Loss of an extensive ciliary connectome induces proteostasis and cell fate switching in a severe motile ciliopathy

Abstract

Motile cilia have essential cellular functions in development, reproduction, and homeostasis. Genetic causes for motile ciliopathies have been identified, but the consequences on cellular functions beyond impaired motility remain unknown. Variants in CCDC39 and CCDC40 cause severe disease not explained by loss of motility. Using human cells with pathological variants in these genes, Chlamydomonas genetics, cryo-electron microscopy, single cell RNA transcriptomics, and proteomics, we identified perturbations in multiple cilia-independent pathways. Absence of the axonemal CCDC39/CCDC40 heterodimer results in loss of a connectome of over 90 proteins. The undocked connectome activates cell quality control pathways, switches multiciliated cell fate, impairs microtubule architecture, and creates a defective periciliary barrier. Both cilia-dependent and independent defects are likely responsible for the disease severity. Our findings provide a foundation for reconsidering the broad cellular impact of pathologic variants in ciliopathies and suggest new directions for therapies.

Keywords: Notch-signaling; airway; cilia; dynein; periciliary barrier; primary ciliary dyskinesia; proteomics.

Figure 1.. Loss of ciliary microtubular integrity of human multiciliated cells in CCDC39 and CCDC40 variants.

(A) Diagram of multiciliated cell, motile cilium cross-section with the nine microtubule doublets (DMT) linked to the major complexes. Dashed rectangle is expanded in Figure 1B. (B) Diagram of a 96-nanometer unit showing the location of the CCDC39/CCDC40 heterodimer (yellow and orange ribbon) on the DMT. The major ciliary complexes are shown. (C) Scheme for obtaining primary airway epithelial cells for culture at air-liquid interface (ALI). Cells are from normal individuals and those with CCDC39 and CCDC40 variants. (D) Immunoblot (IB) detection of CCDC39, CCDC40, and FOXJ1 in normal ALI cultured cells during differentiation. (E) Immunofluorescent (IF) detection of CCDC39 and CCDC40 in normal and variant cells cultured at ALI 28. Bar=10 μm (WU 182, WU146, respectively and in panels H-L). (F) IB detection of CCDC39 and CCDC40 in normal and human variant cells. (G) IB detection of CCDC39 and CCDC40 in Chlamydomonas wild-type and ccdc40 mutant. (H) Cilia beat frequency in normal (n=2) and CCDC39 and CCDC40 variant cells. (I) Cilia transport of microbeads on the surface of well-differentiated normal, CCDC39 and CCDC40 variant cells. Each point represents one bead. (J) IF detection of acetylated α-tubulin (Ace-TUB) in cilia isolated from human normal and variant cells. Detail shows examples of cilia with splaying in variants. Bar=10 μm (left) and 5 μm (right) (K) Quantification of cilia length from Panel J; n=47–50 cilia measured for each genotype from normal (n=3), CCDC39 and CCDC40. (L) Transmission electron microscopy (TEM) of cilia from normal and variant cells showing the microtubules in longitudinal and cross section. Bar=500 nm (top) and 100 nm (bottom). (M) TEM of cross-sections of cilia isolated from Chlamydomonas wild-type, ccdc39, and ccdc40. Asterisks indicate disorganized DMT. Arrows indicate an opening of the B-microtubule. (N) Quantification of open B-microtubules from panel M; n=200 of each genotype. The total loss of the central apparatus in the ccdc39 and ccdc40 mutant strains is significantly different from wild-type (p<0.0001) by chi-squared and Fisher’s exact testing. (O) Diagram of a cross-section of a DMT showing the ciliary complexes and the location of CCDC39/ CCDC40 on the DMT. Structures are as in panels A and B. The box outlined represents the region viewed at 900 in panel P. (P) Three-dimensional structure of the Chlamydomonas axoneme resolved by cryo-EM showing CCDC39 and CCDC40 and indicated proteins on the surface of the DMT. The N terminus of CCDC39/40 is labeled (N). The C-terminus extends from the A- to the B-tubule (C). Bar=10 nm. (Q) Detail of CCDC39/40 hook region from a predicted conformation of the C-terminal region of the CCDC39/40 heterodimer spanning from the A- to the B-microtubule. L845 is the location of the mutant residue in the temperature-sensitive ccdc39 Chlamydomonas mutant. In H, I, K. The bar indicates the medium. Differences between groups were determined using Kruskal-Wallis with Dunn’s Multiple Comparison Test; *p<0.05, **p<0.01 are shown.

Figure 2.. CCDC39/40 is required for assembly of a connectome of ciliary proteins.

(A) Scheme of cilia isolation from cultured human cells from normal (n=3) and CCDC39 variant (WU182) analyzed by mass spectrometry. (B) Immunofluorescent (IF) detection of Ace-TUB and basal body protein centrin in cells after cilia isolation. Bar=10 μm. (C) Heat map of mean levels of proteins normalized for tubulin in the CCDC39 variant compared to normal cilia. Proteins with direct CCDC39/40 contact noted by blue box. (D) CCDC39/40 ciliary address recognition protein (CARP) with associated structure and references to experimental validation. (E) Diagram of the relationship of CCDC39/40 with CARPs and ciliary structures (see Figure 1A for abbreviations).

Figure 3.. Ciliary radial spokes are differentially depleted in CCDC39/40 variants.

(A) Diagram of the radial spokes (RS) in Chlamydomonas and human in a 96-nm repeat along the DMT. Chlamydomonas has two complete and one partial RS. The human has three complete RS proteins at the bases. RSPH3 is shared by all RS. (B) Immunofluorescent (IF) detection of Ace-TUB, CCDC39, and RS proteins in normal and CCDC39 variant airways. Arrow shows indicated protein. Bar=25 μm. (C) IF detection of Ace-TUB, CCDC39, and radial spoke (RS) proteins in normal cells transduced with non-targeted or CCDC96-specific shRNA. Arrow shows indicated protein. Bar=25 μm. (D) TEM of isolated Chlamydomonas wild-type and ccdc39 axonemes. Repeating distances of radial spokes within a 96-nm repeat (32 nm) and between repeats (96 nm) are indicated. (E) Quantification of distance from panel D. (F) Immunoblot detection of RS head protein RSP16 shared by all RS in Chlamydomonas strains.

Figure 4.. Disorganization of microtubule inner proteins (MIPs) in CCDC39/40 variants.

(A) Proteomic analysis of MIPs in normal and CCDC39 variant human cilia (B) Immunofluorescent (IF) detection of TEKT3 in human cells (CCDC39, WU182). Airway, Bar=25 μm; cilia, Bar=2 μm. (C) IF detection of MIPs in cilia cultured cells (CCDC40, WU146). Cells, Bar=5 μm; cilia, Bar=2 μm. (D) MIPs in wild-type and ccdc39–2 Chlamydomonas identified by cryo-EM. DMT structure from wild-type has 48-nm periodicity applied and ccdc39–2 has 16-nm periodicity applied.

Figure 5.. CCDC39/40 traffics independently of connectome proteins.

(A) Scheme of motile ciliogenesis in air-liquid interface (ALI) cultures. (B) Immunofluorescence (IF) detection of basal body protein centrin, CCDC39, and CDH1 during basal body amplification. (C) IF detection of CCDC39 and CCDC40 at pre-cilia stage (arrow indicates co-localization). (D) IF detection of Ace-TUB and CCDC40 with proteins that contact CCDC39/40 in cells with emerging cilia (arrow indicates co-localization). (E) IF detection of Ace-TUB and CCDC40 with proteins that do not directly contact CCDC39/40 in cells with emerging cilia. (F) IF detection of connectome member CFAP100 and basal body protein Centrin during early cilia growth. (G) IF detection of Ace-TUB, CFAP73, and CFAP100 in mature normal and CCDC39 cells. (H) Immunoprecipitation of CCDC40 in cells from early (ALI 14) and late (ALI 40) stages. (I) IB of CCDC39, CCDC73, and CCDC100 in normal, and variant mature cells. Short (S) and long (L) exposure. (J) Proposed schema of cilia assembly in normal and CCDC39 variant. Bar=5 μm in B-G

Figure 6.. CCDC39/40 variant affect differentiation via Notch signaling.

(A) UMAP reductions of normal (n=6) and CCDC39/40 variant (CCDC39, WU182 and WU 157; CCDC40 WU146) differentiated airway epithelial cells showing clusters for basal (bas); secretory (sec); multiciliated (cil) cells. (B) Dotplot shows manually annotated, differentially expressed genes in each cluster that show significantly increased expression in the variant compared to multiciliated cells. Ciliated cells are in purple. (C) Immunofluorescent (IF) detection of Ace-TUB and PSMB6 in normal and CCDC39 (WU182) variant cells. Box identifies an example of PSMB6 in multiciliated cells. Arrow indicates the location of PSMB6 in the cytoplasm. (D) Quantitation of cell numbers in each cluster from A. Comparison of Sec2 normal vs. CCDC39/CCDC40 (39/40) variant, p=0.0025. (E) Immunohistochemistry for detection of PAS-positive mucous cells in the airway from a lung of a normal donor and a CCDC39 variant. (F) IF detection of α-tubulin (α-TUB) and MUC5AC in cell cultures from unique normal donor (n=4), unique CCDC39/40 variants (n=4), and PCD variants DNAAF5, DNAH5, DNAI1 (n=1 per genotype). (G) Quantitation of normal cells with genotypes CCDC39, CCDC40 and other PCD variants from Panels F and L. CCDC39 variant cells expressing control (Cont) and the CCDC39 Transgene (TG) are compared. (H) IF detection of α-TUB and MUC5AC in CCDC39 variant cells. Cil, multiciliated cell; Muc, MUC5AC cell; MC, Multiciliated-MUC5AC cell. (I) IF detection of MUC5AC and Ace-TUB in normal cells transduced with non-targeted CCDC39 or CCDC40 shRNA. (J) Quantitation of panel I. Normal cells transduced with non-targeted, CCDC39, or CCDC40 shRNA are compared. (K) IF detection of α-TUB and MUC5AC in iPSC cells that are cells derived from the control original iPSC from a normal donor, or CRISPR-Cas9 mediated deficiency of DNAH5 or CCDC40 in the normal iPSC. Bar=25 μm (L) Quantitation of panel K. Each point represents a microscope field of cells from the control original iPSC line, DNAH5 CRISPR knockout, or CCDC40 CRISPR knockout. The control cells are compared to the knockout lines. (M) IF detection of α-TUB and MUC5AC in transgene (TG) rescued variant cells transduced with a control (Cont) or FOXJ1-CCDC39 transgene. Quantitation of M is in panel G. (N) IF detection of α-TUB and MUC5AC in unique normal and CCDC39 and CCDC40 variant cultures (ALI >28) day then treated with vehicle or NOTCH inhibitor DAPT for two weeks. (O) Quantitation of panel N. Vehicle cells are compared to DAPT-treated for each group. Each point is the mean of fields of each condition in at least three independent experiments. Normal (n=3); CCDC39 (WU182, n=2); CCDC40 (WU146, n=2). The bar indicates the medium. Differences between groups were determined using Mann-Whitney in D, G, O or Kruskal Wallis with Dunn’s Multiple comparison test in J and L; ns=non-significant, *p<0.05, **p<0.01, ***p<0.001. Bar=5 m in C, H; Bar=25 μm in E, F, K, I, M, N.

Figure 7.. CCDC39/40 variants have disrupted periciliary barrier.

(A) Scanning EM of normal and variant (CCDC39, WU152; CCDC40, WU151) cilia from cells untreated and treated to disrupt the ciliary membrane. Bar=1 μm. (B) Immunofluorescent (IF) detection of Ace-TUB to identify splaying in cilia untreated and treated with buffer to remove cilia membranes from cultured cells. Bar=10 μm. (C) Quantitation of splaying in cilia from cultured cells from normal (n=2) and variant cells (CCDC39, WU152, WU182; CCDC40, WU146, WU151). (D) Scheme for isolation of fresh cilia by nasal brush biopsy and IF staining. (E) Quantitation of IF detection of splaying by Ace-TUB in cilia from normal (n=3) and variant cells (CCDC39, WU120, WU152; CCDC40, WU149, WU151). There was no difference between CCDC39 and CCDC40. (F) Scheme of normal CCDC39 transgene delivery to CCDC39 variant cells (WU182) using control and FOXJ1-CCDC39 rescue transgene lentiviruses (Lenti). (G) IF detection of Ace-TUB to identify cilia length in cells from CCDC39 variant transduced with control (top) and rescue (bottom) lentiviruses in panel F. Bar=10 μm. (H) Quantification of cilia length in CCDC39 variant cells from panel G, transduced with control or transgene (TG) lentivirus; n=150–200 cilia per condition in over 20 fields. (I) Quantitation of cilia beat frequency (CBF) in CCDC39 variant cells and cells transduced with control lentivirus, low or high concentration of the CCDC39 TG. (J) Quantification of cilia splaying in cells transduced with control and TG. n=300–500 cilia per condition from over 20 fields. (K) IF detection of MUC4 in the periciliary region of cultures from normal, DNAH5 (WU165) variant, and two CCDC39 variant cells non-transduced and following transgene rescue. Bar=5 μm. (L) Scheme of periciliary barrier assessment performed by addition of microparticles to the apical surface for 15 min prior to fixation and imaging. (M) Fluorescence particles relative to cilia in cultures from normal and variant (CCDC39, WU182; CCDC40, WU146; DNAAF5, WU108; DNAI1, WU103). Bar=5 μm. (N) Detection of fluorescent particles in CCDC39 variant (WU182) transduced with the rescue CCDC39 transgene as in panel F, Bar=5 μm. (O) Quantitation of fluorescent particles in the periciliary space from panels and L and M. Differences in the indicated conditions were determined compared to normal cells. In C, E, I, N the bar indicates the medium. Difference between groups determined by Kruskal-Wallis and with a Dunn’s Multiple Comparison Test. In H the unpaired two-tailed t-test. In J, the Mann-Whitney test ns=non-significant; ***p<0.001.