Abstract
Congenital hydrocephalus results from cerebrospinal fluid accumulation in the ventricles of the brain and causes severe neurological damage, but the underlying causes are not well understood. It is associated with several syndromes, including primary ciliary dyskinesia (PCD), which is caused by dysfunction of motile cilia. We previously demonstrated that mouse models of PCD lacking ciliary proteins CFAP221, CFAP54 and SPEF2 all have hydrocephalus with a strain-dependent severity. While morphological defects are more severe on the C57BL/6J (B6) background than 129S6/SvEvTac (129), cerebrospinal fluid flow is perturbed on both backgrounds, suggesting that abnormal cilia-driven flow is not the only factor underlying the hydrocephalus phenotype. Here, we performed a microarray analysis on brains from wild type and nm1054 mice lacking CFAP221 on the B6 and 129 backgrounds. Expression differences were observed for a number of genes that cluster into distinct groups based on expression pattern and biological function, many of them implicated in cellular and biochemical processes essential for proper brain development. These include genes known to be functionally relevant to congenital hydrocephalus, as well as formation and function of both motile and sensory cilia. Identification of these genes provides important clues to mechanisms underlying congenital hydrocephalus severity.
Similar content being viewed by others
Introduction
Hydrocephalus is a complex disorder with both genetic and environmental causes1. It results from accumulation of cerebrospinal fluid (CSF) in the ventricles of the brain that typically leads to ventricular enlargement, damage to the underlying ependyma and white matter and thinning of the cerebral cortex2,3. CSF is produced by the choroid plexus and flows from the lateral ventricles through the third ventricle, the aqueduct of Sylvius and the fourth ventricle before entering the subarachnoid space and finally being absorbed into the venous system4,5. Whether the hydrocephalus is genetic or acquired, CSF can accumulate due to an obstruction in the ventricular system, such as a tumor or aqueductal stenosis, or a functional defect in CSF production, flow, or absorption2,3,6,7.
Little is currently known about the genetic basis of congenital hydrocephalus. To date, only three genes have been directly linked to non-syndromic hydrocephalus in human patients: cell adhesion molecule L1CAM8, Wnt pathway inhibitor CCDC88C9,10 and tight junction protein MPDZ11,12. Genetic mouse models have confirmed the role of L1CAM13,14 and CCDC88C15 in development of hydrocephalus and identified several additional genes that have yet to be linked to human hydrocephalus16. Hydrocephalus is also associated with a variety of genetic syndromes, including Dandy-Walker syndrome, Walker-Warburg syndrome, Noonan syndrome, Joubert syndrome and primary ciliary dyskinesia1,7,16.
Primary ciliary dyskinesia (PCD) results from defects in the function of motile cilia and flagella17,18,19. Affected individuals typically suffer from chronic rhinosinusitis, chronic otitis media and male infertility, with situs inversus, hydrocephalus and female infertility present in some cases17,18,19. Motile cilia on the ependymal cells that line the ventricular surface of the brain play an important role in facilitating the flow of CSF through the ventricular system and perturbations in flow can result in hydrocephalus16,17. Although hydrocephalus is only sporadically associated with PCD in patients, it is far more common in mouse models16. We have demonstrated that congenic mouse models of PCD lacking ciliary proteins CFAP221, CFAP54 and SPEF2 all have a severe hydrocephalus on the C57BL6/J (B6) background but not on 129S6/SvEvTac (129) or a mixed (B6x129)F1 background20,21,22, indicating strain specificity in susceptibility to severe PCD-associated hydrocephalus. While there are defects in cilia-driven fluid flow on both backgrounds, ventricular dilatation and secondary damage to surrounding brain tissue are consistently more severe in mutants on the B6 background23. These findings are consistent with reports of severe hydrocephalus in other PCD models on the B6 background and an absence of severe hydrocephalus in models on the 129 background16. Several non-PCD models, including mice with mutations in the genes encoding L1CAM24, the fyn tyrosine kinase25,26, alpha-N-acetylglucosaminidase27 and nuclear receptor NR2E128, also have a more severe hydrocephalus when backcrossed to the B6 strain compared to other genetic backgrounds. In addition, strain-specific differences in brain anatomy and physiology have been observed that could account for variability in disease pathogenesis29. These studies suggest that genetic modifiers of hydrocephalus segregate in certain strains and influence susceptibility to severe hydrocephalus.
To begin to understand the genes and pathways underlying hydrocephalus susceptibility, we performed a DNA microarray analysis to identify strain-specific differences in gene expression in whole brains from wild type mice and nm1054 mice lacking ciliary protein CFAP221 (previously known as PCDP1), as the phenotypic differences have been well characterized and are particularly dramatic in this line20,23. Similar approaches have been effective at identifying strain-specific candidate genes and mechanisms influencing a variety of murine phenotypes, including susceptibility to infection30, craniofacial defects31, hypertension32, eye pigmentation defects33, alcohol sensitivity34, cigarette smoke sensitivity35 and social behavior36. We analyzed gene expression levels in brains from wild type and nm1054 mice on the B6 and 129 backgrounds and identified strain-specific expression levels for a number of genes that cluster into distinct groups based on expression profile and biological function. These genes are implicated in a variety of cellular and biochemical processes essential for proper brain development, including several with known relevance to hydrocephalus and ciliogenesis, providing the first insight into pathways that may underlie susceptibility to severe congenital hydrocephalus.
Results
Microarray analysis uncovers genes with strain-specific expression
Loss of Cfap221 in nm1054 mutant mice results in PCD characterized by hydrocephalus, male infertility and airway abnormalities due to ciliary dysfunction20,23,37,38. Although no strain-specific differences have been observed for the reproductive or respiratory phenotypes, the hydrocephalus is more severe on the B6 background than 129 despite defects in ependymal cilia-driven fluid flow on both backgrounds23. To identify gene expression differences that might influence susceptibility to severe hydrocephalus in nm1054 mice on the B6 and 129 backgrounds, we utilized a DNA microarray approach. As ependymal ciliogenesis occurs during the first week of life39,40, the microarray analysis was performed at P1 to ensure that gene expression data are not influenced by the substantial tissue damage that can occur as a result of CSF accumulation on the B6 background when nm1054 mice age20,23. Gross hydrocephalus was not observed in any mice at P1.
RNA was isolated from whole brain, as the specific mechanisms underlying severe hydrocephalus remain unknown. DNA microarray analysis was performed on WT B6, WT 129, nm1054 B6 and nm1054 129 samples to identify differences in gene expression between WT and nm1054 mice, as well as between the B6 and 129 strains. There were 42,855 entities evaluated after QC analysis. Most genes are expressed at similar levels in WT and nm1054 brains. Line expression plots comparing WT and nm1054 samples on the B6 (Fig. 1a) and 129 (Fig. 1b) backgrounds show most transcripts expressed at nearly the same level in WT and nm1054 samples. While some transcripts show variation between individual samples, only the transcript encoding the acyl CoA-binding protein (ACBP), also known as diazepam binding inhibitor (DBI), is consistently higher in the WT samples than the nm1054 samples in both the B6 and 129 comparisons (Fig. 1a,b). The Acbp gene lies on mouse chromosome 1 within the nm1054 deletion interval20,37, indicating that the microarray approach is effective at identifying expected differences in gene expression. Five additional genes encoding CFAP221, SCTR, STEAP3, TMEM37 and a novel protein represented by GenBank accession number NM_028439 are removed by the nm1054 deletion. Cfap221, Sctr, Steap3 and Tmem37 expression was inconsistently present, marginal, or absent in the individual samples within sample groups, thereby preventing detection in the WT vs nm1054 comparisons and NM_028439 failed to pass QC analysis. To validate that the microarray was accurately detecting expected differences in gene expression, we confirmed that three of the genes deleted by the nm1054 mutation (Cfap221, Acbp and Sctr) are not expressed in the nm1054 brain relative to WT by quantitative RT PCR with RNA from equivalent P1 whole brain lysates (Supplementary Fig. S1).
In contrast to the comparison of WT and nm1054 gene expression, the comparison of WT brains on the B6 and 129 backgrounds revealed that a large number of transcripts are expressed at substantially different levels between the two strains (Fig. 1c). A similar pattern is observed in the comparison of nm1054 brains on the B6 and 129 backgrounds (Fig. 1d). In contrast to the differential expression of Acbp between WT and nm1054 brains, the gene does not show a substantial difference between the B6 and 129 strains.
Statistical analysis of the expression data using ANOVA identified 2,805 transcripts with a corrected p value less than 0.05. To eliminate transcripts that are least likely to have a biologically significant difference in strain-specific expression, the data were filtered by volcano plot analysis to apply a 1.5-fold threshold and all differences in expression that were below 1.5-fold were removed from the pools for both the WT (Fig. 2a) and nm1054 (Fig. 2c) comparison. The line expression plots in Fig. 2 show the comparison of the combined B6 and 129 datasets after filtering out the transcripts below the 1.5-fold threshold. The only remaining transcripts for the WT (Fig. 2b) and nm1054 (Fig. 2d) comparisons are those with consistent, statistically significant expression differences above the applied threshold.
A principal component analysis (PCA) plot shows the individual samples falling consistently into four distinct sample groups: WT B6, nm1054 B6, WT 129 and nm1054 129 (Supplementary Fig. S2). A heat map showing transcript expression profiles for each individual sample after filtering on the volcano plot indicates that, while there are individual transcript differences, similar expression trends are observed within each sample group (Fig. 2e). The heat map in Fig. 2f shows the combined transcript expression profiles for the WT B6, nm1054 B6, WT 129 and nm1054 129 groups. The combined profiles show similar expression trends between WT and nm1054 regardless of genetic background. However, substantial variability in transcript expression is evident in the comparison of B6 and 129 samples, regardless of WT or nm1054 comparison.
Following statistical analysis, a total of 340 transcripts showed greater than 1.5-fold expression differences between WT B6 and WT 129 brains (Supplementary Table S1) and 539 transcripts showed expression differences between nm1054 B6 and nm1054 129 brains (Supplementary Table S2). Of those differentially expressed transcripts, 311 were identified in both the WT and the nm1054 comparison (Fig. 3a), indicating that these expression differences are inherent to the strains and not influenced by the nm1054 mutation. Only Acbp was identified as differentially expressed between WT and nm1054 brains in the B6 and 129 comparisons (Supplementary Tables S3,S4). In addition to the differentially expressed genes, there were a number of genes that were absent in all samples from one strain and present or marginal in all samples from the other, as well as genes present in all samples from one strain and absent or marginal in all samples from the other. A total of 63 transcripts were absent in all B6 samples but present or marginal in the 129 samples (Fig. 3b, Supplementary Table S5) and 69 transcripts were absent in all 129 samples but present or marginal in the B6 samples (Fig. 3c, Supplementary Table S6). These transcripts would not appear in the analysis of differentially expressed genes and indicate that a subset of genes are actively expressed in one strain but not in the other. Similarly, 177 transcripts were present in all B6 samples but absent or marginal in the 129 samples (Fig. 3d, Supplementary Table S7) and 98 transcripts were present in all 129 samples but absent or marginal in the B6 samples (Fig. 3e, Supplementary Table S8). Although the large number of transcripts that are either differentially expressed between the two strains, present in only one strain, or absent in only one strain is likely due to the complexity of the tissue sample, the pool of transcripts is likely to include candidate genes for hydrocephalus susceptibility.
Differentially expressed transcripts fall into distinct clusters and functional groups
The transcripts were clustered using the GeneSpring software into ten groups defined by distinct expression trends (Fig. 4a). Three clusters (2, 4 and 5) show distinct trends of higher expression in B6 brains compared to 129, while two clusters (3 and 9) show distinct trends of higher expression in 129. The remaining groups show only minor expression differences between the two strains. Plotting meta-profiles for each of the clusters into a U-matrix revealed clusters that were similar or dissimilar in expression trend dynamics (Fig. 4b). In this analysis, highly similar profiles are separated by a white hexagon or node indicative of a Euclidean distance close to or equal to 1.0. Highly dissimilar profiles are separated by nodes colored in gray or black, which indicate Euclidean distance metrics that are between 0 and 1 or equal to 0, respectively.
The genes identified as differentially expressed between the B6 and 129 strains encode proteins with a wide variety of functions. Gene network analysis using the Ingenuity Pathway Analysis software showed the most prevalent molecular and cellular functions to be cell morphology, carbohydrate metabolism, small molecule biochemistry and cell-to-cell signaling and interaction in the WT comparison (Table 1). The top functions in the nm1054 comparison were lipid metabolism, nucleic acid metabolism, small molecule biochemistry, cell morphology and carbohydrate metabolism. The analysis also demonstrated a functional enrichment in a diverse spectrum of molecular networks for both the WT and the nm1054 comparison (Table 2, Fig. 5). Network diagrams with gene nodes in a circular layout reveal a higher edge density in the WT comparison (Fig. 5a) than the nm1054 comparison (Fig. 5b), which indicates a greater number of node hubs. These networks suggest that there is greater connectivity between differentially expressed genes in the WT comparison. Neighborhood connectivity, which measures connection between hubs and is an indicator of network resiliency, remains fairly constant relative to the number of neighbors for both comparisons (Fig. 6a,b). The slight negative slope for the nm1054 neighborhood connectivity indicates that the network has a slightly disassortative nature where highly connected nodes tend to connect to nodes with a lower degree, thus reflecting typical biological network architecture41. Analysis of betweenness centrality, which represents the control a node exerts on other nodes, reveals a positive association as the number of neighbors increases (Fig. 6c,d). Both comparisons show several nodes with high value, indicating a substantial effect on other nodes in the network. Analyses of closeness centrality indicate nodes that control the rate at which information spreads throughout the network and also show a positive association with number of neighbors in both the WT and nm1054 comparisons (Fig. 6e,f), further corroborating the hierarchical network architecture of the dataset.
Discussion
In this study, we have demonstrated that a substantial number of transcripts are differentially expressed in brains from B6 and 129 mice at P1. Homozygous deletion mutants lacking CFAP221 (nm1054) have hydrocephalus associated with ciliary dysfunction that is more severe on the B6 background than 129 despite significant differences in ependymal cilia-driven flow on both backgrounds, indicating that genetic modifiers may influence the hydrocephalus phenotype20,23. DNA microarray analysis was performed using the whole brain, as the specific mechanisms underlying hydrocephalus severity remain unknown. Bioinformatics and functional enrichment analyses demonstrate that these genes cluster into distinct groups based on their expression profiles and biological function, many of which are implicated in critical cellular and biochemical processes essential for proper brain development. Because it is possible that the most biologically relevant genes may only be expressed in a small subset of cells, some of those genes could be masked by other genes that are expressed more ubiquitously or at generally higher levels. However, several of the genes identified in this study have a known relevance to hydrocephalus or ciliary function and may either represent those modifiers or be regulated by them. Identification of these genes therefore begins to provide clues to pathways underlying susceptibility to severe hydrocephalus.
Several genes were identified with direct relevance to hydrocephalus, either having at least a 1.5-fold expression difference between the B6 and 129 strains with a p value less than 0.05 or being expressed in only one of the two strains. Two of the known human hydrocephalus genes, CCDC88C and MPDZ, were identified in the microarray. Mutations in the Wnt pathway inhibitor gene CCDC88C (2.57-fold higher expression in B6, but only differentially expressed in the nm1054 comparison) have been shown to result in non-syndromic congenital hydrocephalus in human patients9,10 and a mouse model15. Mutations in the tight junction gene MPDZ (absent in the 129 brain but present or marginal in B6 brains) also result in non-syndromic congenital hydrocephalus in patients11,12. In addition, loss of RND3 (3.21-fold higher in B6, but only differentially expressed in the nm1054 comparison), which regulates cytoskeletal organization and cell adhesion, results in congenital hydrocephalus in mice due to altered Notch signaling42. CDH2 (present in B6 brains but absent or marginal in 129) encodes N-cadherin, which forms cell-cell adherens junctions in the brain43. Blocking N-cadherin junctions ex vivo results in apoptosis of ciliated ependymal cells and damage to the ventricular wall44.
Both motile and primary cilia play a critical role in brain development and physiology16,45,46 and defects in both have been shown to result in hydrocephalus. Centrioles serve as the basis for formation of centrosomes and the basal bodies from which cilia and flagella extend47,48,49 and several genes involved in centriolar and centrosomal function were identified in the microarray analysis. Knockdown of centrosomal gene CEP162 (higher expression in B6 brains, with a 2.20-fold difference in the WT comparison and 2.25-fold in the nm1054 comparison) prevents ciliary transition zone assembly and primary ciliogenesis in cultured cells and results in hydrocephalus in zebrafish50. Knockdown of centrosomal gene SNX10 (2.66-fold higher expression in 129 brains, but only differentially expressed in the WT comparison) also impairs ciliogenesis in both zebrafish and cultured cells51. PCM1 (2.29-fold higher in B6 brains, but only differentially expressed in the nm1054 comparison) encodes a centriolar satellite protein that interacts with several key regulators of centrosomal function and ciliogenesis to promote primary cilia formation and neuronal differentiation and migration52,53,54,55,56,57. Heterozygosity for a targeted allele of PCM1 in mice results in reduced brain volume and behavioral abnormalities58. FGFR1OP, also known as FOP (higher in 129 brains, with a 2.19-fold difference in the WT comparison and 2.08-fold in the nm1054 comparison), also encodes a centriolar satellite protein and knockdown in RPE-1 cells prevents formation of primary cilia59. PLK1 (higher in B6 brains, with a 3.58-fold difference in the WT comparison and 3.27-fold in the nm1054 comparison) encodes a kinase recruited by PCM1 that plays a role in centrosome maturation and primary cilia disassembly57,60,61,62,63,64. MDM1 (present in B6 brains but absent in or marginal in 129) functions as a negative regulator of centriole duplication and is upregulated during ciliogenesis65,66. A nonsense mutation in MDM1 in a mouse model results in retinal degeneration67, a common hallmark of primary ciliopathies. Additionally, NUBP2 (higher in 129 brains, with a 2.17-fold difference in the WT comparison and 2.36-fold in the nm1054 comparison), a nucleotide-binding protein that localizes to the centriole and the basal bodies of primary and motile cilia, functions as a negative regulator of ciliogenesis68.
Several additional genes involved in cilia assembly were identified. ELMO1 (present in B6 brains but absent or marginal in 129) regulates basal body migration and docking at the cell surface and knockdown results in ciliary phenotypes in Xenopus and zebrafish69. EHD3 (higher in B6 brains, with a 2.71-fold difference in the WT comparison and 2.34-fold in the nm1054 comparison) plays a role in promoting ciliary vesicle formation and primary ciliogenesis in cultured cells and zebrafish70. Intraflagellar transport protein IFT74 (absent in 129 but present or marginal in B6 in the nm1054 comparison only) forms a tubulin-binding module with IFT81 required for mammalian ciliogenesis71 and loss in Chlamydomonas reinhardtii perturbs flagellar assembly72. MKKS, also known as BBS6 (higher in B6 brains, with a 3.96-fold difference in the WT comparison and 3.86-fold in the nm1054 comparison), encodes a component of a protein complex that mediates BBSome complex assembly during ciliogenesis73. Mice lacking MKKS possess cilia but have a phenotype resembling the primary ciliopathy Bardet Biedl syndrome74. Biochemical and genetic interactions have been reported with centrosomal protein NPHP6, also known as CEP290, which plays a critical role in promoting primary ciliogenesis and has been implicated in multiple ciliopathies75,76,77. NPHP6 was also found to regulate ATF4 (higher in B6 brains, with a 6.36-fold difference in the WT comparison and 6.33-fold in the nm1054 comparison), a transcription factor involved in multiple cellular stress pathways77,78. NPHP3 (present in 129 but absent or marginal in B6) localizes to primary cilia and human mutations in the Nphp3 gene result in the primary ciliopathy nephronopthisis in humans and mice79. ARL3 (higher in B6 brains, with a 7.10-fold difference in the WT comparison and 8.22-fold in the nm1054 comparison) encodes a GTPase involved in trafficking of proteins to the primary cilium during ciliogenesis80,81,82. WDR92 (2.09-fold higher in B6 brains in the WT comparison only) encodes a cytoplasmic chaperone involved in ciliary assembly and knockdown in planaria perturbs ciliary motility83. Mutations in MYO7A (higher in B6 brains, with a 6.37-fold difference in the WT comparison and 5.85-fold in the nm1054 comparison), which encodes myosin VIIA, result in the primary ciliopathy Usher syndrome84 and loss of MYO7A in the shaker1 mouse model results in abnormal organization of hair cell stereocilia85,86. Mutations in SPAG1 (2.04-fold higher in 129 brains in the WT comparison, but higher in B6 brains in the nm1054 comparison with only a 1.62-fold difference), which is required for assembly of axonemal dynein arms for motile cilia, result in PCD in human patients87. Finally, the extracellular matrix protein SPARC (higher in B6 brains, with a 5.74 fold difference in the WT comparison and 5.97-fold in the nm1054 comparison) interacts with ciliary microtubules in Xenopus embryos88,89 and it has been suggested to play a role in CSF physiology90.
While similar transcript expression trends were observed within each sample group, there are some transcript expression differences between individual samples within each group (Fig. 2e). There are several factors that could contribute to these differences: 1) individual genetic variation, 2) subtle differences in the in utero environment and 3) technical variation due to the sensitivity level of the microarray approach. Despite these differences, the overall similarity of the expression profiles within each group is in stark contrast to the substantially distinct patterns between the B6 and 129 strains (Fig. 2f).
Further studies are required for validation of genes identified in this study and their direct involvement in influencing susceptibility to congenital hydrocephalus. Analysis of more specific brain regions and cell types, as well as additional time points, will aid in the functional refinement of these gene lists. Mapping and identification of modifier polymorphisms will provide additional evidence that, when combined with the gene expression data, will uncover the molecular mechanisms influencing this phenotype. It remains unclear whether the genetic modifiers specifically influence hydrocephalus associated with ependymal cilia dysfunction or whether they more broadly influence congenital hydrocephalus. While strain-specific trends in hydrocephalus severity have been observed for a number of PCD models16, several non-PCD models have exhibited a more severe hydrocephalus on the B6 background than other strains24,26,27,28. Future studies investigating candidate gene expression and sequence in other PCD and non-PCD models on relevant genetic backgrounds will uncover the full spectrum of genetic influence on congenital hydrocephalus. It is possible that specific genetic modifiers segregating in inbred mouse strains also influence susceptibility to severe hydrocephalus in the human population, providing hope that these mouse models will serve as powerful tools to uncover disease mechanisms and provide clues to aid in diagnosis and pharmacological treatment of congenital hydrocephalus.
Methods
Mice
All experiments involving animals were performed in accordance with the Animal Welfare Act and National Institutes of Health (NIH) policies and were approved by the Sanford Research Institutional Animal Care and Use Committee. All methods were carried out in accordance with applicable international, national and institutional guidelines for the care and use of animals. The nm1054 line was maintained on the B6 and 129 backgrounds as previously described20,23,38. The spontaneous and heritable nm1054 mutation is an approximately 400 kb deletion that removes six genes on mouse chromosome 120,37. The PCD phenotype, including hydrocephalus, results exclusively from loss of Cfap221 as previously demonstrated by transgenic rescue20,37. Mice were used for all analyses at postnatal day one (P1). Because mutants also have a severe and lethal anemia on the B6 background due to loss of the Steap3 gene, we analyzed transgenic B6 mutants expressing the RPCI-22 bacterial artificial chromosome 11D19 containing Steap337.
Microarray analysis
Brains were removed from WT and nm1054 homozygous mice on the B6 and 129 backgrounds at P1 (n = 3 in each group), snap frozen in liquid nitrogen and stored at −80 °C until RNA extraction. Each experimental group consisted of a pool of male and female mice, as no sex-specific differences have been observed in the nm1054 hydrocephalus phenotype20,23. Total RNA was extracted from the whole brain with TRIzol and purified using the Maxwell 16 LEV simplyRNA Tissue Kit (Promega, Madison, WI) according to the manufacturer’s instructions. Labeled cRNA was prepared from 500 ng of total RNA using the Illumina RNA Amplification Kit (Ambion, Austin, TX). A total of 1,500 ng of labeled cRNA was hybridized overnight at 58 °C onto the MouseWG-6 Expression BeadChip (Illumina, San Diego, CA) according to the manufacturer’s instructions. These chips contain 45,281 transcripts representing 20,880 unique Entrez genes. Following hybridization, the chips were washed and developed with fluorolink streptavidin-Cy3 (GE Healthcare, Little Chalfont, UK) and scanned with an Illumina BeadArray Reader.
Gene expression data analysis
Intensity values for each probe cell in the hybridized arrays were calculated by GenomeStudio software (Illumina Inc., San Diego, CA) and flags were assigned to each probe set declaring a Present, Marginal, or Absent call (Detection Call Algorithm). Probe cell intensities were used to calculate an average intensity for each set of probe pairs representing a gene, which directly correlated with the amount of cRNA. Further gene expression data analysis and normalization were performed using the GeneSpring GX bioinformatics software suite (Agilent Technologies, Palo Alto, CA). Quality control (QC) filtering was performed on the normalized intensity values, initially excluding the probe sets with an absent call in one hundred percent of the arrays. Alternate analyses were also performed using absent vs. present/marginal call flags or present vs. absent/marginal call flags to analyze genes with absent expression in one group compared to another. After applying QC filtering to diminish background noise created by low-intensity gene probes, genes were clustered into four conditions (WT 129, nm1054 129, WT B6 and nm1054 B6). Data were organized into a hierarchically clustered heat map to elucidate gene expression profiles for each condition and visualized as volcano plots to identify genes significantly up- or down-regulated in each group. Functional enrichment analysis was performed on filtered gene lists of differentially expressed genes using the Ingenuity Pathway Analysis (IPA) software (Qiagen, Venlo, Netherlands).
Quantitative RT PCR
Quantitative reverse transcription polymerase chain reaction (qRT PCR) was performed using the TaqMan approach. Total RNA was extracted from 8 B6 WT, 8 B6 nm1054, 8 129 WT and 6 129 nm1054 brains, all of which were distinct from those used for microarray analysis, as described above. RNA integrity was evaluated using a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA) and cDNA was synthesized from 1 μg RNA using the GoScript Reverse Transcription System (Promega). Quantitative PCR (qPCR) was performed in a Stratagene Mx3000P qPCR system (Agilent Technologies) as previously described22. Commercially available assays (Primetime Assays, IDT, Coralville, IA) were used for Acbp, Cfap221 and Sctr and each was normalized to β-actin. Relative gene expression data were analyzed by the delta-delta Ct method91.
Statistical analysis
Statistical analysis of the microarray gene expression was performed using unpaired t-test and a multiple testing correction formula, the Benjamini Hochberg false discovery rate, which together reported a corrected (Corr) p value for each gene. The hierarchical clustering for groups and entities was performed using Euclidean distance metric and Ward’s linkage algorithm. Statistical significance was set at fold change >1.5 and p (Corr) < 0.05. The qRT PCR data were analyzed by student’s t test using the GraphPad Prism software (GraphPad Software, La Jolla, CA).
Data Availability
The microarray datasets generated and analyzed in this study are available in the NCBI Gene Expression Omnibus (GEO) database under accession number GSE113233.
References
Kahle, K. T., Kulkarni, A. V., Limbrick, D. D. Jr & Warf, B. C. Hydrocephalus in children. Lancet 387, 788–799 (2016).
Del Bigio, M. R. Cellular damage and prevention in childhood hydrocephalus. Brain Pathol 14, 317–324 (2004).
McAllister, J. P. II. Pathophysiology of congenital and neonatal hydrocephalus. Semin Fetal Neonatal Med 17, 285–294 (2012).
Perez-Figares, J. M., Jimenez, A. J. & Rodriguez, E. M. Subcommissural organ, cerebrospinal fluid circulation and hydrocephalus. Microsc Res Tech 52, 591–607 (2001).
Damkier, H. H., Brown, P. D. & Praetorius, J. Cerebrospinal fluid secretion by the choroid plexus. Physiol Rev 93, 1847–1892 (2013).
Rekate, H. L. A consensus on the classification of hydrocephalus: its utility in the assessment of abnormalities of cerebrospinal fluid dynamics. Childs Nerv Syst 27, 1535–1541 (2011).
Tully, H. M. & Dobyns, W. B. Infantile hydrocephalus: a review of epidemiology, classification and causes. Eur J Med Genet 57, 359–368 (2014).
Jouet, M., Rosenthal, A., MacFarlane, J., Kenwrick, S. & Donnai, D. A missense mutation confirms the L1 defect in X-linked hydrocephalus (HSAS). Nat Genet 4, 331 (1993).
Drielsma, A. et al. Two novel CCDC88C mutations confirm the role of DAPLE in autosomal recessive congenital hydrocephalus. J Med Genet 49, 708–712 (2012).
Ekici, A. B. et al. Disturbed Wnt Signalling due to a Mutation in CCDC88C Causes an Autosomal Recessive Non-Syndromic Hydrocephalus with Medial Diverticulum. Mol Syndromol 1, 99–112 (2010).
Al-Dosari, M. S. et al. Mutation in MPDZ causes severe congenital hydrocephalus. J Med Genet 50, 54–58 (2013).
Al-Jezawi, N. K. et al. Compound heterozygous variants in the multiple PDZ domain protein (MPDZ) cause a case of mild non-progressive communicating hydrocephalus. BMC Med Genet 19, 34 (2018).
Fransen, E. et al. L1 knockout mice show dilated ventricles, vermis hypoplasia and impaired exploration patterns. Hum Mol Genet 7, 999–1009 (1998).
Dahme, M. et al. Disruption of the mouse L1 gene leads to malformations of the nervous system. Nat Genet 17, 346–349 (1997).
Takagishi, M. et al. Daple Coordinates Planar Polarized Microtubule Dynamics in Ependymal Cells and Contributes to Hydrocephalus. Cell Rep 20, 960–972 (2017).
Lee, L. Riding the wave of ependymal cilia: Genetic susceptibility to hydrocephalus in primary ciliary dyskinesia. J Neurosci Res 91, 1117–1132 (2013).
Lee, L. Mechanisms of mammalian ciliary motility: Insights from primary ciliary dyskinesia genetics. Gene 473, 57–66 (2011).
Knowles, M. R., Daniels, L. A., Davis, S. D., Zariwala, M. A. & Leigh, M. W. Primary ciliary dyskinesia. Recent advances in diagnostics, genetics and characterization of clinical disease. Am J Respir Crit Care Med 188, 913–922 (2013).
Horani, A., Ferkol, T. W., Dutcher, S. K. & Brody, S. L. Genetics and biology of primary ciliary dyskinesia. Paediatr Respir Rev 18, 18–24 (2016).
Lee, L. et al. Primary ciliary dyskinesia in mice lacking the novel ciliary protein Pcdp1. Mol Cell Biol 28, 949–957 (2008).
Sironen, A. et al. Loss of SPEF2 Function in Mice Results in Spermatogenesis Defects and Primary Ciliary Dyskinesia. Biol Reprod 85, 690–701 (2011).
McKenzie, C. W. et al. CFAP54 is required for proper ciliary motility and assembly of the central pair apparatus in mice. Mol Biol Cell 26, 3140–3149 (2015).
Finn, R., Evans, C. C. & Lee, L. Strain-dependent brain defects in mouse models of primary ciliary dyskinesia with mutations in Pcdp1 and Spef2. Neuroscience 277, 552–567 (2014).
Itoh, K. et al. Brain development in mice lacking L1-L1 homophilic adhesion. J Cell Biol 165, 145–154 (2004).
Grant, S. G. et al. Impaired long-term potentiation, spatial learning and hippocampal development in fyn mutant mice. Science 258, 1903–1910 (1992).
Goto, J., Tezuka, T., Nakazawa, T., Sagara, H. & Yamamoto, T. Loss of Fyn tyrosine kinase on the C57BL/6 genetic background causes hydrocephalus with defects in oligodendrocyte development. Mol Cell Neurosci 38, 203–212 (2008).
Gografe, S. I. et al. Mouse model of Sanfilippo syndrome type B: relation of phenotypic features to background strain. Comp Med 53, 622–632 (2003).
Young, K. A. et al. Fierce: a new mouse deletion of Nr2e1; violent behaviour and ocular abnormalities are background-dependent. Behav Brain Res 132, 145–158 (2002).
Hino, K., Otsuka, S., Ichii, O., Hashimoto, Y. & Kon, Y. Strain differences of cerebral ventricles in mice: can the MRL/MpJ mouse be a model for hydrocephalus? Jpn J Vet Res 57, 3–11 (2009).
Rong, J. et al. Identification of candidate susceptibility and resistance genes of mice infected with Streptococcus suis type 2. PLoS One 7, e32150 (2012).
Mukhopadhyay, P., Brock, G., Webb, C., Pisano, M. M. & Greene, R. M. Strain-specific modifier genes governing craniofacial phenotypes. Birth Defects Res A Clin Mol Teratol 94, 162–175 (2012).
Chiu, C. L. et al. Identification of genes with altered expression in male and female Schlager hypertensive mice. BMC Med Genet 15, 101 (2014).
Trantow, C. M., Cuffy, T. L., Fingert, J. H., Kuehn, M. H. & Anderson, M. G. Microarray analysis of iris gene expression in mice with mutations influencing pigmentation. Invest Ophthalmol Vis Sci 52, 237–248 (2011).
Downing, C. et al. Gene expression changes in C57BL/6J and DBA/2J mice following prenatal alcohol exposure. Alcohol Clin Exp Res 36, 1519–1529 (2012).
Cavarra, E. et al. Early response of gene clusters is associated with mouse lung resistance or sensitivity to cigarette smoke. Am J Physiol Lung Cell Mol Physiol 296, L418–429 (2009).
Ma, L., Piirainen, S., Kulesskaya, N., Rauvala, H. & Tian, L. Association of brain immune genes with social behavior of inbred mouse strains. J Neuroinflammation 12, 75 (2015).
Ohgami, R. S. et al. Identification of a ferrireductase required for efficient transferrin-dependent iron uptake in erythroid cells. Nat Genet 37, 1264–1269 (2005).
McKenzie, C. W. et al. Enhanced response to pulmonary Streptococcus pneumoniae infection is associated with primary ciliary dyskinesia in mice lacking Pcdp1 and Spef2. Cilia 2, 18 (2013).
Banizs, B. et al. Dysfunctional cilia lead to altered ependyma and choroid plexus function and result in the formation of hydrocephalus. Development 132, 5329–5339 (2005).
Spassky, N. et al. Adult ependymal cells are postmitotic and are derived from radial glial cells during embryogenesis. J Neurosci 25, 10–18 (2005).
Faustino, R. S. et al. Systems biology surveillance decrypts pathological transcriptome remodeling. BMC Syst Biol 9, 36 (2015).
Lin, X. et al. Genetic deletion of Rnd3 results in aqueductal stenosis leading to hydrocephalus through up-regulation of Notch signaling. Proc Natl Acad Sci USA 110, 8236–8241 (2013).
Yap, A. S., Brieher, W. M. & Gumbiner, B. M. Molecular and functional analysis of cadherin-based adherens junctions. Annu Rev Cell Dev Biol 13, 119–146 (1997).
Oliver, C. et al. Disruption of CDH2/N-cadherin-based adherens junctions leads to apoptosis of ependymal cells and denudation of brain ventricular walls. J Neuropathol Exp Neurol 72, 846–860 (2013).
Berbari, N. F., O’Connor, A. K., Haycraft, C. J. & Yoder, B. K. The primary cilium as a complex signaling center. Curr Biol 19, R526–535 (2009).
Guemez-Gamboa, A., Coufal, N. G. & Gleeson, J. G. Primary Cilia in the Developing and Mature Brain. Neuron 82, 511–521 (2014).
Tang, T. K. Centriole biogenesis in multiciliated cells. Nat Cell Biol 15, 1400–1402 (2013).
Vladar, E. K. & Stearns, T. Molecular characterization of centriole assembly in ciliated epithelial cells. J Cell Biol 178, 31–42 (2007).
Yan, X., Zhao, H. & Zhu, X. Production of Basal Bodies in bulk for dense multicilia formation. F1000Res 5, https://doi.org/10.12688/f1000research.8469.1 (2016).
Wang, W. J. et al. CEP162 is an axoneme-recognition protein promoting ciliary transition zone assembly at the cilia base. Nat Cell Biol 15, 591–601 (2013).
Chen, Y. et al. A SNX10/V-ATPase pathway regulates ciliogenesis in vitro and in vivo. Cell Res 22, 333–345 (2012).
Gaillard, A. R., Diener, D. R., Rosenbaum, J. L. & Sale, W. S. Flagellar radial spoke protein 3 is an A-kinase anchoring protein (AKAP). J Cell Biol 153, 443–448 (2001).
Ge, X., Frank, C. L., Calderon de Anda, F. & Tsai, L. H. Hook3 interacts with PCM1 to regulate pericentriolar material assembly and the timing of neurogenesis. Neuron 65, 191–203 (2010).
Hori, A. & Toda, T. Regulation of centriolar satellite integrity and its physiology. Cell Mol Life Sci 74, 213–229 (2017).
Insolera, R., Shao, W., Airik, R., Hildebrandt, F. & Shi, S. H. SDCCAG8 regulates pericentriolar material recruitment and neuronal migration in the developing cortex. Neuron 83, 805–822 (2014).
Kim, J., Krishnaswami, S. R. & Gleeson, J. G. CEP290 interacts with the centriolar satellite component PCM-1 and is required for Rab8 localization to the primary cilium. Hum Mol Genet 17, 3796–3805 (2008).
Wang, G. et al. PCM1 recruits Plk1 to the pericentriolar matrix to promote primary cilia disassembly before mitotic entry. J Cell Sci 126, 1355–1365 (2013).
Zoubovsky, S. et al. Neuroanatomical and behavioral deficits in mice haploinsufficient for Pericentriolar material 1 (Pcm1). Neurosci Res 98, 45–49 (2015).
Lee, J. Y. & Stearns, T. FOP is a centriolar satellite protein involved in ciliogenesis. PLoS One 8, e58589 (2013).
Kong, D. et al. Centriole maturation requires regulated Plk1 activity during two consecutive cell cycles. J Cell Biol 206, 855–865 (2014).
Lee, K. H. et al. Identification of a novel Wnt5a-CK1varepsilon-Dvl2-Plk1-mediated primary cilia disassembly pathway. EMBO J 31, 3104–3117 (2012).
Liang, Y., Meng, D., Zhu, B. & Pan, J. Mechanism of ciliary disassembly. Cell Mol Life Sci 73, 1787–1802 (2016).
Seeger-Nukpezah, T. et al. The centrosomal kinase Plk1 localizes to the transition zone of primary cilia and induces phosphorylation of nephrocystin-1. PLoS One 7, e38838 (2012).
Zhang, B. et al. DAZ-interacting Protein 1 (Dzip1) Phosphorylation by Polo-like Kinase 1 (Plk1) Regulates the Centriolar Satellite Localization of the BBSome Protein during the Cell Cycle. J Biol Chem 292, 1351–1360 (2017).
Hoh, R. A., Stowe, T. R., Turk, E. & Stearns, T. Transcriptional program of ciliated epithelial cells reveals new cilium and centrosome components and links to human disease. PLoS One 7, e52166 (2012).
Van de Mark, D., Kong, D., Loncarek, J. & Stearns, T. MDM1 is a microtubule-binding protein that negatively regulates centriole duplication. Mol Biol Cell 26, 3788–3802 (2015).
Chang, B. et al. Age-related retinal degeneration (arrd2) in a novel mouse model due to a nonsense mutation in the Mdm1 gene. Hum Mol Genet 17, 3929–3941 (2008).
Kypri, E. et al. The nucleotide-binding proteins Nubp1 and Nubp2 are negative regulators of ciliogenesis. Cell Mol Life Sci 71, 517–538 (2014).
Epting, D. et al. The Rac1 regulator ELMO controls basal body migration and docking in multiciliated cells through interaction with Ezrin. Development 142, 174–184 (2015).
Lu, Q. et al. Early steps in primary cilium assembly require EHD1/EHD3-dependent ciliary vesicle formation. Nat Cell Biol 17, 228–240 (2015).
Bhogaraju, S. et al. Molecular basis of tubulin transport within the cilium by IFT74 and IFT81. Science 341, 1009–1012 (2013).
Brown, J. M., Cochran, D. A., Craige, B., Kubo, T. & Witman, G. B. Assembly of IFT trains at the ciliary base depends on IFT74. Curr Biol 25, 1583–1593 (2015).
Seo, S. et al. BBS6, BBS10 and BBS12 form a complex with CCT/TRiC family chaperonins and mediate BBSome assembly. Proc Natl Acad Sci USA 107, 1488–1493 (2010).
Fath, M. A. et al. Mkks-null mice have a phenotype resembling Bardet-Biedl syndrome. Hum Mol Genet 14, 1109–1118 (2005).
Rachel, R. A. et al. Combining Cep290 and Mkks ciliopathy alleles in mice rescues sensory defects and restores ciliogenesis. J Clin Invest 122, 1233–1245 (2012).
Rachel, R. A. et al. CEP290 alleles in mice disrupt tissue-specific cilia biogenesis and recapitulate features of syndromic ciliopathies. Hum Mol Genet 24, 3775–3791 (2015).
Sayer, J. A. et al. The centrosomal protein nephrocystin-6 is mutated in Joubert syndrome and activates transcription factor ATF4. Nat Genet 38, 674–681 (2006).
Rutkowski, D. T. & Kaufman, R. J. All roads lead to ATF4. Dev Cell 4, 442–444 (2003).
Olbrich, H. et al. Mutations in a novel gene, NPHP3, cause adolescent nephronophthisis, tapeto-retinal degeneration and hepatic fibrosis. Nat Genet 34, 455–459 (2003).
Hanke-Gogokhia, C. et al. Arf-like Protein 3 (ARL3) Regulates Protein Trafficking and Ciliogenesis in Mouse Photoreceptors. J Biol Chem 291, 7142–7155 (2016).
Kim, H. et al. Ciliary membrane proteins traffic through the Golgi via a Rabep1/GGA1/Arl3-dependent mechanism. Nat Commun 5, 5482 (2014).
Schwarz, N. et al. Arl3 and RP2 regulate the trafficking of ciliary tip kinesins. Hum Mol Genet 26, 2480–2492 (2017).
Patel-King, R. S. & King, S. M. A prefoldin-associated WD-repeat protein (WDR92) is required for the correct architectural assembly of motile cilia. Mol Biol Cell 27, 1204–1209 (2016).
Mathur, P. & Yang, J. Usher syndrome: Hearing loss, retinal degeneration and associated abnormalities. Biochim Biophys Acta 1852, 406–420 (2015).
Holme, R. H. & Steel, K. P. Stereocilia defects in waltzer (Cdh23), shaker1 (Myo7a) and double waltzer/shaker1 mutant mice. Hear Res 169, 13–23 (2002).
Self, T. et al. Shaker-1 mutations reveal roles for myosin VIIA in both development and function of cochlear hair cells. Development 125, 557–566 (1998).
Knowles, M. R. et al. Mutations in SPAG1 cause primary ciliary dyskinesia associated with defective outer and inner dynein arms. Am J Hum Genet 93, 711–720 (2013).
Huynh, M. H., Hong, H., Delovitch, S., Desser, S. & Ringuette, M. Association of SPARC (osteonectin, BM-40) with extracellular and intracellular components of the ciliated surface ectoderm of Xenopus embryos. Cell Motil Cytoskeleton 47, 154–162 (2000).
Huynh, M. H., Sodek, K., Lee, H. & Ringuette, M. Interaction between SPARC and tubulin in Xenopus. Cell Tissue Res 317, 313–317 (2004).
Liddelow, S. A. et al. SPARC/osteonectin, an endogenous mechanism for targeting albumin to the blood-cerebrospinal fluid interface during brain development. Eur J Neurosci 34, 1062–1073 (2011).
Pfaffl, M. W. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 29, e45 (2001).
Acknowledgements
We gratefully thank Kang Liu and Sam Dooyema for technical support, as well as Tyler Bradley for assistance with manuscript preparation. The microarray hybridizations were performed in the Sanford-Burnham-Prebys Medical Discovery Institute Microarray/Q-PCR Core Facility. This research in this study was funded by the Hydrocephalus Association Innovator Award and Sanford Research. The Sanford Research Molecular Biology Core was supported by NIH COBRE grant P20GM103620.
Author information
Authors and Affiliations
Contributions
C.W.M. and R.F. bred the mouse lines, prepared samples for analysis and analyzed the data. C.P., K.M.E. and R.S.F. performed the bioinformatics analyses. L.L. designed the study, oversaw the research and wrote the manuscript. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing Interests
The authors declare no competing interests.
Additional information
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
McKenzie, C.W., Preston, C.C., Finn, R. et al. Strain-specific differences in brain gene expression in a hydrocephalic mouse model with motile cilia dysfunction. Sci Rep 8, 13370 (2018). https://doi.org/10.1038/s41598-018-31743-5
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-018-31743-5
- Springer Nature Limited
This article is cited by
-
Microglia activated by microbial neuraminidase contributes to ependymal cell death
Fluids and Barriers of the CNS (2021)
-
Motile cilia genetics and cell biology: big results from little mice
Cellular and Molecular Life Sciences (2021)
-
Genetic interaction between central pair apparatus genes CFAP221, CFAP54, and SPEF2 in mouse models of primary ciliary dyskinesia
Scientific Reports (2020)
-
An exome-first approach to aid in the diagnosis of primary ciliary dyskinesia
Human Genetics (2020)
-
The year in review: progress in brain barriers and brain fluid research in 2018
Fluids and Barriers of the CNS (2019)