Gut bacterial tyrosine decarboxylases restrict levels of levodopa in the treatment of Parkinson's disease

Abstract

Human gut microbiota senses its environment and responds by releasing metabolites, some of which are key regulators of human health and disease. In this study, we characterize gut-associated bacteria in their ability to decarboxylate levodopa to dopamine via tyrosine decarboxylases. Bacterial tyrosine decarboxylases efficiently convert levodopa to dopamine, even in the presence of tyrosine, a competitive substrate, or inhibitors of human decarboxylase. In situ levels of levodopa are compromised by high abundance of gut bacterial tyrosine decarboxylase in patients with Parkinson's disease. Finally, the higher relative abundance of bacterial tyrosine decarboxylases at the site of levodopa absorption, proximal small intestine, had a significant impact on levels of levodopa in the plasma of rats. Our results highlight the role of microbial metabolism in drug availability, and specifically, that abundance of bacterial tyrosine decarboxylase in the proximal small intestine can explain the increased dosage regimen of levodopa treatment in Parkinson's disease patients.

Fig. 1

Bacteria in jejunal content decarboxylate levodopa to dopamine coinciding with their production of tyramine ex vivo. a Decarboxylation reaction for tyrosine and levodopa. b From left to right coinciding bacterial conversion of tyrosine (TYR) to tyramine (TYRM) and 1 mM of supplemented levodopa (LD) to dopamine (DA) during 24 h of incubation of jejunal content. The jejunal contents are from four different rats ranked form left to right based on the decarboxylation levels of tyrosine and levodopa, showing that tyrosine decarboxylation is coinciding with levodopa decarboxylation

Fig. 2

Gut bacteria harboring tyrosine decarboxylases are responsible for levodopa decarboxylation. a Aligned genomes of E. faecium, E. faecalis, and L. brevis. The conserved tdc-operon is depicted with tdc gene in orange. Overnight cultures of b E. faecalis v583, c E. faecium W54, and d L. brevis W63 incubated anaerobically at 37 °C with 100 µM of levodopa (LD). e Overnight cultures of EFS^WT and EFS^ΔTDC incubated anaerobically at 37 °C with 100 μM levodopa (black line) compared to control (gray line) where no levodopa was added. Curves represent one example of three biological replicates

Fig. 3

Enterococci decarboxylate levodopa in presence of tyrosine despite higher affinity for tyrosine in vitro. Growth curve (gray circle, right Y-axis) of E. faecium W54 (a) and E. faecalis (b) plotted together with levodopa (open square), dopamine (closed square), tyrosine (open triangle), and tyramine (closed triangle) levels (left Y-axis). Concentrations of product and substrate were normalized to the initial levels of the corresponding substrate (100 µM supplemented levodopa and ~500 µM tyrosine present in the medium). pH of the culture is indicated over time as a red line. c Substrate affinity (Km) for levodopa and tyrosine for purified tyrosine decarboxylases from E. faecalis v583 (TDC_EFS), E. faecium W54 (TDC_EFM, ^PTDC_EFM). d–i Michaelis–Menten kinetic curves for levodopa and tyrosine as substrate for TDC_EFS (d, e), TDC_EFM (f, g), and ^PTDC_EFM (h, i). Reactions were performed in triplicate using levodopa concentrations ranging from 0.5 to 12.5 mM and tyrosine concentrations ranging from 0.25 to 2.5 mM. The enzyme kinetic parameters were calculated using nonlinear Michaelis–Menten regression model. Error bars represent the SEM and significance was tested using 2-way-Anova, Fisher LSD test, (*p < 0.02; **p < 0.01; ****<0.0001)

Fig. 4

Human DOPA decarboxylase inhibitor, carbidopa, does not inhibit bacterial tyrosine decarboxylases. a Inhibitory constants (Ki) of bacterial decarboxylases (black) and human DOPA decarboxylase (gray), with fold-difference between bacterial and human decarboxylase displayed on top of the bars. Quantitative comparison of dopamine (DA) production by E. faecium W54, b and E. faecalis v583, c at stationary phase after 15 min, with representative HPLC-ED curve. Bacterial cultures (n = 3) were incubated with 100 µM levodopa (LD) or a 4:1 mixture (in weight) of levodopa and carbidopa (CD) (100 µM levodopa and 21.7 µM carbidopa). Error bars represent SEM (a) or SD (b, c) and significance was tested using a parametric unpaired T-test

Fig. 5

Tyrosine decarboxylase gene abundance correlates with daily levodopa dose and disease duration in fecal samples of Parkinson’s disease patients. a Scatter plot of tdc gene abundance measured by qPCR in fecal samples of PD patients (n = 10) versus daily levodopa/carbidopa dosage fitted with linear regression model. b Scatter plot of tdc gene abundance from the same samples versus disease duration fitted with a linear regression model. Pearson’s r analysis was used to determine significant correlations between tyrosine decarboxylase gene abundance and dosage (r = 0.66, R² = 0.44, P value = 0.037) or disease duration (r = 0.82, R² = 0.68, P value = 0.003)

Fig. 6

Luminal and plasma levels of levodopa are compromised by higher abundance of tyrosine decarboxylase gene in the small intestine of rats. Scatter plot of tdc gene abundance measured by qPCR in jejunal content of wild-type Groningen rats (n = 18) orally supplied with levodopa/carbidopa mixture (4:1) versus a the dopamine: levodopa/carbidopa levels in the jejunal content, the levodopa/carbidopa levels in the jejunal content, b or the levodopa/carbidopa levels in the plasma, fitted with a linear regression model. Intake of levodopa/carbidopa was corrected by using carbidopa as an internal standard. Pearson’s r correlation was used to determine significant correlations between tdc abundance and jejunal dopamine levels (r = 0.78, R² = 0.61, P value = 0.0001), jejunal levodopa/carbidopa levels (r = −0.68, R² = 0.46  P value = 0.021), or plasma levodopa/carbidopa levels (r = −0.57, R² = 0.33, P value = 0.017). No levodopa/carbidopa, dopamine, or DOPAC were detected in the control group (n = 5). c Significant difference in plasma levels of levodopa/carbidopa orally supplied to rats after treatment with EFS^WT (n = 10) or EFS^ΔTDC (n = 10). Significance was tested using parametric unpaired T-test (**p < 0.01)

Fig. 7

Higher abundance of tyrosine decarboxylase can explain increased levodopa administration requirement in Parkinson’s disease patients. A model representing two opposing situations, in which the proximal small intestine is colonized by low (left) or high abundance of tyrosine decarboxylase-encoding bacteria. The latter could result from or lead to increased individual L-DOPA dosage intake