(WHS-I.03) USING MICROBIAL TRANSCRIPTIONAL PROFILES AS A BIOMARKER FOR DIABETIC WOUND HEALING
Wednesday, May 15, 2024
1:45 PM – 4:00 PM East Coast USA Time
Introduction: The diabetic foot ulcer (DFU) microbiome has been associated with healing outcomes and is hypothesized to be a source of biomarkers. Profiling the microbiome by 16S rRNA sequencing is limited by species level classification and functional analysis of these communities, especially over time and in response to treatments. Here we apply metatranscriptomics to profile microbial transcription to identify key metabolic activities of the microbiome and use these to predict clinical outcome.
Methods: 100 patients with Wagner grade 1-3 DFUs were enrolled and matched specimens of debrided tissue and deep ulcer swabs were processed for DNA-based 16S rRNA amplicon sequencing and RNA-based metatranscriptomics. Specimens were collected at clinical visits up to week 12 of enrollment. Comprehensive clinical metadata and wound outcomes were collected for each patient. Data was processed to remove human reads, taxonomically classified, and metabolic pathways reconstructed.
Results: 16S rRNA microbiome composition significantly overlapped between the two specimen types, indicating that both are viable sampling methods. However, the metatranscriptome showed swabs are more diverse for species richness and functional diversity as compared to tissue, suggesting an increased sensitivity of RNA-based methods. We used the ratio of RNA to DNA abundance for each genera as a marker for microbial transcriptional activity. Anaerobes were found to be more transcriptionally active in wounds resulting in an amputation. The summed relative abundance of anaerobic transcripts significantly predicted the likelihood of amputation at the end of study (12 weeks; odds ratio 14.17, 95% CI: 1.14 - 232.95, logistic regression, p<0.05). Notably, anaerobic 16S amplicon sequencing relative abundances were not significantly associated with amputation (logistic regression, p>0.05), supporting the increased sensitivity of RNA-based methods. No individual taxa were associated with this outcome (logistic regression, unadjusted p>0.05), but summed transcript relative abundances of Clostridium, Anaerococcus, Peptoniphilus, and Finegoldia species were sufficient to detect a significant association with amputation (logistic regression, p<0.05), highlighting the importance of Gram-positive anaerobic cocci (GPAC) in this dataset. Anaerobic taxa abundance was not associated with peripheral arterial disease and decreased vascular flow, suggesting that there may be local factors (e.g. biofilms) allowing for anaerobes to persist, and that they may function as a separate indicator for low oxygen microenvironments.
Conclusions: These data support the wound microbiome as a promising clinical biomarkers. Further analysis of these datasets will yield testable microbial biomarkers that are translatable into clinical tools to be validated in a prospective cohort in collaboration with the Diabetic Foot Consortium.