The effect of Dnaaf5 gene dosage on primary ciliary dyskinesia phenotypes

DNAAF5 is a dynein motor assembly factor associated with the autosomal heterogenic recessive condition of motile cilia, primary ciliary dyskinesia (PCD). The effects of allele heterozygosity on motile cilia function are unknown. We used CRISPR-Cas9 genome editing in mice to recreate a human missense variant identified in patients with mild PCD and a second, frameshift-null deletion in Dnaaf5. Litters with Dnaaf5 heteroallelic variants showed distinct missense and null gene dosage effects. Homozygosity for the null Dnaaf5 alleles was embryonic lethal. Compound heterozygous animals with the missense and null alleles showed severe disease manifesting as hydrocephalus and early lethality. However, animals homozygous for the missense mutation had improved survival, with partially preserved cilia function and motor assembly observed by ultrastructure analysis. Notably, the same variant alleles exhibited divergent cilia function across different multiciliated tissues. Proteomic analysis of isolated airway cilia from mutant mice revealed reduction in some axonemal regulatory and structural proteins not previously reported in DNAAF5 variants. Transcriptional analysis of mouse and human mutant cells showed increased expression of genes coding for axonemal proteins. These findings suggest allele-specific and tissue-specific molecular requirements for cilia motor assembly that may affect disease phenotypes and clinical trajectory in motile ciliopathies.


Generation of
Generation Sequencing (NGS)-based validation for cleavage at the target sites. Two single stranded oligo DNA donors (ssODNs) were designed based on the more active gRNA, target sequence 5'-CACACAGCAGCAGATGCTCCAGG. ssODN1 containing the C498F mutation, as well as a silent mutation that abolishes the protospacer adjacent motif (PAM) site for the gRNA, and 60 bases of homology arms on each side was used. A second ssODN2 that only has the silent mutation that abolishes the PAM site for the gRNA, termed the block only ssODN, was also used. The gRNA/Cas9 protein complex was transfected with the ssODNs into N2A cells and validated by next generation sequencing (NGS) for efficient incorporation of the mutations before the reagents were used in for mouse zygote injection. The gRNA/Cas9 complex with both ssODNs were electroporated into single cell C57BL/6 strain embryos. Embryos (20-25) were transferred to each pseudo-pregnant C57BL/6 mother (1). Live born mice were genotyped using NGS, as during validation. Mice were contained in a microorganism barrier facility and all lines were bred in a single room. DNA was extracted from tail biopsies and submitted to NGS to identify genotypes. High speed video-microscopy of multicilated cells. Cells were imaged live and recorded using a Nikon inverted microscope modified with lenses that use phase contrast and Hoffman modulation contrast (NAMC; Nikon). The microscope is enclosed in a customized environmental chamber maintained at 37 °C as described (2,3). Images were captured by a high-speed video CMOS camera and processed with the Sisson-Ammons Video Analysis system (Ammons Engineering, Mount Morris, MI). Cilia beat frequency was analyzed in at least 5 fields obtained from each preparation, after visually confirming ciliated cells in the analyzed areas.
Computer-assisted sperm analysis (CASA). Mice were euthanized and sperm obtain from the cauda epididyma. To assess sperm motility, 3 µL aliquots of sperm suspension was placed into a 20 microns Leja standard count 4 chamber slide, pre-warmed at 37 °C. Approximately 1x10 5 colony-forming units (CFU) was suspended in a total volume of 20 mL of sterile PBS. After oral intubation, two 10 ml volumes of bacteria were delivered into the trachea, 60 seconds apart. After 1 hour, mice were euthanized by cervical dislocation. The lungs and trachea were removed separately, homogenized, and cultured in serial dilutions in triplicates on agar plates. Following incubation at 37 °C for 16 hours, colonies were counted to determine cfu per sample. The burden of bacteria recovered in the tissue sample was determined as the mean CFU recovered, corrected for animal weight.
Brain MRI imaging and ventricle volume quantification. Mice were anesthetized during MRI imaging using isoflurane (2.5% induction, 1.5% maintenance). Magnetic  (4). For brain studies, the diluted microbeads were deposited with a micropipette on the dorsal side, towards the anterior region of the lateral wall of the lateral ventricle.
For both tissues, the flow of microbeads was recorded using a 10x objective lens and10x eyepiece using the no "zoom" function of a Nikon Ti inverted microscope. Image capture was performed at 10 frames per second (fps) for 5 seconds per image, totaling 50 frames per image. Following acquisition, files were imported to ImageJ software (Wayne Rasband, NIH, v 1.52p) and were analyzed using Trackmate (v 5.2.0) (5, 6). Image scale for calibration was set to 0.714 micrometers (pixel width and height), voxel depth 0.00 mm and time interval 0.1 sec. Particle capture was performed using the Laplacian of Gaussian (LoG) detector for detection of individual particles. This algorithm was used for estimation of the size of the particle. Parameters for "blob diameter" and "threshold" were adjusted until capture of at least 80% of beads by visual examination. Most often, a "blob diameter" of 7 mm and "threshold" of 1 were sufficient to capture most beads in the study.
Another algorithm "Linear motion LAP tracker" was used to track the particles/beads with an initial search radius of 15 mm, search radius of 10 mm and 2 as maximum frame gap.
Tracks generated with beads/particles were filtered using the mean quality filter with threshold set to auto. Tracks remaining after this filtration step were used in final statistical analysis and image creation. After detection and analysis, the data was exported using "Track Statistics". The images containing color-coded tracings of the tracks were screen captured and stitched together to create a composite whole ventricle tracking image.
Cilia Isolation for mass spectroscopy. Ciliary axonemes were isolated from the surface of highly ciliated airway cells by application of cilia buffer as described, with some modifications (7). Cells cultured in 6 Proteome Discover. Samples were further normalized and analyzed as previously described (9,10).
Single cell RNA sequence analysis of airway epithelial cells. Cultured primary airway cells were prepared for scRNAseq by dissociating the ALI cultures as previously described (11). Cell viability was maintained above 80% across all samples. Library The samples were pre-processed using Seurat package (12), and filtered to remove stressed or dead cells (those with mitochondrial gene content of more than >25%), potential doublets (cells with more than 7500 genes detected), or low quality cells (those with less than 2500 genes detected). clustering was performed using the FindClusters function in Seurat. Clusters of cells were manually annotated according to their known marker gene, and gene expression and clustering results were displayed on UMAP.