IBD, Crohn's and ulcerative colitis are chronic gut illnesses that affect millions and command substantial budgets from most top pharmaceuticals but, as yet, a viable long-lasting treatment for any of these is still out of reach. This is, in part, due to the limited amount of gut microbiome data available and as such further investigations are required to understand the mechanisms and factors governing the afflictions.
Current single-plex approaches have aided the forward momentum toward treatment for sufferers. Metagenomics using 16S DNA sequencing has given researchers an insight into the various demographics and populations of microbiome communities. Transcriptomic techniques such as RNA-Seq have harnessed the power of the transcriptome to give an insight into the driving factors affecting the genetic control of the various taxa; whereas imaging techniques such as FISH allow investigators the visualisation of a momentary snapshot utilising oligonucleotide capture of targets.
Given the known complexity of the population and the interactions therein of the gut microbiome it is clear that each of the singular approaches has certain drawbacks that limit full exposure of the environment. Metagenomics loses spatial information through homogenisation, transcriptomics misses spatial organisation and imaging loses the taxonomic diversity.
Therefore, a more radical, multi-faceted and above all, multi-disciplinary methodology is required to successfully characterise the population, virulence driving forces and spatial compartmentalisation of the gut microbiome such that targeted and effective drug treatments can be developed.
In this study by Ravi et al., a multiplexed protocol combining the positive aspects of all three approaches is reported. Deploying the same principles as that of ecological Quadratic population studies except instead of one macro square area, thousands of tuneable micron-sized square representative areas are covered in one to give a quantifiable characteristic of the microbiome.
This novel technique, MaPS-seq, fixes the sample into a size-adjustable polyacrylamide matrix (µm-sized) containing a 16S rRNA primer, fractured using bead-beating, lysed, variably filtered for size selection which in turn form the gDNA-containing target that retain the spatial information, known as Clusters. These clusters are co-encapsulated with uniquely barcoded 16S Forward primers contained in gel beads. Triggered degradation discharges the gDNA from the matrix within droplets allowing PCR amplification of the 16S V4 region. Deep sequencing is performed on the dispersed droplets and the Relative Abundance and bacterial Operational Taxonomic Units (OTU) are calculated using the aligned reads that are subsequently categorised into specific groups based on their unique barcodes. Each stage is scrutinised and QC'd to ensure viable input into the proceeding stage.
As proof of principle the technique was able to accurately assign through mapping, with high correlation, mouse fecal bacteria samples and Escherichia coli with minimal cross-contamination and cross-talk.
The technique was then applied to a mixed homogenate of clusters containing the same set of samples. Again MaPS-seq was able to uniquely map and quantify bacterial identity.
Given the positive preliminary testing, the MaPS-seq protocol was then deployed on mouse colonic microbiome, specifically the distal colon.
1406 clusters yielded with 236 OTUs identified. Confirmation by standard 16S sequencing. Abundant taxa dispersed across available space. Some OTUs yielded low amount of clusters indicating clumping likely due to peristaltic action of the gut decreasing spatial distribution.
Frequency of taxa co-occurrence evaluated with pairwise associations statistically significant in 75 of 276 combinations. 72 of the 75 being positive associations possibly attributable to modalities of co-metabolism etc. A few negative associations were also noted possibly due to defined competitive inhibitions of various taxa.
Further confirmation of reproducible data highlighting the afore-mentioned associations noted in a separate mouse cohort meaning that MaPS-seq data is robust and reproducible.
To further investigate the application of MaPS-seq in a more complex microbiome community, the GI tract was assessed. Species abundance was initially categorised by standard 16S profiling as a baseline readout of absolute OTUs across two new mouse cohorts. Variability was noted across the cohorts which exemplified the issues described above of negative aspects of FISH imaging non-targeted probe design.
Three distinct GI areas, the ileum, the cecum and the distal colon were identified for MaPS-seq analysis.
Spatial distribution assessment showed the ileum to be significantly lower in OTUs compared to cecum and distal colon, the cecum being the most densely populated with OTUs and clusters.
t-SNE (t-distributed stochastic neighbour embedding) statistical analysis allowed the visualisation of cell clusters spanning the three GI areas of interest.
Identification of distinct localisation of specific taxa to particular regions. Ileum contained a mixture of both positive and negatively associated clusters, cecum contained all positively associated clusters in its previously mentioned dense compartmentalisation.
Co-associations of taxa identified across all three GI areas.
Even in the dense cecum MaPS-seq analysis identified self-aggregation of specific taxa, specifically Lachnospiricae.
Different regions of the GI tract pre-determine local spatial organisation of some taxa whereas other taxa are evenly distributed across the GI tract.
This was confirmed using imaging utilising previously hybridised FISH probes.
Diet is known to play an important role in the formation, organisation and function of the gut microbiome. Therefore effect of dietary behaviour on spatial structuring was assessed using samples derived from 2 cohorts, one on a Low Fat (LF) and another on a High Fat (HF) diet.
HF diet appeared to decrease the abundance of species in cecum and colon after 10 days.
Although phylogenetic differences were noted, MaPS-seq analysis on the distal colon found similarities in unique OTUs across both diets.
Further statistical analysis to quantify phylogenetic variation was utilised by calculation of the Net Relatedness Index (NRI). NRI measures the mean phylogenetic distances between taxa. Clumping is denoted by positive values and conversely over-dispersion by negative values. Zero values denote a random distribution.
Most clusters were near zero.
Both HF and LF diets registered some highly negative NRI values but LF significantly higher due, in part, to clusters that were only seen in LF diets pushing the values even higher.
tSNE analysis found the cell clusters were distinct which means that the dietary shift alters the spatial organisation of the distal colon.
Given the breadth and granularity of microbiome information being yielded by MaPS-seq, research into almost every microbial plant and animal niches could be advanced leading to faster treatment discoveries based on the increased knowledge garnered by the multiplexed approach.
Ravi U. Sheth, Mingqiang Li, Weiqian Jiang, Peter A. Sims, Kam W. Leong & Harris H. Wang, 2019. Spatial metagenomic characterization of microbial biogeography in the gut. Nature Biotechnology volume 37, pages 877–883