Parkinson's disease (PD) is a progressive neurodegenerative disorder characterized by aggregation of the protein alpha-synuclein in Lewy bodies and neurites, leading to dopaminergic neuron loss in the substantia nigra (1). PD is typically associated with motor symptoms such as bradykinesia, tremor and rigidity, but the importance of non-motor aspects, particularly gastrointestinal problems, is increasingly appreciated. Constipation is a well-recognized risk factor for PD (2), and increased intestinal permeability and inflammation have been described in PD (3). Due to the growing evidence for gastrointestinal involvement in PD, as well as support for the hypothesis that alpha-synuclein may propagate from the gut, along the vagus nerve to the brain (4, 5), the gut microbiome holds promise as a source of biomarkers as well as for potential therapeutic intervention in PD.

To date, a number of studies have reported differences in the gut microbiota composition of PD patients compared to controls (6–18). There has also been significant heterogeneity of results. Only a single study considering disease progression has far been conducted, reporting that low counts of Bifidobacterium may be associated with faster PD progression over 2 years (19).

Medications have been shown to explain a large proportion of variance in the gut microbiota composition (20) and the microbiome has been suggested to impact medication efficacy (21–23). Specific to PD, the microbiome may explain the heterogeneity in the efficacy and side effects of Levodopa, which is not clearly linked with clinical factors (17). A recent report identified E. faecalis as potentially responsible for Levodopa decarboxylation, and potentially, the differential Levodopa response among PD patients (24). It has been shown that bacteria in the rat small intestine express genes encoding for the enzyme tyrosine decarboxylase (TDC) that decarboxylates Levodopa to dopamine, potentially suggesting a role for the microbiome in the pharmacokinetics and effect of Levodopa in individuals with PD (25).

Recent studies have suggested that some of the reported differences in gut microbiota composition of PD patients compared to controls could potentially be related to Levodopa use. In a small study that compared drug naïve (n = 12) and treated (n = 26) PD patients, Levodopa use was significantly associated with the abundances of Bacillaceae (9). In another study, Levodopa dose was inversely associated with abundance of genera Dorea and Phascolarctobacterium (18). A decrease in genus Faecalibacterium was reported in a cross-sectional analysis of microbiota composition of PD patients using Levodopa (26). Notably, in a recent study comparing the microbiota composition of PD patients vs. controls (27), adjustment for Levodopa use, in addition to other covariates, resulted in attenuation of most findings, highlighting the importance of considering medication use in analyses (27).

Understanding the impact of Levodopa on the gut microbiome is crucial to separate disease-related from medication-related impacts on the microbiome. However, to our knowledge, no study to date has examined the impact on the microbiome of starting Levodopa longitudinally in de novo PD. In this study, we evaluated the gut microbiota composition de novo PD patients prospectively and longitudinally, before and after starting Levodopa therapy. We also evaluated associations between the microbiota composition and Levodopa response in PD patients.

Twenty-one patients successfully provided the first two stool samples, prior to starting Levodopa. Two patients stopped Levodopa use because it was clinically ineffective and were thus withdrawn from our study. The remaining 19 patients returned the two post-Levodopa samples (Figure 1), for a total of four stool samples per patient (two pre-Levodopa and two post-Levodopa). Demographic, clinical characteristics, and Levodopa daily dose (LDD) of the patients under study are shown in Table 1. The Levodopa daily dose at 90 days was available for 17 of the 19 patients. The majority (11 of 17) participants were taking 300 mg Levodopa per day. Other doses were as follows: 900 mg (n = 1), 600 mg (n = 2), 450 mg (n = 2), and 200 mg (n = 1). As shown in Table 1, 8 of 19 participants were stable on non-Levodopa PD medications during the study (dopamine agonists, mono-amine oxidase inhibitors, and NMDA antagonists). Participants did not initiate any other PD medication during the study.

There is growing evidence that the gut microbiome plays a role in PD (6, 8, 9, 11–16). Despite the increasing interest in this field, most of the studies conducted thus far have focused on microbiota comparisons between medicated PD patients and controls, using a cross-sectional approach (6, 8, 9, 11, 12, 14, 15). To our knowledge, no study to date has addressed, using a prospective and longitudinal design, whether initiation of Levodopa in de novo PD patients might alter the microbiome or whether the microbiome of PD patients might impact the improvement in motor function in response to initiation of Levodopa by PD patients.

In this longitudinal investigation of de novo PD patients who provided stool samples before and after initiation of Levodopa, we examined whether Levodopa administration alters the PD gut microbiota composition and whether composition of the gut microbiota in PD patients can predict response to Levodopa. In our baseline sample, age, and BMI were the factors most strongly associated with the microbiota (both alpha and beta diversity), which is in agreement with other studies (20). We observed an increase in weighted UniFrac distance, but not other metrics of alpha or beta diversity, when comparing the 2nd vs. 3rd samples, between which Levodopa was initiated, with the distance between 1st vs. 2nd samples (both pre-Levodopa). This suggests that Levodopa initiation is potentially associated with an increase in weighted UniFrac distance. Levodopa initiation was not associated with short-term taxonomic changes in longitudinal analysis. Our finding that only Weighted UniFrac distances, but not other metrics of beta diversity were associated with Levodopa initiation, is not necessarily conflicting with the null results for other metrics, but suggests that these effects may be more nuanced. Both Bray-Curtis and Jaccard consider discrete taxonomic groups. The limitation of Bray-Curtis and Jaccard is that taxonomic groups are equally weighted even though the phylogenetic relationships between taxonomic groups are not equal. UniFrac considers phylogenetic groups, with Unweighted UniFrac emphasizing the presence or absence of these groups and Weighted UniFrac additionally considering their abundance. Thus, Weighted UniFrac, which was the metric significantly associated with Levodopa initiation in our study, is unique among the four metrics used in that it considers the abundance of phylogenetic groups and the association we observed with this metric suggests that the abundance, more so than the presence/absence, of a phylogenetic group is contributing to statistically significant beta diversity differences.

In this study, we observed a marginally lower abundance of Clostridium group IV in PD patients with a good response to Levodopa. Bacteria belonging to Clostridium group IV are commensal, gram positive, and strict anaerobe bacteria. They play an important role in gut homeostasis and ferment dietary fiber to short chain fatty acids (SCFA), particularly butyrate, compounds that provide energy for colonocytes and play a key role in the health of the colon (33). They have also been shown to play an important role in the secondary metabolism of bile acids (34). Relevant to PD, prior work has implicated commensal Clostridia in the production of catecholamines, including norepinephrine and dopamine (35). Clostridia spp., are thought to impact host immunological signaling and immunological development, likely via production of metabolites such as SCFA and secondary bile acid metabolism (36). Specifically, Clostridium group IV have been shown to strongly induce of the accumulation of T regulatory cells, which are key to immune homeostasis, in the colon of mice (37). Immune system dysfunction involving neuroinflammation (38), T cell infiltration (39), and microglia activation (40) have been implicated in PD and suggested to play a role in the degeneration of dopaminergic neurons. The implication of Clostridium group IV in Levodopa response reported here, if confirmed in larger studies, supports the need to investigate the role of SCFA and secondary bile acids in PD. However, due to the very small sample size of this study, this result is preliminary and should be treated with caution.

Prior studies have linked Levodopa use to abundance of Bacillaceae (9) Dorea and Phascolarctobacterium (18). Our study did not identify any taxa significantly associated with Levodopa use but preliminarily implicated Clostridium group IV in Levodopa response, thus adding to an already heterogenous literature. Differences in study design size and patient population may explain some of the differences in findings from prior investigations. There may be several mechanisms by which microbiome changes can impact Levodopa response, including but not limited to impact on gastric emptying, gut mucosa local factors such as inflammation and metabolite concentrations, and direct metabolism of L-dopa via production of decarboxylases. Different mechanisms may be more important in different patients, potentially depending on patient characteristics, disease profile, and Levodopa formulation.

The longitudinal aspect of our study, and our focus on the impact of Levodopa initiation in the microbiota in de novo PD, is a major strength of our study. Examining microbiota composition changes within individuals longitudinally limits potential confounding due to between-person variation in diet, disease severity, and other factors that may impact studies relying on between-subject comparisons.

Our study is preliminary and was limited in power by the low number of patients, which reflected the challenge of identifying de novo PD patients at the time of Levodopa initiation. With adjustment for three covariates (age, sex, and BMI), with 19 total participants, each providing four samples, our study had 80% power to detect a 0.279—unit difference in common taxa (~3,792 sequences and 0.0758421 relative abundance at 100% read usage rate), and a 0.108—unit difference in rare taxa (~581 sequences and 0.0116187 relative abundance at 100% read usage rate). Most realistic microbiome effects are likely to be much smaller, and we had limited power to detect those. Larger, better-powered studies are thus needed to examine the microbiota changes after initiation of Levodopa in PD patients. Due to constrains on power, we kept our statistical models parsimonious, adjusting our analyses for age, sex, and BMI only. Several additional confounding or stratification factors could be of potential interest in future, better powered studies, including initial disease severity or disease phenotype, which could both influence extent of response to Levodopa. While this study could not take these factors into account, they would be important to consider in future analyses. Another important limitation of our study is that it did not include a healthy control group, and we were thus not able to reproduce results from prior case-control studies that made such comparisons.

Furthermore, the study focused on short-term Levodopa use and thus cannot address the question of whether longer term use Levodopa could be associated with changes in the gut microbiome. Although we did not have an untreated PD control group, it is less likely that substantial microbiota changes would have occurred, without treatment or intervention, during the 90-day study period. An additional limitation is the use of 16S sequencing, which can limit taxonomic identification. Larger, longer-term studies, with more detailed specimen analysis, such as metagenomics or multi-omic approaches are needed to validate our results and better understand the mechanism of the connections between the gut microbiome and PD pathophysiology, progression, and response to Levodopa and other PD medications.

In summary, we did not observe significant shifts in microbial alpha or beta diversity with Levodopa treatment in this short-term study of de novo PD patients. The lower abundance of Clostridium group IV in PD patients who had a good motor response to Levodopa compared to those who had no response is preliminary and should be confirmed in future studies.

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

The studies involving human participants were reviewed and approved by Institutional Review Board (IRB) of the University of Massachusetts Medical School (Docket #H00013990). The patients/participants provided their written informed consent to participate in this study.

AH conceived of the research project, organized and executed the project, including participant clinical and biospecimen data collection. He contributed to data analysis and interpretation, and contributed to writing the first draft and revising the manuscript. NP designed and executed statistical analyses, and reviewed and critiqued results, and contributed to writing the first draft and revising the manuscript. JF contributed to the design and execution of the statistical analysis. DW performed sample extraction, microbiome sequencing and analysis, and contributed to design and review/critique of the statistical analysis, and the revising of the manuscript. KG contributed to organization and execution of the research project, and revising of the manuscript. AD contributed to execution of the research project (participant recruitment), and review and critique of the results and the manuscript. KS contributed to study design and organization, participant recruitment and data collection, reviewed and critiqued results, and revised the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.