Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Cilia are specialized cellular structures restricted in location and function by highly regulated epithelial cell differentiation. Mammalian cilia have been characterized as "sensory" or "motor" on the basis of structure and function (1). Sensory and other "primary" type cilia have a cross-sectional microtubule ultrastructure of "9 + 0," describing the configuration of nine outer pairs of microtubules (3). Motile cilia and flagella have a classic "9 + 2" microtubule ultrastructure due to an additional inner pair of microtubules and include dynein motors directly linked to the outer pairs that, together with other proteins, drive movement (1, 2). For both cilia types, mammalian ciliogenesis has been conceptualized in four stages: (1) generation of the centriole (centriologenesis), (2) migration of the centriole to the apical aspect of the cell, (3) docking the centriole to the apical membrane web as a basal body, and (4) axoneme formation (1). Recently, mammalian proteins involved in cilia formation, including kinesins required for axoneme assembly and dyneins required for motor function, have been isolated on the basis of homology to genes initially identified in Chlamydamonas and sea urchin (5, 6). However, factors important for the regulation of each step of ciliogenesis are not well understood.

The winged helix/forkhead transcription factor, hepatocyte nuclear factor-3 (HNF-3)/forkhead homologue-4 (HFH-4, FKHL-13, Winged Helix Consortium name foxj1) was known to be expressed in late embryonic through adult epithelial cells of the choroid plexus, proximal lung epithelial cells, oviduct, and testis (7)all tissues containing ciliated cells. Members of this family of transcription factors have been shown to play fundamental roles in development and tissue-specific differentiation (10, 11). Immunohistochemical localization of HFH-4 expression by us and others has demonstrated that HFH-4 is specifically expressed in ciliated epithelial cells in the late embryo through adult tissues and that HFH-4 expression is temporally related to ciliogenesis in the developing airway (12, 13). We have also found that HFH-4 expression is limited to epithelial cells containing classical motile cilia (9 + 2) and is not present in cells with cilia lacking the central microtubule pair and dynein arms (9 + 0) (12). Although HFH-4 recognizes a DNA consensus sequence in genes expressed in epithelial cells of the bronchi and choroid plexus (8), the function of HFH-4 has not been established. The restricted pattern of HFH-4 expression suggests that one function of this protein is regulation of ciliogenesis.

Using targeted deletion of the Hfh4 gene in mouse embryonic stem (ES) cells, we found that HFH-4 mutant mice had randomized left-right axis determination and absent cilia in the airway, confirming an observation made while our work was in progress (14). The ability of HFH-4 to regulate ciliogenesis in the lung was also consistent with the observation that overexpression of HFH-4 by the surfactant protein C promoter in the lungs of transgenic mice could induce ciliogenesis in undifferentiated alveolar epithelial cells, indicating that HFH-4 could direct the ciliated cell phenotype (15). However, a stage-specific function for HFH-4 in ciliogenesis has not been identified.

Here, we demonstrate that HFH-4 acts during stages of basal body migration and apical membrane docking. Further, we also demonstrate that HFH-4 is expressed in the postimplantation stage at the embryonic node prior to left-right asymmetry development, as might be predicted by the presence of situs inversus in half of the HFH-4 mutant mice. This is consistent with recent findings that have identified a structurally and functionally unique type of motile monocilia in the early embryo in cells located in the embryonic node (16, 17). Despite the lack of cilia in the airway, we found that at the node, cilia are present in the HFH-4 mutant mice, indicating that regulation of ciliogenesis at the node does not parallel mechanisms in the airway. Together, these observations indicate that there are alternative regulatory pathways for ciliogenesis in the embryonic node cell cilium, classic motor 9 + 2 cilia, and in 9 + 0 sensory cilia.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Gene Targeting and Genotyping

The targeting vector contained a 5' 4.5-kb Hfh4 HincII-HincII fragment upstream from the translation start site (Figure 1A) (18). The Hfh4 3' flanking region was a 1.8-kb HincII-EcoRV fragment containing exon I and a portion of intron I. The gene was interrupted with the neomycin resistance gene driven by the phosphoglycerol kinase promoter and fused with the lacZ gene. ES cells (RW4 line, 129/SvJ) were electroporated with the linearized targeting vector and clones were selected in the presence of G418 (19). ES cell genomic DNA was digested with HindIII or DraI and analyzed by Southern hybridization using probes located outside of the targeting construct. Clones that underwent homologous recombination were analyzed with a neomycin resistance gene probe to confirm single-site integration. Correctly targeted clones were injected into C57BL/6J blastocysts and implanted into receptive females. Two chimeric mice transmitted the mutant Hfh4 allele to the germline. Heterozygous animals were bred to produce homozygous deficient mice and were genotyped using Southern analysis. To genotype embryos younger than embryonic day (E) 10, yolk sac membranes were washed in phosphate-buffered saline and incubated for 3 h or overnight at 50°C in 100 µl of proteinase K lysis buffer (proteinase K, 2 mg/ml; KCl, 50 mM; Tris, pH 8.3, 10 mM; MgCl₂, 2 mM; gelatin, 0.1 mg/ml; Nonidet P-40, 0.45%; and Tween-20, 0.45%). The lysate was then heated (94°C, 10 min) to denature proteinase K. An aliquot (1 or 2 µl) of the lysate was used for polymerase chain reaction (PCR) genotyping with Taq polymerase (Red Taq; Sigma, St. Louis, MO), and primers HGT1 (5'-TTCAAGGGCAGATG-GAGAG- AGG) and HGT2 (5'-TGGCATAGTCCAGTCAGG) to amplify a 661-base pair (bp) wild-type allele-specific sequence; or HGT1 and HGT-lacZ1 (5'-CTCTTCGCTATTACGCCAGC-TGG) to amplify a 416-bp mutant allele sequence (94°C for 60 s, 55°C for 30 s, and 72°C for 60 s; 35 cycles). For timed pregnancies, the presence of a morning vaginal plug was considered E0.5 at noon. Embryo age was confirmed using established criteria (20).

View larger version (52K): [in this window] [in a new window]   Figure 1.   Targeted disruption of Hfh4. (A) Schematic map of the mouse Hfh4 gene, targeting vector (containing the lacZ-PGKNeo cassette), and mutant locus. Solid boxes indicate exons, with the DNA-binding domain indicated by stippling. Indicated are the location of external 5' and 3' probes, the expected fragment sizes for Southern analysis, and restriction enzyme sites (H, HindIII; H2, HincII; E, EcoRI; Rv, EcoRV; D, DraI). (B) Southern analysis of DraI digested genomic DNA from littermates produced by mating HFH-4+/ animals, hybridized with the 3' probe generating a 9-kb wild-type and a 13-kb disrupted Hfh4 allele. (C) PCR, allele-specific amplification of genomic DNA isolated from E7.5 yolk sacs generating a 661-bp wild-type (wt) or 416-bp mutant knockout (ko) DNA product. (D) Northern analysis of RNA isolated from HFH-4+/+ or HFH-4/ mouse tissues hybridized with Hfh4 and glyceraldehyde-3-phosphate dehydrogenase cDNA probes. (E) Western analysis of protein from HFH-4+/+ or HFH-4/ mouse tissues incubated with HFH-4 antibody. The control (con) lysate is from cells transfected with an HFH-4 expression plasmid (lane 7).

RNA Expression

Total cellular RNA was purified from tissues by guanidinium lysis and anion-exchange resin separation (Qiagen, Santa Clarita, CA) for small samples or by cesium chloride gradient separation. RNA blots were analyzed using a complementary DNA (cDNA) probe containing sequences for the rat HFH-4 amino acids 1 to 117 (7).

Whole-Mount In Situ Hybridization

To detect Hfh4 messenger RNA (mRNA) in whole embryos, a probe containing nucleotide sequences for the HFH-4 amino acids 167 to 420 was used. cRNA sense and control antisense probes were labeled with digoxigenin and hybridized with embryos according to the method described by Wilkinson and Nieto (21). For each experiment three to seven embryos of each embryonic stage were used, and analysis of each stage was repeated two to eight times. Hfh4 expression was evaluated by the presence of blue color approximately 60 min after incubation in alkaline phosphatase substrates (nitroblue tetrazolium; 5-bromo-4-chloro-3- indoylphosphatate) for color development. Embryos were postfixed and photographed, then genotyped. For cross-sectional analysis after color development, embryos were embedded in tissue-freezing medium (Triangle Biomedical Sciences, Durham, NC), then cut by cryostat in 20-µM sections.

Immunodetection

For immunodetection of HFH-4, an affinity-purified rabbit antirat HFH-4 polyclonal antibody was used (12). Western analysis was performed using supernatants of tissues (50 µg protein/sample), separated by polyacrylamide gel (7.5%) electrophoresis, transferred to nitrocelluose, and incubated (4°C, overnight) in blocking solution (Tris-buffered saline: nonfat milk, 5%; and Tween-20, 0.2%), then incubated (room temperature, 1 h) in anti-HFH-4 antibody (1:500 dilution). Lysate from cells transfected with an HFH-4 cDNA expression plasmid was used as a positive control in these assays.

Electron Microscopy and Magnetic Resonance Imaging

Tissues for transmission and scanning electron microscopy (EM) were fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate and postfixed in 1.25% osmium tetroxide. For transmission EM, samples were further fixed in 4% uranyl acetate, thin-sectioned (90 nm) in Polybed 812 (Polysciences, Warrington, PA), poststained in uranyl acetate and lead citrate, then visualized on a Zeiss 902 microscope (Zeiss, Thornwood, NY). For scanning EM, embryos were critical point-dried from liquid carbon dioxide, gold sputter-coated, then visualized on a Hitachi S-450 (Hitachi, Tokyo, Japan) instrument. Proton magnetic resonance images were acquired using standard spin echo or fast spin echo techniques for brain and body scans, respectively. Imaging parameters used were (for brain/body, respectively): repetition time 1.5/2.0 s, echo time 80/21 ms, field-of-view 1.5 × 1.5/3.0 × 2.0 cm, and data matrix 128 × 128/512 × 256 pixels.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Gross Phenotype of HFH-4 Mutant Mice

To understand the role of HFH-4, the Hfh4 gene was interrupted in exon I by homologous recombination in ES cells and transmitted to the germline of mice (Figures 1A, 1B, and 1C). Homozygous mutant mice did not express Hfh4 mRNA or protein, indicating that targeted deletion resulted in a null mutation (Figures 1D and 1E). Genotype analysis of litters from heterozygous crossings at E7.5 to E19.5 revealed that 29 of 114 embryos (25.4%) were homozygous mutant, indicating normal intrauterine survival. At birth, genotype analysis of 107 mice showed 20 to be homozygous mutant. Surprisingly, approximately half (32 of 74, 43%) of the mutant pups (E11.5 to adult) had situs inversus, as evidenced by reversal of the position of intra-abdominal viscera and a right-sided heart (dextrocardia) (Figures 2A and 2B). Some mutant mice had heterotaxia characterized by a normally positioned left-sided stomach and spleen (and liver on the right) but with dextrocardia (n = 3) or a right-sided viscera with a left-sided heart (n = 1). In other mice, a central heart with right- or left-sided viscera was present (n = 3). Mutant mice also had abnormalities of the great vessels of the heart (e.g., transposition) alone or concurrent with abnormalities of situs (n = 4). Most mutant animals died at birth to postnatal day (P)3, possibly due to cardiovascular abnormalities associated with defects of left-right formation as described for other mutant mice with randomized left-right axis (22). Thus, offspring of heterozygous crossings at P14 had only 40 of 535 homozygous mutant mice (8%; 25% is predicted).

View larger version (114K): [in this window] [in a new window]   Figure 2.   Gross morphology of HFH-4 mutant mice (/) and wild-type mice (+/+). (A) Newborn littermates from heterozygous breeding show total situs inversus in the mutant mouse. Arrows indicate milk-filled stomachs. Visceral dissection demonstrates the location of stomach (st) and spleen (sp). (B) Magnetic resonance images show situs inversus of abdominal viscera and heart (*) in HFH-4/ mice. (C) Brains from HFH-4 mutant and wild-type adult mice littermates. (D) Magnetic resonance images of mutant and wild-type brains demonstrate ventricular dilation in the mutant mouse.

Homozygous mutant animals that survived to P7 developed a characteristic hydrocephalic head shape at P7 to P14, were runted, and died at P12 to P40. The brains of mice that survived to this age had distended hemispheres with markedly dilated lateral ventricles and a thin anterior cortex (Figure 2C). Brains evaluated microscopically or by magnetic resonance imaging (Figures 2C and 2D) confirmed the presence of lateral ventricle dilation and demonstrated an absence of ventricular obstruction and minimal third- and fourth-ventricle dilation, suggesting a localized anatomic or physiologic defect. An association between nonobstructive hydrocephalus and situs inversus is consistent with reports in humans and other species (23).

Absent 9 + 2 Ciliogenesis in HFH-4 Mutant Mice

In humans, defects in left-right asymmetry occur as part of Kartagener's syndrome, characterized by situs inversus, respiratory infections, and infertility, and is also associated with cardiac malformations and hydrocephalus (26, 27). Individuals with this syndrome have abnormalities in cilia ranging from immotility with absent dynein arms to total absence of cilia (27). Nearly identical syndromes have been observed in dogs and rats with cilia defects, but without identification of a specific genetic defect (24, 25). Therefore, we evaluated the cilia in HFH-4 mutant mice using light microscopy and EM. Microscopic examination of the heterogeneous population of airway epithelial cells (Figure 3A) of the lung from HFH-4 mutant mice (ages E18 through P35, n = 10) revealed a specific absence of cilia in airway cells (Figure 3B). A similar defect was present in epithelial cells of the brain and oviduct (data not shown) of HFH-4 mutant mice. We have previously found that HFH-4 expression in nose and paranasal sinuses is restricted to ciliated respiratory epithelial cells and is not present in the 9 + 0 ciliated olfactory neuroepithelial cells (12). Consistent with this observation, the respiratory epithelial cells of the nose and paranasal sinuses of mutant mice lacked cilia (Figure 3D). However, cilia were present in the olfactory neuroepithelial cells in the roof of the nasal sinuses (Figure 3E). This indicates a selective defect in ciliogenesis in HFH-4 mutant mice affecting a subpopulation of ciliated epithelial cells restricted to those cilia that have a 9 + 2 microtubule configuration.

View larger version (118K): [in this window] [in a new window]   Figure 3.   Selective defect in ciliogenesis in epithelial cells of HFH-4 mutant mice. Light micrographs of sections of respiratory epithelium from adult wild-type (A and C) and HFH-4 mutant mice (B, D, and E). Tracheas (A and B) from wild-type mice (A) have a highly differentiated population with epithelial cells, including those with cilia (arrows) that are absent in the mutant mice (B). Respiratory epithelial cells lining the nasal septum contain a uniform population of ciliated (arrows) cells (C) that lack cilia in mutant mice (D). Cilia are present on olfactory neuroepithelial cells in the superior region of the nasal cavity of mutant mice (E). Toludine blue (A and B) or hematoxylin and eosin (C and E) counterstain. Bar in E represents 19 (A and B), 24 (C and D), and 9 µm (E).

We also sought to determine whether the defect in airway epithelial cell ciliogenesis affected all cells with a similar uniformity and whether the density of cilia might be affected in heterozygous animals. Scanning electron micrographs of the tracheas showed no morphologic differences between wild-type (Figures 4A and 4B) and heterozygous animals (data not shown). The defect in ciliogenesis resulted in almost complete absence of cilia. Careful analysis of multiple regions revealed that the defect in ciliogenesis was not absolute and that rare cilia were present in the airway (arrows, Figure 4D).

View larger version (182K): [in this window] [in a new window]   Figure 4.   Tracheas of HFH-4 mutant mice. Scanning electron micrographs of tracheas from wild-type (A and B) or HFH-4 mutant (C and D) mice. Wild-type micrographs demonstrate that approximately half of the epithelial cells are ciliated, containing microvilli and 100 to 200 cilia/cell. Mutant epithelium is devoid of cilia except for the presence of rare cilia on some cells (arrows). Bar in D represents 7.1 µm(A and B) and 3.6 µm (C and D).

To further evaluate the defect in ciliogenesis, tracheas (n = 4) from adult and newborn mutant mice were examined by transmission EM (Figure 5). These studies confirmed an absence of cilia and revealed that centrioles and basal bodies were present, but that most basal bodies failed to reach the apical membrane of the epithelial cell and develop cilia (Figure 5C). Compared with the epithelial cells from wild-type mice, where basal bodies were neatly aligned beneath the apical membrane (Figure 5A), the cilia precursors in the HFH-4 mutants appeared disorganized within the apical aspect of the cytoplasm (Figure 5C). Rather than the characteristic array of nine peripheral doublets plus two central microtubules (9 + 2) seen in the cilia protruding from cells of wild-type mice, the apical membrane of airway epithelial cells from mutant mice had only microvilli, structures not containing microtubules (Figure 5D). Centriologenesis in HFH-4 mutant mice appeared to be normal (Figure 5C, insert), indicating a downstream defect in the ciliogenesis pathway. These findings suggest that the absence of HFH-4 leads to impaired centriole migration, docking, and subsequent axoneme elongation. Thus, it is possible that HFH-4 regulates expression of a protein that is important for migration or stabilization of basal bodies at the apical membrane. This may include proteins that interact with the centriole directly or indirectly through a complex of proteins that activates binding or binds directly to microtubules.

View larger version (114K): [in this window] [in a new window]   Figure 5.   Defective centriole position in epithelial cells of HFH-4 mutant mice. Transmission electron micrographs of tracheas from adult wild-type (A and B) and HFH-4 mutant mice (C and D) show that cilia (arrow) are absent in airway epithelial cells of the mutant mice visualized by transmission EM. Cilia precursor basal bodies (bb) and centrioles (cen) are indicated. The normal three-fiber centriole structure is present in the mutant mouse (insert in C). Detail cross-section of apical structures from wild-type (B) and HFH-4 mutant mice (D) shows only microvilli in mutant mice. Bar in D represents 360 (A and C), 106 (insert in C), and 60 nm (B and D).

Hfh4 Is Transiently Expressed at the Node

To further evaluate the relationship of ciliogenesis to the defect in left-right axis formation, we evaluated Hfh4 expression in early wild-type embryos. Studies of Xenopus and zebrafish demonstrated that the left-right axis is established at the earliest developmental stages, when the dorsal-ventral axis is also determined (28). In mice, development of organ asymmetry occurs between E7 and E10, suggesting that left-right axis regulatory factors must be present at these early stages of embryogenesis (29). Accordingly, wild-type mouse embryos were evaluated for Hfh4 expression using in situ hybridization. Hfh4 expression was present only at the node in embryos from E7.0 (n = 13 of 13 embryos evaluated) to E7.5 (n = 23 of 28) (Figure 6). Expression was transient during this defined period from the midprimitive streak stage to the early headfold stage and was not detected in E6.5 (n = 17), E8.0 (n = 11), or E8.5 (n = 17) embryos. Asymmetric expression or expression outside of the region of the node was not observed. Expression at the node was limited to the columnar population of ventral epithelial cells of the node that are known to have monocilium (Figure 6E) (32, 33).

View larger version (104K): [in this window] [in a new window]   Figure 6.   Expression of Hfh4 in embryo development. Localization of Hfh4 expression at E7.5 by whole-mount in situ hybridization in embryos shows examples of early (A and C) and later (B and D) node development. Lateral view (A and B) shows Hfh4 expression at the ventral node (anterior neural fold of the head process is en face, left is up). Ventral view shows Hfh4 expression at the midline node structure (midline anterior neural fold is down). Sagittal section (E) of the embryo shown in D demonstrates expression in the columnar cells of the ventral node. Expression was detected only at the node, and no signal was detected with the sense probe.

Cilia at the Node of HFH-4 Mutant Mice

The ciliogenesis defect and randomized left-right asymmetry in the HFH-4 mutant mouse suggested that cilia at the node might be absent. It has been postulated that the cilia located at the node also play a role in determining left- right axis by establishing a directional flow-driven gradient of a yet-unidentified morphogen across the node (16, 17, 32, 34). In contrast to the structure of 9 + 2 cilia, transmission electron micrographs of the cilia at the node, performed by others, revealed a unique ultrastructure composed of a 9 + 0 microtubule configuration but with inner and outer dynein arms and no central doublet apparatus (17). Although cilia in respiratory tract and oviduct epithelial cells were absent in the HFH-4 mutant mice, examination of E7.5 to E7.7 HFH-4 mutant embryos (n = 7 from three different litters) by scanning EM revealed that cilium at the node are present (Figure 7). The shape, size, and location of the HFH-4 mutant embryo nodes were similar to heterozygous or wild-type embryos. The height and density of the cilia in wild-type and mutant embryos are also similar. Thus, it appears that ciliogenesis is differentially affected in the mutant mice such that the regulation of ciliogenesis at the node is through a different pathway than that utilized for classical motile cilia in other tissues.

View larger version (166K): [in this window] [in a new window]   Figure 7.   Cilia at the node in HFH-4 mutant mice. Scanning electron micrographs of the node HFH-4 heterozygous (A, B, and C) and mutant (D, E, and F ) embryos at early headfold stage (E7.7). Embryos are viewed dorsally (A and D). Ciliated cells are shown at higher magnification, demonstrating that both heterozygous (B and C) and mutant (E and F ) embryo node cells contain microvilli and monocilium. Bar in F represents 12 (A and D), 2.5 (B and D), or 1 µM (C and F ).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

HFH-4, a member of the winged helix/forkhead family of transcription factors is critical for the determination of both left-right axis and classical motile ciliogenesis. The specific pattern of expression of HFH-4 in ciliated epithelial cells and in cells of the embryonic node appears to be unique. The pattern of HFH-4 expression in mouse 9 + 2 ciliated cells and at the node parallels the expression of Hfh4 homologue XFKH5 in Xenopus 9 + 2 ciliated epidermal cells of the animal third of the embryo (stage 33) and in Spemann's organizer (stage 11) region (35). The pattern of HFH-4 expression and phenotype of the HFH-4 mutant mouse provides further insight regarding mechanisms of ciliogenesis important for understanding human disorders such as primary ciliary dyskinesis and Kartagener's syndrome (26, 27).

In the HFH-4 mutant mouse, the presence of node cilia and the absence of 9 + 2 cilia indicate that cilia are assembled through different mechanisms. The findings also underscore that there are three types of mammalian ciliated cells: (1) nonmotile, sensory 9 + 0, not expressing HFH-4; (2) motile 9 + 2, multiciliated with a wavelike beating pattern, expressing HFH-4 and requiring HFH-4 for axoneme growth; and (3) motile 9 + 0 monocilium with dynein arms and a rotary beating pattern expressing HFH-4, but not requiring HFH-4 for axoneme growth. This third type of cilia appears unique in mammals, but may be similar to motile sperm flagella from the eel Anguilla that lack the central microtubule pair (and outer dynein arms) but have distinct inner dynein arms to function as motors (36). The HFH-4 mutant mouse suggests that there is a shared defect related to a role in the regulation of motile ciliogenesis in the airway and at the node necessary for left-right axis formation.

A role for HFH-4 at the node in the regulation of the left-right pathway axis formation is supported by several findings. In the mouse node, Hfh4 expression at E7.0 to 7.5 occurs before the onset of expression of left-right pathway effector proteins nodal, lefty1, and Pitx2, suggesting that HFH-4 acts upstream in this molecular cascade (29- 31, 37, 38). Thus, within the left-right pathway, HFH-4 joins HNF-3 as a proximal, node-specific transcription factor, indicating that transcription is an important mechanism for regulation of the left-right pathway (38, 39).

It is possible that through a regulatory defect the cilia of the HFH-4 mutant node, although present, have a functional (motility) defect, leading to a failure of directed left-right determination and therefore randomization of left-right axis. Several observations have implicated the embryonic node, and specifically the cilia, as a potential site for left-right signal regulation. First, stage-specific surgical alteration or excision of the node in the chick can alter left-right downstream signaling (37). Second, absence of the node in the HNF-3 mutant embryos results in alteration of asymmetric patterns of downstream left-right gene expression and failure to establish normal left-right asymmetry (39). Third, absence of cilia at the node in the KIF3A and KIF3B mutant mice is also associated with randomized left-right asymmetry (16, 17). And fourth, inversus viscerum (iv) mice with mutation of axonemal dynein gene left/right dynein (lrd) that is expressed at the mouse node at E7.5 have randomized left-right axis development (40) and cilia that do not beat (34). These observations have recently led Nonaka and colleagues to the finding that node cilia have vectoral beating that generates flow across the node (16). These investigators hypothesize that node cilia flow directs a yet undefined morphogen to establish an asymmetric gradient that triggers a defined downstream cascade to direct anatomic left-right asymmetry (16, 34).

The phenotype of HFH-4 mutant mice is consistent with a node ciliogenesis defect. In contrast to KIF3A and KIF3B mutant embryos that survive only to E10, the HFH-4 mutant mice share several features with iv mice. Animals deficient in HFH-4 have randomized left-right asymmetry characterized by either situs solitus or situs inversus and a small percentage (< 10%) of mice have heterotaxia, similar to iv mice. Also similar to iv mice, some HFH-4 mutant mice likely die from cardiovascular defects, reflecting incomplete situs development (22). The shared randomized left-right asymmetry and cardiovascular defects, and the similar temporal and spatial expression of lrd and Hfh4 suggested that HFH-4 might regulate lrd. However, using reverse transcriptase-PCR of RNA from E7.5 embryos and lungs, we found that lrd gene expression was present in HFH-4 mutant embryos (Yan and Brody, unpublished observation), in contrast to a previous report (14). The presence of lrd in HFH-4 mutant embryos and the knowledge that lrd codes for a dynein suggest that HFH-4 regulates an essential pathway parallel to lrd, possibly an axoneme assembly program in the node utilizing previously created LRD protein. The possibility remains that LRD protein expression might be altered in the HFH-4 mutant mice, but that it is less likely the result of direct transcriptional regulation by HFH-4. Thus, the relationship of lrd to HFH-4 in left-right asymmetry remains undefined. However, the absence of cilia in the airway, brains, and oviduct of HFH-4 mutant mice, and normal 9 + 2 cilia structure and function in trachea and sperm and normal survival in iv mice (41), suggest that HFH-4 has a greater spectrum of function than LRD.

A specific function of HFH-4 at the node and in postembryonic ciliated epithelial cells is unclear, and a unified biomolecular pathway for a role for HFH-4 in ciliogenesis and left-right asymmetry determination has not been defined. In addition to a direct effect on ciliogenesis, there are other possibilities. First, it is possible that in early development HFH-4 regulates the expression of gene(s) important for left-right determination in node cells not directly related to ciliogenesis (e.g., related to morphogen function) and later, other gene(s) important for ciliogenesis in specific organs. This would be analogous to HNF-3, which, in early embryogenesis, regulates node development and later regulates adult liver-specific gene expression (10, 39). Second, HFH-4 mutant mice have a defect that specifically affects the position of the axonemal (motor) centriole, and centriole positioning has been implicated in development of asymmetry in early embryogenesis of Caenorhabditis elegans (42). Therefore, an additional possibility is that in HFH-4 mutants a defect in centriole positioning within the node results in randomized left-right cell polarity assignment. Further characterization of function, structure, and transport of proteins in HFH-4 mutant mice may lead to the identification of molecular targets for HFH-4 transcriptional activation responsible for proper left-right asymmetry and cilia formation.