Characterization of Subtle Brain Abnormalities in a Mouse Model of Hedgehog Pathway Antagonist-Induced Cleft Lip and Palate

Subtle behavioral and cognitive deficits have been documented in patient cohorts with orofacial clefts (OFCs). Recent neuroimaging studies argue that these traits are associated with structural brain abnormalities but have been limited to adolescent and adult populations where brain plasticity during infancy and childhood may be a confounding factor. Here, we employed high resolution magnetic resonance microscopy to examine primary brain morphology in a mouse model of OFCs. Transient in utero exposure to the Hedgehog (Hh) signaling pathway antagonist cyclopamine resulted in a spectrum of facial dysmorphology, including unilateral and bilateral cleft lip and palate, cleft of the secondary palate only, and a non-cleft phenotype marked by midfacial hypoplasia. Relative to controls, cyclopamine-exposed fetuses exhibited volumetric differences in several brain regions, including hypoplasia of the pituitary gland and olfactory bulbs, hyperplasia of the forebrain septal region, and expansion of the third ventricle. However, in affected fetuses the corpus callosum was intact and normal division of the forebrain was observed. This argues that temporally-specific Hh signaling perturbation can result in typical appearing OFCs in the absence of holoprosencephaly—a condition classically associated with Hh pathway inhibition and frequently co-occurring with OFCs. Supporting the premise that some forms of OFCs co-occur with subtle brain malformations, these results provide a possible ontological basis for traits identified in clinical populations. They also argue in favor of future investigations into genetic and/or environmental modulation of the Hh pathway in the etiopathogenesis of orofacial clefting.


Introduction
Clefts of the lip with or without palate (CL/P) and cleft palate only (CPO) are commonly occurring human birth defects that cause significant morbidity and require extensive medical intervention [1]. Even when comprehensive treatment is available, these malformations bring significant individual, familial, and societal burden [1,2]. The mental health of individuals born with ''non-syndromic'' orofacial clefts (OFCs) has been extensively studied and psychosocial impairment, particularly in relation to social interaction, has been well documented [3][4][5]. Cognitive impairment, with specific deficits in verbal fluency, has also been described [6][7][8][9]. While the majority function within the normal IQ range, some studies have found that affected individuals score lower than peers without clefts [7,10]. Of particular importance, the prevalence of learning disabilities in populations with OFCs has been observed to be nearly ten-fold higher than that of the general population [6,9,[11][12][13].
Conventional wisdom has held that these behavioral and cognitive traits are secondary to the speech and hearing-related complications frequent in this population or even to social stigma related to physical appearance [14,15]. However, an alternative hypothesis has been advanced, which holds that cognitive and behavioral traits identified in populations with OFCs are, in fact, a primary problem resulting from abnormal brain development [16]. Recent neuroimaging studies support this premise with the demonstration of consistent patterns of subtle structural brain abnormalities in adolescent and adult populations with OFCs. These studies have found that clinical populations exhibit disproportionate volume reductions of the frontal lobe, subcortical nuclei, and cerebellum [16,17]. In addition to these overall size reductions, non-uniform shifts in cerebral and cerebellar volumes have also been found [18].
Development of the face and brain is an intimately-interrelated process [19]. Along with serving as a structural scaffold, the brain provides inductive molecular signals that guide development of the adjacently developing face [20]. The Hedgehog (Hh) signaling pathway has been identified as a key molecular mediator of brainface development. Sonic Hedgehog (Shh) is initially detected in the neuroectoderm of the ventral midbrain, with expression expanding caudally into the hindbrain, and rostrally into the diencephalon and telencephalon. Hh signaling drives medial forebrain expansion and induces ventral progenitor domains. Expression of Shh in the neuroectoderm of the ventral forebrain indirectly induces a parallel field of expression in the facial ectoderm [21], inducing the expression of Hh target genes, including Gli1, in the intervening neural crest-derived mesenchymal cell population. This forebrain signaling center is critical for growth and differentiation of the midfacial primordia [22]. Shh null mice exhibit profound developmental defects [23], including holoprosencephaly (HPE), a condition defined by incomplete division of the forebrain, characterized by medial forebrain deficiency, and which commonly occurs with CL/P in clinical populations [23,24]. In humans, mutations in SHH are the most commonly identified cause of non-chromosomal HPE, accounting for approximately 12% of such cases [25,26]. However, in a recent analysis only 36% of SHH mutation carriers were found to have true HPE, with the remaining carriers classified as unaffected or as having microform HPE (i.e. facial abnormalities in the absence of detectable neuroanatomical anomalies) [26].
When Hh signaling from the forebrain neuroectoderm is blocked, Shh expression is not established in the facial ectoderm. This results in attenuated growth of the frontonasal prominence, which, in the chick, causes truncation of the upper beak [21]. In the mouse, we have shown that temporally-specific exposure to the Hh pathway antagonist cyclopamine results in a deficiency of the frontonasal prominence-derived medial nasal processes, which contribute to the philtrum of the upper lip, the alveolar ridge, the primary palate, and the median nose. This manifests as clefts of the lip and palate that appear to mimic human clinical phenotypes [27].
Here, employing a refined model of cyclopamine-exposure, we set out to determine whether Hh antagonist-induced facial dysmorphology is associated with abnormal brain development. High resolution magnetic resonance microscopy (MRM) was applied for concurrent visualization and measurement of facial and brain features, while diffusion tensor imaging was used to visualize white matter fiber tracts. The findings described herein illustrate that temporally-specific inhibition of the Hh signaling pathway results in clinically-relevant facial dysmorphology in the absence of the severe brain malformations that define HPE. Importantly however, subtle volumetric differences were identified in several forebrain regions of cyclopamine-exposed fetuses, providing a potential developmental basis for characteristics previously observed in affected clinical populations. Along with human neuroimaging studies, these results support the premise that some forms of typically appearing OFCs co-occur with primary structural brain abnormalities.

Timed Mouse Mating
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All procedures involving animals were approved by the University of North Carolina at Chapel Hill Institutional Animal Care and Use Committee (protocol number 13-081.0). C57BL/6J mice were purchased from The Jackson Laboratory. Two female mice were placed with a single male for 2 hrs in the light cycle and subsequently examined for the presence of copulation plugs, marking gestational day (GD)0.

Cyclopamine Exposure
Cyclopamine (LC Labs) was dissolved in a sodium phosphate/ citrate buffer containing 30% (wt/wt) 2-hydropropyl-b-cyclodextrin and administered to timed pregnant animals by subcutaneous infusion using Alzet (Durect) microosmotic pumps as previously described [28]. Model 2001D pumps (227 ml volume capacity, dispensed at 8.4 ml/h for 27 h) were loaded with cyclopamine solution to achieve a dispensation rate of 120 mg/kg/d and surgically implanted subcutaneously at GD8.25. Control animals were administered vehicle alone following the same paradigm. To assess potential effects of the surgical procedure, pump implantation, and vehicle exposure on brain development, control animals from this study were compared to a separate control population of GD17 C57BL/6J fetuses exposed to two intraperitoneal injections of lactated Ringer's solution given four hrs apart beginning at either GD7 or GD8.5 [29]. Both control populations were examined using the same methodology. Neither total brain volume, nor regional volume percentages were significantly different between the two control groups (Fig. S1).

Magnetic Resonance Imaging
On GD 17, dams were euthanized and fetuses were removed in ice-cold phosphate-buffered saline. Following fixation in a 20:1 Bouin's fixative:Prohance (Bracco Diagnostics) solution for 9 hours, fetuses were stored at 4uC in a 200:1 solution of PBS:Prohance until imaging. Magnetic Resonance Microscopy (MRM) was performed on a 9.4 T vertical bore magnet interfaced to a GE console running Epic 12.46 (GE Medical Systems). The system is equipped with Resonance Research gradients (Resonance Research, Inc.), which achieve peak gradients of 2000 mT/ m. Two fetal heads were placed in an acrylic sample holder and immersed in fomblin. 3D volume images were acquired in a 10 mm diameter 625 mm long solenoid radiofrequency coil using a radiofrequency refocused spin echo sequence (TR = 50 ms, TE = 5.2 ms, field of view = 20610610 mm, matrix size = 51262566256), resulting in isotropic spatial resolution of 39 mm. A novel acquisition strategy that amplifies the highfrequency information by selectively altering the receiver gain during the phase-encoding steps was applied to extend the dynamic range of the system, capture the higher-frequency components, and limit saturation in the central k-space [30]. Total scan time for each pair of specimens was approximately 1.1 hrs.

Brain Segmentation
Pituitary, cerebellum, and forebrain regions were manually segmented using ITK-Snap (Version 2.1.4) as previously described [29,31]. Automated skull-stripping [32] was used to generate whole brain volumes as previously described [29]. As shown in

Linear Measurements
Linear measurements of snout width (SW), snout length (SL), and mandible length (ML) were obtained from 3D head reconstructions using netfabb Studio Basic (Version 4.7). For animals with UL-CLP, length measurements were taken from the side of the face with the lip cleft. Length measurements were taken from the right side for all other fetuses. Interocular distance (IOD) measurements were captured from coronal MRM images using ImageJ (Version 1.47t; http://rsb.info.nih.gov/ij/). Volumetric brain measurements were obtained with ITK-Snap (Version 2.1.4).

Statistics
Multivariate analyses of variance (MANOVAs) were used to determine significant group differences in linear and volumetric measurements. Significant between-subject effects were followed by Student Newman Keuls posthoc tests when appropriate. An alpha value of 0.05 was maintained for all analyses.

Supporting Information
Validation of vehicle exposure paradigm. The control population for this study was exposed in utero to the cyclopamine vehicle (sodium phosphate/citrate buffer containing 30% (wt/wt) 2-hydropropyl-b-cyclodextrin) using Alzet microosmotic pumps subcutaneously implanted in timed-pregnant dams. Total brain volume and regional volume percentages from this control population were compared to those of a previously described control population exposed to a vehicle for ethanol (lactated Ringer's solution) by intraperitoneal injection at either GD7 or GD8.5 [29]. Image acquisition and processing, brain segmentation, and volumetric analysis were performed using the same methodologies for both data sets.
Gross Visualization of the Corpus Callosum. Following fixation in Bouin's solution, GD19 fetuses were hemisected in the sagittal plane with a scalpel. 20 mls of undiluted Harris Modified Hematoxylin solution was placed on the cut surface of the brain and allowed to penetrate for 20 min. Bright-field images were captured with a MicroPubisher 5.0 camera using QCapture Suite software and subsequently converted to grayscale.

Results
To examine face-brain dysmorphology, GD17 fetuses exposed to cyclopamine or vehicle alone were imaged by MRM (Fig. 1). For each fetus, facial morphology was assessed by gross inspection under a light microscope and examination of serial coronal MRM images. While vehicle-exposed control fetuses were grossly normal, facial dysmorphology was observed in each of the six cyclopamineexposed litters. For subsequent analyses, affected fetuses were grouped by facial phenotype, including: unilateral cleft lip and palate (UL-CLP; n = 10), bilateral cleft lip and palate (BL-CLP; n = 10), cleft of the secondary palate only (CPO; n = 5), and noncleft (NC; n = 20) (table 1). In the NC group, midfacial hypoplasia of varying severity was observed (Fig. 2). Relative to the vehicleexposed control group, snout width was reduced in the NC and CPO groups but increased in the UL-CLP and BL-CLP groups. In each cyclopamine-exposed group, snout length and mandible length was decreased, while interocular distance was increased. Primary brain morphology was then examined in MRM sections and 3D brain reconstructions as illustrated in Fig. 3 for  representative animals from each group. Along with illustrating normal division of the cerebral cortices in all groups, coronal sections confirmed clefts of the secondary palate in the CPO, UL-CLP, and BL-CLP groups. 3D reconstructions and transverse sections illustrate normal division of the forebrain with an intact septal region in vehicle and cyclopamine-exposed animals. However, relative to the vehicle control, several brain regions appear dysmorphic in cyclopamine-exposed fetuses. Indeed, quantitative analysis revealed that total brain volume was reduced in each of the cyclopamine-exposed groups with overt clefts, while disproportionate changes in regional brain volumes were detected in each of the cyclopamine-exposed groups (Fig. 4). Specifically, pituitary volume was reduced in each of the groups, while olfactory bulb volume was reduced in the UL-CLP and BL-CLP groups. Increased volume of the forebrain septal region was found in each of the cyclopamine-exposure groups, and volume of the third ventricle was significantly increased in groups with OFCs. These affected regions can be visualized in the transverse sections and reconstructions shown in Fig. 3. With its complete or partial agenesis being a hallmark neuroanatomical feature of HPE [34,35], we next examined the morphology of the corpus callosum in cyclopamine-exposed fetuses with UL-CLP and BL-CLP, along with vehicle-exposed controls. Diffusion tensor imaging was applied to generate directionally encoded color maps of white matter fiber tracts in GD19 fetuses. This stage was chosen because increased myelination and size facilitated better visualization of white matter fiber tracts. In each of three fetuses with cyclopamine-induced UL-CLP that were examined, the corpus callosum, as well as the hippocampal and anterior commissures were present and not grossly hypoplastic relative to the vehicle-exposed controls (Fig. 5). This result was paralleled in three vehicle-and three cyclopamineexposed fetuses with BL-CLP examined by gross histological staining (Fig. S2).

Discussion
Along with the findings of human neuroimaging studies and a recent examination of brain morphology in Irf6 heterozygous adult mice [36], the results presented here support the hypothesis that some forms of OFCs co-occur with brain dysmorphology. These largely congruent observations from clinical cohorts and highly controlled animal models should prompt increased clinical awareness of primary brain abnormalities that preferentially occur in patients with OFCs. As described herein, observed abnormalities of the pituitary, olfactory bulbs, and septal region may provide an ontological basis for traits previously described in clinical populations. One such clinical study described olfaction deficits in adult males born with OFCs [37], while short stature has been well characterized and linked to possible pituitary dysfunction in affected patient populations [38][39][40]. While a recent study found no significant difference in pituitary volume between adult patients with OFCs and controls without clefts, pituitary structure and function were not specifically examined [41]. Consistent with Transverse sections show normal division of the cerebral cortices with an intact septal region. These images also illustrate deficiency of the pituitary (arrow in G) and olfactory bulbs, and enlargement of the third ventricle (arrow head in M) and septal region in cyclopamine-exposed fetuses. Color-coding in F-O is shown in Fig. 1, where dark red = cerebral cortices, light green = diencephalon, dark blue = septal region, yellow = lateral ventricles, orange = third ventricle, pink = olfactory bulbs, light purple = pituitary. doi:10.1371/journal.pone.0102603.g003 Figure 2. Cyclopamine-induced facial dysmorphology. From MRM images, extracted facial surfaces are shown for a vehicle-exposed control fetus (A,B), along with three fetuses representative of the phenotypic spectrum observed in the cyclopamine-exposed non-cleft (NC) group (C-H). Snout width (SW), snout length (SL), and mandible length (ML) were measured from facial reconstructions, while interocular distance was measured from coronal MRM sections (inset). For each cyclopamine-exposed group, linear measurements are reported as percent difference relative to the vehicle exposure group. Values represent the mean + S.E.M. * p,0.05 compared to vehicle-exposed control group. doi:10.1371/journal.pone.0102603.g002 increased volume of the forebrain septal region as described here, neuroimaging studies have identified shifts in tissue distribution in which the anterior cerebrum is enlarged [16,42]. Given the developmental and anatomical relationship to the septum pellucidum, the observation of septal region hyperplasia is developmentally consistent with the finding of increased incidence of cavum septum pellucidum in adult males with OFCs [43]. Enlargement of the third ventricle, as observed in cyclopamineexposed fetuses with OFCs, has been described in human clinical populations with neurocognitive deficits, as well as in fetal mice exposed prenatally to ethanol [29,31,44,45]. While it is thought that ventricular enlargement may be indicative of cognitive dysfunction, a causative relationship has not been clearly established.
Previously, we demonstrated that the brains of mouse fetuses with cyclopamine-induced CL/P are not grossly holoprosencephalic [27]. However, the throughput and resolution of the employed methodology limited analysis to a small sample population and was incompatible with regional brain segmentation and volumetric analysis. The imaging platform used for this study enabled near-histologic resolution of anatomical sections and visualization of white matter fiber tracts. With a sample size sufficient to examine phenotypic subpopulations, we demonstrate that the applied cyclopamine exposure paradigm results in facial dysmorphism in the absence of the defining characteristics of HPE. Specifically, cyclopamine-exposed fetuses exhibited separate cerebral hemispheres, an intact corpus callosum, and a surprising increase in forebrain septal region volume. Moreover, a slight but significant increase in interocular distance was found in each of the cyclopamine-exposed groups with clefts. While HPE is typically associated with reduced interocular distance [35], clinical populations with non-syndromic OFCs have been described as hyperteloric or as having normal interocular distance [46].
We have also shown that cyclopamine-induced cleft lip results from a medial nasal process tissue deficiency [27]. This facial structure receives molecular, cellular, and structural input from the developing forebrain. While forebrain abnormalities have been described in clinical populations with CL/P, how these may relate to this initial pathogeneses of clefting is less clear because direct observation of early morphogenesis in affected human embryos is not readily done. However, several lines of evidence suggest that the morphology of the upper lateral incisor in patients with cleft lip provides important clues regarding the genesis of the clefting defect. The lateral incisor has dual origins, with two distinct dental epithelial thickenings present on the medial nasal and maxillary processes uniting following closure of the upper lip to form this tooth [47]. In individuals with cleft lip and palate, four distinct upper lateral incisor patterns occur and entail each permutation of the presence or absence of a small tooth on either side of the cleft site. One study of primary dentition in a large sample size found that a single tooth lateral to the cleft was present in 82% of cases [48]. The absence of a small lateral incisor component medial to the cleft site implies a deficiency of the medial nasal process was operational in the pathogenesis of clefting in these individuals. The apparent commonality of medial nasal process deficiency and subtle forebrain abnormalities bolsters the translational potential of the data acquired from the mouse model examined here to the clinical condition.
In addition to technological advances, this study also benefited from practical methodological refinements of a previously used cyclopamine exposure paradigm [27,28]. Specifically, timedmating periods were reduced from overnight to two hours, allowing for increased stage specificity, and cyclopamine was purchased from LC Labs following in-house testing demonstrating increased purity and solubility of the compound compared to previously used sources. This allowed us to reduce the rate of cyclopamine infusion to 120 mg/kg/d while maintaining teratogenic efficacy. While circumventing dam toxicity, these refine- . Volumetric brain abnormalities in cyclopamineexposed fetuses. Total brain volumes (inset) were derived following automated skull stripping. Values represent the mean + S.E.M. * p,0.05 compared to the vehicle-exposed control group. For determination of disproportionate differences, the volume of each manually segmented brain region was calculated as a percentage of total brain volume for each animal. Remaining volume includes mid-and hindbrain regions. To illustrate relative changes on the same scale, percent volumes are normalized to mean control values. Values represent the mean 6 the S.E.M. * p,0.05 compared to the control group. doi:10.1371/journal.pone.0102603.g004 Figure 5. Corpus callosum integrity in fetuses with UL-CLP. Diffusion tensor imaging was used to visualize white matter fiber tracts in GD19 vehicle-exposed (top row) and cyclopamine-exposed fetuses with UL-CLP (bottom row). For each group, one fetus was selected at random from each of three independent litters. The directionality of the brain fiber tracts in the color-coded maps are indicated as follows: red, left/right; blue, inferior/superior; and green, anterior/posterior. In both controls and cyclopamine-exposed fetuses, the corpus callosum (dashed outline), hippocampal commissure (arrow) and anterior commissure (arrowhead) are evident. doi:10.1371/journal.pone.0102603.g005 ments led to a marked increase in interlitter penetrance, with fetuses with clefts present in all six cyclopamine-exposed litters. In addition to previously described cleft lip and palate phenotypes, animals with clefts of the secondary palate only were also observed, albeit infrequently. While it is unclear how cyclopamine exposure causes these phenotypes that are traditionally considered disparate, a role for Hh signaling in secondary palate development and CPO has been well described in mouse models [49].
Slightly more than half of all cyclopamine-exposed fetuses did not present with clefts. However, the hypertelorism seen in animals with clefts was recapitulated in this group, as were most of the identified volumetric brain changes. Presenting with midfacial hypoplasia, this group could be considered as having a ''sub-cleft'' phenotype, which appears to be clinically relevant. Weinberg et al. found significant disparity between ''unaffected'' parents of children with overt clefts, and controls, with the former exhibiting midfacial retrusion and excess interorbital width [50]. Together, these findings argue that clefting is not a simple binary outcome and that intermediate phenotypes may provide important etiological insights.
Prevention strategies for non-syndromic CL/P and CPO are limited because our current understanding of causative factors is inadequate. The findings presented here argue that temporallyspecific perturbation of the Hh signaling pathway results in a spectrum of clinically-relevant facial dysmorphology, including cleft lip and palate, cleft palate only, and a sub-cleft phenotype. While not meeting the criteria for HPE, affected animals exhibited subtle volumetric brain abnormalities that provide a possible ontological basis for traits described in clinical cohorts with nonsyndromic OFCs. Taken together, these findings argue that efforts to identify genetic and environmental Hh signaling pathway modifiers could provide new insights and potential preventative strategies for these common and morbid human birth defects. Figure S1 Exposure paradigm validation. The vehicle exposure group for this study [Vehicle (cyclopamine)] was compared to that of a previously examined control group exposed to lactated ringers by intraperitoneal injection at GD7 or GD8. 5 [Vehicle (ethanol)]. For determination of disproportionate differences, the volume of each manually segmented brain region was calculated as a percentage of total brain volume for each animal. To illustrate relative changes on the same scale, percent volumes are normalized to mean control values. Values represent the mean 6 the S.E.M. Neither total brain volume (inset) nor percent regional brain volume was significantly different between vehicle exposure groups. * p,0.05 compared to control group. (TIF) Figure S2 Corpus callosum integrity in fetuses with BL-CLP. Vehicle exposed control (top row) and cyclopamine-exposed GD19 fetuses (bottom row) were hemisected in the sagittal plane. The corpus callosum (dashed outline) was visualized using concentrated hematoxylin stain. For each group, one fetus was selected at random from each of three independent litters. (TIF)

Supporting Information
Author Contributions