Angiomotin mutation causes glomerulopathy and renal cysts by upregulating hepatocyte nuclear factor transcriptional activity

Dear editor, We identified angiomotin (AMOT) as a novel candidate gene for X-linked recessive nephropathy associated with glomerular disease, tubulopathy and progressive kidney cystic disease. Transgenic rats carrying this mutation recapitulated the human phenotype to some extent. Orthogonal methods implicate the hepatocyte nuclear factor (Hnf) family of transcription factors, particularly Hnf4α and Hnf1β, in the development of Amot mutation-induced nephropathy. Monogenic causes involving more than 50 genes have been identified in 25%–30% of young patients with steroid-resistant nephrotic syndrome (SRNS).1, 2 We report a family with X-linked recessive early-onset SRNS and Fanconi syndrome (Figure 1A–D). Exome sequencing identified NM_001113490.1:c.148A>G (p. S50G) variant in exon 1 of AMOT, which encodes AMOT. This variant causes the substitution of serine to glycine at the 50th position of AMOT-P130 (Figures 1E,F and S1). To study the effects of this mutation in rats, we used the CRISPR/Cas9 system to substitute both serine and threonine residues at positions 49–50 with proline and glycine (termed ‘PG’ rat) or proline and serine (termed ‘PS’ rat), respectively (Figure S2). PG rats developed higher body weights and albuminuria (Figure 1G,H). At 6 months old, PG rats developed focal segmental glomerulosclerosis, tubular dilatation, tubulointerstitial inflammation and fibrosis (Figure 1J–L). Systematic scoring of the light microscopic changes revealed no kidney abnormalities in wild type (WT) and PS rats, while the PG rats had minimal to moderate changes (Figure 1M). We then examined the ultrastructural changes at 3 months old. While WT rats had normal morphology, the podocytes in PG rats showed extensive foot process effacement and detachment, revealing areas with nude glomerular basement membranes (Figure 1N). At 21 days old, macroscopic cysts were also noted on the kidney surfaces in PG rats (8/11 = 72.7%; Figure 1O). In ex vivo podocytes, the PG mutation disrupted the expression of F-actin, causing abnormal formation of stress fibers (Figures 2A and S3), and reduced cell stiffness (Figure S4). Abnormal zonula occludens-1 (Zo-1) expression and distribution were observed in mutant rat kidneys (Figure 2B), ex vivo podocytes (Figure 2C) and proximal convoluted tubule cells (PCT; Figure 2D). In addition, the expression of occludin, another tight junction protein, was reduced in Amot mutant PCT cells (Figure 2E). In fluorescein isothiocyanate (FITC)-albumin flux analysis, FITC-albumin flux across the PCT monolayer was increased (Figure 2F,G) in the PG mutant cells, implying that the PG mutation resulted in perturbation in the tight junctions. AMOT interacts with multiple proteins of the Hippo signalling pathway, including the transcriptional effector Yes-associated protein.3 Our studies, however, suggested that the Hippo signalling pathway is not the crucial mechanism downstream of the Amot mutation (Figure S5). Single-cell RNA sequencing (RNA-Seq) revealed relatively high transcriptional levels of AMOT in podocyte and proximal tubule segment S1 cells. Thus, to decipher disease pathogenesis, RNA-Seq was performed on freshly isolated PCT cells since they are more easily isolated than podocytes. Hypergeometric Optimisation of Motif Enrichment (HOMER) was then applied to detect the transcription factor binding motifs in the promoters of the up- and down-regulated genes (Figure 3A–D). Here, we showed that Hnfs, namely Hnf1, Hnf1β and Hnf4α, were enriched in the promoters of PG-upregulated genes, suggesting that the Amot mutation caused an upregulation of Hnfs. Of note, HNF1A and HNF1B are associated with Fanconi syndrome and kidney cystic diseases in humans.4, 5 Gene set enrichment analysis (GSEA) of Hnf4α target genes (Table S3) corroborated the motif enrichment analysis, as it suggested that Hnf4α target genes were upregulated in PG rats (Figure 3E). To confirm this finding, the differential expression of four upregulated genes (Acox2, Aldh2, Aldob and Otc) known to be Hnf4α target genes was validated by quantitative polymerase chain reaction (Figure 3F). The genome-wide chromatin accessibility landscape was profiled by Assay for Transposase-Accessible Chromatin using sequencing (ATAC-Seq) in freshly isolated PCT cells (Figures 3G and S6). Putative transcription factor motifs in the ATAC-gained regions were identified with HOMER. Again, the nuclear transcription factor Hnf4α was enriched for both known motifs and de novo motifs. Another Hnf family member, Hnf1β, was also among the top 10 enriched motifs (Figure 3H,I). On further analysis of ATAC-seq data around Hnf4α motif sites, we confirmed its increased chromatin accessibility across the whole genome in the PG rat (Figure 3J). HNF4α is the master regulator of many hepatocyte-specific genes.6 Neither the WT nor the mutant AMOT bound directly to the HNF4α protein (Figure S7). Activation of the Hnf4α pathway in PG rat kidneys occurred in the absence of upregulated Hnf4α protein levels (Figure 3K). Instead, the AMOT mutation resulted in nuclear recruitment of HNF4α in PG rat tubules, corroborating findings in the patient's (II-4) kidney cyst tissue (Figure 3L,M). The importance of HNF4α activation to the PG phenotype is further supported by stress fiber formation and altered ZO-1 distribution when kidney cell lines are treated with Benfluorex, a known HNF4α activator (Figure S8).7 Considering that the GSEA of the differentially expressed genes revealed metabolic pathway alterations, we performed metabolic profiling of the rat plasma as an exploratory analysis to identify the metabolic pathways involved (Figure 4A). A total of 37 metabolites were identified as critical metabolites as a result of the Amot mutation (Table. S4). Pathway enrichment of the critical metabolites with MetaboAnalyst showed that the pentose and glucuronate interconversion pathway and the tricarboxylic acid (TCA) cycle were the most impactful metabolic pathways (Figures 4B and S9). The main altered molecules that account for these metabolic pathways are shown in Figure 4C,D. Both of these metabolic pathways were also significantly altered in the KEGG pathway analysis of the differentially expressed genes identified in RNA-Seq (Figure 4E). Rodent models of several genes associated with glomerulotubular nephropathy (PAX2, CRB2 and FAT1) have revealed different pathophysiological mechanisms.8-10 By implicating HNFs and downstream metabolic changes in AMOT nephropathy, we add to the current pathogenetic understanding of glomerulotubular nephropathies. In conclusion, we report a putative novel role of AMOT in causing glomerulotubular nephropathy in patients and rats, possibly by regulating the HNF4α pathway and the subsequent metabolic pathways. This study was supported by academic grants NMRC/CSA/0057/2013 and NMRC/CSA-INV/0015/2017 from the National Medical Research Council (NMRC), Singapore. The authors declare no conflicts of interest. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.