Authors' Reply: Nephron-Segment Composition and Apparent Fatty Acid Oxidation Suppression in Pkd1KO+Glis3KO Kidneys

We thank Chen et al. for their thoughtful and constructive comments1 regarding our JASN study2 and for highlighting the important issue of nephron-segment composition in the interpretation of bulk kidney transcriptomic data. To begin to address the concern that the observed suppression of fatty acid metabolic process in Pkd1KO+Glis3KO kidneys compared with Pkd1KO kidneys2 may reflect differences in proximal tubule composition rather than genotype-dependent transcriptional regulation, we first examined the expression of the canonical segment markers suggested by the authors in our bulk RNA-Seq data. None of these marker genes (Lrp2 and Slc34a1 for proximal tubule, Slc12a1 for thick ascending limb, and Aqp2 for collecting duct) were significantly differentially expressed on a per-gene basis when comparing Pkd1KO+Glis3KO to Pkd1KO kidneys. We next performed an additional enrichment analysis using a curated proximal tubule marker gene set from a published segment-resolved mouse kidney transcriptome.3 This proximal tubule marker gene set contains 32 genes, of which six showed mild downregulation Pkd1KO+Glis3KO kidneys compared with Pkd1KO kidneys. However, when this gene set was tested as a custom term alongside gene ontology terms using a unified enrichment framework, the enricher function in clusterProfiler, “fatty acid metabolic process” remained the top-ranked enriched term, with a strong adjusted P value of 2.08×10−10. By contrast, the proximal tubule marker gene set showed only modest enrichment (adjusted P value of 9.89×10−4), significantly weaker than that observed for fatty acid metabolic pathways. In addition, gene set enrichment analysis did not identify significant enrichment of the proximal tubule marker gene set (adjusted P value > 0.05). Together, these results suggest that the prominent downregulation of fatty acid metabolic process in Pkd1KO+Glis3KO kidneys compared with Pkd1KO alone is unlikely to be primarily driven by changes in proximal tubule abundance and instead is more consistent with genuine transcriptomic changes within the tubular epithelium. We thank the authors again for their insightful comments and appreciate the opportunity to clarify this point. We believe that these additional analyses further strengthen the interpretation that impaired fatty acid metabolism represents a molecular consequence of Glis3 loss in the context of Pkd1 deficiency.