In vitro Characterization of Gut Microbiota-Derived Bacterial Strains With Neuroprotective Properties

Abstract

Neurodegenerative diseases are disabling, incurable, and progressive conditions characterized by neuronal loss and decreased cognitive function. Changes in gut microbiome composition have been linked to a number of neurodegenerative diseases, indicating a role for the gut-brain axis. Here, we show how specific gut-derived bacterial strains can modulate neuroinflammatory and neurodegenerative processes in vitro through the production of specific metabolites and discuss the potential therapeutic implications for neurodegenerative disorders. A panel of fifty gut bacterial strains was screened for their ability to reduce pro-inflammatory IL-6 secretion in U373 glioblastoma astrocytoma cells. Parabacteroides distasonis MRx0005 and Megasphaera massiliensis MRx0029 had the strongest capacity to reduce IL-6 secretion in vitro. Oxidative stress plays a crucial role in neuroinflammation and neurodegeneration, and both bacterial strains displayed intrinsic antioxidant capacity. While MRx0005 showed a general antioxidant activity on different brain cell lines, MRx0029 only protected differentiated SH-SY5Y neuroblastoma cells from chemically induced oxidative stress. MRx0029 also induced a mature phenotype in undifferentiated neuroblastoma cells through upregulation of microtubule-associated protein 2. Interestingly, short-chain fatty acid analysis revealed that MRx0005 mainly produced C1-C3 fatty acids, while MRx0029 produced C4-C6 fatty acids, specifically butyric, valeric and hexanoic acid. None of the short-chain fatty acids tested protected neuroblastoma cells from chemically induced oxidative stress. However, butyrate was able to reduce neuroinflammation in vitro, and the combination of butyrate and valerate induced neuronal maturation, albeit not to the same degree as the complex cell-free supernatant of MRx0029. This observation was confirmed by solvent extraction of cell-free supernatants, where only MRx0029 methanolic fractions containing butyrate and valerate showed an anti-inflammatory activity in U373 cells and retained the ability to differentiate neuroblastoma cells. In summary, our results suggest that the pleiotropic nature of live biotherapeutics, as opposed to isolated metabolites, could be a promising novel drug class in drug discovery for neurodegenerative disorders.

Keywords: gut microbiota-derived bacterial strains; gut-brain axis; microbiome; neurodegenerative diseases; neuroinflammation; neuroprotection; oxidative stress; short-chain fatty acids.

FIGURE 1

Differential mode of action between MRx0005 and MRx0029: implication for neuroinflammation. U373 cells were pre-treated with 1 μg/ml LPS followed by 10% bacterial cell-free supernatants (BCFS) or media alone. Cell-free supernatants were collected 24 h after treatment from U373 cells and secretion of IL-6 was measured by ELISA. Data are mean ± SEM (n = 2) (A). U373 cells were treated with 1 μg/ml LPS followed by 10% BCFS from MRx0005 and MRx0029 and secretion of IL-6 (B) and IL-8 (C) measured by ELISA. Controls included also cells treated with BCFS alone. Data are mean ± SEM (n = 5 for IL-6, n = 4 for IL-8). HMC3 cells were pre-treated with TNF-α and then with 10% BCFS from MRx0005 and MRx0029 or media alone for 24 h and IL-6 (D) and IL-8 (E) measured by ELISA. Controls included also cells treated with BCFS alone. Data are mean ± SEM (n = 3). (F) HEK-Blue hTLR4 cells were treated with 10 ng/ml LPS in the presence of 10% BCFS from MRx0005 and MRx0029 or media as control. Controls included also cells treated with BCFS alone. After 22 h, NF-κB-induced SEAP activity was measured using QUANTI-Blue at 655 nm. Data are mean ± SEM (n = 5).

FIGURE 2

Antioxidant capacity of MRx0005 and MRx0029: Implication for neuro-protection. MRx0005 and MRx0029 bacteria were grown to stationary phase. Biochemical assays were performed as described in section “Material and Methods.” (A) Indole concentration (mM); (B) DPPH scavenging activity assay reported as percentage related to positive control compound U83836E; (C) Total antioxidant capacity relative to Trolox concentration (mM). Each value in the graph represents the mean of three experiments corresponding to three cultures of MRx0005 and MRx0029. Data are mean ± SEM (n = 3). (D) U373 cells, (E) HMC3 cells, and (F) differentiated SH-SY5Y cells were stained with DCFDA probe, then treated with 10% BCFS from MRx0005 and MRx0029 or media in the presence of TBHP for 2 h, followed by measurement of fluorescence intensity at Ex/Em = 485/530 nm. Data are mean ± SEM (respectively n = 4, n = 3, and n = 4).

FIGURE 3

MRx0029 induced SH-SY5Y cells differentiation. Undifferentiated SH-SY5Y cells were treated with 10% MRx0005, MRx0029, media, 10 μM RA for 24 h. (A) Representative images using phase contrast microscope to show differences in morphology. Magnification 400×. Scale bar = 10 μm (B) Treated cells as described above were collected, total RNA isolated for qPCR to quantify MAP2 and SYP gene expression. Data are mean ± SEM (n = 3). (C) (a–c) Representative images of immunolabelled cells with Phalloidin (red) and MAP2 (green), (d–f) and merged with DAPI (blue) images; (g–i) β3-tubulin immunolabelled cells (green) and (j–l) merged with DAPI (blue) images. Magnification 630×. Scale bar = 5 μm. (D,E) Western blot analysis of effects of MRx0005 and MRx0029 treatment on SH-SY5Y cells. Western blot membranes were probed with antibodies to (D) MAP2 and (E) β3-tubulin. Actin was used as a loading control. Lower panels: representative blots from one of five separate experiments; upper panels: relative densitometric intensity. Data are mean ± SEM (n = 5).

FIGURE 4

Difference in metabolites produced by MRx0005 and MRx0029. Bacteria were grown to stationary phase and processed as described in section “Materials and Methods.” YCFA⁺ was processed alongside MRx0005 and MRx0029. Authentic chemical standards were compared to retention time and mass spectra for the identified metabolites and/or based on library matching of the acquired MS spectra with the NIST library. Only metabolites that differed between the two strains are reported in this figure. (A) SCFA and MCFA profile (mM). (B) Succinic acid (mM) and 4-hydroxy-phenyl acetic acid (ratio sample/media). The data presented in (A,B) are from one experiment (n = 1). (C) Representative chromatograms for MRx0005 (left panel in blue) and MRx0029 (right panel in red) overlaid to a standard mix of SCFAs and MCFAs (Formic acid 10 mM, Acetic acid 20 mM, Propionic acid 10 mM, Butyric acid 10 mM, Valeric acid 10 mM, and Hexanoic acid 10 mM in black).

FIGURE 5

Short-chain fatty acids effect on neuroinflammation and neuroprotection. U373 cells were treated with increasing concentrations of Sodium Butyrate (SB, 500 μM–2 mM), Sodium Valerate (SV, 250 μM–1 mM), or Hexanoic acid (HA, 100–500 μM) in the presence or absence of 1 μg/ml LPS as reported in section “Materials and Methods.” Cell-free supernatants were collected 24 h after treatment and secretion of IL-6 (A) and IL-8 (B) were measured by ELISA. Differentiated SH-SY5Y were treated with 2 mM SB, 0.65 mM SV, and 0.2 mM HA or a combination of the three SCFA at the same concentration found in 10% BCFS from MRx0029. For ROS evaluation, cell treatment in the presence of 100 μM of TBHP for 2 h. Fluorescence intensity was measured at Ex/Em = 485/530 nm. (D) Undifferentiated cells were treated with 10% BCFS from MRx0005, MRx0029, MRx0071, and MRx1342. After 24 h, cells were collected, total RNA isolated for qPCR to quantify MAP2 gene expression. Data are mean ± SEMSH-SY5Y cells were treated with 2 mM SB, 0.65 mM SV, 0.2 mM HA or a combination of the three SCFA as described above. After 24 h, cells were collected, total RNA isolated for qPCR to quantify MAP2 gene expression. Data are mean ± SEM (n = 3). (E) U373 cells were treated with 10% BCFS from MRx0071 and MRx1342, two butyrate-producing strains from 4D pharma culture collection. Cell-free supernatants were collected 24 h after treatment and secretion of IL-6 was measured. (F) Differentiated SH-SY5Y cells and were stained with DCFDA and then treated with 10% BCFS from MRx0005, MRx0029, MRx0071, and MRx1342 in the presence of TBHP for 2 h, followed by measurement of fluorescence intensity at Ex/Em = 485/530 nm. (G) Undifferentiated SH-SY5Y (n = 3).

FIGURE 6

Biochemical characterization and in vitro evaluation of the biological activity of MRx0005 and MRx0029 culture supernatant fractions. BCFS from three biological replicates from MRx0005 and MRx0029 were mixed with different solvents of increasing polarity and extracts used in cell-based assays. 10% BCFS, media alone, five different fractions and blank media treated with the same solvents used to prepare the fractions were added to U373 cells in the presence or absence of LPS as described in section “Materials and Methods.” Cell-free supernatants were collected 24 h after treatment and secretion of IL-6 and IL-8 was measured by ELISA. IL-6 secretion measured in U373 cells (A) pre-treated with 1 μg/ml LPS or in the absence of LPS (B); IL-8 secretion measured in U373 cells (C) pre-treated with 1 μg/ml LPS or in the absence of LPS (D). (E,F) The MeOH fractions from MRx0005 and MRx0029 were analyzed by HPLC for SCFAs band MCFAs. Representative examples of HPLC chromatograms are shown for MRx0005 and MRx0029. The methanolic fractions for MRx0005 (E) and MRx0029 (F) were extracted for SCFAs and MCFAs and then overlaid to a standard mix of SCFAs and MCFAs (Formic acid 10 mM, Acetic acid 20 mM, Propionic acid 10 mM, Butyric acid 10 mM, Valeric acid 10 mM, and Hexanoic acid 10 mM in black). The chromatograms highlight the lack of butyric, valeric and hexanoic acid in the MRx0005 methanolic fraction and their presence in the MRx0029 methanolic fraction.