/\mathrm{mL}$) and dispase ($0.2\mathrm{mg}/\mathrm{mL}$) at 37°C for 30 min, which was stopped by adding 10% v/v FBS and centrifuging the cells at $250\times g$ for 5 min. The supernatant was aspirated, and the pellet was triturated in 10% v/v DMEM/F-12 until a homogenous cell suspension was obtained. Cells were filtered through a $40\text{-}\mathrm{\mu }\mathrm{m}$ cell strainer and plated at a density of $1\times {10}^5{/\mathrm{cm}}^2$ on poly D-lysine- (PDL-) and laminin-coated plates. After 24 h, media were changed to 1% DMEM/F-12 containing $50\text{\hspace{0.17em}ng}/\mathrm{mL}$ GDNF and N2 and B27 supplements and cultured for 3 wk at 37°C and 5% ${\mathrm{CO}}_2$. Half of the media was changed every 2 d.

MTS Assay

Cell viability was measured by Cell Titer 96® aqueous nonradioactive cell viability assay (MTS assay; Promega). Briefly, 25,000 EGCs were seeded onto 96-well plates and allowed to attach overnight before exposure to 0, 1, 3, 10, 30, 100, 300, or $\mathrm{1,000}\mathrm{\mu }\mathrm{M}$ Mn for 24 h. MTS solution ($20\mathrm{\mu }\mathrm{L}$) was added to each well, and the plates were incubated at 37°C and 5% ${\mathrm{CO}}_2$ for 1.5 h. Measurements were made at $490\mathrm{\hspace{0.17em}nm}$ and $670\mathrm{\hspace{0.17em}nm}$ (reference wavelength) using a fluorescence microplate reader (SpectraMax® Gemini™ XS; Molecular Devices). Experiment was repeated three times, each time in triplicate.

SYTOX™ Green Assay

Cell death was determined by the cell-impermeable dye SYTOX™ green (Song et al. 2010). EGCs were grown in 96-well plates at a density of 25,000 cells/well and treated with 0, 1, 3, 10, 30, 100, 300, or $\mathrm{1,000}\mathrm{\mu }\mathrm{M}$ Mn in 2% v/v DMEM for 24 h. After treatment, $1\mathrm{\mu }\mathrm{M}$ SYTOX™ green was added to each well for 20 min. Cell death was quantitatively measured at excitation/emission (ex/em) wavelengths of $485/525\mathrm{\hspace{0.17em}nm}$ using a fluorescence microplate reader (SpectraMax® Gemini™). Nuclei were stained with Hoechst 33342 ($5\mathrm{\mu }\mathrm{g}/\mathrm{mL}$) for 15 min to determine total cell number in each well. The resulting green fluorescence readings were normalized to cell number. Photomicrographs showing cell death and cell morphology were also taken using an EVOS FLoid™ cell imager (Invitrogen). Experiment was repeated three times, each time in duplicate.

Caspase-3 Activity

Following 24-h exposure to 0, 1, 3, 10, 30, 100, 300, or $\mathrm{1,000}\mathrm{\mu }\mathrm{M}$ Mn, media was discarded, and the adherent EGCs were washed once with 1× phosphate-buffered saline (PBS), trypsinized, collected in individual tubes, centrifuged at $500\times g$ for 10 min and lysed in $230\mathrm{\mu }\mathrm{L}$ lysis buffer (10% wt/vol sucrose, $5\text{\hspace{0.17em}mM}$ dithiothreitol, $100\text{\hspace{0.17em}mM}$ HEPES, 0.1% wt/vol CHAPS, and $20\text{\hspace{0.17em}mM}$ EDTA in ${\text{dH}}_2\mathrm{O}$; pH 7.4). The samples were incubated for 20 min at 37°C to lyse the cells. Thereafter, $190\mathrm{\mu }\mathrm{L}$ of lysate was added to individual wells of a 96-well, black, clear-bottomed plate (Cat. No. Greiner 655,079) containing $10\mathrm{\mu }\mathrm{L}$ caspase-3 substrate ($50\mathrm{\mu }\mathrm{M}$ Ac-DEVD-AFC) at 37°C for 60 min. Formation of 7-amino-4-methylcoumarin (AFC), resulting from caspase-3 activity, was measured at $460/510\mathrm{\hspace{0.17em}nm}$ (ex/em), respectively, using a fluorescence plate reader. The caspase-3 activity was normalized to protein concentration as determined by the Bradford protein assay (Harischandra et al. 2015). The experiment was repeated three times.

Determination of Mitochondrial Morphology and Mass

EGCs (40,000) were plated on PDL-coated coverslips and treated with 0, 1, 3, 10, or $30\mathrm{\mu }\mathrm{M}$ Mn for 24 h. After treatment, cells were washed with Hanks’ balanced salt solution (HBSS; Invitrogen) containing ${\mathrm{Ca}}^{+2}$ and ${\mathrm{Mg}}^{+2}$ to remove any residual media. Next, $100\text{\hspace{0.17em}nM}$ MitoTracker™ Red CMXRos probe (Invitrogen) was added to each well for 15 min to stain the cells. Cells were triple-washed with sterile HBSS and fixed for 20 min in 4% wt/vol paraformaldehyde (PFA). Alexa Fluor®488 phalloidin (Invitrogen) was used at a 1:100 dilution for 30 min to stain F-actin, whereas nuclei were stained using Hoechst 33342 at a final concentration of $0.2\mathrm{\mu }\mathrm{g}/\mathrm{mL}$ for 7 min. The wells were washed with distilled water, and coverslips were mounted onto precleaned glass slides using Fluoromount™ (Sigma). Images were obtained using an inverted fluorescence Leica DMIRE2 confocal microscope with 63× magnification, and image analysis was performed using ImageJ. Quantification of mitochondrial length and degree of circularity was accomplished using a macro text file plug-in for ImageJ (version 1.53, National Institute of Health) software (Dagda et al. 2009). For image analysis, two separate experiments were performed and eight images per group were quantified.

To determine mitochondrial mass, 25,000 EGCs were plated per well in a 96-well plate. Cells were treated with 0, 10, or $100\mathrm{\mu }\mathrm{M}$ Mn for 24 h. After treatment, the media was removed, and $100\mathrm{\mu }\mathrm{L}$ of $200\text{\hspace{0.17em}nM}$ MitoTracker™ Green dye diluted in serum-free DMEM media was added to each well and incubated at 37°C for 15 min. Following incubation, green fluorescence was measured at $485/520\text{\hspace{0.17em}\hspace{0.17em}nm}$ (ex/em), respectively, using a fluorescence microplate reader. Nuclei were stained with Hoechst 33342 ($5\mathrm{\mu }\mathrm{g}/\mathrm{mL}$) for 15 min to determine total cell number in each well. The resulting green fluorescence readings were normalized to cell number. The experiment was repeated twice in duplicates or triplicates.

Determination of Mitochondrial Superoxide Production

To determine whether Mn exposure leads to mitochondrial superoxide production, 25,000 EGCs were plated per well in a 96-well plate. Following 24-h exposure to 0, 1, 3, 10, 30, 100, 300, or $\mathrm{1,000}\mathrm{\mu }\mathrm{M}$ Mn, media was removed, and cells were incubated with $5\mathrm{\mu }\mathrm{M}$ MitoSox™ Red and Hoechst 33342 ($5\mathrm{\mu }\mathrm{g}/\mathrm{mL}$) in HBSS with calcium and magnesium (HBSS/Mg/Ca) for 10 min and protected from light. Following incubation, cells were washed gently twice with HBSS/Mg/Ca, and fluorescence was measured at $510/580\mathrm{\hspace{0.17em}nm}$ (ex/em), respectively, using a fluorescence microplate reader. The resulting red fluorescence readings were normalized to cell numbers. Experiment was repeated twice, each time in duplicate.

Aconitase Assay

Mitochondrial aconitase (m-aconitase) levels in EGCs were measured using a commercial aconitase assay kit (Abcam, Cat. No. ab83459), as described by Langley et al. (2017), with minor modifications. Briefly, $1\times {10}^6$ EGCs were grown in a T25 cell culture flask and exposed to $10\mathrm{\mu }\mathrm{M}$ or $100\mathrm{\mu }\mathrm{M}$ Mn for 24 h. Cells were collected after treatment and resuspended in the supplied assay buffer on ice, followed by centrifugation at $2000\times g$ for 5 min at 4°C. The resulting supernatant was collected and further centrifuged at $\mathrm{20,000}\times g$ for 15 min to collect the mitochondrial fraction. The pellet was thoroughly mixed with $100\mathrm{\mu }\mathrm{L}$ of assay buffer and sonicated for 25 s at 4°C. Citrate was added as a substrate to each well. The plate was read at $450\mathrm{\hspace{0.17em}nm}$ using a microplate reader. The OD values were normalized to the protein concentration of each sample, and the results were expressed as a percentage of control. Experiment was repeated three times, each time in duplicate.

Mitochondrial Oxygen Consumption and Extracellular Flux Analysis

We used the XFe-24 Extracellular Flux Analyzer (Seahorse Bioscience) and the vendor’s Mito Stress kit for measuring the mitochondrial oxygen consumption rate (OCR) and adenosine triphosphate (ATP) production in EGCs treated with different doses of Mn. EGCs (30,000/well) were treated with 0, 10, or $100\mathrm{\mu }\mathrm{M}$ Mn for 24 h. The experiment was repeated two times in triplicate. No significant interassay differences were detected between the two repetitions regarding the appearance and viability of cells or the observed values of ${\mathrm{O}}_2$ consumption.

${}^3\left[H\right]$ Glutamate Uptake Assay

EGCs (30,000/well) were treated in triplicate with 0, 10, or $100\mathrm{\mu }\mathrm{M}$ Mn for 24 h. To remove media posttreatment, cells were washed in Krebs solution containing (in mM) 32 NaCl, 4 KCl, $1.2{\mathrm{Na}}_2{\text{HPO}}_4$, $1.4{\text{\hspace{0.17em}MgCl}}_2$, 6 glucose, 10 HEPES, and $1{\text{\hspace{0.17em}CaCl}}_2$ at pH 7.4. Cells were incubated with $2\mathrm{\mu }\text{Ci}$ L-[3,4 ${}^3\mathrm{H}$] glutamic acid for 15 min at 37°C. Media was removed, and cells were washed twice with ice-cold Krebs solution to stop the uptake. Cells were lysed in $0.5\mathrm{mL}$ 1N NaOH, and the total radiolabeled glutamate uptake was measured via liquid scintillation counting (Tri-Carb 4,000; Packard). Experiments were repeated thrice, each time in triplicate.

Animal Studies

We purchased 53 male 8- to 10-wk old C57BL/6 mice from Charles River Laboratories for all animal experiments. Mice were housed on a 12-h light cycle with ad libitum access to food [Teklad S-2335 mouse breeder diet containing low concentration of Mn $(75\mathrm{mg}/\mathrm{kg}$ Mn) Envigo] and water. After 2 weeks of acclimation, one-half of the mice ($n=27$) received $15\text{mg}/\mathrm{kg}$ body weight/d Mn (as ${\text{MnCl}}_2·4{\mathrm{H}}_2\mathrm{O}$), whereas the other half ($n=26$) received an equal volume of vehicle (water) via oral gavage for 30 d. It should be noted that the $15\mathrm{mg}/\mathrm{kg}$ dose has been adjusted for Mn, not ${\text{MnCl}}_2·4{\mathrm{H}}_2\mathrm{O}$. Our extensive review of the current literature to determine the dose and route of Mn administration (Crossgrove and Zheng 2004; Li et al. 2006; Taylor et al. 2020) determined that, for mice, Mn exposure paradigms range from $5–30\mathrm{mg}/\mathrm{kg}/\mathrm{d}$ for 21–120 d (Krishna et al. 2014; Moreno et al. 2009). Thus, we chose to use an oral dose of $15\mathrm{mg}/\mathrm{kg}$ Mn over a 4-wk period to avoid the excessive stress and overt toxicity associated with higher doses and different routes of administration. Animal care and all procedures strictly followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Iowa State University IACUC. GI function was measured every 10 d. At the end of the 30-d treatment regimen, animals ($n=6$) designated in each group for biochemical and neurochemical analyses were euthanized by ${\mathrm{CO}}_2$, and then $3\text{-cm}$ strips of the duodenum, ileum, and colon were collected and emptied of their contents by rinsing thoroughly in ice-cold PBS. Tissues were further sectioned into $1\text{-cm}$ segments for various biochemical tests and stored at $-80°\mathrm{C}$ until further processing. For immunohistological analyses, animals in each group ($n=5$) were transcardially perfused, and the ileum and colon were kept in 4% PFA. After 24 h, the mucosa and strips of circular muscle were removed to expose the myenteric neurons.

GI Functional Assay

Whole-gut transit time (WGTT).

On days 10, 20, and 30, WGTT was measured by ingestion of carmine red, a red dye that cannot be absorbed by the GI tract. Briefly, each mouse was administered $0.3\mathrm{mL}$ of 6% wt/vol carmine dye in 0.5% wt/vol methylcellulose via intragastric gavage as previously described (Ghaisas et al. 2019). WGTT was defined as the time taken from administration of the dye until the first appearance of a red stool pellet. Animals whose stool had no dye even after a maximum observation time of 8 h were recorded as $>8\mathrm{h}$.

Bead Latency

Distal colonic transit time (CTT) was measured in each mouse by the bead expulsion test (Vilz et al. 2012) on day 30. Mice were anesthetized with isoflurane (1–2 min), and a $2\text{-mm}$ diameter lubricated glass bead was introduced into the distal colon ($2\mathrm{cm}$ from the anus margin) using a fire-polished glass rod. After insertion of the bead, mice were isolated in a clear plastic cage for observation. Bead latency as a measure of CTT was defined as the time required for expulsion of the glass bead, which is typically less than 15 min.

Fecal Water Content

Thirty days post Mn exposure, each mouse was moved to a clear, empty plastic cage for 1 h to collect fecal pellets. Pellets were transferred to $1.5\text{-mL}$ centrifuge tubes and capped, and the total wet weight of each mouse’s fecal pellets was recorded. Thereafter, the tubes were placed uncapped in a 55°C incubator for 24 h to dry out the pellets. The tubes were recapped and weighed after drying to obtain dry weights. Water content was calculated by the formula $\left[\left(\text{wet\hspace{0.17em}weight}-\text{dry\hspace{0.17em}weight}\right)/\text{wet\hspace{0.17em}weight}\right]\times 100$.

Quantitative Real-Time Polymerase Chain reaction (qPCR)

Of the mice ($n=6$) designated in each group for biochemical and neurochemical analyses, distal colon from each control and Mn-treated group were used for qPCR. The samples were homogenized in a bullet blender (Next Advance) using DNase- and RNase-free stainless steel beads, and then RNA was extracted using RNeasy Plus Mini kit (Qiagen). Total RNA ($1\mathrm{\mu }\mathrm{g}$) was converted to cDNA using the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Inc.) following manufacturer’s instructions. All reactions were performed in triplicate. Mouse 18S rRNA was used as an internal standard for normalization. Validated QuantiTect primer sets for mouse 18S rRNA (Cat. No. QT02448075), $\text{TNF}\mathrm{\alpha }$ (Cat. No. QT00104006), and iNOS (Cat. No. QT00100275) were also used. The PCR cycling conditions contained an initial denaturation at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 30 s. Data were analyzed using the comparative threshold cycle (Ct) method. RNA extraction from primary enteric cultures ($n=6$) was done using the RNeasy Mini kit (Cat. No. 74104; Qiagen) and $1\mathrm{\mu }\mathrm{g}$ RNA was converted to cDNA (Cat. No. 4368814; ThermoFisher Sci.) according to the manufacturer’s instructions (Harischandra et al. 2020).

Western Blots

Whole-cell lysates or tissue lysates were prepared using modified radioimmunoprecipitation assay (RIPA) buffer containing (in mM) 20 Tris–HCl (pH 8.0), 2 EDTA, 10 EGTA, 2 dithiothreitol (DTT), 1 phenylmethylsulfonyl fluoride (PMSF), 20 sodium orthovanadate, 1 sodium fluoride and protease, and phosphatase inhibitor cocktail (Thermo Scientific), as described by Harischandra et al. (2014, 2015). The supernatants were obtained after centrifuging the homogenates at $\mathrm{15,000}\text{\hspace{0.17em}rpm}$ for 1 h. Protein concentration was measured using Bradford dye (Bio-Rad). Lysates containing equal amounts of protein were separated on a 12% or 15% wt/vol SDS-polyacrylamide gel. After separation, proteins were electro-blotted onto a nitrocellulose membrane, and nonspecific-binding sites were blocked with LI-COR blocking buffer. PGP9.5 (1:1000), GFAP (1:1200; Abcam, Cat. No. ab4648), iNOS (1:500; Abcam, Cat. No. ab15323), ferroportin (Fpn; 1:500; Alpha Diagnostics, Cat. No. MTP11-A), DMT-1 (1:500; Abcam, Cat. No. ab55812), DJ-1 (1:500; Santa Cruz, Cat. No. sc-55572), and $\mathrm{\beta }\text{-actin}$ (1:10,000; Sigma-Aldrich, Cat. No. A544) primary antibodies were used to blot the membranes for 16 h at 4°C. Thereafter, blots were washed in PBS containing 0.01% v/v Tween-20, incubated for 1 h in respective secondary antibodies IR-800–conjugated goat antimouse IgG (1:20,000; LI-COR, Cat. No. 926-68070) and IR-700–conjugated goat anti-rabbit IgG (1:20,000; LI-COR, Cat. No. 926-32211). Fluorescent bands corresponding to the protein of interest were observed by scanning membranes on a LI-COR Odyssey® scanner.

Immunohistochemistry

Immunohistochemistry was performed at the end of the treatment regime on all mice ($n=5$) in each group not designated for biochemical and neurochemical analyses. These mice were perfused first with PBS followed by 4% PFA. Colons were removed and postfixed in 10% formalin for an additional 24 h. Tissues were embedded in paraffin and sectioned at $7\mathrm{\mu }\mathrm{m}$ at Iowa State University’s veterinary pathology lab. Paraffin-embedded sections of mouse tissues were deparaffinized in decreasing alcohol grades. Heat-mediated antigen retrieval was performed using $10\text{\hspace{0.17em}mM}$ citrate buffer (pH 6.0) for 30 min followed by blocking antibody incubation steps as described by Harischandra et al. (2018). Tissues were incubated with the primary antibodies GFAP (1:1,000) and iNOS (1:500) for 16 h, washed three times in 1× PBS, and incubated with the respective secondary antibodies (1:1,500; Alexa Fluor® 488, Invitrogen, Cat. No. A-11008 or Alexa Fluor® 555, Cat. No. A-21422) for 1.5 h. After extensive washing and counterstaining with Hoechst 33342 ($5\mathrm{\mu }\mathrm{g}/\mathrm{mL}$) for 5 min, sections were dehydrated and mounted in DPX.

For alcian blue–periodic acid Schiff (AB-PAS) staining, the slides were deparaffinized in xylene and rehydrated in a decreasing ethanol gradient before final rinsing of the slides in distilled water. Slides were stained with 1% alcian blue in 3% acetic acid for 2.5 h, washed in running tap water and then distilled water for 2 min each, and then stained with the PAS staining method described by Lindén et al. (2008). Once washed in running tap water for 5 min, the slides were serially dehydrated, cleared, and mounted in DPX.

Inductively Coupled Plasma Mass Spectroscopy (ICP-MS)

Colon samples were carefully excised from mice designated for biochemical and neurochemical analyses and weighed prior to sample preparation. Samples were analyzed for cadmium (Cd), Ca, cromium (Cr), cobalt (Co), copper (Cu), iron (Fe), Mg, Mn, molybdenum (Mo), phosphorus (P), potassium (K), selenium (Se), sodium (Na), and zinc (Zn), using ICP-MS (Analytik Jena, Inc.) in CRI mode with hydrogen as the skimmer gas. Standards for elemental analyses were obtained from Inorganic Ventures, and digestion vessels, trace mineral grade nitric acid, and hydrochloric acid were obtained from Fisher Scientific. Briefly, samples were digested in 70% v/v nitric acid at 60°C for $\ge 12\mathrm{h}$. Preweighed samples were transferred to $15\text{-mL}$ tubes, and $0.25\mathrm{mL}$ of 70% nitric acid was added. All samples were digested overnight at 60°C. After digestion, all samples were diluted to $5\mathrm{mL}$ using 1% v/v nitric acid with 0.5% v/v hydrochloric acid and then analyzed by ICP-MS. For quality control, bismuth (Bi), scandium (Sc), indium (In), lithium (Li), yttrium (Y) and terbium (Tb) were used as internal standards for the ICP-MS.

Genomic DNA Isolation

Mice were kept in clean cages without bedding for 1 h, and fecal pellets were collected on days 10, 20, and 30 of dosing. Each time, the fecal pellets were promptly collected in $1.5\mathrm{mL}$ Eppendorf tubes and stored at $-80°\mathrm{C}$ until analysis. Genomic DNA isolation was performed using the PowerSoil® DNA Isolation kit (MO BIO) on samples from 10 mice per group for each time point. The purified genomic DNA extracts were quantified using a Qubit 2.0 Fluorometer (Life Technologies) and stored at $-20°\mathrm{C}$ in $10\text{\hspace{0.17em}mM}$ Tris buffer.

DNA Sequencing and Analysis

PCR amplification of the V4 variable region of the 16S rRNA gene using V4 region-specific barcoded primers (515F-816R) and amplicon sequencing were performed by the Institute for Genomics & Systems Biology at the Argonne National Laboratory (Argonne, Illinois) on the Illumina MiSeq platform. The sequences were analyzed using QIIME™ (Quantitative Insights into Microbial Ecology version 1.9; QIIME Development Team) as described by Caporaso et al. (2010). Briefly, reads were first demultiplexed and quality-filtered, and then nucleotide sequences with $>6$ homopolymer runs, $>6$ ambiguous bases, nonmatching barcodes, barcode errors, or quality scores $<25$ were removed. Operational taxonomic units (OTUs), defined by sequence similarities, were called using UCLUST and the closed reference OTU picking strategy in QIIME™ (Edgar 2010). Sequences were aligned to the Greengenes (13_8) database using PyNAST at 97% similarity (DeSantis et al. 2006). Taxonomic assignments were made based on the Greengenes reference database using the RDP classifier 2.2 in QIIME™ (Wang et al. 2007). Alpha diversity, beta diversity, principal coordinate analysis (PCoA) plots, analysis of similarity (ANOSIM), and Adonis tests were also generated using QIIME™. Wilcoxon rank-sum tests were performed on taxonomic summaries, obtained from the QIIME™ pipeline, using a custom R script (R Project for Statistical Computing) obtained from the Institute for Genome Sciences at the University of Maryland School of Medicine.

Gas Chromatography and Mass Spectroscopy Analysis (GC-MS)

To isolate the metabolites for the GC-MS analysis, we followed the methods described by Noutsos et al. (2015). Identification of fecal metabolites was done by comparing the mass spectral to the NIST08 Library and using retention indices, all relative to internal standards. Prior to metabolite extraction, $\sim 20–30\mathrm{mg}$ of fecal matter was weighed, spiked with internal standards ($20\mathrm{\mu }\mathrm{g}$ of ribitol and $20\mathrm{\mu }\mathrm{g}$ of nonadecanoic acid), and homogenized with $0.35\mathrm{mL}$ of hot methanol (60°C). The mixture was immediately incubated for 10 min at 60°C and sonicated for 10 min, and $0.3\mathrm{mL}$ each of chloroform and water was added to this mixture and vortexed for 3 min. After centrifugation to separate phases, $0.2\mathrm{mL}$ of the upper polar phase and $0.2\mathrm{mL}$ of the lower nonpolar phase were removed and transferred into $2\text{-mL}$ glass vials. Both fractions were dried in a SpeedVac concentrator. The extracts were subjected to methoximation with methoxyamine hydrochloride at 30°C for 90 min. Samples were silylated with BSTFA/TMCS at 60°C for 30 min and then subjected to GC-MS (Agilent Technologies) on a 6890N gas chromatograph in tandem with a 5973MSD detector equipped with a gas ionization detector (gas flow: UHP Helium, $1.0\mathrm{mL}/\text{min}$; ionization mode: El; polarity: positive; skimmer/focusing lens voltages: 70 eV). Samples were added to the instrument with a 7683B automatic liquid sampler. The mass range was set from $40\text{\hspace{0.17em}to\hspace{0.17em}}\mathrm{1,000}m/z$. The separation column was an HP5MSI ($30\mathrm{m}$ long, $0.250\text{-mm}$ ID, $0.25\text{-}\mathrm{\mu }\mathrm{m}$ film thickness, Agilent Technologies). The gas chromatography was operated by ChemStation software (version A.10.02).

Data Analysis

Data analysis was performed using Prism 6.0 Software (GraphPad Prism). In vitro data with $n<6$ were assessed by nonparametric Wilcoxon rank test, and Student’s $t$-test was used for data with $n>6$. One-way analysis of variance (ANOVA) (with Bonferroni’s correction for comparison of multiple means) or, where appropriate, the respective nonparametric Kruskal-Wallis test (with Dunn’s correction for comparison of multiple means) was applied for comparative analyses. Significance was defined as *$p<0.05$, **$p<0.01$, or ***$p<0.001$, respectively. The metabolites’ log-concentrations were analyzed using mixed-effects models that included the fixed main effects due to group (two levels: Control and Mn), fixed main effects due to the time points (3 levels: 10, 20, 30) interaction between the two factors, and mouse-specific random effects. A false discovery rate (FDR) adjustment was used to account for multiplicity of the overall F-tests for each of the metabolites. The mixed-effects models were fitted and analyzed by R statistical software using lmerTest and emmeans packages.

Results

Cytotoxicity and Mitochondrial Viability in EGCs Exposed to Mn for 24 h

To determine the effect of Mn on EGCs, we performed dose–response cytotoxicity and mitochondrial viability assays in an immortalized EGC line. EGCs treated with increasing doses of Mn ($0–\mathrm{1,000}\mathrm{\mu }\mathrm{M}$) for 24 h had lower metabolic activity than those treated with vehicle. This difference appeared to be dose-dependent. The ${\text{IC}}_{50}$ was calculated by a three-parameter nonlinear regression curve and found to be about $5\mathrm{\mu }\mathrm{M}$ (Figure 1A). EGCs treated with 1, 3, 10, and $30\mathrm{\mu }\mathrm{M}$ Mn for 24 h showed lower mitochondrial mass by about 10%, 30%, 32%, and 60%, respectively, in comparison with untreated controls as measured by MitoTracker™ Green (Figure 1B). We performed a MitoTracker™ Red assay to contrast the mitochondrial morphology of treated and untreated EGCs. Untreated cells exhibited the characteristics of healthy mitochondria, with long threadlike strands densely surrounding the nucleus, whereas those treated with Mn became rounded and began to disintegrate (Figure 1C). Cells treated with 10 and $100\mathrm{\mu }\mathrm{M}$ Mn, respectively, showed a 19.1% and 28.1% increase in circularity when compared with untreated controls (Figure 1D). Similarly, lower mitochondrial aconitase (m-aconitase) activity was observed in EGCs treated with $10\mathrm{\mu }\mathrm{M}$ and $100\mathrm{\mu }\mathrm{M}$ Mn (Figure 1E). We also observed higher dose-dependent mitochondrial superoxide production (Figure 1F). Next, we hypothesized that Mn exposure impairs basal mitochondrial respiration as well as spare respiratory capacity. Hence, we measured the OCR of $100\mathrm{\mu }\mathrm{M}$ Mn-treated and untreated EGCs as an indicator of mitochondrial respiration (Figure 1G) and found that Mn-treated EGCs had significantly lower ATP-linked respiration in comparison with controls (Figure 1H).

Figure 1.

Cell viability and mitochondrial dynamics following 24-h Mn exposure. (A) Dose–response (${\text{IC}}_{50}$) curve indicating the viability of EGCs treated with increasing concentrations of Mn for 24 h as measured by MTS assay. (B) Mitochondrial mass of control or 1, 3, 10, or $30\mathrm{\mu }\mathrm{M}$ Mn-treated cells measured by MitoTracker™ Green FM. (C) Mitochondrial morphology of 0, 10, 100, or $1000\mathrm{\mu }\mathrm{M}$ Mn-treated cells as measured by MitoTracker™ Red CMXRos. $\text{Scale\hspace{0.17em}bar}=10\mathrm{\mu }\mathrm{m}$. (D) Mitochondrial circularity expressed as percent control determined by ImageJ analysis of mitochondria from seven different images/group. (E) Mitochondrial aconitase activity of EGCs treated with 10 or $100\mathrm{\mu }\mathrm{M}$ Mn for 24 h as measured using a commercial kit per manufacturer’s instructions. (F) Dose-dependent increase in mitochondrial superoxide on Mn treatment as measured by MitoSox™ dye fluorescence. (G) OCR of EGCs treated with 0 and $100\mathrm{\mu }\mathrm{M}$ Mn for 24 h as measured by Seahorse $\text{XF}{24}^{\mathrm{e}}$. (H) ATP production of untreated and Mn-treated EGCs after 24-h exposure. Data represented as $\text{mean}\pm \text{SD}$ (*$p<0.05$, **$p<0.01$, ***$p<0.001$; ns) of 4 to 10 replicates as determined by either one-way analysis of variance or Student’s unpaired $t$-test. Summary data for (B), (D), (E), (F), and (H) can be found in Table S2. Note: ATP, adenosine triphosphate; EGC, enteric glial cell; Mn, manganese; MTS, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium; ns, nonsignificant; OCR, oxygen consumption rate; SD, standard deviation.

Cell Viability in EGCs Exposed to Mn for 24 h

Interestingly, when the same Mn treatment paradigm was used to measure cytotoxicity in EGCs, we found that only EGCs treated with Mn doses $\ge 300\mathrm{\mu }\mathrm{M}$ stained positively for SYTOX™ green, a dye that only passes through the compromised cell membranes of dead cells (Figure 2A,B). We further validated this observation by demonstrating that only EGCs exposed to $\ge 300\mathrm{\mu }\mathrm{M}$ Mn for 24 h exhibited significant caspase-3 activity (Figure 2C). Moreover, EGCs treated with $100\mathrm{\mu }\mathrm{M}$ Mn had $>50\%$ lower glutamate uptake in comparison with untreated controls (Figure 2D). Even Mn doses of $5–10\mathrm{\mu }\mathrm{M}$ resulted in $\sim 10\%$ lower glutamate uptake.

Figure 2.

Enteric glial cytotoxicity, caspase-3 activity, and glutamate uptake following 24-h Mn exposure. (A–B) Phase contrast and green fluorescence superimposed images of EGCs treated with different doses of Mn ($1–\mathrm{1,000}\mathrm{\mu }\mathrm{M}$) for 24 h. (C) Caspase-3 activity in EGCs treated with increasing doses of Mn for 24 h. (D) Radiolabeled glutamate uptake assay following EGC exposure to 0, 10, or $100\mathrm{\mu }\mathrm{M}$ Mn. Data represented as the group $\text{mean}\pm \text{SD}$ from at least 3 experiments. Asterisks (***$p<0.001$, **$p<0.01$ and *$p<0.05$) indicate significant differences between Mn-treated and control groups as determined by one-way analysis of variance. Summary data for (B), (C), and (D) can be found in Table S3. Note: EGC, enteric glial cell; Mn, manganese; SD, standard deviation.

Markers of Inflammation in Primary Enteric Culture Exposed to Mn for 24 h

Primary enteric mixed cultures treated with increasing doses of Mn ($0–\mathrm{1,000}\mathrm{\mu }\mathrm{M}$) for 24 h showed lower metabolic activity, and the ${\text{IC}}_{50}$ was calculated to be about $100\mathrm{\mu }\mathrm{M}$ (Figure 3A). At this dose, Mn-treated cells had $\sim 3\text{-fold}$ higher expression of $\text{TNF}\mathrm{\alpha }$ and $\sim 36\text{-fold}$ higher expression of iNOS mRNA transcripts in comparison with control cells (Figure 3B). Lower expression of the enteric neuronal marker PGP9.5 was observed after the 24-h treatment (Figure 3C,E). Mn treatment did not significantly alter the expression of DMT1 ($n=6$; $p<0.94$)—a transporter involved in the influx of divalent metals (including Mn)—or Fpn ($n=6$; $p<0.38$), a Mn-responsive protein that transports Mn out of the cell (Figures 3C,E). A trend of lower GFAP expression was revealed ($n=3$; $p<0.068$) by Western blot (Figures 3D,E).

Figure 3.

Effect of 24-h Mn exposure on primary enteric mixed culture. (A) Cell viability of primary enteric mixed cultures treated with increasing doses of Mn for 24 h as measured by MTS assay, with values normalized to control and expressed as percent control ($n=8$). (B) Markers of inflammation ($\text{TNF}\mathrm{\alpha }$ and iNOS) in primary enteric culture exposed to $100\text{-}\mathrm{\mu }\mathrm{M}$ Mn for 24 h normalized to 18S rRNA expression ($n=6$). (C) Western blot analysis of the enteric neuron marker PGP9.5 and the Mn transporters DMT1 and Fpn expression in untreated and $100\text{-}\mathrm{\mu }\mathrm{M}$ Mn-treated primary enteric mixed cultures ($n=6$). (D) Representative Western blot image of GFAP expression following $100\text{-}\mathrm{\mu }\mathrm{M}$ Mn exposure ($n=3$). (E) Densitometry of Western blots are expressed as arbitrary units relative to control. Data represented as the group $\text{mean}\pm \text{SEM}$ from at least 3 experiments. Asterisks (***$p<0.001$, **$p<0.01$ and *$p<0.05$) indicate significant differences between Mn-treated and control groups as determined by either one-way analysis of variance or Student’s unpaired $t$-test. Summary data for (B) and (E) can be found in Table S4. Note: Fpn, ferroportin; GFAP, glial fibrillary acidic protein; Mn, manganese; MTS, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium; SEM, standard error of the mean.

Smooth Muscle Function in Primary Enteric Culture Exposed to Mn for 24 h

Primary enteric mixed cultures grown for approximately 3 wk on PDL- and laminin-coated plates show extensive neuronal networks (red arrows) growing over a bed (yellow arrow) of smooth muscle and epithelial cells (see Figure S1). Other cell types, possibly endothelial cells, smooth muscle cells and fibroblasts were also present in the culture. At 3–4 wk post plating, isolated contractions were seen in different parts of the culture (Supplementary Video A). As shown in the video, these spontaneous contractions gradually reduced in cultures treated with $100\mathrm{\mu }\mathrm{M}$ Mn for 24 h (Supplementary Video B.).

GI Transit Time and Microscopic Analysis of Tissue Architecture in Vehicle- and Mn-Treated Ileum and Colon

To assess GI changes following increased ingestion of Mn, mice were given a dose of $15\mathrm{mg}/\mathrm{kg}/\mathrm{d}$ for 30 d (Figure 4A). Chronic administration of low-dose Mn transiently lowered peristalsis. Mn-treated mice showed a trend for WGTT that differed from controls from day 20 onward (Figure 4B). On day 30, these mice took an average of $357.1\pm 11.76$ min to expel the carmine red meal fed to them in comparison with the $328.1\pm 13.57$ min it took for control mice. Despite WGTT averaging longer in Mn-treated mice, we did not observe any significant difference in fecal water content between the groups at day 30 (Figure S2). With regard to CTT, Mn-treated mice took $560.0\pm 87.18$ seconds to expel the glass bead as opposed to $394.7\pm 53.77$ seconds for control mice (Figure 4C). Microscopic analysis of tissue architecture in Mn-treated ileum and colon by hematoxylin and eosin (H&E) staining showed no signs of cell necrosis and immune cell infiltration (Figure 4D). Qualitative assessment of colon tissues from both groups revealed less intense PAS staining as well as fewer goblet cells in the Mn group, indicating a lower production of these mucopolysaccharides (Figure 4E).

Figure 4.

GI transit time and microscopic analysis of tissue architecture in control and Mn-treated gastrointestinal tracts. (A) Schematic representation of the Mn trt paradigm in C57 mice. (B) Whole-gut transit time as measured by carmine-red meal over the course of the study ($n=27$) and analyzed by two-way analysis of variance. (C) Bead latency test to determine distal colon transit time. Data represented as the group $\text{mean}\pm \text{SEM}$ from $n=17–18$ per group and analyzed by Student’s unpaired $t$-test. (D) Representative 20× [20 times] magnification images of H&E-stained colon sections from untreated and Mn-treated groups ($n=3$). $\text{Scale\hspace{0.17em}bar}=150\mathrm{\mu }\mathrm{m}$. (E) Representative $20\times$ magnification images of AB-PAS-stained colon sections. $\text{Scale\hspace{0.17em}bar}=150\mathrm{\mu }\mathrm{m}$. Summary data for (B) and (C) can be found in Table S5. Note: AB-PAS, alcian blue–periodic acid Schiff; GI, gastrointestinal; H&E, hematoxylin and eosin; Mn, manganese; Mn trt, Mn treatment; SEM, standard error of the mean.

ICP-MS and Biochemical Analysis of Colon and Liver Samples from Control and Mn-Treated Mice

ICP-MS analyses of colon samples (Figure 5A) and liver samples (Figure S3) from both groups showed significantly higher tissue levels of Mn in the Mn-treated group over control group. Tissue levels of other metals did not differ significantly between groups (see Supplemental Table, “ICP-MS analysis of colon tissue”). We performed qRT-PCR and Western blot analyses to probe for inflammatory markers. Our qRT-PCR analysis showed 2.8- and 2.6-fold higher mRNA transcripts of $\text{TNF}\mathrm{\alpha }$ and iNOS, respectively, in the colons of Mn-treated animals (Figure 5B). Furthermore, Western blotting showed that Mn exposure resulted in higher expression of iNOS (Figure 5C,D; $n=5$, $p<0.04$) in colon lysates. To ascertain whether this Mn-induced iNOS expression in part occurs in the ENS, whole-mount sections containing the myenteric plexus were probed for GFAP and iNOS. Indeed, immunofluorescence images revealed a greater punctate presence of iNOS in the myenteric ganglia of Mn-treated animals (Figure 5E). Western blot analysis of full-thickness colons ($n=6$) from both groups showed a higher expression of the EGC marker GFAP ($p<0.007$) as well as lower expression of the mitochondrial protein DJ-1 ($p<0.034$) (Figure 5F,G). Levels of the enteric neuronal marker—PGP9.5—did not differ significantly ($p<0.17$). The mRNA expression of the major Mn exporter Slc30a10 as well as two other metal ion transporters, Slc39a8 and Slc39a14 (ZIP14), did not show any significant differences between groups (Figure S4). However, we found that although the protein levels of DMT1 remained relatively similar ($p<0.85$), Fpn levels were relatively lower in the Mn group (Figure 5F,G, $p<0.08$), suggesting a possible dysfunction in the export of metal from the cell.

Figure 5.

Biochemical analysis of colons of control and Mn-treated mice. (A) ICP-MS data showing Mn content in colon tissue. (B) qRT-PCR analysis of $\text{TNF}\mathrm{\alpha }$, $\text{IL-}1\mathrm{\beta }$, and iNOS mRNA transcripts, normalized to 18S rRNA expression. (C) Western blot and (D) densitometry of iNOS in colon samples from vehicle and Mn-treated mice. Results expressed as arbitrary units relative to control. (E) Representative 40× immunofluorescence images of the myenteric plexus of a vehicle- and a Mn-treated mouse probed for GFAP and iNOS. $\text{Scale\hspace{0.17em}bar}=100\mathrm{\mu }\mathrm{m}$. Arrows indicate presence of punctate iNOS staining. (F) Representative Western blot showing expression of the EGC marker GFAP, mitochondrial protein DJ-1, the Mn transporters Fpn and DMT-1, and enteric neuronal marker PGP9.5. Results expressed as arbitrary units relative to control. (G) Densitometry of Western blots. Data represented as the group $\text{mean}\pm \text{SEM}$ from $n=6–8$ per group. Asterisks (*$p<0.05$ and **$p<0.01$) indicate significant differences between control and Mn-treated groups as determined by Student’s unpaired $t$-test. Summary data for (A), (B), (D), and (G) can be found in Table S6. Note: EGC, enteric glial cell; Fpn, ferroportin; GFAP, glial fibrillary acidic protein; ICP-MS, inductively coupled plasma mass spectroscopy; Mn, manganese; qRT-PCR, quantitative real-time polymerase chain reaction; SEM, standard error of the mean; $\text{TNF}\mathrm{\alpha }$, tumor necrosis factor alpha.

Taxonomic Abundances of Gut Bacteria in Control and Mn-Treated Mice

To determine whether Mn exposure perturbs gut bacteria, we analyzed potential differences in the gut bacterial population between age-matched untreated and Mn-treated mice using 16S rRNA gene sequencing. A total of 7,152,846 reads passed quality filtering with an average of $\mathrm{89,410.58}\pm \mathrm{21,937.2}$ sequences per sample. Wilcoxon rank-sum tests were used to compare same-day taxonomic abundances between the control and Mn-treatment groups (i.e., Day 0 control vs Day 0 Mn-treated) with 10 samples per treatment group and results summarized at the phylum and genus levels (Figure 6A,B). At the phylum level, Bacteroidetes was the dominant phylum (control: 63.75%–73.83%; Mn: 64.87%–73.41%), followed by Firmicutes (control: 24.42%–34.41%; Mn: 23.17%–32.27%). Comparing same-day taxonomic abundances across phyla, no statistically significant differences emerged between control and Mn-treated groups.

Figure 6.

Taxonomic abundances of gut bacteria in control and Mn-treated mice. Taxonomic summaries at the (A) phylum and (B) genus levels. (C) Rarefaction plots of observed species richness and (D) Faith’s phylogenetic diversity. (E) Weighted UniFrac PCoA plots where each point is a single sample, and each color represents a group. Note: Mn, manganese; PCoA, principal coordinates analysis.

At the genus level (Figure 6B), the dominant genera were Bacteroides, Parabacteroides, and an unknown genus in the family S24-7 (phylum Bacteroidetes). The two most dominant genera in the phylum Firmicutes were Lactobacillus and an unknown genus in the order Clostridiales. In both groups, S24-7 increased over the 30-d period, whereas Bacteroides decreased, although none of the differences between the control and Mn-treated mice were statistically significant at any time point.

Alpha diversity metrics were used to compare within-sample diversity. Nonparametric two-sample $t$-tests using Monte Carlo permutations were used to compare the observed OTUs (Figure 6C) and Faith’s (1992) phylogenetic diversity (Figure 6D). Comparisons of alpha diversity values showed that control and Mn-treated mice did not differ significantly at any time point. Mn treatment had no effect on microbial diversity over the 30-d period. The observed OTU rarefaction plot shows that the curves plateaued, indicating the sequencing depth was sufficient to exclude the discovery of new or rare OTUs.

Weighted (Figure 6E) UniFrac PCoA plots were generated using QIIME™ Groupings. Adonis and ANOSIM tests found no obvious clustering based on treatment and sample date, whereas statistical tests indicated no differences between groups at any time point. More specifically, adonis tests ($p=0.06$, $F\text{\hspace{0.17em}value}=1.51$, and $R^2=0.1293$) suggest the clustering accounted for only 12.93% of the variation seen, whereas ANOSIM tests ($p=0.19$; $\text{test\hspace{0.17em}statistic}=0.0262$) indicated weak grouping based on treatment and sample date.

Fecal Metabolite Analysis of Mice Treated with or without Mn

GC-MS was applied to screen fecal metabolites from mice in this study. A total of 479 metabolites belonging to 13 categories were identified using the National Institute of Standards and Technology (NIST) 08 library. From 10 to 30 d post Mn exposure, metabolites such as certain long-chain fatty acids were significantly lower at days 20 and 30 post exposure in comparison with the control group (Figure 7A,C). Sugars such as erythrose and 6-phosphate D-fructose and amino acids such as threonine and glycine showed significant temporal variations in comparison with the control group (Figure 7B,D). Of the various classes of differentially expressed metabolites altered by 30 d of Mn exposure, the vast majority were amino acids, sugars, and fatty acids, as well as glycolytic intermediates (Figure 7E–F).

Figure 7.

Fecal metabolite analysis following Mn exposure. (A) Top 18 Polar and (B) top 12 nonpolar heat maps of the Mn effects on metabolites across three time points. Paneled plots of selected polar (C) and nonpolar (D) metabolites affected by Mn exposure across the three time points. ($n=5-6$). Pie charts denoting global changes in different classes of fecal metabolites at day 30 in (E) control samples and (F) Mn-exposed samples. Note: Mn, manganese.

Discussion

Since the discovery by Braak et al. (2006) that aggregated proteins—a hallmark of many neurodegenerative diseases including PD—exist in the ENS of a number of early PD patients, scientific interest in the possibility of enteric dysfunction preceding CNS dysfunction has increased. Studies have reported that exposure (via ingestion or inhalation) to environmental toxins is associated with the development of neurological disorders (Gomez-Mejiba et al. 2009; Rahman et al. 2017). Metals like Mn, aluminum, mercury, and lead are known to be neurotoxic when present in high amounts in the body (Farina et al. 2013). Although Mn toxicity in the brain has been studied extensively, scant attention has been paid to its possible toxic influence on enteric neurons despite cases of GI disturbances having been reported following chronic Mn exposure (Chandra and Imam 1973; Huang and Lin 2004). The present study revealed—for the first time, to our knowledge—that subchronic exposure to low doses of Mn can result in intestinal inflammation and metabolic dysregulation as evaluated in vitro and in a mouse model.

To investigate the effects of chronic exposure to low-dose Mn on the GI tract’s complex network of enteric neurons and glia, we first examined its effects in vitro through the lens of mitochondrial stress in pure EGC cultures and the inflammatory response in primary enteric mixed cultures. EGCs are known to project into the mucosal villi, playing a crucial role in maintaining the intestinal barrier as well as cross-talking with immune cells and enteric neurons (Gulbransen and Sharkey 2012; Neunlist et al. 2013; Savidge et al. 2007). Recent evidence suggests that some GI disorders occur partly due to a dysregulation of EGCs (von Boyen and Steinkamp 2010; Yu and Li 2014). Because EGCs can come into direct contact with Mn as it is absorbed via the enterocytes (Ye et al. 2017) on the villi, we focused on the effect of Mn on EGCs. We observed that EGCs, and in particular their mitochondria, appeared to be particularly sensitive to Mn. Although mitochondria showed signs of damage induced by doses $\ge 30\mathrm{\mu }\mathrm{M}$ Mn, the cells per se did not undergo apoptosis until prolonged exposure to $\ge 300\mathrm{\mu }\mathrm{M}$ Mn. At lower, environmentally relevant doses, a window possibly exists wherein cells, although stressed, do not succumb quickly. Do stressed enteric glia function as effectively as normal enteric glia? To answer this question, we analyzed the glutamate uptake capacity of Mn-exposed EGCs. Glutamate is a major neurotransmitter produced by the ENS and glia in particular and plays a role in peristalsis as well as in the bi-directional signaling of the gut–brain axis (Filpa et al. 2016). Recent evidence has shown the presence of glutamate receptors on EGCs whose uptake of extracellular glutamate prevents neuronal excitotoxicity (Panickar and Norenberg 2005; von Boyen et al. 2006). Mn-treated EGCs’ ability to recover glutamate declined in a dose-dependent manner. We also found that stressed EGCs accelerated iNOS production and released proinflammatory factors. We postulate that stressed EGCs in turn affect surrounding nonglial cells, activating or suppressing immune cells or contributing to stressed enteric neurons (Bogdan 2001; Capoccia et al. 2015), although more research is needed in this area.

The ENS, in conjunction with different neuropeptides, Cajal cells, and enteric smooth muscle cells, controls the peristaltic movement of ingested food through the GI system. This synchrony is necessary for absorption of food and micronutrients (in the small intestine) as well as water and salt retention and fecal production (in the large intestine) (Huizinga 1999). Our in vivo data showed that enteric neurons were more resistant to Mn toxicity. However, they can be susceptible to stress induced by proinflammatory factors released by activated and stressed EGCs (Brown et al. 2016). Also, both primary enteric cultures and mice subchronically treated with low-dose Mn showed altered smooth muscle activity. Although slower peristalsis can lead to bacterial overgrowth and lower nutrient absorption, in our study, mice given $15\mathrm{mg}/\mathrm{kg}/\mathrm{d}$ Mn showed only a nominal increase in intestinal transit time after day 20 in comparison with controls. Mice were randomized based on weight, yet we observed an initial nonsignificant trend of faster GI transit time in the Mn-treated mice in comparison with that of controls. Intriguingly, on day 20 of the exposure paradigm, this trend was reversed and remained so until the end of the study. We are not certain why that happened, yet the sudden shift in transit time only occurred in the Mn exposure group and not in the control group. This difference may have confounded our statistical analysis.

As with many epidemiological studies that detail human exposure to various environmental factors, it is difficult to state a definite dose beyond which toxicity occurs. A definite dose is difficult to ascertain because multiple factors such as genetics, age, diet, a patient’s health status, and surrounding environment can play a role in determining susceptibility to that particular environmental toxin. In the case of Mn toxicity, particularly via ingestion, researchers have documented that continued ingestion of $1.8–14\mathrm{mg}\mathrm{Mn}/\mathrm{L}$ by individuals can result in Mn toxicity (Aschner et al. 2005). In a preclinical setting, Mn overexposure leading to significant motor impairment was observed in mice treated with either $59\mathrm{mg}/\mathrm{kg}$ Mn for 10 wk (Vezér et al. 2007) or $100\mathrm{mg}/\mathrm{kg}$ Mn for 8 wk (Liu et al. 2006). These mice also showed weight loss and other signs of sickness. However, because the aim of our study was to characterize changes in gut physiology, weight loss would confound any inference we derived from this study, which is why a nontoxic dose of $15\mathrm{mg}/\mathrm{kg}/\mathrm{d}$ for 30 d was chosen. In this study, mice accumulated significantly more Mn in the colon in comparison with vehicle (water)-treated mice. Indeed, this impairment in Mn homeostasis was confirmed by probing for the common Mn cellular transporters DMT1 and Fpn. DMT1 is involved in the influx, and Fpn is involved in the efflux of divalent ions from cells. Our findings of lower Fpn (protein) and Slc30a10 (mRNA) but no differences in DMT1 (protein) expression in the colons of Mn-exposed mice could account for the increase in tissue Mn content in this group. Interestingly, because no statistically significant group differences were observed with respect to other trace metals, it is likely that excess Mn in intestinal tissues is the result of a disrupted enteric Mn homeostasis.

Western blot, qRT-PCR, and myenteric plexus immunofluorescence all showed increased iNOS production in the colons of Mn-treated mice. Given that repeated toxin exposure is thought to generate oxidative stress, intestinal inflammation, and to subsequently potentiate protein aggregation (House and Tansey 2017), future experiments involving long-term exposure to Mn and probing for protein aggregation in the enteric neurons need to be conducted.

In recent years, researchers have increasingly recognized the importance of gut microbiota in driving enteric inflammation as well as maintaining host health. In particular, researchers are exploring the theory that the gut microbiome plays an important role in certain neurological disorders such as autism and PD (Ghaisas et al. 2016; Mulle et al. 2013; Scheperjans et al. 2015). Although it is not known exactly how the enteric microbiota initiate or aggravate neurological dysfunctions, it is thought that ingesting or inhaling environmental toxins affects microbial populations either by selectively enriching certain species of pathogenic bacteria or by reducing the number of beneficial bacteria, thus driving inflammation. Because we found inflammation in the colon, we characterized the colonic bacteria of control and Mn-treated mice. Although we were unable to detect significant differences early during the treatment regime, a trend for increased Gammaproteobacteria was observed around 20 d post Mn exposure. An even longer exposure to Mn could result in statistically significant changes in the composition of the gut microbiota. Indeed, Chi et al. (2017) reported that mice exposed to a higher dose of Mn for 13 wk had significantly altered the amounts of bacterial genes and key metabolites involved in neurotransmitter synthesis. Changes in fecal metabolite composition indicate either altered bacteria-to-host signaling or vice versa. Despite the low-dose exposure, the Mn-treated group showed changes in certain fatty acids, amino acids, and sugars relative to the control group.

To conclude, in our in vitro exposure models, Mn exposure affected the ENS even at low doses, affecting mainly the EGCs by causing mitochondrial damage. In vivo, Mn exposure triggered inflammation in the gut, leading to inflammation and slower peristalsis. Owing to the communication between the ENS and CNS via the gut–brain axis, future experiments need to be carried out to ascertain whether ENS dysregulation contributes to the gradual impairment in CNS functioning that precedes many neurological disorders.