RESEARCH ARTICLE

Impaired neural differentiation and glymphatic CSF flow in theCcdc39 rat model of neonatal hydrocephalus: genetic interactionwith L1camA. Scott Emmert1,*, Eri Iwasawa1,*, Crystal Shula1, Preston Schultz1, Diana Lindquist2, R. Scott Dunn2,Elizabeth M. Fugate2, Yueh-Chiang Hu3, Francesco T. Mangano1 and June Goto1,‡

ABSTRACTNeonatal hydrocephalus affects about one child per 1000 births andis a major congenital brain abnormality. We previously discovered agene mutation within the coiled-coil domain-containing 39 (Ccdc39)gene, which causes the progressive hydrocephalus (prh) phenotypein mice due to lack of ependymal-cilia-mediated cerebrospinal fluid(CSF) flow. In this study, we used CRISPR/Cas9 to introduce theCcdc39 gene mutation into rats, which are more suitable for imagingand surgical experiments. The Ccdc39prh/prh mutants exhibited mildventriculomegaly at postnatal day (P)5 that progressed into severehydrocephalus by P11 (P<0.001). After P11, macrophage andneutrophil invasion along with subarachnoid hemorrhage wereobserved in mutant brains showing reduced neurofilament density,hypomyelination and increased cell death signals compared withwild-type brains. Significantly more macrophages entered the brainparenchyma at P5 before hemorrhaging was noted and increasedexpression of a pro-inflammatory factor (monocyte chemoattractantprotein-1) was found in the cortical neural and endothelial cells in themutant brains at P11. Glymphatic-mediated CSF circulation wasprogressively impaired along the middle cerebral artery from P11 asmutants developed severe hydrocephalus (P<0.001). In addition,Ccdc39prh/prh mutants with L1 cell adhesion molecule (L1cam) genemutation, which causes X-linked human congenital hydrocephalus,showedan accelerated early hydrocephalus phenotype (P<0.05-0.01).Our findings inCcdc39prh/prhmutant rats demonstrate a possible causalrole of neuroinflammation in neonatal hydrocephalus development,which involves impaired cortical development and glymphatic CSF flow.Improved understanding of inflammatory responses and the glymphaticsystem in neonatal hydrocephalus could lead to new therapeuticstrategies for this condition.

This article has an associated First Person interview with the joint firstauthors of the paper.

KEY WORDS: Neonatal hydrocephalus, Genetic model,Neocorticogenesis, Glymphatic, CRISPR-Cas, Rat model

INTRODUCTIONNeonatal hydrocephalus is a devastating condition defined by theabnormal accumulation of cerebrospinal fluid (CSF) in the brain thatmay arise from both genetic and acquired causes and can lead tobrain damage and neurocognitive and motor skill problems(Schrander-Stumpel and Fryns, 1998; Tully and Dobyns, 2014;Vinchon et al., 2012). Neonatal hydrocephalus affectsapproximately one child in every 1000 live births (Haverkampet al., 1999; Schrander-Stumpel and Fryns, 1998; Stoll et al., 1992)and contains a genetic etiology in nearly 40% of cases (Haverkampet al., 1999; Schrander-Stumpel and Fryns, 1998; Zhang et al.,2006). Recent advancements in human genetic approaches such aswhole-exome sequencing in familial and consanguineous forms ofcongenital hydrocephalus have identified a limited number of genesrelated to the development of this disease in humans,including L1CAM (Rosenthal et al., 1992), MPDZ (Al-Dosariet al., 2013), CCDC88C (Ekici et al., 2010), EML1 (Shaheen et al.,2017),WDR81 (Shaheen et al., 2017), TRIM71 (Furey et al., 2018a),SMARCC1 (Furey et al., 2018a) and PTCH1 (Furey et al., 2018a).Collectively, mutations in these genes account for approximately15% of all congenital hydrocephalus cases (Furey et al., 2018b).However, the cellular events that are essential for neonatalhydrocephalus development and its interactions with thesurrounding physiological networks of CSF circulation andabsorption are poorly understood. For example, the role of therecently discovered glymphatic system, a brain-wide CSF andsoluble compounds distribution system, in neonatal hydrocephalusis unknown. Defined as the glial-associated lymphatic system(glymphatic system) for its dependence on astrocytic aquaporin-4(AQP4) channels and lymphatic-like role in the brain, theglymphatic system of perivascular spaces, which are lined byastrocytic foot processes with AQP4 channels and endothelialabluminal membranes, drains the brain of interstitial fluid and wasteproducts into extracranial lymphatics (Iliff et al., 2012). Associatedwith the meningeal lymphatic network (Aspelund et al., 2015;Louveau et al., 2015), the glymphatic system is an area of promisingresearch for its role in draining the CSF contents in healthy brains aswell as those affected by neurological disorders such as Alzheimer’sdisease (Iliff et al., 2012), traumatic brain injury (Iliff et al., 2014)and hydrocephalus (Ringstad et al., 2017).

A genetic model of neonatal hydrocephalus involving thecoiled-coil domain-containing 39 (Ccdc39) gene may helpelucidate cellular mechanisms leading to the abnormalaccumulation of cerebrospinal fluid in neonatal hydrocephalus,as this model shows a robust and 100% penetrant hydrocephalusphenotype in mice, which is not common in other rodent models.Ccdc39 is selectively expressed in embryonic choroid plexus andependymal cells on the medial wall of the ventricular forebrainReceived 19 June 2019; Accepted 14 October 2019

1Division of Pediatric Neurosurgery, Cincinnati Children’s Hospital Medical Center,Cincinnati, OH 45229, USA. 2Division of Radiology, Cincinnati Children’s HospitalMedical Center, Cincinnati, OH 45229, USA. 3Developmental Biology, CincinnatiChildren’s Hospital Medical Center, Cincinnati, OH 45229, USA.*These authors contributed equally to this work

‡Author for correspondence (june.goto@cchmc.org)

R.S.D., 0000-0001-8501-8290; J.G., 0000-0002-4000-9750

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

1

© 2019. Published by The Company of Biologists Ltd | Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

(Abdelhamed et al., 2018), and the protein is localized to theaxoneme of motile cilia (Merveille et al., 2011). In mice, Ccdc39gene mutation leads to ependymal cells with shorter cilia, withmicrotubules lacking the axonemal inner arm dynein, resulting inimpaired ependymal cilia beating and intraventricular CSF flow(Abdelhamed et al., 2018). Although substantial strides have beenmade in characterizing the mechanisms of ciliary dysfunction(Lee, 2013) and CSF flow abnormalities (Date et al., 2019; Olstadet al., 2019) caused by ciliary gene mutations, thepathophysiologic downstream processes whereby impaired CSFflow leads to hydrocephalus are still unsolved. The small size ofmurine models inhibits the use of surgical procedures conducive tostudying these processes at early developmental time points;however, such procedures could be performed on the brains oflarger mammalian models of neonatal hydrocephalus generatedusing the CRISPR/Cas9 genome editing system.The emergence of CRISPR/Cas9 technology provides an

accessible method for generating transgenic rat models of congenitalhydrocephalus (Emmert et al., 2019) that were unfeasible withprevious genetic techniques (Mashimo, 2014; Sander and Joung,2014). Furthermore, CRISPR/Cas9 offers the opportunity to testgenetic modifiers and possible genetic interactions that determinedisease severity in congenital hydrocephalus. For instance, X-linkedhydrocephalus (XLH), which can result from mutations in the L1 celladhesion molecule (L1cam) gene in mice (Dahme et al., 1997;Demyanenko et al., 1999), rats (Emmert et al., 2019) and humans(Adle-Biassette et al., 2013; Dahme et al., 1997; Rosenthal et al.,1992), varies in severity from hydrocephalus with multiple structuralabnormalities and prenatal death to a milder phenotype with cognitiveimpairment or isolated symptoms even within the same family (Frynset al., 1991; Serville et al., 1992). CRISPR/Cas9-generated rodentmodels of congenital hydrocephalus resulting from mutations indifferent hydrocephalus-related genes, such as L1cam and Ccdc39, canbe interbred to investigate epistatic interactions previously believed toaffect mutation penetrance and ventricular size in other models ofhydrocephalus (Weller and Gärtner, 2001; Zhang et al., 2006).Inflammation related to neonatal hydrocephalus has been

investigated primarily in posthemorrhagic hydrocephalus both inanimal models (Gram et al., 2014; Yung et al., 2011) and in patients(Heep et al., 2004; Klebe et al., 2019; Sävman et al., 2002). Thesestudies show an increase in several cytokines in the CSF, whicheventually causes oxidative stress and further damages the braintissue. Although little is known about the inflammatory response incases of neonatal hydrocephalus without preceding hemorrhage,inflammation has been shown to be causal in hydrocephalus models(Abdi et al., 2018; Botfield et al., 2013; Lattke et al., 2012).Periventricular white matter has been reported as a specificallyvulnerable region to the hydrocephalus insult in neonates both inhuman and animal models of hydrocephalus (Del Bigio et al., 2003;Hanlo et al., 1997), possibly because of the direct physical stress fromperiventricular distention or indirectly by hypoxic-ischemic stress(du Plessis, 1998) or other mechanisms, including inflammation.Using CRISPR/Cas9 to model neonatal hydrocephalus in rats,

we generated a Ccdc39 knockout line in Sprague Dawley rats.Here, we studied the genetic interaction of two hydrocephalus-related genes, L1cam and Ccdc39, through genetic, survival andgrowth characterization of the Ccdc39prh/prh mutant rat in thepresence and absence of an L1cam-null allele. We also examinedspatiotemporal inflammatory reactions along with corticaldevelopment in this novel neonatal hydrocephalus model. Toconclude, we investigated the pattern of CSF circulation throughthe glymphatic system for the first time in neonatal hydrocephalus,

using Evans Blue dye injected into the cisterna magna of controland Ccdc39prh/prh mutant rats.

RESULTSCRISPR/Cas9-mediated modeling of the prh mutation(Ccdc39c.916+2T>A) in ratsWe previously identified a Ccdc39prh mutation (Abdelhamed et al.,2018) in the progressive hydrocephalus (prh) mouse mutant withneonatal hydrocephalus (Stottmann et al., 2011). To efficiently andspecifically induce the same mutation in rats, CRISPR guide RNA(gRNA), primers and oligonucleotide donor repair templates weredesigned to introduce the homozygous chr2:g.120305679A>Tchange that creates a splice site (Ccdc39c.916+2T ) mutation in the ratCcdc39 gene (Fig. 1A, Table 1). Guide RNA sequences wereselected based upon favorable on-target and off-target scoresaccording to CRISPR guide design tools Benchling version 1(https://benchling.com/academic) and CRISPOR (http://crispor.tefor.net/). Of the rats born from CRISPR-modified embryos(n=26), two rats exhibited the intended Ccdc39c.916+2T>A mutationupon Sanger sequencing, whereas other edited offspringdemonstrated insertions and deletions (n=10) around the targetedsite. F0 animals with mosaicism were bred to wild-type SpragueDawley rats to generate F1 heterozygous rats (Ccdc39wt/prh), whichshow the Ccdc39c.916+2T>A change with an adenine peak ofapproximately half the intensity of the wild-type thymine peak inthe sequencing chromatogram (Fig. 1B). The F1 Ccdc39wt/prh

heterozygous rats were subsequently bred to generate homozygousCcdc39prh/prh rat mutants (Ccdc39prh/prh). The homozygousmutation was confirmed by Sanger sequencing of F1-F2 pups.Subsequent generations were genotyped with TaqMan probesdetecting the difference between T and A at Ccdc39c.916+2T (seeMaterials and Methods). In western blotting, CCDC39 proteinexpression (approximately 110 kDa; black arrow in Fig. 1C) wassignificantly reduced in Ccdc39wt/prh (n=2) rats and completelyeliminated in Ccdc39prh/prh (n=2) rats relative to wild-typelittermates (n=2) (Fig. 1C), which aligns with our previousfinding that the prh mouse mutant exhibits loss of CCDC39protein as a result of abnormal mRNA splicing (Abdelhamedet al., 2018).

Progressive hydrocephalus in Ccdc39prh/prh mutant ratsCcdc39prh/prh rats exhibited dome-shaped heads (Fig. 1D) anddeveloped progressive postnatal hydrocephalus (Fig. 2A-M) over 2weeks. Male and female Ccdc39prh/prh mutant pups demonstratedgradual growth delays (nwt/wt=8-14, nprh/prh=8; P<0.01-0.001)and early mortality in life (nwt/wt=22, nprh/prh=17; P<0.001)(Fig. 3A-C). In histology, P5-P8, but not P1, Ccdc39prh/prh ratsexhibited dilation of the lateral ventricles, smaller subventricularzone and thinning of the cortical mantle relative to wild-type andheterozygous littermates (Fig. 2A). Three-dimensional (3D)volumetric T2-weighted MR images were acquired through thebrains of wild-type and Ccdc39prh/prh rats. Volumetric analysis ofthe lateral ventricles, third ventricle, fourth ventricle and pinealrecess revealed that Ccdc39prh/prh rats demonstrate mildenlargement of the lateral ventricles beginning at P5 (Fig. 2A,B,F,G) that progressed into severe ventriculomegaly relative to controllittermates by P11 (nwt/wt=4-6, nwt/prh=3-6, nprh/prh=4; P<0.001)(Fig. 2C,L,M). However, no differences were found in the volumes ofthe third ventricle, fourth ventricle or pineal recess in Ccdc39prh/prh

(n=4) and L1cam heterozygous/Ccdc39 mutant (L1camwt/−;Ccdc39prh/prh, n=2) rats relative to control (n=4-6) andheterozygous (n=3-6) littermates at either P5 or P11.

2

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

Interaction between hydrocephalus-causing Ccdc39 andL1cam gene mutations in L1camy/−;Ccdc39prh/prh double-mutant rats worsens the survival, growth andventriculomegaly of Ccdc39prh/prh mutant ratsTo investigate the epistatic interactions between knownhydrocephalus-causing gene mutations, we crossed the Ccdc39prh

allele with the L1cam− allele that we generated recently (Emmertet al., 2019). L1camy/− single mutant rats show delayed growth anddevelop mild enlargement of the fourth and lateral ventricles by 3 and6 weeks postnatally, respectively (Emmert et al., 2019), modeling amild form of XLH. Survival rate and body weight analyses of maleand female postnatal wild-type, L1cam heterozygous/Ccdc39heterozygous (L1camwt/−;Ccdc39wt/prh), L1cam mutant/Ccdc39

heterozygous (L1camy/−;Ccdc39wt/prh), double-mutant (L1camy/−;Ccdc39prh/prh), and L1cam heterozygous/Ccdc39mutant (L1camwt/−;Ccdc39prh/prh) rats revealed that the stunted growth and prematuremortality ofCcdc39prh/prh rats were exaggerated in the presence of theL1cam-null allele, as evidenced by the decreased survival and growthof male L1camy/−;Ccdc39prh/prh rats (n=6 and n=2, respectively;P<0.001 betweenwild-type and L1camy/−;Ccdc39prh/prh andP<0.01between Ccdc39prh/prh and L1camy/−;Ccdc39prh/prh for survival;P<0.05 between Ccdc39prh/prh and L1camy/−;Ccdc39prh/prh for bodyweight) (Fig. 3A,B). Furthermore, L1camy/−;Ccdc39prh/prh doublemutants exhibited mild dilation of the lateral ventricles as early as P1in histology (n=4; Fig. 2A), which was not seen in single mutants.These data suggest that mutations in L1cam and Ccdc39 could

Fig. 1. CRISPR/Cas9-mediated modeling of the prh mutation(Ccdc39c.916+2T>A) in rats. (A) Wild-type and prh (Ccdc39prh) DNA sequenceharboring the critical thymine nucleotide (black arrowhead) required for propersplicing of the Ccdc39 gene in rats (chr2:g.120305679). Capital letters, exon;boxed letters, guide RNA sequence; black underline, protospacer adjacentmotif; blue underline, engineeredHindIII restriction site; red, T>Amutation. (B)Genomic DNA Sanger sequencing traces of wild-type (Ccdc39wt/wt),heterozygous (Ccdc39wt/prh) and mutant (Ccdc39prh/prh) rat offspringgenerated with CRISPR/Cas9 genome editing. Ccdc39prh/prh rats exhibit theintended homozygous (Ccdc39c.916+2T>A) change. (C) Immunoblot analysison P11 rat brain lysate confirms the loss of CCDC39 in the homozygousmutant. CCDC39 protein expression (approximately 110 kDa; black arrow)was reduced to approximately 50% and 0% of the normal expression level inheterozygous and homozygous mutant rats, respectively. β-tubulin was theloading control. (D) Gross picture of wild-type (Ccdc39wt/wt) and mutant(Ccdc39prh/prh) rats show the dome-shaped head of a mutant rat at P25.

Table 1. Sequence and genomic description of gRNA, ssDNA donor oligonucleotide repair templates and sequencing primers used to generate theprh rat model of neonatal hydrocephalus with CRISPR/Cas9

Sequence (5′-3′) Genomic location Strand

gRNA GAGGTGTGCCACACAAGCACTGG Chr2: 120305661-120305683 −ssDNA donor oligo #1 CACGAAGACTTAAGGGAAAGAGCAGCAGGCCTCCCCAGTCAAACAGGCC-

TTTGCCATTGACTGAGGAATCTGTATGACCGCAGGCCAAAGCTT-GTGTGGCACTCCTCATCCTTCAGCTGATTTCTGTTGG

Chr2: 120305577-120305706 +

ssDNA donor oligo #2 CCAACAGAAATCAGCTGAAGGATGAGGAGTGCCACACAAGCTTTGGCCTG-CGGTCATACAGATTCCTCAGTCAATGGCAAAGGCCTGTTTGACTGGGGA-GGCCTGCTGCTCTTTCCCTTAAGTCTTCGTG

Chr2: 120305577-120305706 −

Forward sequencing primer CTCCTCTGCACAGCTTAGGAC Chr2: 120305110-120305130 +Reverse sequencing primer CACTCCCCACAACAGCTTACA Chr2: 120305977 120305997 −

Italics, gRNA. Bold underline, chr2:g.120305679A>T change in the prh sequence. Bold, HindIII restriction site. Underline, PAM. Chr, chromosome; gRNA, guideRNA; oligo, oligonucleotide; PAM, protospacer adjacent motif; prh, progressive hydrocephalus; ssDNA, single-stranded DNA.

3

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

converge phenotypically in hydrocephalus and that mutation in bothgenes accelerates the development of neonatal hydrocephalus(Fig. 2A). However, the heterozygous loss of the Ccdc39 allele didnot worsen the survival (Fig. 3A) or growth phenotype (Fig. 3D,F) ofL1camy/− (n=23 for survival, n=5-6 for weight) in L1camy/−;Ccdc39wt/prh rats (n=18 for survival, n=9 for weight) either in theearly postnatal period or through 3months of age. Also, heterozygousloss of L1camwt/− did not affect the early lethality of Ccdc39prh/prh

single mutants (L1camwt/−;Ccdc39prh/prh n=3; Fig. 3A,). Similarly,L1camwt/−;Ccdc39wt/prh rats (n=11) demonstratedmoderate decreasesin growth compared with wild-type animals (n=8; P<0.05) but didnot differ in weight from L1camwt/− rats (n=5) (Fig. 3E).Additionally, L1camwt/−;Ccdc39wt/prh rats (n=26) did not exhibitthe hydrocephalus (Fig. 2A) or early lethality (Fig. 3A) ofCcdc39prh/prh

or L1camy/−;Ccdc39prh/prh rats.We further examined the genetic interaction between L1cam and

Ccdc39 gene mutations in brain development by performing 3Dvolumetry on T2-weighted magnetic resonance imaging. At P5,L1camwt/−;Ccdc39prh/prh rats (n=2) exhibited mild hydrocephalus

relative to wild-type littermates (n=6, P<0.001) but did not differ inthe volume of their lateral ventricles compared with Ccdc39prh/prh

rats (n=4) (Fig. 2B,D-G). At P90, L1camy/−;Ccdc39wt/prh rats (n=6)exhibited dilation of the lateral ventricles relative to wild types (n=3,P<0.05) and a similar level of ventriculomegaly to that of theL1camy/− rat (n=8) (Fig. 2N-T). Magnetic resonance imaging(MRI) of L1camy/−;Ccdc39prh/prh rats was difficult to acquirebecause of their early lethality.

Inflammation and subarachnoid hemorrhage are seen inCcdc39prh/prh mutant ratsTo describe the molecular events of hydrocephalus in Ccdc39prh/prh

mutant rats, we first characterized the developmental time course ofintracranial hemorrhage in serial hematoxylin and eosin (H&E)staining of the postnatal Ccdc39prh/prh mutant rats. A small amountof bleeding was seen in the subarachnoid area of P8 Ccdc39prh/prh

mutant rats and became massive in both the subarachnoid space andsubpial area, and later into the lateral ventricles by P30 (Fig. 4A). Nobleeding was seen in wild-type rats by P30 (nprh/prh=16, nwt/wt=8).

Fig. 2. Postnatal hydrocephalus in Ccdc39mutant rats and genetic interaction with theL1cam-null allele. (A) H&E staining of P1 andP5-P8 wild-type (Ccdc39wt/wt) (n=5), Ccdc39heterozygous (Ccdc39wt/prh) (n=2), Ccdc39homozygous mutant (Ccdc39prh/prh) (n=6 atP1, n=4 at P5-8), L1cam mutant (L1camy/−)(n=5 at P1, n=2 at P5-8) and L1cam;Ccdc39double-mutant (L1camy/−;Ccdc39prh/prh) rats(n=4 at P1, n=3 at P5-8). Ventricular dilationwas observed in P1 L1camy/−;Ccdc39prh/prh

doublemutants and P5Ccdc39prh/prhmutants,with the most severe dilation observed in P5L1camy/−;Ccdc39prh/prh double mutants. Eacharea of the dashed square overlaying P1 andP5 Ccdc39prh/prh mutants is shown with ahigher magnification image. (B,C) Volumetricanalysis of the lateral ventricles (LV), thirdventricle (3V), fourth ventricle (4V) and pinealrecess (PR) shows that Ccdc39prh/prh (n=4)and L1camwt/−;Ccdc39prh/prh (n=2) mutantsdemonstrate significantly dilated lateralventricles beginning at P5, which developsinto severe hydrocephalus at P11 relative toCcdc39wt/prh (n=3) and wild-type (n=4) rats.***P<0.001; ns, not significant determined bytwo-way ANOVA and Holm–Sidak’s multiplecomparisons test. (D-S) Representative three-dimensional T2-weighted MR images of wild-type (D,E,H,I,N,O), Ccdc39prh/prh (F,G,L,M),Ccdc39wt/prh (J,K), L1camy/− (P,Q) andL1camy/−;Ccdc39wt/prh (R,S) rats at P5 (D-G),P11 (H-M) and P90 (N-S). Red arrows indicatelateral ventricles. Yellow arrows indicate thirdventricle. (T) L1camy/− (n=8) and L1camy/−;Ccdc39wt/prh (n=6) rats demonstrate a similarlyenlarged third ventricle and lateral ventricles atP90. Results are presented as individualreplicates overlaying the mean±s.d. #P<0.05between wild type and L1camy/−;Ccdc39wt/prh

determined as for B,C. *P<0.05 between wildtype and L1camy/− determined as for B,C.Scale bars: 2 mm (A, D-S); 500 μm (magnifiedimages in A).

4

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

Some amount of hemorrhage was located under pia mater. Becauseeither pia mater or subarachnoid membrane seemed to bedissociated from the brain surface along with fluid accumulationin the space between the membrane and brain parenchyma(arrows in Fig. 4A), which was not observed wild type rats, weinvestigated possible cellular processes that might cause pial/arachnoid membrane dissociation from the brain surface andeventual bleeding. Staining for the activated form of matrixmetalloproteinase 9 (MMP9) showed its upregulation near thebrain surface of Ccdc39prh/prh mutant rats at P11, and it wasexpressed in glial fibrillary acidic protein (GFAP)-positiveastrocytes (Fig. 4B). We hypothesized that this abnormallyactivated MMP9 and possibly the eventual bleeding occurredsecondarily to the inflammatory reaction in hydrocephalus. When

assessing inflammatory reactions in Ccdc39prh/prh mutant rats,significantly more CD68-positive activated macrophages firstappeared in the periventricular white matter and striatum at P5compared with wild-type rats (n=3 in each group, P=0.046;Fig. S1A) and were subsequently seen in the subarachnoid area atP11. Myeloperoxidase (MPO)-positive neutrophils filled thesubarachnoid space and partly filled the perivascular space at P30(Fig. 5A). This inflammatory cell invasion into the perivascularspace was not seen prior to P30. Interestingly, there was more signalof a candidate factor of inflammatory cell migration, monocytechemoattractant protein 1 (MCP1), as early as P5 in Ccdc39prh/prh

rats, although not statistically significant (n=3 in each group,P=0.14; Fig. 5B). The upregulation of MCP1 expression wasstatistically significant at P11 (n=3 in each group, P<0.01; Fig. 5B)

Fig. 3. Reduced survival and growth in Ccdc39 mutant rats with the L1cam-null allele. (A) Survival rate of male and female postnatal wild-type (WT; lightblue), Ccdc39 heterozygous (wt/prh; purple), Ccdc39mutant (prh/prh; orange), L1cam heterozygous (wt/−; dark blue), L1cammutant (y/−; lime green), L1camheterozygous;Ccdc39 heterozygous (wt/−;wt/prh; forest green), L1cammutant;Ccdc39 heterozygous (y/−;wt/prh; brown), L1cam heterozygous;Ccdc39mutant(wt/−; prh/prh; pink) and L1cam;Ccdc39 double-mutant (y/−; prh/prh; red) rats reveals that Ccdc39 mutant (n=17) and L1cam mutant;Ccdc39 heterozygous(n=18) rats recapitulate the pre-weaning phasemortality ofCcdc39mutant mice. The survival phenotype inCcdc39mutant rats is exaggerated when this model iscrossed with the L1cam-null allele, as shown by L1cam;Ccdc39 double mutants (n=6) (P=0.0013). (B-F) Body weight analysis of male (B,D,F; circles) and female(C,E; squares) postnatal wild-type, Ccdc39 mutant, double-mutant, Ccdc39 heterozygous, L1cam mutant, L1cam mutant;Ccdc39 heterozygous, L1camheterozygous and double-heterozygous rats younger (B-E) and older (F) than 3 weeks (wks) of age demonstrates that Ccdc39mutants (n=8) exhibit mild growthdelays throughout the first 3 weeks of life, which are exaggerated by the presence of a L1cam-null allele in L1cam;Ccdc39 double mutants (n=2). L1camheterozygous (n=5) and L1cam heterozygous;Ccdc39 heterozygous (n=11) rats exhibit similar reductions in growth over 3 weeks (E). L1cammutant rats (n=5-6)similarly exhibit decreased body weight from birth through 3 months of age, including when harboring a Ccdc39-null allele, as shown by L1cam mutant;Ccdc39heterozygous rats (n=8) (F). Male and female rats were analyzed separately for body weight due to sex differences in feeding behavior in rats (Fukushima et al.,2015). Results are presented as mean±s.d. P-values, which were determined by log-rank test (survival) or nonlinear regression (body weight) between eachgenotype and the wild type (unless otherwise noted) are indicated on the graph and color-coded by genotype. Allele abbreviations and color coding for eachgenotype presented in B-F are the same as for A.

5

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

and was mostly seen in neuronal nuclei (NeuN)-positive neurons,CD68-positive macrophages and endothelial cells labeled with IB4(Fig. 5C; Fig. S1B).

Impaired neural differentiation and increased cell death inCcdc39prh/prh ratsTo assess the development of the cerebral cortices, we evaluatedmyelin basic protein (MBP) staining in cerebral white matter. TheMBP-positive area was significantly reduced in Ccdc39prh/prh

mutant rats compared with Ccdc39wt/wt wild-type rats at P11(Fig. 6A). At P5, myelination had barely started, even inCcdc39wt/wt

wild-type rats (Fig. S2A), and there was no significant difference inthe number of MBP-positive mature oligodendrocytes betweenCcdc39prh/prh mutant and Ccdc39wt/wt wild-type rats (n=3 in eachgroup, P=0.18; Fig. S2A). The impairment in myelin formation atP11 was accompanied by impaired neurofilament expression, asassessed with neurofilament-H (NF-H) subunit staining, whichshowed sparse and shortened neurofilaments in Ccdc39prh/prh

mutant rats at P11 (Fig. 6B). There was a tendency for reducedneurofilament formation in Ccdc39prh/prh mutant rats at P5,although this phenomenon was not statistically significant (n=3in each group, P=0.11; Fig. S2B). We next evaluated cell deathsignals using TUNEL staining in Ccdc39prh/prh mutant rats at

P1-P5. Increased cell death, as measured by TUNEL-positivecells, was detected around the periventricular region inCcdc39prh/prh mutant rat brains compared with wild-type rats(Fig. 6C).

Impaired CSF flow through the glymphatic system in theCcdc39prh/prh mutant rat model of neonatal hydrocephalusWe investigated the CSF circulation path within the glymphaticsystem at different stages of hydrocephalus in the Ccdc39prh/prh

mutant rat model by injecting Evans Blue dye into the cisternamagna of mutants (n=7) ranging in age from P5 to P27 and age-matched controls (n=5). Evans Blue is a low molecular weighttracing dye with a high affinity for serum albumin thataccumulates in perivascular spaces, such as the Virchow–Robinspaces of the glymphatic system, after intracerebral injection(Maloveska et al., 2018; Stoelinga and van Munster, 1967). Ithas been used as a marker for blood–brain barrier integrity,meningeal lymphatic vasculature and glymphatic flux because ofits colorimetric, fluorescent and protein-binding qualities(Maloveska et al., 2018; Stoelinga and van Munster, 1967;Wolf et al., 2019). Likewise, various groups have usedfluorescently labeled protein tracers injected into the cisternamagna to show flux of labeled proteins in the CSF through the

Fig. 4. Bleeding in the Ccdc39prh/prh mutant rat model ofneonatal hydrocephalus. (A) The Ccdc39 homozygousmutant (Ccdc39prh/prh) rat shows subtle subarachnoidbleeding along with arachnoid membrane dissociation frombrain parenchyma (arrows) starting from P9, and the rate ofbleeding reaches 100% by P11. The hemorrhage is found inboth subarachnoid/subpial area (arrows) and lateral ventricles(arrowheads) at P30 (graph of bleeding rate; n=16 of Ccdc39mutants, P1-P30). (B) The activated form of MMP9 staining atP11 showsMMP9 activation near the brain surface containingthe pia-dissociated area in Ccdc39 homozygous mutant(Ccdc39prh/prh) rats and is expressed in GFAP-positiveastrocytes (white arrows) (n=1 in each group). Scale bars:2 mm (A, left column); 200 μm (A, middle and right columns);100 μm (B).

6

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

glymphatic system (Iliff et al., 2012; Ma et al., 2017). From P5(ncontrol=2, nprh/prh=1), when enlargement of the lateral ventricleswas first observed by MRI in Ccdc39prh/prh rats (Fig. 2B,D-G),to P12-P13 (ncontrol=3, nprh/prh=3), after the onset of severehydrocephalus in this model (Fig. 2C,H-M), mutants exhibitweakened Evans Blue staining of the perivascular spacesaround the ventral and lateral surfaces of the middle cerebralartery (MCA) compared with age-matched controls (P<0.001;Fig. 7A-D,I,J). At later stages of hydrocephalus, such as P27(ncontrol=1, nprh/prh=1), Ccdc39prh/prh rats exhibit no staining ofthe perivascular spaces around the ventral or lateral surfaces ofthe MCA compared with controls (P<0.001) (Fig. 7E-J).Nonlinear regression analysis of the length of tracer stainingdemonstrated a progressive decline in staining of the perivascularspaces surrounding the MCA of Ccdc39prh/prh mutants as thehydrocephalus phenotype progressed from P5 to P27 (P<0.001)

(Fig. 7I,J). Based on this data, we conclude that a progressiveimpairment in glymphatic-mediated CSF circulation, as traced byEvans Blue dye, exists in this model as Ccdc39prh/prh mutantsdevelop progressive hydrocephalus and enlargement of the lateralventricles over the first 4 weeks of life.

DISCUSSIONThe accessibility and affordability of CRISPR/Cas9 genomeengineering have allowed multiple scientific groups to generatenovel models of congenital hydrocephalus in a variety of organisms,includingmice (Morimoto et al., 2019), rats (Emmert et al., 2019) andfrogs (Date et al., 2019). The current study demonstrates oursuccessful generation of a robust rat model of neonatal hydrocephalusvia the use of CRISPR/Cas9 genome editing to introduce ahomozygous splice-site mutation, Ccdc39c.916+2T>A, in the Ccdc39gene that we previously identified in the prh mouse mutant.

Fig. 5. Neuroinflammation emerges during theearly stage of ventriculomegaly in Ccdc39mutant rats. (A) CD68-positive macrophagesmigrate to the subarachnoid space at P11 (n=2) andinvade into parenchyma at P30 (n=1) in Ccdc39homozygous mutants (Ccdc39prh/prh). MPO-positive neutrophils are barely seen in wild typesand Ccdc39prh/prh mutants at P11 (n=2, in eachgroup), but distinctly accumulate over the brainsurface with the parenchymal invasion at P30 inCcdc39prh/prh mutant rats (n=1, in each group).(B) Greater signal of MCP1 expression was seenbeginning at P5 in Ccdc39prh/prh mutant rats, whichreached statistical significance at P11 (mean±s.d.,P=0.14 for P5, **P<0.01 for P11; n=3 in eachgroup). (C) MCP1 signals are seen in NeuN-positiveneurons, CD68-positive macrophages andendothelial cells labeled with isolectin B4 (IB4) inCcdc39prh/prh mutants at P11 (n=3). Scale bars:200 μm (A,B); 100 μm (C, top and middle rows);50 μm (C, bottom row).

7

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

Using the L1cam mutant rat allele in a rat model of XLH, whichwe previously generated, we demonstrate potential geneticinteractions between the L1cam and Ccdc39 genes withphenotypic similarity. Epistatic influences on the expression ofgenes implicated in the pathogenesis of hydrocephalus have longbeen of interest to researchers attempting to explain the variablepresentation of this condition in laboratory models as well ashumans. For example, the phenotype of L1cam knockout mice,which includes cognitive deficits, hypoplasia of the corpuscallosum, and ventricular enlargement, varies greatly dependingon the genetic background of the model (Cohen et al., 1998; Dahmeet al., 1997; Itoh et al., 2004; Rolf et al., 2001; Rünker et al., 2003).Patients with L1CAM gene mutations present with severalconditions that vary considerably both between and withinfamilies, suggesting that the penetrance of L1 gene mutations isdependent upon the epistatic influences of other genes (Weller andGärtner, 2001). We found that rats containing mutations in bothgenes (L1camy/−;Ccdc39prh/prh) are smaller, develop earlier andmore severe ventriculomegaly and die earlier than either L1camy/−

or Ccdc39prh/prh rat mutants. Due to the essential functions of theL1CAM protein in neuronal migration, axon growth and guidanceand synaptic plasticity (Itoh and Fushiki, 2015), the disruptedneurogenesis that occurs as a result of the loss of functional L1CAMmay exacerbate the impaired motile cilia function and CSF flownecessary for neural cell growth and differentiation in this rat model.Indeed, we found that the lack of motile cilia results in delayed

maturation of cerebral cortical neurons, impaired myelinationand accelerated neural cell death in this model (Fig. 6). Theneurogenesis and neuronal developmental origins of hydrocephalusalso have been suggested by a recent study that found multipleneural stem cell fate gene mutations in the largest exome sequencingcohort in human congenital hydrocephalus (Furey et al., 2018a).Further study of neocortical neural development in L1cam;prhdouble-mutant rats may clarify the molecular mechanisms ofepistasis in hydrocephalus, leading to a better understanding of theheterogeneous presentations of this condition in both rodents andhumans. Likewise, future experiments should aim to elucidatethe epistatic influences of hydrocephalus-causing gene mutations onthe cognitive development of the hydrocephalic brain, as theL1cam;Ccdc39 double-mutation model is amenable to a variety ofbehavioral and nonbehavioral assays that were not performed in thecurrent study.

We found that the extent of interaction between L1cam andCcdc39 is not significant in the double-heterozygote condition. Ourdata on survival, growth and ventricular volume between L1camy/−;Ccdc39wt/prh and L1camy/− rats as well as L1camwt/−;Ccdc39wt/prh

and L1camwt/− rats suggest that the heterozygous level of CCDC39or L1CAM protein expression is sufficient to maintain ependymalcilia beating or neuronal cell growth, respectively, and does notaffect the XLH phenotype or primary ciliary dyskinesia phenotype.As such, our rat genetics data exclude the human heterozygousCCDC39 mutant allele and heterozygous L1CAM allele from the

Fig. 6. Myelination and neurofilamentdevelopment are impaired and cell death signalsare increased in Ccdc39prh/prh mutant brains.(A) MBP staining shows significantly reducedmyelination in the brain cortex in the rat Ccdc39homozygous mutant (Ccdc39prh/prh) compared withwild-type (Ccdc39wt/wt) rats at P11 (mean±s.e.m.,n=3 in each group). (B) NF-H staining showssignificantly reduced neurofilament densities inCcdc39 homozygous mutant (Ccdc39prh/prh)compared with wild-type (Ccdc39wt/wt) rats at P11(mean±s.e.m., n=3 in each group). (C) In Ccdc39homozygous mutant (Ccdc39prh/prh) rats, comparedwith wild type, more TUNEL-positive cells (arrows)are seen in the periventricular region at P1-P5(mean±s.e.m., n=3 in each group). Scale bars:100 μm (A); 200 μm (B,C). *P<0.05, **P<0.01.IV, cortical layer 4; V, cortical layer 5; VI, corticallayer 6; WM, white matter.

8

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

candidate genetic modifier elements that alter the disease severity ofXLH and primary ciliary dyskinesia.In previous experiments using the hydrocephalus rodent model, a

variety of inflammatory responses were observed, such as astrocyticactivation in hyh (hydrocephalus with hop gait) mutant mice (Roales-Buján et al., 2012) and reactive microgliosis in hydrocephalic H-Txrats, which can eventually inhibit neurite outgrowth and brain recovery(Mangano et al., 1998; Miller and McAllister, 2007). In kaolin-induced hydrocephalus rats, altered inflammatory gene expression dueto hydrocephalus was documented (Deren et al., 2010) in addition tothe reactive astrocytes and microglial response (Deren et al., 2010;Khan et al., 2006), although there has been criticism that kaolin itselfcould produce a global inflammatory responses in the brain(Oreškovic and Klarica, 2011). In a mouse model ofposthemorrhagic hydrocephalus, Toll-like receptor 4–nuclear factorkB (NF-kB) signaling has been described as an important pathway inmediating hydrocephalus development by causing further damage tothe ependymal cells or CSF hypersecretion from the choroid plexusepithelium (Karimy et al., 2017; Simard et al., 2011). The activation ofNF-kB is also known to interfere with ependymal ciliogenesis (Lattkeet al., 2012). However, little is known about the inflammatorymolecular pathway that mediates and can exacerbate hydrocephalusdevelopment from a relatively early phase in a model with no

preceding hemorrhage. In this studyof neonatal hydrocephalus causedby dysfunction of motile cilia in the ependyma and choroid plexus,and thus without pre-existing hemorrhage, we have shown thatinflammatory macrophage migration occurs at the early stage ofventricular expansion (P5) and that pro-inflammatory cytokineexpression, as well as further immune cells migration, follows inthe middle to later disease stages (P11-P30) mainly around thesubarachnoid/perivascular space as well as the periventriculararea in our model. To elucidate subtypes of pro-inflammatorysignals recruiting macrophages to the brain, we evaluated MCP1expression in our rat model of neonatal hydrocephalus. We foundsignificantly increased MCP1 signals at P11 that were expressedin neurons and endothelial cells, as well as macrophage invasionstarting from P5 with relatively mild ventriculomegaly. IncreasedMCP1 expression has been commonly reported in the CSF ofpatients with posthemorrhagic (Killer et al., 2010) or idiopathicnormal pressure hydrocephalus (Pfanner et al., 2018). Our datasuggest that, in neonatal hydrocephalus, MCP1 is one of the keymolecules that recruits the initial leukocyte infiltration andaccelerates neuroinflammation with macrophage and neutrophilextravasations. Future investigations in this model will elucidatewhether MCP1 inhibitors can ameliorate neuroinflammation and,potentially, hydrocephalus progression or functional outcomes, as

Fig. 7. Cerebrospinal fluid circulation throughthe glymphatic system is impaired inCcdc39prh/prh mutants. (A-F) Representativeimages of P5 control (A,B) and mutant(Ccdc39prh/prh) (C,D) rat brains and P27 control(E,F) and mutant (G,H) rat brains injected withEvans Blue dye into the cisterna magna to tracethe circulation of CSF through the glymphaticsystem. (I,J) Quantification of the length of EBstaining along the ventral (I) and lateral (J)surfaces of the middle cerebral artery in control(n=7) and mutant (n=5) rats. Results arepresented as individual replicates. The legend inJ also applies to I. ***P<0.001 determined bycomparing fits of nonlinear regression curves andbest-fit values. CM, cisterna magna; EB, EvansBlue; MCA, middle cerebral artery.

9

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

shown in other disease models such as experimental autoimmuneencephalomyelitis (Ge et al., 2012) or ethanol-inducedneurodegeneration (Zhang et al., 2018).Subarachnoid/subpial hemorrhagewas observed at P11, followed

later by hemorrhage into the lateral ventricles, possibly as a result ofinflammatory insult or distention of the neural tissues, whicheventually ruptured the vessels. MMP9 activation in astrocytes wasalso seen in the neocortical area, with arachnoid/pial membranedissociation from the parenchymal tissue with accumulated fluidand occasional bleeding. MMP9 in astrocytes has been reported tobe induced by many pro-inflammatory cytokines such as tumornecrosis factor-α, interleukin-1β or Toll-like receptor 2 along withreactive oxygen species in the setting of CNS inflammation (Minet al., 2015; Rosenberg, 2002; Yang et al., 2015). MMP9 is knownto degrade collagen IV and collagen V, leading to tissue destructionand eventually blood–brain barrier disruption (Könnecke andBechmann, 2013); therefore, it could exacerbate pathogenesis ofthe hydrocephalus by inducing bleeding and more inflammatorycell migration.Although it is challenging to differentiate pia mater from the

subarachnoid membrane by histology, some fluid accumulation inCcdc39prh/prh mutant rats was seen under the layer containingleptomeningeal vessels over the brain surface (arrows in Fig. 4A),which implies that fluid accumulated under pia mater. Pia materlacks tight junctions and is known to be permeable to water(Alcolado et al., 1988). Thus, it is possible that water comes acrossthe pia mater as a result of over-accumulation of CSF in thesubarachnoid space in our model of neonatal hydrocephalus. BecauseCD68-positive macrophage invasion into the periventricular regionwas seen in P5 mutants before the subarachnoid hemorrhage started,the inflammatory reaction is primarily a result of the hydrocephalusinsult. Conversely, this eventual subarachnoid hemorrhage andexacerbated inflammatory cell recruitment in the subarachnoidspace could have affected and further impaired the glymphatic flowalong perivascular spaces by hampering the absorption of CSF.An increased number of cells positive for cell death signal

(TUNEL) were observed in the periventricular region in ourCcdc39prh/prh pups at P1-P5. This distribution was very similar tothat of CD68-positive macrophages at this age. Because there was nosignificant difference in the number of cleaved caspase-3-positivecells to label apoptotic cells at P11 in our model (data not shown), it islikely that significant cell death in the periventricular region occurs atan earlier phase (P1-P5 in our model) rather than at the later stage ofthe disease in our model. Del Bigio and colleagues reported thatalthough there were significantly more overall TUNEL-positive cellsin kaolin-induced hydrocephalus rats 3-4 weeks after kaolin injectioninto the cisterna magna compared with controls, most of the TUNEL-positive cells (more than 95%) morphologically looked like non-neuronal cells (Del Bigio and Zhang, 1998). In the evaluation ofaxonal damage in kaolin-induced hydrocephalus rats, Ding et al.found that therewas little evidence for neuronal cell death at any stageof hydrocephalus, irrespective of the severity of axonal degeneration(Ding et al., 2001). Therefore, although periventricular axons areprimarily damaged by ventriculomegaly, neurons rarely die initiallyuntil the periventricular white matter is completely eroded (Del Bigio,2010) or thalamic retrograde axonal degeneration leads to cellapoptosis (Mori et al., 2002). The TUNEL-positive cells in theperiventricular region in our Ccdc39prh/prh rats may have been glia orother non-neuronal cells, as suggested in a previous report (Del Bigioand Zhang, 1998), although more investigation is needed.Impaired myelination and poorly developed neurofilaments at

P11 were also observed in this study. Although myelination had

barely started at P5, even in control rats, and there was no significantdifference in the number of mature oligodendrocytes in Ccdc39prh/prh mutant and control rats (Fig. S2A), there was a tendency fordecreased neurofilament density in Ccdc39prh/prh mutants at P5(Fig. S2B). Both periventricular axons and the myelin sheath arereported to be the major sites of injury in the hydrocephalic braindue to tissue extension, hypoxic insult (Del Bigio, 2001, 2010) orinflammatory insult, as observed in our hydrocephalus model.Oligodendrocytes are reported to be vulnerable to thehydrocephalus insult at young ages, such as up to P35 for ferrets(Di Curzio et al., 2013), and the impairment of myelination has beendescribed in neonatal hydrocephalus both in human (Hanlo et al.,1997) and animal models (Del Bigio et al., 1997, 2003). In 3-week-old rats with hydrocephalus induced by kaolin injection, the myelinsheath around axons greater than 0.4 μm in diameterbecame significantly thinner 1 week after injection, followed byirreversible axon loss if left untreated (Del Bigio et al., 1997).Whether dysmyelination precedes axonal loss or vice-versa remainscontroversial, but our result at P5 implies that axonal immaturityprecedes delayed myelination or demyelination. Brain geneexpression in hydrocephalic rats related to synaptogenesis,myelination, cell cycles and other important signaling seems to bealtered, depending on age (Balasubramaniam and Del Bigio, 2002).Therefore, our result might only apply to the sequelae of a specificage of the hydrocephalus development in which ventriculomegalystarts from P5 in rats.

The glymphatic system of humans and rodents serves as a CSFconduit from the subarachnoid space through periarterial spaces intothe brain parenchyma for drainage of parenchymal waste and solutesthrough perivenous spaces into meningeal and cervical lymphaticvessels (Klebe et al., 2019; Rasmussen et al., 2018). Impairedglymphatic-mediated fluid exchange has been associated withmultiple neurological disorders, including Alzheimer’s disease (Iliffet al., 2012), traumatic brain injury (Iliff et al., 2014) and stroke(Gaberel et al., 2014). The glymphatic system is hypothesized to playa role in dementia experienced by patients with normal pressurehydrocephalus (Rasmussen et al., 2018; Ringstad et al., 2017) as wellas inflammation and motor and cognitive defects in mice undergoingcraniectomy (Plog et al., 2019). In concordance with the recentdiscovery of the perinatal pattern of development of the glymphaticsystem (Munk et al., 2019), we found impaired circulation of CSFtracers along the glymphatic system in our neonatal hydrocephalusmodel, which constitutes the first report of the glymphatic system’sinvolvement in neonatal hydrocephalus. Specifically, we found thatCcdc39prh/prhmutant rats show reduced uptake of Evans Blue tracingdye along the perivascular spaces of the ventral and lateral surfaces ofthe middle cerebral artery as the progressive hydrocephalusphenotype becomes more severe from P5 to P27. Given theimportance of CSF pressure gradients between the subarachnoidspace and venous sinuses in driving CSF outflow (Klebe et al., 2019),we believe that reduced perivascular fluid bulk movement, whichcould be explained by altered CSF hydrodynamics or intracranialpressure in this model, causes reduced glymphatic uptake of CSF inCcdc39prh/prh rats. Alternatively, it is also possible that the lack ofmotile cilia-dependent CSF flow affects the structural developmentand maturation of this system. Due to the crucial solute-clearance roleof the glymphatic system, its reduced function in our model ofneonatal hydrocephalus may partly explain the cognitive impairmentof pediatric hydrocephalus patients if left untreated (Mangano et al.,2016).

Although understanding of the organization and function of theglymphatic pathway has quickly advanced since its original

10

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

identification in the rodent brain in 2012 (Iliff et al., 2012),mechanisms of glymphatic dysfunction in conditions such ashydrocephalus are predominantly theoretical (Klebe et al., 2019;Rasmussen et al., 2018). In line with emerging information that themouse glymphatic system begins developing in the hippocampus atP1 and is fully established in the cortex by 2 weeks of age (Munket al., 2019), future studies using shunt treatments such as theventriculo-subcutaneous shunt (Harris et al., 1994, 1996; Joneset al., 1995) in the Ccdc39prh/prh rat model can address whetherneonatal hydrocephalus impairs normal development of theglymphatic system. Therefore, we present the Ccdc39prh/prh rat asa strong model for studying the molecular basis of impairedneocorticogenesis, glymphatic fluid exchange and neurocognitiveand motor skill development in neonatal hydrocephalus.A primary limitation of the study was the small sample size of

L1camwt/−;Ccdc39prh/prh mutants used in ventricular volumeanalysis at P5 (n=2) and Ccdc39prh/prh mutants used inhistological analysis of MCP1 (n=3), MBP (n=3), NF-H (n=3)and cell death (n=3) at P1-P11. The primary challenge of generatinga new transgenic Ccdc39prh/prh rat model of hydrocephalus andinterbreeding it with a pre-existing transgenic L1camy/− rat model ofXLH is obtaining and maintaining mutants at ages ranging from P1to P90, which is confounded by the unpredictable death of rodentswith hydrocephalus (Di Curzio, 2018). Although our findings ofventricular enlargement, reduced survival and growth, increasedMCP1 expression, higher number of TUNEL-positive cells,downregulation of MBP and NF-H, and reduced Evans Blue dyeuptake into the glymphatic system in the Ccdc39prh/prh mutant ratdemonstrate statistically significant differences between groups, thepost-hoc power analysis of these studies demonstrates statisticalpower ranging from 62.1% to 100%. Future experiments ofventricular dilation and expression of neuroinflammatory markersin the early hydrocephalic brain in this model will further elucidatethe pathogenesis of hydrocephalus in Ccdc39prh/prh mutants in theabsence and presence of other genetic modifiers.In comparison to prh mutant mice (Abdelhamed et al., 2018),

Ccdc39prh/prhmutant rats similarly exhibited loss of brain CCDC39,decreased growth and survival throughout the weaning phase, andprogressive hydrocephalus with dilation of the lateral ventriclesfrom P5 that progressed into severe hydrocephalus by P11.Ccdc39prh/prh mutant rats showed slightly slower hydrocephalusdevelopment than prhmutant mice, as hydrocephalus in Ccdc39prh/prh mice progressed from P1 to a severe form by P7. We found thatCcdc39prh/prh rats developed dome-shaped heads, which indicatesthat the unfused skull suture might alleviate the elevation inintracranial pressure (ICP) at this neonatal stage (Jones et al., 2000),although ICP is expected to increase once the skull sutures fusebetween P12 and P20 (Grova et al., 2012; Roth et al., 1997) and theventriculomegaly reaches moderate proportions (Jones et al., 2000).Although the prhmutant mouse was useful for identifying the directdefects of motile cilia (Abdelhamed et al., 2018), the smallanatomical size of the murine model constrains its use for surgicalprocedures such as ICP recording (Hiploylee and Colbourne, 2014)and ventricular–subcutaneous shunting (Santos et al., 2016) that canmore easily and accurately be performed in rats. Ultimately, the useof these advanced surgical procedures in our larger rodent model ofneonatal hydrocephalus can provide new insights into the role ofintracranial pressure change in neonatal hydrocephalus.In conclusion, we report our successful generation of a novel rat

model of neonatal hydrocephalus using CRISPR/Cas9 to introducea recessive splice donor site mutation into the rat Ccdc39gene. Ccdc39prh/prh rat mutants recapitulate the progressive

hydrocephalus and decreased growth and survival of prh mice.We further found evidence for a novel genetic interaction betweentwo gene mutations in L1cam and Ccdc39 in neonatalhydrocephalus development, which supports their potentiallycommon physiological roles in normal brain development, suchas in neuronal growth, survival and maturation that occurdownstream of ependymal cilia-mediated CSF flow. Inflammatoryreactions via impeded CSF flow as demonstrated in our modelsuggest the active rather than passive roles of CSF flow retardationin altering neocortical maturation, which can cause further braindamage in neonatal hydrocephalus. Impaired glymphatic-mediatedCSF circulation in our model provides new insights into theglymphatic system, which may be involved in normal braindevelopment and the pathogenesis of neonatal hydrocephalus.Future experiments should investigate mechanisms of theinflammatory response as well as those regulating glymphaticflow and clearance using a combination of molecular techniques andadvanced imaging and surgery. Ultimately, improved understandingof CSF exchange mechanisms under conditions of inflammationcould lead to new therapeutic strategies for hydrocephalus byaltering CSF circulation.

MATERIALS AND METHODSCRISPR/Cas9-based rat Ccdc39 mutant generationCRISPR/Cas9-based genome editing was performed as describedpreviously (Emmert et al., 2019) with the following modifications: Singleguide (sg)RNAs (42 ng/µl of each) were mixed with 200 ng/µl Cas9 protein(Thermo Fisher Scientific), and single-stranded (ss)DNA donoroligonucleotides (oligos) containing the Ccdc39c.916+2T>A mutation andengineered HindIII restriction site (AAGCTT) were incubated at 37°C for15 min to form a ribonucleoprotein complex. The CRISPR/Cas9 complexmixture and donor oligos were injected into the cytoplasm of single-cellstage embryos (n=38) of Sprague Dawley rats using a piezo-drivenmicroinjection technique. Following embryo transfer into the oviducalampulla of pseudopregnant females, a founder generation (F0) of rats (n=26)was born and Sanger sequenced. F1 mutants were obtained throughbackcrossing of F0 mosaic animals to wild-type Sprague Dawley rat. Allfollowing generations of offspring were genotyped using a TaqManSample-to-SNP probe-based assay with custom probes for the ratCcdc39c.916+2T>A mutation (assay ID #AN322UG for Ccdc39prh).Breeding with and genotyping of the rat L1cam mutant allele wereperformed as described previously (Emmert et al., 2019). Rats analyzedusing MRI, histology and glymphatic-mediated CSF tracing were of F2 orF3 generations. Only male rats homozygous for L1cammutation (L1camy/−,L1camy/−;Ccdc39wt/prh, L1camy/−;Ccdc39prh/prh) were included becausefemale homozygous L1cam mutants cannot be obtained with our breedingtechniques due to sterility of male L1cam mutants with XLH. Rats werehoused in specific pathogen-free conditions and all experiments wereperformed according to the Institutional Animal Care and Use Committeeguidelines of the Cincinnati Children’s Hospital Medical Center.

Western blottingWestern blotting was performed as previously described (Abdelhamedet al., 2018; Emmert et al., 2019). Briefly, whole P11 rat brain lysates inRIPA buffer [50 mM Tris-Cl pH 7.4, 150 mM NaCl, 5 mM EDTA, 1%Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, 1% proteinase inhibitorcocktail (Thermo Fisher Scientific)] were separated, transferred to aPVDF membrane and probed with anti-CCDC39 (1:1000, #HPA035364,Sigma-Aldrich), anti-β-tubulin (1/1000, #T8660, Sigma-Aldrich), and anti-rabbit/mouse IgG-IRDye680RD/800CW (LI-COR) antibodies. Fluorescentsignals were detected using the Odyssey Imaging System (LI-COR).

Magnetic resonance imagingMRI datawere acquired on a Bruker 7TAvance horizontal bore small animalMRI scanner (Bruker, Billerica, MA). Control and mutant rats ranging inage from P1 to P90 were scanned. Rat pups at P1-P7 were anesthetized using

11

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

hypothermia because isoflurane is not effective at this age. Rats at P11 andolder were anesthetized with 2.5–3.5% isoflurane in air, positioned supineand scanned. Rat pups at P20 and older were secured on the MRI animal bedwith their teeth on a bite bar. For animals under isoflurane anesthesia,respiration was monitored and body temperature was maintained at 36°C–38°C using an animal monitoring system (SA Instruments, Stony Brook,NJ). Animals were positioned in the coil and centered in the bore of themagnet. Fluid-sensitive images were acquired with a fat-saturated 3D T2RARE sequence (Hennig et al., 1986) using the following parameters:repetition time 2 s, echo time 264 ms, echo spacing 11 ms, RARE factor 60,receiver bandwidth 104 kHz, averages 2, matrix 320×108×96, field of view48×16×144 mm and total scan time 4 min 40 s. DICOM images of controlsand mutants were imported into the ImageJ package Fiji. Volumes (mm3) ofthe lateral ventricles, third ventricle, fourth ventricle and pineal recess weremeasured using the Surfaces feature of the Imaris software (BitplaneScientific Software) (voxel size: x=0.150 mm, y=0.148 mm, z=0.150 mm).

Tracing of glymphatic-mediated CSF fluxThe pattern of CSF flow along the glymphatic system was assessed asdescribed previously (Iliff et al., 2012). Briefly, P5-P27 rats were deeplyanesthetized using ketamine (100 mg/kg) and xylazine (10 mg/kg). About2-10 μl of 4% Evans Blue dye in PBS (0.96 kDa; Sigma-Aldrich) wasinjected into the cisterna magna at a rate of 1-2 μl/min with a 33 G needleconnected to the micropump 11 elite (Harvard Apparatus). To measure theuptake of Evans Blue dye into the brain over 60 min, the rats were suturedand administered carprofen according to the survival surgery guidelines ofCincinnati Children’s Hospital Medical Center. Rats were perfusion fixed for60 min after the injection. Brains were fixed in 4%PFA overnight and imagedusing a stereomicroscope. The length (mm) of Evans Blue staining of theventral and lateral surfaces of the middle cerebral artery of controls andCcdc39prh/prh mutants was measured using the freehand line tool on Fiji.

Histology, immunohistochemical/immunofluorescence stainingand cell quantificationBrains in the skull removed fromCcdc39prh/prh and control rats aged P1-P30(n=28) were fixed in formalin for 24 h without systemic perfusion andembedded in paraffin after decalcification for two overnights (ShandonTBD-2, Thermo Fisher Scientific) and ethanol dehydration. Microtomesections (5 µm thick) after deparaffinization and rehydration were used inH&E staining or immunohistochemical staining following antigen retrievalin citrate buffer (pH 6) for 45 min. For the immunohistochemical/immunofluorescent staining, sections were incubated with primaryantibodies of either anti-rabbit MMP9 (N-terminus) (1:100; Proteintech,10375-2AP), anti-mouse GFAP (1:500; Sigma-Aldrich, G3893), anti-mouse CD68 (1:100; Abcam, ab31630), anti-rabbit MPO (1:50; Abcam,ab9535), anti-rabbit MCP1 (1:100; Bioss, BS1101R), anti-mouse NeuN(1:1000; Chemicon, MAB377B), anti-rabbit MBP (1:500; Abcam,ab40390) or anti-mouse NF-H (1:500; BioLegend, 801701) overnightafter blocking in 10% normal donkey serum and 0.1% Triton X-100 in PBSfor 1 h. After stringent washing and subsequent incubation withfluorophore- or horseradish peroxidase-conjugated secondary antibodiesand isolectin GS-IB4, the Alexa Fluor 594 Conjugate (IB4) (Thermo FisherScientific) was incubated along with secondary antibodies to labelendothelial cells. Sections were counterstained with DAPI (Sigma-Aldrich) or hematoxylin (Vector Laboratories), respectively. The terminaldeoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay wasperformed according to the manufacture’s protocol using Apop TagFluorescein In Situ Apoptosis Detection Kit (Millipore).

Fluorescently stained sections were observed under a confocal laserscanning microscope (Nikon A1R Ti-E inverted microscope). ImageJsoftware (NIH) was used to estimate the MBP-positive (P11) and NF-H-positive (P5, P11) areas out of the total brain area in a section, as well asMBP-positive cells per area at P5 in comparable sections. To quantifyTUNEL-positive cells relative to the number of DAPI-positive nuclei,comparable sections were selected to show periventricular regions. Thenumber of TUNEL-positive cells was manually counted and divided by thenumber of nuclei counted using ImageJ software. The MCP1immunohistochemistry signal reacted with DAB substrate at the surface of

the brain cortex, deep parenchyma and around the lateral ventricles wasobserved using a Nikon A1R Ti-E inverted microscope with a 40× objectivelens for P11 and an Olympus DP71 microscope with 40× objective lens forP5 evaluations. Signal intensity was analyzed using ImageJ software.

Statistical analysisAll values are expressed as the mean±standard deviation (s.d.) or standarderror of the mean (s.e.m.), as indicated. Justification for the use of parametrictwo-tailed statistical analyses, including two-way analysis of variance(ANOVA), Kaplan–Meier survival analysis and the Student’s t-test, isderived from normality testing of the data using Shapiro–Wilk tests.Statistical computation of group differences among more than two groupswas performed with two-way ANOVA with Holm–Šidák control formultiple comparisons. Survival data (defined as the number of days untildeath) were analyzed using the log-rank procedure of Kaplan–Meiersurvival analysis. Body weights and length of Evans Blue staining along theventral and lateral surfaces of the middle cerebral artery were analyzed bycomparing fits of nonlinear regression curves and best-fit values. Thedifferences between the two groups were compared using the Student’st-test. P<0.05 was considered statistically significant. All statisticalcomputations were performed in GraphPad Prism.

AcknowledgementsThe authors thank Prof. Kenneth Campbell (Cincinnati Children’s Hospital MedicalCenter) for providing critical comments and suggestions on this study.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsConceptualization: A.S.E, E.I, F.T.M., J.G.; Methodology: A.S.E., E.I., E.M.F., J.G.,R.S.D.; Software: A.S.E., R.S.D.; Validation: C.S., D.L., J.G., Y.-C.H.; FormalAnalysis: A.S.E., C.S., E.I., P.S.; Investigation: A.S.E., C.S., E.I., J.G., P.S.;Resources: C.S., D.L., Y.-C.H.; Data curation: E.M.F., R.S.D.; Writing - original draftpreparation: A.S.E., E.I., J.G.; Writing – review and editing: A.S.E., C.S., D.L., E.I.,E.M.F., F.T.M., J.G., R.S.D.; Visualization: A.S.E., E.I.; Supervision: F.T.M, J.G;Project administration: J.G; Funding acquisition: F.T.M, J.G.

FundingThis work was supported by the Hydrocephalus Association Innovator Award, theCenter for Clinical and Translational Science and Training Trustee Award, andMayfield Education and Research Foundation.

Data availabilityAll data are available upon request.

Supplementary informationSupplementary information available online athttp://dmm.biologists.org/lookup/doi/10.1242/dmm.040972.supplemental

ReferencesAbdelhamed, Z., Vuong, S. M., Hill, L., Shula, C., Timms, A., Beier, D.,

Campbell, K., Mangano, F. T., Stottmann, R.W. andGoto, J. (2018). Amutationin Ccdc39 causes neonatal hydrocephalus with abnormal motile ciliadevelopment in mice. Development 145, dev154500. doi:10.1242/dev.154500

Abdi, K., Lai, C.-H., Paez-Gonzalez, P., Lay, M., Pyun, J. and Kuo, C. T. (2018).Uncovering inherent cellular plasticity of multiciliated ependyma leading toventricular wall transformation and hydrocephalus.Nat. Commun. 9, 1655. doi:10.1038/s41467-018-03812-w

Adle-Biassette, H., Saugier-Veber, P., Fallet-Bianco, C., Delezoide, A.-L., Razavi,F., Drouot, N., Bazin, A., Beaufrere, A.-M., Bessieres, B., Blesson, S. et al.(2013). Neuropathological review of 138 cases genetically tested for X-linkedhydrocephalus: evidence for closely related clinical entities of unknown molecularbases. Acta Neuropathol. 126, 427-442. doi:10.1007/s00401-013-1146-1

Al-Dosari, M. S., Al-Owain, M., Tulbah, M., Kurdi, W., Adly, N., Al-Hemidan, A.,Masoodi, T. A., Albash, B. and Alkuraya, F. S. (2013). Mutation in MPDZ causessevere congenital hydrocephalus. J. Med. Genet. 50, 54-58. doi:10.1136/jmedgenet-2012-101294

Alcolado, R., Weller, R. O., Parrish, E. P. and Garrod, D. (1988). The cranialarachnoid and pia mater in man: anatomical and ultrastructural observations.Neuropathol. Appl. Neurobiol. 14, 1-17. doi:10.1111/j.1365-2990.1988.tb00862.x

Aspelund, A., Antila, S., Proulx, S. T., Karlsen, T. V., Karaman, S., Detmar, M.,Wiig, H. and Alitalo, K. (2015). A dural lymphatic vascular system that drains

12

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

brain interstitial fluid and macromolecules. J. Exp. Med. 212, 991-999. doi:10.1084/jem.20142290

Balasubramaniam, J. and Del Bigio, M. R. (2002). Analysis of age-dependantalteration in the brain gene expression profile following induction of hydrocephalusin rats. Exp. Neurol. 173, 105-113. doi:10.1006/exnr.2001.7831

Botfield, H., Gonzalez, A. M., Abdullah, O., Skjolding, A. D., Berry, M.,McAllister, J. P. and Logan, A. (2013). Decorin prevents the development ofjuvenile communicating hydrocephalus.Brain 136, 2842-2858. doi:10.1093/brain/awt203

Cohen, N. R., Taylor, J. S. H., Scott, L. B., Guillery, R.W., Soriano, P. and Furley,A. J. W. (1998). Errors in corticospinal axon guidance in mice lacking the neuralcell adhesion molecule L1. Curr. Biol. 8, 26-33. doi:10.1016/S0960-9822(98)70017-X

Dahme, M., Bartsch, U., Martini, R., Anliker, B., Schachner, M. and Mantei, N.(1997). Disruption of the mouse L1 gene leads to malformations of the nervoussystem. Nat. Genet. 17, 346-349. doi:10.1038/ng1197-346

Date, P., Ackermann, P., Furey, C., Fink, I. B., Jonas, S., Khokha, M. K., Kahle,K. T. and Deniz, E. (2019). Visualizing flow in an intact CSF network using opticalcoherence tomography: implications for human congenital hydrocephalus. Sci.Rep. 9, 6196. doi:10.1038/s41598-019-42549-4

Del Bigio, M. R. (2001). Pathophysiologic consequences of hydrocephalus.Neurosurg. Clin. North Am. 12, 639-649, vii. doi:10.1016/S1042-3680(18)30022-6

Del Bigio, M. R. (2010). Neuropathology and structural changes in hydrocephalus.Dev. Disabil. Res. Rev. 16, 16-22. doi:10.1002/ddrr.94

Del Bigio, M. R. and Zhang, Y. W. (1998). Cell death, axonal damage, and cell birthin the immature rat brain following induction of hydrocephalus. Exp. Neurol. 154,157-169. doi:10.1006/exnr.1998.6922

Del Bigio, M. R., Kanfer, J. N. and Zhang, Y. W. (1997). Myelination delay in thecerebral white matter of immature rats with Kaolin-induced hydrocephalus isreversible. J. Neuropathol. Exp. Neurol. 56, 1053-1066. doi:10.1097/00005072-199709000-00010

Del Bigio, M. R., Wilson, M. J. and Enno, T. (2003). Chronic hydrocephalus in ratsand humans: white matter loss and behavior changes. Ann. Neurol. 53, 337-346.doi:10.1002/ana.10453

Demyanenko, G. P., Tsai, A. Y. and Maness, P. F. (1999). Abnormalities inneuronal process extension, hippocampal development, and the ventricularsystem of L1 knockout mice. J. Neurosci. 19, 4907-4920. doi:10.1523/JNEUROSCI.19-12-04907.1999

Deren, K. E., Packer, M., Forsyth, J., Milash, B., Abdullah, O. M., Hsu, E. W. andMcAllister, J. P. (2010). Reactive astrocytosis, microgliosis and inflammation inrats with neonatal hydrocephalus. Exp. Neurol. 226, 110-119. doi:10.1016/j.expneurol.2010.08.010

Di Curzio, D. L. (2018). Animal models of hydrocephalus.Open J. Mod. Neurosurg.8, 57-71. doi:10.4236/ojmn.2018.81004

Di Curzio, D. L., Buist, R. J. and Del Bigio, M. R. (2013). Reduced subventricularzone proliferation and white matter damage in juvenile ferrets with kaolin-inducedhydrocephalus. Exp. Neurol. 248, 112-128. doi:10.1016/j.expneurol.2013.06.004

Ding, Y., McAllister, J. P., Yao, B., Yan, N. and Canady, A. I. (2001).Axona damage associated with enlargement of ventricles duringhydrocephalus: a silver impregnation study. Neurol. Res. 23, 581-587. doi:10.1179/016164101101199045

du Plessis, A. J. (1998). Posthemorrhagic hydrocephalus and brain injury in thepreterm infant: dilemmas in diagnosis and management. Semin. Pediatr. Neurol.5, 161-179. doi:10.1016/S1071-9091(98)80032-6

Ekici, A. B., Hilfinger, D., Jatzwauk, M., Thiel, C. T., Wenzel, D., Lorenz, I.,Boltshauser, E., Goecke, T. W., Staatz, G., Morris-Rosendahl, D. J. et al.(2010). Disturbed Wnt signalling due to a mutation in CCDC88C causes anautosomal recessive non-syndromic hydrocephalus with medial diverticulum.Mol. Syndromol. 1, 99-112. doi:10.1159/000319859

Emmert, A. S., Vuong, S. M., Shula, C., Lindquist, D., Yuan, W., Hu, Y.-C.,Mangano, F. T. and Goto, J. (2019). Characterization of a novel rat model ofX-linked hydrocephalus by CRISPR-mediated mutation in L1cam. J. Neurosurg.[Epub ahead of print] doi:10.3171/2018.10.JNS181015

Fryns, J. P., Spaepen, A., Cassiman, J. J. and van den Berghe, H. (1991).X linked complicated spastic paraplegia, MASA syndrome, and X linkedhydrocephalus owing to congenital stenosis of the aqueduct of Sylvius: variableexpression of the same mutation at Xq28. J. Med. Genet. 28, 429-431. doi:10.1136/jmg.28.6.429-a

Fukushima, A., Hagiwara, H., Fujioka, H., Kimura, F., Akema, T. and Funabashi,T. (2015). Sex differences in feeding behavior in rats: The relationship withneuronal activation in the hypothalamus. Front. Neurosci. 9, 1-7. doi:10.3389/fnins.2015.00088

Furey, C. G., Choi, J., Jin, S. C., Zeng, X., Timberlake, A. T., Nelson-Williams, C.,Mansuri, M. S., Lu, Q., Duran, D., Panchagnula, S. et al. (2018a). De novomutation in genes regulating neural stem cell fate in human congenitalhydrocephalus. Neuron 99, 302-314.e4. doi:10.1016/j.neuron.2018.06.019

Furey, C. G., Zeng, X., Dong, W., Jin, S. C., Choi, J., Timberlake, A. T., Dunbar,A. M., Allocco, A. A., Gunel, M., Lifton, R. P. et al. (2018b). Human genetics andmolecular mechanisms of congenital hydrocephalus. World Neurosurg. 119,441-443. doi:10.1016/j.wneu.2018.09.018

Gaberel, T., Gakuba, C., Goulay, R., De Lizarrondo, S. M., Hanouz, J.-L., Emery,E., Touze, E., Vivien, D. and Gauberti, M. (2014). Impaired glymphatic perfusionafter strokes revealed by contrast-enhanced MRI. Stroke 45, 3092-3096. doi:10.1161/STROKEAHA.114.006617

Ge, S., Shrestha, B., Paul, D., Keating, C., Cone, R., Guglielmotti, A. andPachter, J. S. (2012). The CCL2 synthesis inhibitor bindarit targets cells of theneurovascular unit, and suppresses experimental autoimmune encephalomyelitis.J. Neuroinflamm. 9, 171. doi:10.1186/1742-2094-9-171

Gram,M., Sveinsdottir, S.,Cinthio,M., Sveinsdottir, K., Hansson, S.R., Morgelin,M., Åkerstrom, B. and Ley, D. (2014). Extracellular hemoglobin - mediator ofinflammation and cell death in the choroid plexus following preterm intraventricularhemorrhage. J. Neuroinflamm. 11, 200. doi:10.1186/s12974-014-0200-9

Grova, M., Lo, D. D., Montoro, D., Hyun, J. S., Chung, M. T., Wan, D. C. andLongaker, M. T. (2012). Models of cranial suture biology. J. Craniofac. Surg. 23,1954-1958. doi:10.1097/SCS.0b013e318258ba53

Hanlo, P.W., Gooskens, R. J. H. M., van Schooneveld, M., Tulleken, C. A. F., vander Knaap, M. S., Faber, J. A. J. and Willemse, J. (1997). The effect ofintracranial pressure on myelination and the relationship with neurodevelopmentin infantile hydrocephalus. Dev. Med. Child Neurol. 39, 286-291. doi:10.1111/j.1469-8749.1997.tb07433.x

Harris, N. G., Jones, H. C. and Patel, S. (1994). Ventricle shunting in young H-Txrats with inherited congenital hydrocephalus: a quantitative histological study ofcortical grey matter.Childs. Nerv. Syst. 10, 293-301; discussion 301. doi:10.1007/BF00335166

Harris, N. G., McAllister, II, J. P., Conaughty, J. M. and Jones, H. C. (1996). Theeffect of inherited hydrocephalus and shunt treatment on cortical pyramidal celldendrites in the infant H-Tx rat. Exp. Neurol. 141, 269-279. doi:10.1006/exnr.1996.0161

Haverkamp, F., Wolfle, J., Aretz, M., Kramer, A., Hohmann, B., Fahnenstich, H.and Zerres, K. (1999). Congenital hydrocephalus internus and aqueductstenosis: aetiology and implications for genetic counselling. Eur. J. Pediatr. 158,474-478. doi:10.1007/s004310051123

Heep, A., Stoffel-Wagner, B., Bartmann, P., Benseler, S., Schaller, C., Groneck,P., Obladen, M. and Felderhoff-Mueser, U. (2004). Vascular endothelial growthfactor and transforming growth factor-β1 are highly expressed in the cerebrospinalfluid of premature infants with posthemorrhagic hydrocephalus. Pediatr. Res. 56,768-774. doi:10.1203/01.PDR.0000141524.32142.53

Hennig, J., Nauerth, A. and Friedburg, H. (1986). RARE imaging: a fast imagingmethod for clinicalMR.Magn.Reson.Med.3, 823-833. doi:10.1002/mrm.1910030602

Hiploylee, C. and Colbourne, F. (2014). Intracranial pressure measured in freelymoving rats for days after intracerebral hemorrhage. Exp. Neurol. 255, 49-55.doi:10.1016/j.expneurol.2014.02.017

Iliff, J. J., Wang, M., Liao, Y., Plogg, B. A., Peng, W., Gundersen, G. A.,Benveniste, H., Vates, G. E., Deane, R., Goldman, S. A. et al. (2012). Aparavascular pathway facilitates CSF flow through the brain parenchyma and theclearance of interstitial solutes, including amyloid β.Sci. Transl. Med. 4, 147ra111.doi:10.1126/scitranslmed.3003748

Iliff, J. J., Chen, M. J., Plog, B. A., Zeppenfeld, D. M., Soltero, M., Yang, L., Singh,I., Deane, R. and Nedergaard, M. (2014). Impairment of glymphatic pathwayfunction promotes tau pathology after traumatic brain injury. J. Neurosci. 34,16180-16193. doi:10.1523/JNEUROSCI.3020-14.2014

Itoh, K. and Fushiki, S. (2015). The role of L1cam in murine corticogenesis, and thepathogenesis of hydrocephalus. Pathol. Int. 65, 58-66. doi:10.1111/pin.12245

Itoh, K., Cheng, L., Kamei, Y., Fushiki, S., Kamiguchi, H., Gutwein, P., Stoeck, A.,Arnold,B., Altevogt, P. andLemmon, V. (2004). Brain development inmice lackingL1–L1 homophilic adhesion. J. Cell Biol. 165, 145-154. doi:10.1083/jcb.200312107

Jones, H. C., Harris, N. G., Briggs, R. W. and Williams, S. C. R. (1995). Shunttreatment at two postnatal ages in hydrocephalic H-Tx rats quantified using MRimaging. Exp. Neurol. 133, 144-152. doi:10.1006/exnr.1995.1017

Jones, H. C., Harris, N. G., Rocca, J. R. and Andersohn, R. W. (2000).Progressive tissue injury in infantile hydrocephalus and prevention/reversal withshunt treatment. Neurol. Res. 22, 89-96. doi:10.1080/01616412.2000.11741041

Karimy, J. K., Zhang, J., Kurland, D. B., Theriault, B. C., Duran, D., Stokum, J. A.,Furey, C. G., Zhou, X., Mansuri, M. S., Montejo, J. et al. (2017). Inflammation-dependent cerebrospinal fluid hypersecretion by the choroid plexus epithelium inposthemorrhagic hydrocephalus. Nat. Med. 23, 997-1003. doi:10.1038/nm.4361

Khan, O. H., Enno, T. L. and Del Bigio, M. R. (2006). Brain damage in neonatal ratsfollowing kaolin induction of hydrocephalus. Exp. Neurol. 200, 311-320. doi:10.1016/j.expneurol.2006.02.113

Killer, M., Arthur, A., Al-Schameri, A. R., Barr, J., Elbert, D., Ladurner, G., Shum,J. andCruise, G. (2010). Cytokine and growth factor concentration in cerebrospinalfluid from patients with hydrocephalus following endovascular embolization ofunruptured aneurysms in comparison with other types of hydrocephalus.Neurochem. Res. 35, 1652-1658. doi:10.1007/s11064-010-0226-z

Klebe, D., McBride, D., Krafft, P. R., Flores, J. J., Tang, J. and Zhang, J. H.(2019). Posthemorrhagic hydrocephalus development after germinal matrixhemorrhage: established mechanisms and proposed pathways. J. Neurosci.Res. [Epub ahead of print] doi:10.1002/jnr.24394

13

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms

Konnecke, H. and Bechmann, I. (2013). The role of microglia and matrixmetalloproteinases involvement in neuroinflammation and gliomas. Clin. Dev.Immunol. 2013, 914104. doi:10.1155/2013/914104

Lattke, M., Magnutzki, A., Walther, P., Wirth, T. and Baumann, B. (2012). Nuclearfactor B activation impairs ependymal ciliogenesis and links neuroinflammation tohydrocephalus formation. J. Neurosci. 32, 11511-11523. doi:10.1523/JNEUROSCI.0182-12.2012

Lee, L. (2013). Riding the wave of ependymal cilia: genetic susceptibility tohydrocephalus in primary ciliary dyskinesia. J. Neurosci. Res. 91, 1117-1132.doi:10.1002/jnr.23238

Louveau, A., Smirnov, I., Keyes, T. J., Eccles, J. D., Rouhani, S. J., Peske, J. D.,Derecki, N. C., Castle, D., Mandell, J. W., Lee, K. S. et al. (2015). Structural andfunctional features of central nervous system lymphatic vessels. Nature 523,337-341. doi:10.1038/nature14432

Ma, Q., Ineichen, B. V., Detmar, M. and Proulx, S. T. (2017). Outflow ofcerebrospinal fluid is predominantly through lymphatic vessels and is reduced inaged mice. Nat. Commun. 8, 1434. doi:10.1038/s41467-017-01484-6

Maloveska, M., Danko, J., Petrovova, E., Kresakova, L., Vdoviakova, K.,Michalicova, A., Kovac, A., Cubinkova, V. and Cizkova, D. (2018). Dynamics ofEvans blue clearance from cerebrospinal fluid into meningeal lymphatic vesselsand deep cervical lymph nodes. Neurol. Res. 40, 372-380. doi:10.1080/01616412.2018.1446282

Mangano, F. T., McAllister, J. P., II, Jones, H. C., Johnson, M. J. and Kriebel,R. M. (1998). The microglial response to progressive hydrocephalus in a model ofinherited aqueductal stenosis. Neurol. Res. 20, 697-704. doi:10.1080/01616412.1998.11740586

Mangano, F. T., Altaye, M., McKinstry, R. C., Shimony, J. S., Powell, S. K.,Phillips, J. M., Barnard, H., Limbrick, D. D., Holland, S. K., Jones, B. V. et al.(2016). Diffusion tensor imaging study of pediatric patients with congenitalhydrocephalus: 1-year postsurgical outcomes. J. Neurosurg. Pediatr. 18,306-319. doi:10.3171/2016.2.PEDS15628

Mashimo, T. (2014). Gene targeting technologies in rats: zinc finger nucleases,transcription activator-like effector nucleases, and clustered regularly interspacedshort palindromic repeats. Dev. Growth Differ. 56, 46-52. doi:10.1111/dgd.12110

Merveille, A.-C., Davis, E. E., Becker-Heck, A., Legendre, M., Amirav, I., Bataille, G.,Belmont, J., Beydon,N., Billen, F., Clement, A. et al. (2011). CCDC39 is required forassembly of inner dynein arms and the dynein regulatory complex and for normalciliary motility in humans and dogs. Nat. Genet. 43, 72-78. doi:10.1038/ng.726

Miller, J. M. and McAllister, J. P.II. (2007). Reduction of astrogliosis andmicrogliosis by cerebrospinal fluid shunting in experimental hydrocephalus.Cerebrospinal Fluid Res. 4, 5. doi:10.1186/1743-8454-4-5

Min, H., Hong, J., Cho, I.-H., Jang, Y. H., Lee, H., Kim, D., Yu, S.-W., Lee, S. andLee, S. J. (2015). TLR2-induced astrocyte MMP9 activation compromises theblood brain barrier and exacerbates intracerebral hemorrhage in animal models.Mol. Brain 8, 23. doi:10.1186/s13041-015-0116-z

Mori, F., Tanji, K., Yoshida, Y. and Wakabayashi, K. (2002). Thalamic retrogradedegeneration in the congenitally hydrocephalic rat is attributable to apoptotic celldeath. Neuropathology 22, 186-193. doi:10.1046/j.1440-1789.2002.00445.x

Morimoto, Y., Yoshida, S., Kinoshita, A., Satoh, C., Mishima, H., Yamaguchi, N.,Matsuda, K., Sakaguchi, M., Tanaka, T., Komohara, Y. et al. (2019). Nonsensemutation in CFAP43 causes normal-pressure hydrocephalus with ciliaryabnormalities. Neurology 92, e2364-e2374. doi:10.1212/WNL.0000000000007505

Munk, A. S., Wang,W., Bechet, N. B., Eltanahy, A. M., Cheng, A. X., Sigurdsson,B., Benraiss, A., Mae, M. A., Kress, B. T., Kelley, D. H. et al. (2019). PDGF-B isrequired for development of the glymphatic system. Cell Rep. 26, 2955-2969.e3.doi:10.1016/j.celrep.2019.02.050

Olstad, E. W., Ringers, C., Hansen, J. N., Wens, A., Brandt, C., Wachten, D.,Yaksi, E. and Jurisch-Yaksi, N. (2019). Ciliary beating compartmentalizescerebrospinal fluid flow in the brain and regulates ventricular development. Curr.Biol. 29, 229-241.e6. doi:10.1016/j.cub.2018.11.059

Oreskovic, D. and Klarica, M. (2011). Development of hydrocephalus and classicalhypothesis of cerebrospinal fluid hydrodynamics: facts and illusions. Prog.Neurobiol. 94, 238-258. doi:10.1016/j.pneurobio.2011.05.005

Pfanner, T., Henri-Bhargava, A. and Borchert, S. (2018). Cerebrospinal fluidbiomarkers as predictors of shunt response in idiopathic normal pressurehydrocephalus: a systematic review. Can. J. Neurol. Sci. 45, 3-10. doi:10.1017/cjn.2017.251

Plog, B. A., Lou, N., Pierre, C. A., Cove, A., Kenney, H. M., Hitomi, E., Kang, H.,Iliff, J. J., Zeppenfeld, D. M., Nedergaard, M. et al. (2019). When the air hits yourbrain: decreased arterial pulsatility after craniectomy leading to impairedglymphatic flow. J. Neurosurg., 1-14. [Epub ahead of print] doi:10.3171/2019.2.JNS182675

Rasmussen, M. K., Mestre, H. and Nedergaard, M. (2018). The glymphaticpathway in neurological disorders. Lancet Neurol. 17, 1016-1024. doi:10.1016/S1474-4422(18)30318-1

Ringstad, G., Vatnehol, S. A. S. and Eide, P. K. (2017). Glymphatic MRI inidiopathic normal pressure hydrocephalus. Brain 140, 2691-2705. doi:10.1093/brain/awx191

Roales-Bujan, R., Paez, P., Guerra, M., Rodrıguez, S., Vıo, K., Ho-Plagaro, A.,Garcıa-Bonilla, M., Rodrıguez-Perez, L.-M., Domınguez-Pinos, M.-D.,Rodrıguez, E.-M. et al. (2012). Astrocytes acquire morphological and functionalcharacteristics of ependymal cells following disruption of ependyma inhydrocephalus. Acta Neuropathol. 124, 531-546. doi:10.1007/s00401-012-0992-6

Rolf, B., Kutsche, M. andBartsch, U. (2001). Severe hydrocephalus in L1-deficientmice. Brain Res. 891, 247-252. doi:10.1016/S0006-8993(00)03219-4

Rosenberg, G. A. (2002). Matrix metalloproteinases in neuroinflammation.Glia 39,279-291. doi:10.1002/glia.10108

Rosenthal, A., Jouet, M. and Kenwrick, S. (1992). Aberrant splicing of neural celladhesion molecule L1 mRNA in a family with X-linked hydrocephalus [publishederratum appears in NatGenet 1993 Mar;3(3):273].Nat. Genet. 2, 107-112. doi:10.1038/ng1092-107

Roth, D. A., Longaker, M. T., McCarthy, J. G., Rosen, D. M., McMullen, H. F.,Levine, J. P., Sung, J. andGold, L. I. (1997). Studies in cranial suture biology: partI. increased immunoreactivity for TGF-β isoforms (β1, β2, and β3) during rat cranialsuture fusion. J. Bone Miner. Res. 12, 311-321. doi:10.1359/jbmr.1997.12.3.311

Runker, A. E., Bartsch, U., Nave, K.-A. and Schachner, M. (2003). The C264Ymissensemutation in the extracellular domain of L1 impairs protein trafficking in vitroand in vivo. J. Neurosci. 23, 277-286. doi:10.1523/JNEUROSCI.23-01-00277.2003

Sander, J. D. and Joung, J. K. (2014). CRISPR-Cas systems for editing, regulatingand targeting genomes. Nat. Biotechnol. 32, 347-355. doi:10.1038/nbt.2842

Santos, M. V., Garcia, C. A. B., Jardini, E. O., Romeiro, T. H., da Silva Lopes, L.,Machado, H. R. and de Oliveira, R. S. (2016). Ventricular-subcutaneous shuntfor the treatment of experimental hydrocephalus in young rats: technical note.Child’s Nerv. Syst. 32, 1507-1511. doi:10.1007/s00381-016-3042-1

Savman, K., Blennow, M., Hagberg, H., Tarkowski, E., Thoresen, M. andWhitelaw, A. (2002). Cytokine response in cerebrospinal fluid from preterminfants with posthaemorrhagic ventricular dilatation. Acta Paediatr. 91,1357-1363. doi:10.1111/j.1651-2227.2002.tb02834.x

Schrander-Stumpel, C. and Fryns, J.-P. (1998). Congenital hydrocephalus:Nosology and guidelines for clinical approach and genetic counselling.Eur. J. Pediatr. 157, 355-362. doi:10.1007/s004310050830

Serville, F., Lyonnet, S., Pelet, A., Reynaud, M., Louail, C., Munnich, A. and LeMerrer, M. (1992). X-linked hydrocephalus: clinical heterogeneity at a single genelocus. Eur. J. Pediatr. 151, 515-518. doi:10.1007/BF01957757

Shaheen, R., Sebai, M. A., Patel, N., Ewida, N., Kurdi, W., Altweijri, I., Sogaty, S.,Almardawi, E., Seidahmed, M. Z., Alnemri, A. et al. (2017). The geneticlandscape of familial congenital hydrocephalus. Ann. Neurol. 81, 890-897. doi:10.1002/ana.24964

Simard, P. F., Tosun, C., Melnichenko, L., Ivanova, S., Gerzanich, V. andSimard, J. M. (2011). Inflammation of the choroid plexus and ependymal layer ofthe ventricle following intraventricular hemorrhage. Transl. Stroke Res. 2,227-231. doi:10.1007/s12975-011-0070-8

Stoelinga, G. B. and van Munster, P. J. (1967). The behaviour of Evans blue (azo-dye T-1824) in the body after intravenous injection. Acta Physiol. Pharmacol.Neerl. 14, 391-409.

Stoll, C., Alembik, Y., Dott, B. and Roth, M. P. (1992). An epidemiologic study ofenvironmental and genetic factors in congenital hydrocephalus.Eur. J. Epidemiol.8, 797-803. doi:10.1007/BF00145322

Stottmann, R. W., Moran, J. L., Turbe-Doan, A., Driver, E., Kelley, M. and Beier,D. R. (2011). Focusing forward genetics: a tripartite ENU screen forneurodevelopmental mutations in the mouse. Genetics 188, 615-624. doi:10.1534/genetics.111.126862

Tully, H. M. and Dobyns, W. B. (2014). Infantile hydrocephalus: a review ofepidemiology, classification and causes. Eur. J. Med. Genet. 57, 359-368. doi:10.1016/j.ejmg.2014.06.002

Vinchon, M., Rekate, H. and Kulkarni, A. V. (2012). Pediatric hydrocephalusoutcomes: a review. Fluids Barriers CNS 9, 18. doi:10.1186/2045-8118-9-18

Weller, S. and Gartner, J. (2001). Genetic and clinical aspects of X-linkedhydrocephalus (L1 disease): mutations in the L1CAM gene.Hum.Mutat. 18, 1-12.doi:10.1002/humu.1144

Wolf, M. S., Chen, Y., Simon, D. W., Alexander, H., Ross, M., Gibson, G. A.,Manole,M. D., Bayır, H., Kochanek, P. M. and Clark, R. S. B. (2019). Quantitativeand qualitative assessment of glymphatic flux using Evans blue albumin.J. Neurosci. Methods 311, 436-441. doi:10.1016/j.jneumeth.2018.09.031

Yang, C.-M., Hsieh, H.-L., Yu, P.-H., Lin, C.-C. and Liu, S.-W. (2015). IL-1β inducesMMP-9-dependent brain astrocytic migration via transactivation of PDGFreceptor/NADPH oxidase 2-derived reactive oxygen species signals. Mol.Neurobiol. 52, 303-317. doi:10.1007/s12035-014-8838-y

Yung, Y. C., Mutoh, T., Lin, M.-E., Noguchi, K., Rivera, R. R., Choi, J. W.,Kingsbury, M. A. and Chun, J. (2011). Lysophosphatidic acid signaling mayinitiate fetal hydrocephalus. Sci. Transl. Med. 3, 99ra87. doi:10.1126/scitranslmed.3002095

Zhang, J., Williams, M. A. and Rigamonti, D. (2006). Genetics of humanhydrocephalus. J. Neurol. 253, 1255-1266. doi:10.1007/s00415-006-0245-5

Zhang, K., Wang, H., Xu, M., Frank, J. A. and Luo, J. (2018). Role of MCP-1and CCR2 in ethanol-induced neuroinflammation and neurodegeneration inthe developing brain. J. Neuroinflammation 15, 197. doi:10.1186/s12974-018-1241-2

14

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm040972. doi:10.1242/dmm.040972

Disea

seModels&Mechan

isms