Microbial Sequences in Multiple Sclerosis Brain Specimens

Background:  Our group has developed RNA sequencing and analysis for the identification of microbes in diseased brain tissue. We have previously identified several viruses in MS and encephalitis-affected frozen brain tissue. Two other research groups have shown evidence for a “brain microbiome” by demonstrating microbial antigens and RNA in both normal and diseased human brain tissue.

Objectives:  to identify specific microbial signatures within early human MS brain lesions that may be driving inflammation and demyelination

Methods:  Nine primary demyelination (MS) biopsy and 14 control epilepsy surgical human brain specimens were studied. The study was reviewed and approved by the University of Utah Health Sciences IRB. RNA was extracted and quantified from the formalin-fixed paraffin-embedded specimens. Extracted RNA was subjected to quality control, then deep sequenced on the Illumina 2500 platform. The resulting reads were filtered for quality and mapped against a panmicrobial database. Normalized hit rates (HRs) were determined for each specimen against each of 5835 microbial taxa. HRs were compared between each diseased specimen and the set of controls using 2 complementary statistical techniques.

Results:  Between 1.06 – 2.96 x 10⁸ high-quality 125 bp paired-end reads were derived from the brain specimens. 6.4M read pairs mapped to the panmicrobial database, representing about 0.1% of the total reads. Microbial reads were present in both the MS and control specimens and were not significantly different between the groups. Outlier analysis, corrected for multiple comparisons, identified significantly increased microbial sequence in 7 of the 9 MS brain specimens. The microbial origin of 98% of these reads was confirmed by remapping to the human genome and transcription, with assembly across the entire microbial genome of interest. Thirty-five different candidate microbial taxa were significantly overrepresented in at least one of the MS brain samples. The MS candidate microbes included a variety of aerobic and anaerobic bacteria, but no fungi, protists, or viruses. Akkermansia muciniphila, Bifidobacterium adolescentis, Haemophilus parainfluenzae, Lactobacillus fermentans, Propionibacterium acnes, and Pseudomonas aeruginosa are the candidates currently selected for further study by in situ hybridization.

Conclusions: The sequencing data shows that most of the MS brain specimens studied contain a set of microbial sequences that are significantly different than the controls. This suggests that MS lesions may be related to the invasion of a diverse set of microbes into the brain, or to a disturbance of preexisting microbes within brain tissue. The data also supports the existence of a brain microbiome. Studies are in progress to localize and visualize specific MS candidate microbes within the brain tissue specimens.