A gene network to predict the clinical response to whole lung lavage (WLL), in pulmonary alveolar proteinosis (PAP)

Rationale: PAP is a ultra-rare disease, where surfactant accumulation within alveolar spaces could lead to respiratory failure and death. WLL is the current standard treatment for PAP, nevertheless disease persistence after treatment is reported in 30% of cases. Aim: This work is aimed to identify a common gene co-expression network able to predict the outcome of the WLL: total resolution/persistent improvement vs transient resolution/progressive deterioration. Methods: We applied a weighted genes co-expression network analysis to transcriptional profiles of total RNA, collected from peripheral blood mononuclear cells (PBMC) of 16 PAP patients who underwent WLL. After a 24-month follow up, they were dichotomized in positive vs negative outcome and a microarray analysis, with around 20000 genes screened (SurePrint G3 Human Gene, Agilent) was performed. Finally Gene Ontology (GO), Reactome, KEGG, and Disease Ontology enrichment analysis was performed for differentially expressed genes (DEGs). Results: Four biologically meaningful modules of genes highly connected were identified. In particular, one of these (comprehending 111 genes with a p value Conclusion: For the first time we provide a modular profile of PAP which could explain the regulatory mechanism underlying the clinical outcome.