Relationships of gut microbiota, short-chain fatty acids, inflammation, and the gut barrier in Parkinson's disease

Abstract

Background: Previous studies have reported that gut microbiota, permeability, short-chain fatty acids (SCFAs), and inflammation are altered in Parkinson's disease (PD), but how these factors are linked and how they contribute to disease processes and symptoms remains uncertain. This study sought to compare and identify associations among these factors in PD patients and controls to elucidate their interrelations and links to clinical manifestations of PD.

Methods: Stool and plasma samples and clinical data were collected from 55 PD patients and 56 controls. Levels of stool SCFAs and stool and plasma inflammatory and permeability markers were compared between patients and controls and related to one another and to the gut microbiota.

Results: Calprotectin was increased and SCFAs decreased in stool in PD in a sex-dependent manner. Inflammatory markers in plasma and stool were neither intercorrelated nor strongly associated with SCFA levels. Age at PD onset was positively correlated with SCFAs and negatively correlated with CXCL8 and IL-1β in stool. Fecal zonulin correlated positively with fecal NGAL and negatively with PD motor and non-motor symptoms. Microbiota diversity and composition were linked to levels of SCFAs, inflammatory factors, and zonulin in stool. Certain relationships differed between patients and controls and by sex.

Conclusions: Intestinal inflammatory responses and reductions in fecal SCFAs occur in PD, are related to the microbiota and to disease onset, and are not reflected in plasma inflammatory profiles. Some of these relationships are distinct in PD and are sex-dependent. This study revealed potential alterations in microbiota-host interactions and links between earlier PD onset and intestinal inflammatory responses and reduced SCFA levels, highlighting candidate molecules and pathways which may contribute to PD pathogenesis and clinical presentation and which warrant further investigation.

Keywords: Inflammation; Intestine; Microbiota; Parkinson’s disease; Short-chain fatty acids.

Fig. 1

Differences in analyte levels in PD by sex. Levels of stool butyric acid (p=0.101 female, p=0.003 male) and calprotectin (p=0.019 female, p=0.489 male) and plasma CXCL8 (p=0.017 female, p=0.145 male) in male and female PD patients (n=26 female, 29 male) and control subjects (n=29 female, 27 male) compared by Wilcoxon rank sum test

Fig. 2

Correlations among SCFAs and inflammatory and permeability markers. Pearson correlations in a full data, b Control subjects (n=56), and c PD patients (n=55)

Fig. 3

Correlations of SCFAs, immune markers, and clinical variables. Pearson correlations of a SCFAs, and immune markers in b plasma and c stool with clinical variables in full data and in only control subjects (n=56) or PD patients (n=55)

Fig. 4

Correlations of alpha diversity for SCFAs and inflammatory and permeability markers. Pearson correlations with Shannon and inverse Simpson diversity indices, n=56 controls, n=55 PD patients; p-value is marked as follows: ***: p ≤ 0.001 / **: p ≤ 0.01 / *: p ≤ 0.05 / .: 0.1 > p > 0.05

Fig. 5

Linear modeling for alpha diversity showing effects of PD/control status and sex. a Stool calprotectin and inverse Simpson index and b propionic acid and Shannon index for male and female PD patients (n=26 female, 29 male) and control subjects (n=29 female, 27 male)

Fig. 6

Butyric acid, NGAL, and zonulin levels by enterotype among PD patients and control subjects. Levels of butyric acid (p=0.013 for PD patients), stool NGAL (p=0.002 for control subjects), and stool zonulin (p< 0.001 for control subjects) compared across enterotypes by Kruskal-Wallis test. Asterisks indicate results of pairwise Wilcoxon rank sum post-hoc tests (Bacteroides – 8 control, 8 PD; Firmicutes – 25 control, 35 PD; Prevotella – 22 control, 10 PD) with ***: p ≤ 0.001 / **: p ≤ 0.01 / *: p ≤ 0.05

Fig. 7

Associations of SCFAs and stool inflammatory and permeability markers with bacterial genera. Associations determined using differential expression analysis for sequence count data for full data, control subjects (n=56), and PD patients (n=55). Confounders included in confounder-corrected models were Rome III 9–15 sum score and sex. P-value is marked as follows: ***: p ≤ 0.001 / **: p ≤ 0.01 / *: p ≤ 0.05 / .: 0.1 > p > 0.05