Last year, we reported that the C-terminal tail (CTT) of PC1 is cleaved and localizes to the mitochondria. While we could detect the cleavage product by immunoblot (IB) in a cell line with modest over-expression of PC1, and easily show by immunofluorescence (IF) that CTT localizes to the mitochondria when the short fragment was expressed as a recombinant protein, we could only visualize the mitochondrial pattern in a subset of cells that had over-expression of the full-length cDNA. In an effort to improve our ability to track the endogenous protein, we have generated a mouse line with a 3X HA epitope tag and eGFP knocked into the C-terminus of PC1 by CRISPR. Offspring of Pkd1-eGFP-3HA X Pkd1-KO crosses are born at the expected frequency, and Pkd1-eGFP-3HA homozygotes aged up to 18 months are healthy and have normal renal and liver histology, indicating that the modified allele functions normally. The PC1-eGFP-3HA fusion proteins (full-length and multiple PC1-cleavage products (N-terminal fragment NTF, C-terminal fragment CTF, and C-terminal tail CTT) can be reliably detected by IP and IB in various tissues as well as in cultured murine embryonic fibroblasts (MEFs) and renal epithelial cells. We also can detect the protein unambiguously in primary cilia of renal epithelial cells by IF, though we have struggled to detect the protein unambiguously anywhere in MEFs, or in any other cell location in renal epithelial cells despite testing multiple antibodies, fixation methods and confocal/super-resolution microscopes. Attempts at live cell imaging to date also have been disappointing despite partnering with various NIH microscopy cores including one that specializes in single molecule imaging. The problem appears to be that endogenous expression levels of PC1 are extremely low, at the level of background fluorescence. We continue to explore strategies to better visualize the protein using GFP nanobodies, tissue-clearing methods, and additional single-molecule imaging methods. We also have been using the epitopes to affinity purify PC1 and its interactome from mouse tissues for analysis by mass-spectrometry, though the very low level of endogenous expression has also made this a technically challenging endeavor. We have made steady progress, however, and we are currently performing multiple replicates to confirm identify of putative positive hits. Given that prior studies had reported impaired fatty acid metabolism in PKD cells and tissues, and that congenital diseases of peroxisome biogenesis cause various developmental/metabolic phenotypes including renal cysts and aberrant mitochondrial morphology, we hypothesized PC1 might directly/indirectly affect peroxisomal activity and thereby alter mitochondrial behavior. We therefore assessed peroxisome biogenesis in control and Pkd1-/- cells and whether PC1 CTT targeted peroxisomes. We found no difference in the rates of biogenesis and that the CTT localizes exclusively to the mitochondria. We also assessed levels of long-chain/very long-chain fatty acids in kidney tissues of control and Pkd1 mutant mice by mass spectrometry and the rate of peroxisomal-specific -oxidation in control and Pkd1-/- cells using isotope-labeled docosanoic acid followed by mass-spectrometry. We found little difference in peroxisomal -oxidation rates between WT and Pkd1 -/- cells, though two out of the three isogenic Pkd1 mutant cells did have a higher composition of very long chain fatty acids (VLCFA). However, the accumulation of VLCFA was not observed in kidney tissues and the global levels of fatty acids were lower in Pkd1 mutant mice. In sum, these data suggest that the peroxisome is not associated with ADPKD pathogenesis, thereby implicating mitochondrial change as a primary factor in dysregulated cellular metabolism in PKD. Last year, we had described our efforts to characterize a putative PC1 binding partner identified by mass spectrometry which was encoded by a gene that was among the most highly differentially expressed in our Pkd1 mutant transcriptomic studies. Lacking a rigorous cell culture model for assessing an interaction between PC1 and this protein, we opted to generate a mutant mouse by CRISPR. We have confirmed successful generation of several unique mutant lines and are currently breeding to produce homozygous mutants. As noted last year, we had previously generated a Pkhd1 mouse model that had lox P sites flanking exons 3 and 4 and used it to show that homozygous deletion of the exons resulted in renal cystic disease and hepatic fibrocystic disease, mimicking the human phenotypes. While our earlier studies had suggested that the intronic neomycin cassette inserted during gene targeting was benign, we subsequently found that the mice developed mild fibrocystic disease as they became aged. We hypothesized that the delayed onset of the phenotype in this mouse line indicated that the fibrocystic hepatic disease could be an acquired trait, challenging the current model that it is a strictly developmental condition. To test this hypothesis, we crossed the Pkhd1flox3-4 line to FlpeR to remove the neomycin cassette, confirmed that their liver histology remained normal up to 1 year of age, and then induced deletion of Pkhd1 at day P40 in Pkhd1delneo.flox34 mice. We found that all Cre positive animals had enlarged livers with visible cysts starting on day 200 and this progressed with age. We are currently characterizing the tissues with various markers to confirm that the apparent similarities between the developmental and acquired forms of liver disease are similar. On a histologic level, these studies show that adult inactivation of Pkhd1 mimics germline inactivation in causing an ARPKD liver phenotype in a mouse model, and suggests that adult carrier parents likely develop occasional liver cysts as a result of acquired inactivation of the germline WT allele. This finding also suggests that additional mechanisms, independent of developmental ductal plate remodeling and involution, may play a role in the ARPKD polycystic liver. While orthologous mouse models of ARPKD consistently reproduce the liver phenotype, the kidney disease is usually mild, of low penetrance and strain-dependent. We found that our Pkhd1del3-4 line, which initially developed renal cystic disease when in a mixed background, lost that property when backcrossed into both the C57Bl6 and 129Sv backgrounds. Among the proposed explanations, it was suggested that alternative splice variants of Pkhd1 could compensate for the lack of full length Pkhd1 transcript in the mouse models generated so far. Therefore, we tested the hypothesis that genomic deletion of most of the Pkhd1 locus, from exons 3 to 67, could bring out a more pronounced and penetrant kidney phenotype. We used a unique Cre-based strategy to recombine lox P sites on separate chromosomes of our two previously modified Pkhd1 alleles (lox 3-4; lox 67) to produce offspring with deletions between exon 2 and the 3UTR of Pkhd1 on the same haplotype. We confirmed the correct trans-chromosomal recombination by PCR and sequence analysis of the junction fragment and assessed for other large-scale rearrangements by 10X genomic sequencing. Pkhd1del3-67 het/het matings produce offspring at expected Mendelian ratios, and the homozygotes develop fibrocystic liver disease similar to that observed in our Pkhd1del3-4 mouse line. They have not developed significant kidney disease, even when aged up to one year, though they do develop an unusual eye phenotype. Studies are currently underway with collaborators in the intramural program of NEI to characterize the phenotype. These findings show that compensation by alternative splice variants cannot explain the low penetrance and mild renal phenotype variably reported in orthologous mouse models.
Showing the most recent 10 out of 33 publications
