Abstract Background and Aims The hormone FGF23 is a biomarker associated with morbidity and mortality in individuals with and without kidney disease, but the role of FGF23 excess as a pathogenic factor is incompletely understood. Here, we investigated renal FGF23 signaling in kidney disease models. Method This study used three independent disease models: i) Male C57BL/6 mice (n = 16) were treated with Freud's adjuvant and nephrotoxic serum (NTS) to induce anti-glomerular basement membrane (anti-GBM) disease. ii) Female BALB/c mice (n = 8) were treated with adriamycin (doxorubicin) 10.5 mg/kg to induce focal and segmental glomerulosclerosis (FSGS). iii) Male DBA/2J mice (n = 20) were fed a diet containing adenine 0.2%. Control mice of the respective same strain and sex received vehicle injections. Mice of the anti-GBM and FSGS models and controls received intravenous injections of recombinant FGF23 1 µg or vehicle for six consecutive days (anti-GBM) or once (FSGS), with dissection 24 h after the last injection. Mice of the adenine model were dissected after 15 weeks. Following dissection, ex vivo precision-cut kidney slices (PCKS) were obtained from n = 4 mice and maintained for 24 at 37°C, 95% O2 and 5% CO2, followed by a 1 h treatment with recombinant FGF23 200 ng/ml or vehicle. Glomerular filtration rate (GFR) or renal retention parameters in serum and albuminuria were assessed. Kidneys underwent histological assessment using hematoxylin & eosin or Masson's trichrome. Renal RNA was isolated for bulk RNAseq. Differential gene expression was assessed to determine the effect of FGF23 treatment in disease-free controls, in each disease model, and the interaction between FGF23 X disease effect for the anti-GBM and FSGS groups. Gene set enrichment analysis was performed with a focus on the immune system. Results Anti-GBM mice showed a 45 ± 27% GFR decline (mean ± SD) compared to baseline, whereas controls showed an GFR increase of 11 ± 36% (p < 0.01). Kidney histology revealed an interstitial infiltrate. The FSGS model showed elevated serum creatinine and albuminuria, and hypercellularity in some glomeruli. Mice of the adenine model showed progressively increased blood urea nitrogen concentrations starting after 2 weeks, up to a 3-fold increase at week 14. Histology revealed interstitial fibrosis and tubular atrophy. Bulk RNAseq revealed proinflammatory transcriptional signatures after FGF23 treatment in all disease models, regardless of which timepoint or model: In the anti-GBM model, Hallmark gene sets of inflammatory response, interferon-α response, interferon-β response and allograft rejection were significantly increased after 6 days of FGF23 treatment, whereas disease-free controls showed no significantly changed immune gene sets after 6 days of FGF23. In FSGS mice, none of the Hallmark gene sets were changed after 1 FGF23 injection, but top individual gene expression affected by 1 FGF23 injection were an increased proinflammatory transcription factor Cebpb and an increased DNA damage transcript Dclre1b. Finally, the PCKS of adenine model revealed no significant Hallmark gene sets, but the two genes with differential expression were a proinflammatory macrophage/fibroblast chemokine Ccl4 (increased) and an immunoglobulin heavy variable Ighv2-2 (decreased). Conclusion FGF23-driven patterns of proinflammatory gene expression were observed in three experimental models of kidney disease, both in vivo and ex vivo assessment of FGF23 signaling, thus excluding effects of FGF23 on other organs. As these findings were limited to transcriptional analyses, the present data warrant further confirmation and exploration of FGF23 as a pathogenic agent in kidney disease.
Read full abstract