Structural defects in cilia of the choroid plexus, subfornical organ and ventricular ependyma are associated with ventriculomegaly

Wild- type, Bbs1^M390R/M390R, Bbs2^−/−, Bbs4^−/−, and Bbs6^−/− mice were generated and maintained as described previously [25]. All studies adhered to guidelines established for the care and use of experimental animals and were approved by the Animal Care and Use Committee of the University of Iowa.

Seven month-old wild- type and Bbs4 ^−/− mice (n = 3 for each genotype) were anesthetized with a mixture of ketamine (91 mg/kg) and xylazine (9.1 mg/kg) intraperitoneally (i.p.). The fur, skin, membrane and musculature on the surface of the skull were removed and a small hole was drilled into the skull at Bregma level 33. Evans Blue dye (5–10 μl, 2% in 1X PBS; phosphate-buffered saline minus Ca⁺² and Mg⁺²) was injected slowly into the lateral ventricle using a 25 μl Hamilton syringe to visualize the movement of CSF. The syringe was left in the ventricle post-injection to prevent CSF loss and the dye was allowed to circulate with the CSF for 20 min. Mice were euthanized by an overdose of ketamine and xylazine followed by cervical dislocation and whole mice were immediately frozen at −20°C. The next day, frozen heads were cut in the axial, coronal and sagittal planes and photographed with an Olympus SZX12 stereomicroscope while the tissue was still frozen to prevent dye diffusion.

To examine the ultrastructure of ependyma from the lateral and third ventricles and choroid plexuses from the lateral ventricles prior to the onset of ventriculomegaly seen at P9, newborn (P0) and P2 wild- type and Bbs1^M390R/M390R animals (n = 3 for each age and genotype) were euthanized and the skin surrounding the skull was removed. The intact skull was placed in 4% paraformaldehyde in 1X PBS for 4–5 hr at 4°C, and transferred to 2.5% glutaraldehyde-0.1 M cacodylate buffer overnight at 4°C. The next day, intact skulls were stabilized in 2% agarose and 100 μm-thick sections were cut with a vibratome, post-fixed with 1% OsO4, rinsed, dehydrated in a series of alcohol and flat embedded in Eponate-12 epoxy resin (Ted Pella, Redding, CA). Tissues were sectioned (85 nm thickness) with a Leica UC-6 ultramicrotome (Wein, Austria). Sections were counterstained with uranyl acetate and lead citrate, and photographed with a JEOL JEM-1230 (Tokyo, Japan) transmission electron microscope. Electron microscopic images were taken with a Gatan UltraScan 1000 (Pleasonton, CA) 2kx2k CCD digital camera.

To study the progression of ventriculomegaly in P9 Bbs1 ^M390R/M390R and Bbs2 ^−/− , Bbs4 ^−/− and Bbs6 ^−/− mice up to 2 years of age (n = 3 for each age and genotype), the protocol was modified to include transcardial perfusion of anesthetized mice with 1X PBS, followed by a solution of 4% paraformaldehyde-0.25% glutaraldehyde. Brains were removed and post-fixed overnight with 2.5% glutaraldehyde-0.1 M cacodylate buffer. The next day, 100 μm-thick sections were cut with a vibratome and processed for TEM analysis of the lateral and third ventricles, choroid plexus, SCO and SFO as described above.

Sagittal MRI images of wild- type and Bbs1 ^M390R/M390R , Bbs2^−/−, Bbs4^−/−, and Bbs6^−/− mouse brains (n = 3 for wild type and each BBS mutant strain) was performed as described [25]. All images were collected from 6 month-old mice with the exception of Bbs1^M90R/M390R mice which were 3.5 months old.

To test whether obstruction of intra- or extra-ventricular CSF flow contributed to ventriculomegaly in BBS mutant mice, we visualized in vivo CSF flow in 7 month-old wild -type and Bbs4 ^−/− mice with pronounced enlargement of the lateral and third ventricles. Evans Blue dye was injected into the lateral ventricle of anesthetized mice, allowed to circulate with the CSF throughout the brain for 20 min., following which animals were euthanized. Axial, coronal or sagittal sections from mutant mouse brains revealed no obstruction of intra- or extra-ventricular CSF flow, as seen by the presence of dye in the enlarged lateral ventricles, the third and fourth ventricles, the cisterns, and subarachnoid space (Figure 1A-E). Similar results were observed in comparably aged Bbs1^M390R/M390R mice (data not shown). Absence of CSF flow obstruction was further documented in sagittal plane MRIs of 3.5 month-old Bbs1^M390R/M390R mice, and 6 month-old Bbs2 ^−/− , Bbs4 ^−/− , and Bbs6 ^−/− mice (Figure 1F). As reported earlier [25] there is no significant difference in the degree of ventriculomegaly in BBS mutant mice between 3.5-6 months of age. Altogether, these data indicated a communicating form of hydrocephalus in these BBS mutant mouse strains.

CSF is formed by the net directional transport of sodium, chloride, bicarbonate and water from the rich choroidal blood supply across the choroid plexus epithelium via basal membrane transporters and apical membrane ion channels and aquaporin water channels into the ventricles [28]. Integrity of the choroid plexus epithelium, therefore, is critical for maintaining the blood:CSF barrier and for CSF homeostasis. To determine whether choroid plexus ultrastructure is affected in BBS mutant mice, we used TEM to examine the brains of newborn (P0) wild- type and BBS mice and during the early stages of ventriculomegaly at P9 (the earliest stage other than P0 used in the study) (Figure 2A-D) and during its progression at 15 weeks. At birth, there was no evidence of ventriculomegaly (data not shown) and the Bbs1^M390R/M390R choroid plexus epithelium appeared normal and contained numerous microvilli together with solitary primary cilia and clusters of primary cilia on the apical membrane surface of some of the cells (Figure 3 A,B). Tight intracellular junctions of the apical epithelium were intact, as were the basolateral membranes. Comparable results were observed in 15 week-old Bbs4^−/− and Bbs6 ^−/− mice (Figure 3C-F). Based on these findings, there appeared to be no gross disruption of the blood:CSF barrier at the ultrastructural level.

In light of the growing body of evidence that choroid plexus primary cilia play an important role in signaling and CSF homeostasis [7‐9], we used TEM to examine their ultrastructure in young (P0 and P9) BBS mutant mice as well as in 15 month-old mice when ventriculomegaly had progressed. Seen in cross-section, the choroid plexus epithelium of Bbs1 ^M390R/M390R mice had multiple intact basal bodies (Figure 4Ac). The Y-shaped links of the ciliary necklace that connect the microtubule doublets to the ciliary membrane in the transition zone; the region of the basal body where the triplet microtubules transition to the doublet microtubules of the axoneme appeared to be intact in the P0 mutant mice (Figure 4B). Compared to the abundant ependymal cilia lining the ventricles, primary cilia on the apical surface of the choroid plexus were rare and occurred in clusters. Examination of cross-sections of 15 month-old wild- type and Bbs6 ^−/− mouse choroid plexus cilia along the ciliary shaft distal to the transition zone showed occasional clusters of primary cilia with typical derivatives of the 9 + 0 arrangement of microtubule doublets including 8 + 1 axonemes with displacement of one pair of peripheral doublets, and rarely, 7 + 2 axonemes with displacement of 2 pairs of peripheral doublet (Figure 4C, D) in agreement with previous studies [7, 9, 29, 30].

When clusters of choroid plexus cilia from P0, P2, and P10 mutant animals were viewed longitudinally, a sub-population of cilia contained vesicle-like inclusions and electron-dense material along the ciliary shaft distal to the transition zone. This material was larger than the IFT-like particles typically seen in the choroid plexus cilia (Figure 4F-H). The mutant cilia also exhibited disruption of the ciliary axoneme that was not seen in comparably aged wild- type mice (Figure 4 E-H). In some cilia, the electron-dense material was found between the peripheral microtubule doublets of the axoneme and the ciliary plasma membrane where IFT particles are typically located (Figure 4F, G), while in others it was found in the ciliary lumen, surrounded by the peripheral microtubule doublets (Figure 4H).

The structural abnormalities observed in some choroid plexus cilia as early as the time of birth prompted us to examine more closely the ependymal cilia of BBS mutant mice during the progression of ventriculomegaly. Similar to earlier studies [31], we observed in Bbs1 ^M390R/M390R newborn mice the normal stages of primary ciliogenesis in the radial glia progenitor cells that subsequently transform into ventricular ependymal cells: (1) the mother centriole surrounded by the secondary centriolar vesicle, forming the ciliary pocket (Figure 5A, B); (2) docking of the mother centriole at the apical cell surface and fusion of the ciliary membrane with the ependymal plasma membrane (Figure 5C, D) and (3) elongation of the mature centriole (basal body) to form the cilium with the help of secondary ciliary vesicles and intraflagellar transport (IFT) (Figure 5E, F). Based on these observations, the BBS mutant ependymal primary cilia appeared to have assembled correctly.

In mice, ependymal primary cilia are replaced by multiple motile cilia during the second week of postnatal life [32, 33]. As early as P9, we observed ultrastructural ciliary defects in the Bbs1 ^M390R/M390R motile cilia and modest enlargement of the lateral ventricles (Figure 2A-D). We observed vesicle-like inclusions and electron-dense material along the ciliary shaft distal to the transition zone in approximately 20% of ependymal cilia lining the lateral and 3^rd ventricles. These structural abnormalities were also seen in mutant mice up to 2 years of age, but not in their wild-type littermates. Small IFT-like particles [34] were seen in the transition zone and proximal axoneme of wild-type cilia (Figure 2E) as well as in some BBS mutant cilia (Figure 2H). In mutant cilia, electron-dense material larger than ITF-like particles was frequently observed just distal to the transition zone between the microtubules and the ciliary membrane where IFT particles were typically found (Figure 2F), while in other cilia this material was located more distally along the ciliary shaft (Figure 2G). In some extreme cases, the region just beyond the transition zone appeared to be bloated and filled with electron-dense vesicular material (Figure 2 H-J). Seen in cross-section, the presence of electron-dense material coincides with a disrupted axoneme in some cilia (Figure 2K) while in others the axoneme appeared to be intact (Figure 2G). We observed similar structural abnormalities in the ependymal cilia lining the lateral and third ventricles of 2 to 7 month-old Bbs2 ^−/− , Bbs4 ^−/− , and Bbs6 ^−/− mice (Figure 6A-H).

By 2 years of age, the extent of ventriculomegaly had progressed in BBS mutant mice (Figure 7A-D). At this age, wild- type ependymal cells looked normal and had numerous motile cilia (Figure 7E, G) while the ependymal layer of Bbs4 ^−/− mutant mice was significantly thinner in some regions (Figure 7H). Other areas of ependyma appeared normal with intact zonae adherentes between cells, yet the underlying neuropil showed evidence of edema (Figure 7H, I). The BBS mutant ependyma had very few microvilli or cilia on its surface and some cilia appeared structurally abnormal (Figure 7F). Comparable changes were also observed in 2 year-old Bbs1 ^M390R/M390R mice (data not shown).

No gross malformations in the SCO of Bbs1 ^M390R/M390R mice compared to their wild-type littermates were observed (Figure 8). The apical zonae adherentes between cells appeared to be intact. Single cilia were identified on the apical protrusions of the goblet-shaped ependymal cells that project into the lumen of the third ventricle and are in contact with CSF. Previous studies have shown that these cilia have 9 + 2 axonemes [35, 36]. In light of the apparent intact SCO ultrastructure and the patency of the aqueduct of Sylvius and fourth ventricle in BBS mutant mice, it is unlikely that abnormalities in the SCO contribute to ventriculomegaly yet we cannot rule out the possibility of absent or defective Reissner’s fibers.

We did not observe any gross morphological changes in the cellular ultrastructure of the SFO in wild-type, Bbs1 ^M390R/M390R or Bbs6 ^−/− mutant mice and the apical zonae adherentes between cells appeared to be intact (Figure 9A, B). In the coronal plane, the ventral-most region of the ependymal surface of the SFO that faces the lumen of the third ventricle was sparsely ciliated in both wild-type and mutant mice, in agreement with previous studies [37, 38]. The transition zone at the base of the SFO cilia appeared to be intact in BBS mutant mice. However, axonemal disruption and the presence of electron-dense vesicle-like material along the ciliary shaft was observed in a sub-population of motile SFO cilia in BBS mutant mice similar to that seen in the cilia of the choroid plexus and ventricular ependyma of these mice (Figure 9C-H).

The most common causes of communicating hydrocephalus in humans are increased CSF synthesis due to choroid plexus papillomas and impaired CSF resorption by the arachnoid granulations [4]. To evaluate possible choroidal anomalies, we analyzed the ultrastructure of the choroid plexuses of the lateral and third ventricles in BBS mutant mice using TEM and found no evidence of choroidal papillomas. The normal appearance of the choroid plexus epithelia with intact junctions between cells argues against the possibility of passive diffusion of molecules from the choroid plexus vasculature into the CSF that could have led to breakdown of the blood:CSF barrier and a loss of CSF homeostasis. Still, we cannot rule out the possibility that the mice produce an increased volume of CSF as a compensatory response to hydrocephalus ex vacuo due to atrophy of underlying brain tissue or incomplete brain development. Future studies of the temporal appearance of ventriculomegaly by MRI analysis in conjunction with cell proliferation and cell death assays will address this issue. Another avenue of future research would be to examine the presence and localization of the aquaporin channels in the choroid plexus that control water transfer [[43]; and references therein].

Dysfunction of the SCO, a specialized zone of ependyma located on the roof of the third ventricle at the entrance to the aqueduct of Sylvius, has been shown to result in hydrocephalus in a number of animal models [44]. The ependymal cells of the SCO secrete glycoproteins that aggregate to form the threadlike Reissner’s fibers that maintain patency of the aqueduct and central canal of the spinal cord [5, 44]. In the absence of a correctly functioning SCO, patency of the aqueduct is compromised, leading to aqueductal stenosis and non-communicating hydrocephalus [45]. Given the apparent intact SCO ultrastructure and patency of the aqueduct of Sylvius and fourth ventricle in BBS mutant mice as old as 2 years of age, it appears unlikely that abnormalities in the SCO contribute to ventriculomegaly. Further, aqueductal stenosis does not appear to be a secondary effect of ventriculomegaly in the BBS mutant mice. Still, the lack of evidence to support the presence of Reissner’s fibers in the mutant mice remains to be addressed.

The SFO, which protrudes into the midline anterior wall of the third ventricle at the junction of the intraventricular foramina of Munro, is anatomically well positioned for its role in osmosensation and the regulation of body and CNS water balance. Although there is currently no known role for the primary and motile cilia that line the SFO [37, 38], we noted structural defects in some of these cilia in BBS mutant mice. They exhibited the axonemal abnormalities and electron-dense vesicle-like material as seen in primary cilia of the choroid plexus and motile ventricular ependymal cilia.

BBS has been described as a degenerative disease of the cilium in which certain proteins accumulate in the cilia over time, leading to progressive ciliary dysfunction [46]. Signaling proteins have been shown to accumulate in BBS mutant Chlamydomonas reinhartii flagella [46] and in zebrafish the absence of BBS proteins leads to delayed retrograde transport [47, 48]. The BBSome functions in the trafficking of membrane proteins between the plasma and ciliary membranes. Several G-protein coupled receptors including MCHR, SSTR3 and dopamine receptor 1 accumulate abnormally within neuronal cilia of BBS mutant mice [49, 50] and ciliary trafficking of the hedgehog signal transducer Smoothened is controlled by the BBSome [18, 22, 51]. Recently, the BBSome has been shown to control IFT assembly and IFT turnaround at the ciliary tip [52].

It is possible that the electron-dense vesicle-like material found in BBS mutant mouse choroid plexus, SFO and ependymal cilia are the result of defective cilia maintenance. The dimensions of the electron-dense, vesicle-like material are too large to have entered the cilia though the transition zone [53]. A more likely scenario is that the material accumulates in the cilia due to defective anterograde and/or retrograde IFT caused by damage to the axoneme that has been observed in some of the defective cilia or that turnaround at the ciliary tip is impaired in the mutants [52].

The presence, at birth, of electron dense vesicle-like inclusions in some of the BBS mutant mouse choroid plexus primary cilia may result from defective ciliary maintenance could lead to impaired signaling between the choroid plexus cilia and epithelium and cause an ionic imbalance in the CSF resulting in its overproduction. The choroid plexuses develop during the early stages of mammalian embryogenesis and are fully formed, ciliated and functional at birth [54]. Interestingly, studies of the Tg^737orpk mouse and in vitro primary cultures of choroid plexus epithelia have underscored the importance of the choroid plexus and its primary cilia in the regulation of CSF production. It has been proposed that protein mislocalization caused by defective IFT in the Tg^737orpk mouse choroid plexus primary cilia impairs their ability to signal to the underlying epithelium via a cAMP-regulated mechanism in order to modulate CSF production [7, 8]. Furthermore, clusters of primary cilia present on the apical surface of porcine choroid plexus primary epithelial cell cultures have been shown to act as negative regulators of fluid transcytosis by decreasing intracellular cAMP levels. These cilia also express neuropeptide FF (NPFF) receptor 2 thought to play a role in chemosensory function and regulation of fluid transport [9].

There has been a resurgence of interest in the study of higher vertebrates for the role of motile cilia in chemosensation or mechanosensation, functions thought to be ascribed solely to primary, immotile cilia [55‐58] following the initial studies in Paramecium and Chlamydomonas reviewed in [55]. Taste receptors have been localized in the ciliary membranes of mouse tracheal epithelial cells [59], progesterone receptors are present in the motile cilia of the mouse oviduct [60], and the Px2 receptor has been localized in rabbit tracheal motile cilia [61]. The appearance of structurally defective ependymal cilia in BBS mutant mice as early as P9, which is shortly after the replacement of primary cilia by motile cilia in mice [32, 33], may impair a sensory mechanism that is relayed to the choroid plexus to regulate CSF synthesis, resulting in a loss of CSF homeostasis and ventriculomegaly.

In conclusion, abnormalities in BBS mutant mouse cilia structure and function have the potential to influence ciliary intraflagellar transport (IFT), ciliary beat frequency, cilia maintenance, protein trafficking, and regulation of CSF production. Ciliary structural defects are the only consistent pathological features associated with CSF-related structures in BBS mutant mice. These defects are observed from an early age, and may contribute to the underlying pathophysiology of ventriculomegaly. Additional research is necessary to establish a causal relationship between the ciliary abnormalities and the development of ventriculomegaly in BBS mutant mice. Continued study of the these mice will add to the growing body of knowledge of the roles played by cilia of the choroid plexus, SFO, and ventricular ependyma in CSF homeostasis as well as in understanding the underlying pathophysiologies of BBS.