Multi-Omic Data Integration Suggests Putative Microbial Drivers of Aetiopathogenesis in Mycosis Fungoides.

Background: Mycosis fungoides (MF) represents the most prevalent entity of cutaneous T cell lymphoma (CTCL). The MF aetiopathogenesis is incompletely understood, due to significant transcriptomic heterogeneity and conflicting views on whether oncologic transformation originates in early thymocytes or mature effector memory T cells. Recently, using clinical specimens, our group showed that the skin microbiome aggravates disease course, mainly driven by an outgrowing, pathogenic strain carrying the virulence factor spa, which was shown by others to activate the T cell signalling pathway NF-κB.

Methods: To explore the role of the skin microbiome in MF aetiopathogenesis, we here performed RNA sequencing, multi-omic data integration of the skin microbiome and skin transcriptome using Multi-Omic Factor Analysis (MOFA), virome profiling, and T cell receptor (TCR) sequencing in 10 MF patients from our previous study group.

Results: We observed that inter-patient transcriptional heterogeneity may be largely attributed to differential activation of T cell signalling pathways. Notably, the MOFA model resolved the heterogenous activation pattern of T cell signalling after denoising the transcriptome from microbial influence. The MOFA model suggested that the outgrowing strain evoked signalling by non-canonical NF-κB and IL-1B, which in turn may have fuelled the aggravated disease course. Further, the MOFA model indicated aberrant pathways of early thymopoiesis alongside enrichment of antiviral innate immunity. In line with this, viral prevalence, particularly of Epstein-Barr virus (EBV), trended higher in both lesional skin and the blood compared to nonlesional skin. Additionally, TCRs in both MF skin lesions and the blood were significantly more likely to recognize EBV peptides involved in latent infection.

Conclusions: First, our findings suggest that with its virulence factor spa fuels MF progression through non-canonical NF-κB and IL-1B signalling. Second, our data provide insights into the potential role of viruses in MF aetiology. Last, we propose a model of microbiome-driven MF aetiopathogenesis: Thymocytes undergo initial oncologic transformation, potentially caused by viruses. After maturation and skin infiltration, an outgrowing, pathogenic strain evokes activation and maturation into effector memory T cells, resulting in aggressive disease. Further studies are warranted to verify and extend our data, which are based on computational analyses.