NEW YORK – A group of researchers at the State University of New York, Upstate Medical University has identified biomarkers in the oral microbiome of patients with Parkinson's disease (PD) that could serve as evidence for the presence of early-stage disease.
The researchers are now partnering with Quadrant Biosciences to explore the development of a saliva-based diagnostic assay for early-stage PD based on the microbial biomarkers.
Multiple studies have demonstrated the presence of molecules called Lewy bodies — abnormal deposits of the alpha-synuclein proteins in the brain — in a large proportion of PD subjects. Researchers have also recently found that gut disturbances in patients with PD can rapidly expedite disease progression by contributing to Lewy body development.
"However, most studies have focused only on lower GI or fecal samples and looked at subjects with established motor symptoms or advanced states," SUNY Upstate neuroscience and physiology associate professor Frank Middleton explained. "We looked in the oral microbiome because there's compelling data showing dysfunction in the entire GI that occurs quite commonly in PD."
In a pilot study published last month in PLOS One, Middleton and his team examined phenotypic and genotypic information in a group of 48 patients with PD, in addition to 36 healthy controls. Most subjects had previously been observed to have resting tremor and were on PD medication. In addition, 95 percent of the patients had prior evidence of upper or lower gastrointestinal disturbance.
Middleton and his team began by collecting phenotypic data on the patients, who underwent detailed assessments of motor, cognitive, balance, autonomic, and chemosensory functions to determine their disease stage.
Compared to healthy control subjects, the researchers found that the PD cohort exhibited significant changes in several tests regarding motor, cognitive, and sensory function. The PD group demonstrated slower reaction speeds, as well as poorer senses of smell and taste. While the team could not separate motor and cognitive performance changes in the PD subject cohort, the team believes increasing cognitive demands of task led to reduced performance.
The researchers then collected between 1 and 4 ml of saliva from each participant. After purifying and extracting microbial RNA and host mRNA, the team assessed the yield and quality of the RNA samples before building sequencing libraries. The team then performed shotgun sequencing on the samples using Illumina's NextSeq 500 platform.
Middleton explained that his team used shotgun sequencing over standard 16S sequencing because the tool yielded a more comprehensive scope of microbial coverage. In addition, shotgun sequencing helped the researchers identify beyond the genus level down to species and sometimes subspecies level of bacteria.
"We could also obtain information about yeast, such as Candida or Saccharomyces, and viruses, which you can't do with 16s sequencing," Middleton said.
To investigate the potential for host mRNA interactions with the microbiome, Middleton and his team aligned NGS data to the human transcriptome. The study authors said that they also investigated potential differences in host mRNA abundance in subjects by filtering out the raw data "to only include human mRNAs present with at least 10 raw reads in 10 percent of the samples."
While Middleton and his team did not find any differences in overall alpha and beta diversity between subject groups, they did identify changes in specific microbial taxa. The researchers noticed 50 different taxa that exhibited significant differences in PD subjects compared to controls, including 16 genera and 34 species from bacteria, phage, and eukaryotic taxa.
Microbial changes included both increases and decreases in abundance. At the species level, Middleton and his team identified several members of the Lactobacillus and Bifidobacterium bacterial genera. The researchers also noticed a decrease in a Streptococcus phage, and increases in three yeast species (Candida albicans, Candida dubliniensis, and Saccharomyces cerevisiae) in PD patients.
To further examine the microbiome differences in the group, the team ran genus- and species-level data through logistic regression classification and area under the curve analysis, finding a potentially strong diagnostic performance with as few as 11 taxa that could distinguish early-stage PD subjects from controls, achieving an AUC of .90.
In parallel with the microbiome analysis, the team also examined the data for changes in human salivary mRNAs in the same subjects. After filtering out the most commonly expressed human mRNAs, the team tested for group differences, revealing nine mRNAs that differed significantly in the PD population and also showed a strong ability to distinguish subjects in the two groups. The host mRNAs were also linked to various brain functions and showed correlations to significantly changed microbiota taxa.
"We identified some miRNA that are interesting because they are related to brain functions, from which we might be able to infer that there might be a host response in additional to a microbial response," Middleton said. "Therefore, we might be able to identify a response from the host's transcriptional side."
According to Middleton, his team then analyzed significant links among the microbial data and the full set of medical, demographic, and functional outcome measures. The team identified 10 robust correlations with nine positive associations at the species level and a single negative association at the genus level.
Middleton noted that the abundance of Lactobacillius reuteri, often considered a probiotic in nature, instead positively correlated with sluggish movement in the PD cohort. His team therefore believes that members of the bacterial family may represent a disadvantageous yet consistent event in the disease.
"While [L. reuteri] is thought to help reduce constipation, IBS, and diarrhea, there are studies showing that it might do that because it affects the activity of a specific nerve, which is potentially thought to be a major conduit for the pathological spread of PD," Middleton explained. "You have this idea that the bacteria would normally do good things for a normal non-diseased person, but it actually may be ramping up activity of the nerve in PD, which might actually be detrimental to the ongoing disease process."
James Beck, CSO of the Parkinson's Foundation who was not involved with the study, believes that the SUNY researchers need more samples from additional PD patients to accurately interpret and establish definitive conclusions. He pointed out that the researchers collected high amounts of phenotypic data, which "is a lot for a cross-sectional study," in order to link potential correlations with patients' genetic data.
Middleton acknowledged that his team encountered several limitations in the study. The biggest issue the researchers had to contend with was the small cohort size. In addition, Middleton noted that the study's results may have been affected by the potential influence of medications taken by patients before and during the study. He therefore hopes to eventually build a cohort comprising patients in the earliest, pre-medicated stage of PDs.
"Since it's a small study, we don't have a lot of subjects that were not taking medications for PD," Middleton said. "It's difficult to find early-stage patients, since you need a neurologist to confirm and diagnose PD, and by the time those patients get to neurologists, they might already be on medication for other issues."
Like Middleton, Beck argues that the study would have had more success if the researchers had grouped individuals by their different disease states, as well as separated patients based on their medication status prior to the study. While Beck noted that medications are not always successful, they can often alter the patient's physiology and produce mixed results.
"I don't think much is going to come from [the study] at this particular stage," Beck said. "It needs to include a variety of different patients as well, [since] the question is if other people with similar neurological diseases, like Alzheimer's, show similar behaviors or information."
Clinical applications
Middleton envisions using the biomarkers to eventually create a point-of-care assay for PD. However, he doesn't believe such a tool could act as a full substitute for a PD patient's normal functional and biological assessment at a clinician's office.
At the same time, he noted that there is a 15 percent misdiagnosis rate of PD among medical specialists, which "is even worse in non-specialists." In addition, clinicians often misdiagnose PD in patients for several other disorders. Middleton noted that clinicians can perform a procedure called single-photon emission computed tomography (SPECT) imaging that scans for a radiopharmaceutical agent previously injected into a patient's veins to potentially detect cases of PD; however, he argued that the SPECT tests can be very expensive, are overused, involve radioactivity, and don't always provide definitive results.
Middleton therefore believes that clinicians could use a diagnostic tool based on the microbial biomarkers as a complementary diagnostic with a standard neurology exam.
"This could enable a finer diagnosis to be made and would easily identify dysbiosis (imbalance in the gut microbiome) that is present and pro-inflammatory to an extent," Middleton said. "This would be a means of intervening, since changing the microbiome of the mouth is easier than changing the GI tract."
Middleton said his team is interested in examining the consistency of microbial changes in different geographic areas. Believing that a person's diet could have a larger impact on the oral microbiome on the mouth depending on their location, the researchers would aim to directly relate the microbiome changes to early stages of PD.
"We want to see if certain patients have a higher rate of risk if they live in areas that have pesticides," Middleton said. "Since we know that certain pesticides, herbicides, and antibiotics used in agriculture can have a major impact on the microbiome, we're interested if this could be known to have links to PD in areas with potential triggers."
In addition, Middleton said that his team is currently in the middle of a study that looks at "the human side of the story," rather than focusing on the microbial taxa.
Middleton noted that SUNY Upstate is collaborating with other academic sites to increase the size of the patient cohort for future studies, including Allegheny General, the University of Texas at San Antonio, and Florida State University.
Quadrant Biosciences, a spinout of SUNY Upstate that has been developing molecular and other tests for several neurological diseases and disorders, may be interested in the future direction of the research. According to Middleton, who has served on Quadrant's scientific advisory board since 2017, Quadrant funded the initial study with a $61,880 pilot award. While Quadrant did not have a role in study design, data analysis, or initial writing of the study, Middleton said the firm did assist with data collection through software preparation and administrative support.
In addition, SUNY Upstate and Quadrant share joint intellectual property for the technology developed in the study. While initial patient recruitment was handled through SUNY Upstate,Quadrant is now helping the research team in its ongoing PD patient recruitment for future multi-side studies. As a Quadrant SAB member, Middleton provides technical input on PD, autism, and several other research projects.
Beck, of the Parkinson's Foundation, noted that while he is unaware of other studies that have linked patients' oral microbiomes' to PD, several research groups have previously examined the gut microbiome's relationship with PD.
"It's a real chicken and egg problem, where you're looking at the microbiome and asking about the relationship between it and the disease," Beck said. "If you have someone with a disease and a unique biome, is that the cause of the biome, or did the biome itself cause the disease?"
However, Beck highlighted that the SUNY study may be useful as a longitudinal assessment, where Middleton's team could follow up patients in two years to identify genetic alterations and see if the oral microbiome had predicted said genetic change. While Beck sees potential in establishing links between PD and the oral microbiome, he believes that "more work needs to be done … to ask interesting questions and follow up with a study to definitively answer those questions."