Progressive liver, kidney, and heart degeneration in children and adults affected by TULP3 mutations

Abstract

Organ fibrosis is a shared endpoint of many diseases, yet underlying mechanisms are not well understood. Several pathways governed by the primary cilium, a sensory antenna present on most vertebrate cells, have been linked with fibrosis. Ciliopathies usually start early in life and represent a considerable disease burden. We performed massively parallel sequencing by using cohorts of genetically unsolved individuals with unexplained liver and kidney failure and correlated this with clinical, imaging, and histopathological analyses. Mechanistic studies were conducted with a vertebrate model and primary cells. We detected bi-allelic deleterious variants in TULP3, encoding a critical adaptor protein for ciliary trafficking, in a total of 15 mostly adult individuals, originating from eight unrelated families, with progressive degenerative liver fibrosis, fibrocystic kidney disease, and hypertrophic cardiomyopathy with atypical fibrotic patterns on histopathology. We recapitulated the human phenotype in adult zebrafish and confirmed disruption of critical ciliary cargo composition in several primary cell lines derived from affected individuals. Further, we show interaction between TULP3 and the nuclear deacetylase SIRT1, with roles in DNA damage repair and fibrosis, and report increased DNA damage ex vivo. Transcriptomic studies demonstrated upregulation of profibrotic pathways with gene clusters for hypertrophic cardiomyopathy and WNT and TGF-β signaling. These findings identify variants in TULP3 as a monogenic cause for progressive degenerative disease of major organs in which affected individuals benefit from early detection and improved clinical management. Elucidation of mechanisms crucial for DNA damage repair and tissue maintenance will guide novel therapeutic avenues for this and similar genetic and non-genomic diseases.

Keywords: ciliopathy; genetic disease burden; hypertrophic cardiomyopathy; internal medicine genetics; kidney failure; liver fibrosis; organ fibrosis; tubby-like proteins.

Graphical abstract

Figure 1

Identification of variants in TULP3 as a cause of progressive organ fibrosis in 15 affected individuals from eight unrelated families (A) Pedigrees for each of the eight reported families. Different massively parallel sequencing (MPS) approaches and GeneMatcher were used for identification of TULP3 variants. Affected individuals (black symbols) presented with progressive fibrotic liver disease and variable kidney and heart disease (full details in Table 1). TULP3 genetic changes are shown below symbols of individuals. Notably, the clinical features segregate with bi-allelic mutations in TULP3 (homozygous or compound heterozygous), implicating variants in TULP3 in autosomal recessive progressive fibrotic disease. (B) Graphical timeline showing the age of identification of liver, kidney, and heart phenotypes for all affected individuals (full details in Table 1). All affected individuals presented with complications of liver disease, and initial disease manifestations ranged from 2 to 33 years of age. Signs of chronic kidney disease were predominantly observed starting at the 2^nd decade. End-stage kidney disease was observed in three affected individuals at the age of 7, 51, and 55 years (family 6 [II.1] and family 1 [II.1 and II.2], respectively). Three affected individuals were affected by hypertrophic non-obstructive cardiomyopathy (HNCM) in their 6^th and 7^th decades of life (family 3 [II.2] and family 1 [II.1 and II.2], respectively). eGFR, estimated glomerular filtration rate; LTx, liver transplantation; RTx, renal transplantation; y, years.

Figure 2

Clinical imaging and histopathological evaluation of affected individuals (A) Imaging data from affected individuals. A computed tomography (CT) of an individual from family 1 (II.2) showing small renal cysts (white arrow). CT image of family 2 (II.1) showing large renal cysts (red arrow); note: transplant liver. Magnetic resonance imaging (MRI) scan from an individual from family 6 (II.1) shows splenomegaly (blue arrow) and end-stage atrophic fibrocystic kidneys (yellow arrow). MRI of affected individuals from family 7 (II.2) showing splenomegaly (blue arrow). Abdominal ultrasound pictures from family 8 (II.1) show renal microcysts with increased cortical echogenicity and hepatomegaly with increased tissue echogenicity. (B) Histopathology phenotypes observed in selected affected individuals. Family 1 (II.2), histological analyses of liver (H&E staining) showing hepatic fibrosis-dependent expansion of portal spaces and only little inflammatory infiltrate (scale bar, 100 μm). Family 2 (II.1) and family 1 (II.1), histological analyses of liver (H&E staining) compatible with hepatic cirrhosis (2 [II.1]: scale bar, 500 μm; 1 [II.1]: scale bar, 200 μm). Family 5 (II.1), orcein and H&E stains highlighting broad fibrous septa crossing the biopsy (scale bar, 200 μm). Family 6 (II.1), fibrosis pattern compatible with either biliary type cirrhosis or congenital hepatic fibrosis-like pattern but without ductal plate malformation, with only minimal portal inflammatory infiltrate and a moderate unspecific ductular reaction (scale bar, 200 μm). Family 8 (II.1), architectural disturbance of the hepatic parenchyma due to portal bridging fibrosis as assessed by Masson`s trichome staining (MTC). CK7 (cytokeratin 7) staining shows proliferation of dysmorphic bile ductules (scale bar, 200 μm). Family 6 (II.2), kidney sections: MG (May-Gruenwald) staining (scale bar, 1,000 μm), and H&E staining (scale bar, 100 μm) showing renal fibrosis and kidney cysts. (C) Cardiovascular magnetic resonance (CMR) images of the affected individuals from family 3 (II.2) that were acquired at the age of 53 years. The upper panel shows cine-CMR images obtained in diastole. Middle panel illustrates corresponding systolic cine-images. CMR reveals a severe, concentric, septally pronounced pattern of LV hypertrophy with additional intraventricular obstruction due to a kissing-wall phenomenon (blue arrows) and subsequent apical wall thinning with regional akinesia (yellow arrow). Corresponding late-gadolinium-enhancement (LGE) images are illustrated in the lower panels. Hyperintense areas are indicative of myocardial fibrosis (red arrows). A progressive and extensive non-ischemic pattern of LGE was depicted not only at the right ventricular (RV) insertion points (basal anteroseptal and inferoseptal LV wall) but also in the mid- to apical LV free wall and whole LV apex. This extensive and peculiar pattern of myocardial fibrosis cannot be explained by classical hypertrophic cardiomyopathy (HCM) but in contrast indicates a systemic disease with cardiac involvement.

Figure 3

Inactivation of zebrafish tulp3 causes adult liver and kidney disease (A) Left: schematic showing conservation between human TULP3 and zebrafish Tulp3. Amino acid sequences aligned with the Clustal Omega MView tool. Tulp3 shows 54.5% overall homology with TULP3 particularly within the IFT-A-binding (red) and Tubby (purple) domains (61% and 75%, respectively). The lower part of the figure shows mapping of the exons onto Tulp3 protein structure and shows the position of CRISPR-Cas9-mediated deletion in exon 5 leading to an early stop codon, p.Asp106Glyfs2Ter. Right: Semiquantitative RT-PCR and qPCR revealed a strong reduction in tulp3 mRNA expression in MZtulp3 mutant embryos compared to the respective control indicating a functional tulp3 null mutation. Sanger sequencing confirmed the 5 bp deletion in exon 5. Error bar represents SEM; ∗p < 0.05 (one-sample t test). (B) Semiquantitative RT-PCR analysis of tulp3 expression during development (left) and in isolated adult zebrafish tissues (right). tulp3 is expressed in various adult tissues, including liver, kidney, and heart; ef1α was used as housekeeping gene. Hpf, h post fertilization; dpf, days post fertilization; 15 S, 15-somites stage. (C) Histological analyses of liver samples (H&E stain) isolated from adult (18 months) tulp3 +/+ wild-type and tulp3 m/m mutant zebrafish clutchmates. Liver sections of tulp3 m/m zebrafish show cytoplasmic clearing of hepatocytes indicating steatosis (scale bar upper panel, 100 μm; scale bar lower panel, 10 μm). (D) Histological analyses of kidney samples (H&E stain) isolated from adult (18 months) tulp3 +/+ wild-type and tulp3 m/m mutant zebrafish clutchmates. Kidney sections of tulp3 m/m zebrafish show mild cysts (black arrow), observed in both proximal and distal kidney tubules (scale bar upper panel, 50 μm; scale bar lower panel, 10 μm). CD: collecting duct. (E) Histological analyses of heart samples (H&E stain) isolated from adult (18 months) tulp3 +/+ wild-type and tulp3 m/m mutant zebrafish clutchmates (scale bar, 100 μm). Microtome sections of the adult zebrafish ventricle. We did not note any fibrotic events or cellular changes in the hearts of these animals at 18 months. (F) Boxplots for liver and kidney phenotypes observed in tulp3 m/m adult zebrafish. For indirect quantification of cytoplasmic clearing, the nuclei of hepatocytes in visual fields of tulp3 m/m compared to tulp3 +/+ clutchmates were analyzed, showing a significant reduction in nuclei in tulp3 m/m zebrafish. ^∗p = 0.0159 (two-tailed, unpaired Student’s t test). An increased cystic index score was observed in tulp3 m/m zebrafish kidney compared to tulp3 +/+ clutchmates (n = 5). ^∗∗∗p < 0.001 (two-tailed, unpaired Student’s t test).

Figure 4

TULP3-affected individuals’ cells show defects in ciliary composition and increased DNA damage (A) Representative confocal micrographs assessing the effect of TULP3 mutations on ciliary composition in urine-derived renal epithelial cells (URECs) (family 2 [II.1], c.1223G>A [p.Arg408His]) and compared to age- and sex-matched control cells. Serum-starved cells were stained with antibodies against acetylated tubulin, GPR161, ARL13B, and INPP5E. Cell nuclei were counterstained with DAPI. Affected-individual-derived URECs showed significantly reduced ciliary localization of GPR161, ARL13B, and INPP5E. Left panel: staining of control and affected individual (family 2 [II.1)]) URECs (scale bar, 5 μm). Right panel: corresponding quantification of GPR161-positive cilia and ciliary signal intensities of ARL13B and INPP5E. (B) Representative confocal micrographs assessing the effect of TULP3 mutations on ciliary composition in fibroblasts derived from an affected individual (family 6 [II.4], c.1023+1G>A/c.70C>T [p.Arg24Ter]). Compared to age- and sex-matched control cells affected-individual-derived fibroblasts showed significantly reduced ciliary localization of GPR161, ARL13B, and INPP5E (scale bar, 5 μm). Right panel: corresponding quantification of ciliary signal intensity levels. For (A) and (B): ^∗p < 0.05; error bars show SEM (two-tailed, unpaired Student’s t test performed on the means of three independent experiments). (C) Interaction of human TULP3 with SIRT1. FLAG-tagged full-length human TULP3 was co-transfected with V5-tagged full-length SIRT1 in HEK293T cells. SIRT1 was detected in TULP3 precipitates (FLAG-M2 beads for immunoprecipitation [IP], anti-V5 for immunoblotting [IB]), and correspondingly TULP3 was detected in SIRT1 precipitates (V5 beads for IP, anti-FLAG for IB); kDa, kilodalton. (D) Increased DNA damage response (DDR) in affected-individual-derived URECs (family 2 [II.1]). γH2AX was used as an immunocytochemical marker of DDR and intensity of nuclear signal was compared in the affected individual and sex- and age-matched control URECs at the same passage number (scale bar, 20 μm). A significant increase in γH2AX signal was detected in affected-individual-derived URECs compared to control cells. ^∗p < 0.05; error bars represent SEM (two-tailed, unpaired Student’s t test on three independent experiments).

Figure 5

Increased levels of WNT-signaling-, TGF-β-signaling-, and cardiomyopathy-associated genes in TULP3-affected individual’s cells (A) RNA sequencing results performed in healthy control and fibroblasts derived from affected individual (family 3 [II.2]). Differentially regulated genes (DEGs) were identified by the gene set analysis method GAGE (generally applicable gene-set enrichment). The most dysregulated genes associated with WNT signaling or cardiac disease are labeled. (B) Enrichment analysis of signaling pathways. Pathways were considered significant with adjusted p values (Benjamini-Hochberg) p < 0.05. Among these pathways, significantly dysregulated genes associated with WNT signaling, TGF-β signaling, and cardiac muscle contraction/microRNAs in cardiomyocyte hypertrophy were identified. (C) Gene expression heatmaps for differentially regulated genes from the indicated GSEA terms “cardiac muscle contraction/microRNAs in cardiomyocyte hypertrophy” (left) and WNT signaling (right). Each column represents an individual sample from control or affected-individual-derived cells.

Figure 6

Converging pathomechanisms of fibrosis in TULP3-affected individuals A schematic of the hypothesized disease mechanism of fibrosis in TULP3-affected individuals. Shown are cartoons of the primary cilium with axoneme, intraflagellar transport machinery and membrane bound receptors and ciliary proteins, transcriptional changes, affected pathways, and clinical outcomes. On the left is the TULP3 wild-type healthy state versus TULP3 mutant disease state (right), separated by the vertical dotted line. At the cilium, TULP3 mutations results in defective trafficking of ARL13B, INPP5E, and GPR161 (and likely other GPCRs). Disrupted ciliary composition leads to a loss of regulatory signals from the cilium, causing dysregulated profibrotic WNT and TGF-β signaling pathways, either directly or indirectly. Green crosses represent correctly localized receptors and proteins. Red crosses represent loss of receptors or proteins from the cilium. We demonstrate that TULP3 interacts directly with SIRT1, a key regulator of fibrosis. We hypothesize that disruptive mutations in TULP3 lead to reduced SIRT1 modulation of profibrotic signaling pathways through a yet unknown mechanism. We propose that TULP3 is a key regulator of fibrosis that functions at multiple levels. Disruption of these regulatory mechanisms converge and results in chronic activation of profibrotic signaling cascades leading to progressive fibrosis in TULP3-affected individuals. Black arrows represent activation, bar headed lines represent inhibition, and red Xs represent loss of elements from the network. Created with https://biorender.com/.