The authors have declared that no competing interests exist.
Conceived and designed the experiments: RJL KKS PH GAJ. Performed the experiments: RJL PH JJA SJP YJ FB EAG JLE SEP DBD. Analyzed the data: SKO-M PH RJL. Contributed reagents/materials/analysis tools: MS IO MAS HTH. Wrote the paper: RJL KKS SKO-M PH.
Prenatal ethanol exposure is the leading preventable cause of congenital mental disability. Whereas a diagnosis of fetal alcohol syndrome (FAS) requires identification of a specific pattern of craniofacial dysmorphology, most individuals with behavioral and neurological sequelae of heavy prenatal ethanol exposure do not exhibit these defining facial characteristics. Here, a novel integration of MRI and dense surface modeling-based shape analysis was applied to characterize concurrent face-brain phenotypes in C57Bl/6J fetuses exposed to ethanol on gestational day (GD)7 or GD8.5. The facial phenotype resulting from ethanol exposure depended upon stage of insult and was predictive of unique patterns of corresponding brain abnormalities. Ethanol exposure on GD7 produced a constellation of dysmorphic facial features characteristic of human FAS, including severe midfacial hypoplasia, shortening of the palpebral fissures, an elongated upper lip, and deficient philtrum. In contrast, ethanol exposure on GD8.5 caused mild midfacial hypoplasia and palpebral fissure shortening, a shortened upper lip, and a preserved philtrum. These distinct, stage-specific facial phenotypes were associated with unique volumetric and shape abnormalities of the septal region, pituitary, and olfactory bulbs. By demonstrating that early prenatal ethanol exposure can cause more than one temporally-specific pattern of defects, these findings illustrate the need for an expansion of current diagnostic criteria to better capture the full range of facial and brain dysmorphology in fetal alcohol spectrum disorders.
Prenatal ethanol exposure causes a range of structural and functional abnormalities, collectively termed fetal alcohol spectrum disorder (FASD). At the severe end of this spectrum is fetal alcohol syndrome (FAS), which is characterized by growth retardation, CNS dysfunction, and a specific pattern of craniofacial dysmorphology. While prenatal ethanol exposure is already recognized as a leading known cause of mental disability
Development of the face and brain is intimately interrelated, as the brain provides structural, cellular, and molecular input that the guides development of the adjacently developing face
Forebrain and pituitary regions were manually segmented from transverse 39 µm MRM sections (A). 3D brain reconstructions were generated by overlaying manually segmented regions with whole-brain masks (B). From the same MRM scans, 3D head reconstructions were created featuring detailed facial surfaces (C). The brain and face can be visualized concurrently
Animal-based studies have been instrumental in elucidating the complex pathogenesis of ethanol-mediated teratogenicity
(Left) Mean surface shape of the GD7 and GD8.5 exposure groups relative to the control group is shown in the first and second columns respectively, while the GD7 group is directly compared to the GD8.5 ethanol-exposed group in the third column. (A–C) Color-map comparisons reflecting the displacement of mean surface shape for the indicated groups, where red indicates regions most distant and internal, while blue indicates regions most distant and external. Other colors shown in the scales identify intermediate positions. (D–L) Color-map comparisons reflecting the displacement of the indicated mean surface shapes parallel to the three orthogonal axes. Red and blue color intensities reflect displacement in the direction indicated by the corresponding color-coded arrow. Changes are shown at a scale of 1.2 standard deviations. (Right) Snout width (SW) was measured between the most lateral 3rd row of vibrissae; Median upper lip length (ULL) was measured from the lower edge of the nostrils to the bottom of the upper lip; Facial depth (FD) was measured from the middle of the ear to the top of the philtrum. Values represent the means + the S.E.M. *p<0.05 compared to control group. ∧p<0.05 compared to counterpart ethanol exposure group.
The studies described herein applied a novel integration of high-resolution MRI - magnetic resonance microscopy (MRM) - and dense surface modeling (DSM)-based shape analysis to define and compare the face/brain sequelae resulting from acute ethanol insult occurring during early gastrulation (GD7) or neurulation (GD8.5) stages of embryogenesis in mice. These stages of exposure were chosen because they have been shown to cause unique patterns of cell death in the developing forebrain
(Right) Total brain volumes were derived from automated skull stripping. Values represent the mean + S.E.M. Letters above each bar indicate group classes; the same letter above a subset of bars denotes lack of statistical difference, whereas different letters represent statistically different classes (p<0.05). (Right) For determination of disproportionate differences, the volume of each manually segmented forebrain 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 ± the S.E.M. *p<0.05 compared to control group. ∧p<0.05 compared to counterpart ethanol exposure group.
For examination of dysmorphology patterns resulting from stage-specific insult, GD17 C57Bl/6J fetuses exposed to ethanol on GD7 or GD8.5, and vehicle-exposed controls, were imaged by MRM (
The comparisons shown in
Along the lateral axis, severe midfacial narrowing is apparent in the GD7 group (
Mean shape comparisons are shown for the facial surface along with the olfactory bulbs (o), septal region (s), and pituitary (p). Comparisons parallel those shown in Fig. 2. Changes are shown at a scale of 1.5 standard deviations. In this anterior-oblique view, differences in palpebral fissure length are readily apparent.
Group differences in the shape of the upper lip midline and nose were also apparent. In the GD7 ethanol exposure group, depth reduction of the upper lip midline groove is less than that of the surrounding areas of the upper lip, indicative of a deficient philtrum (
(A–F) Images of specimens hemisected in the coronal (frontal) plane illustrate posterior and anterior aspects of the embryonic brain in control and ethanol-affected groups. (G–I) Additional GD12 embryos were cut to provide a sagittal view of the brain. Notable abnormalities include differences in width of the third ventricle (dashed calipers), and the area from which the septal region will develop (dashed outline in the anterior view and solid calipers in sagittal view). Ganglionic eminences (*).
Comparisons of mean facial surfaces also highlighted differences in palpebral fissure length between treatment groups. In the GD7 ethanol exposure group, red coloring of the area anterior-medial to both eyes (endocanthus) is indicative of markedly reduced depth compared to controls (
In addition to ethanol-induced changes in facial shape, stage of exposure-dependent alterations in brain volume and shape were examined. As shown in
Regression lines were plotted for the control group (solid line) and for the ethanol exposure groups (dashed line). In the exposed groups, a significant negative correlation between philtrum length and septal region volume, along with a significant positive correlation between snout width and olfactory bulb volume, was found. Correlations in the control group were not significant (p = 0.55, p = 0.30, respectively).
Next, DSM was used to define shape changes contributing to overall volumetric differences in selected brain regions.
Along with a control animal (A), representative examples of fetuses severely affected by ethanol exposure on GD7 (B) and GD8.5 (C) are shown. The elongated upper lip with deficient philtrum of the GD7 exposed mouse mimics that seen in children with full-blown FAS
In the GD7 exposure group, the septal region is narrow, short, and anteriorly displaced (
Treatmentgroup | LittersGenerated | Animalsscanned | Brainsegmentation | Densesurfacemodeling |
Vehicle(control) | 5 | 37 | 18 | 17 |
Ethanol –GD7 | 5 | 40 | 24 | 19 |
Ethanol –GD8.5 | 4 | 29 | 18 | 18 |
|
|
|
|
|
To investigate the embryogenesis of identified fetal brain abnormalities, scanning electron microscopy was performed on ethanol and vehicle-exposed GD12.0 embryos (
The DSM-based shape analysis described above demonstrates that developmental stage-specific ethanol exposure results in distinct facial phenotypes that correspond to unique brain abnormalities. However, comparisons of mean values do not capture how these changes correlate across the spectrum of severity present in each group. Therefore, the relationship between selected face and brain features was examined across the range of phenotypic severity. Among animals exposed to ethanol prenatally, there was a significant negative correlation between median upper lip length and septal region volume and a positive correlation between snout width and olfactory bulb volume (
The integration of advanced technologies for small animal imaging and shape analysis implemented here represents an innovative tool for defining concurrent abnormalities of the face and brain in 3D. The MRM imaging parameters utilized provided a high-throughput capability (35 min scan/animal), allowing examination of a large number of animals encompassing the phenotypic range resulting from each of two acute ethanol exposure times. Acquired 3D image sets allowed segmentation of regional brain structures as well as extraction of remarkably accurate and detailed facial surfaces from fetal mice measuring just 15 mm from the crown to rump. In addition to linear and volumetric measurements, this enabled application of DSM and assessment of 3D changes in facial surfaces and brain regions. While this technique has proven useful in modeling and delineating face shape differences across populations with various genetic syndromes, this study represents its first application in defining concurrent face-brain shape changes. These results will be readily comparable to the findings of ongoing clinical studies using DSM to define facial shape in populations with FAS, as well as those not meeting diagnostic criteria but having documented heavy prenatal ethanol exposure
The results of this study have important conceptual and practical implications for diagnostic and interventional strategies in FASD. Conceptually, they demonstrate that early prenatal ethanol exposure can cause more than one pattern of facial dysmorphology. Practically, these results should encourage new avenues for clinical FASD research by highlighting abnormalities of specific facial features and brain regions that have previously received little attention. The pattern of craniofacial abnormalities resulting from insult during early gastrulation, including microcephaly, shortened palpebral fissures and an elongated upper lip with a smooth/deficient philtrum, is consistent with that currently recognized for FAS. Median forebrain deficiencies have been well described in populations with FAS
Among the GD7 ethanol-exposed subjects, increased upper lip length was associated with hypoplasia of the olfactory bulbs and septal region. Remarkably, ethanol insult just 1.5 days later yielded a unique, and partially opposing pattern of dysmorphic features, including an abnormally shortened upper lip associated with increased septal region volume. As illustrated in
Cumulatively, the results of this study provide a proof of principle that early prenatal ethanol exposure can cause more than one temporally-dependent pattern of defects involving the face and brain and illustrate the predictive nature of the facial features for associated brain abnormalities. Numerous mechanisms for ethanol teratogenesis have been proposed. While outside the scope of the work presented here, the finding that ethanol elicits stage-specific teratogenic effects should provide new opportunities to examine these mechanisms within specific embryological contexts. Of clinical importance, these results suggest that an expansion of current diagnostic criteria to better represent the range of facial phenotypes induced by prenatal ethanol exposure could significantly advance early diagnosis and intervention strategies that are critical for the management of FASD.
All animal treatment protocols were approved by the University of North Carolina at Chapel Hill Institutional Animal Care and Use Committee. C57Bl/6J mice were purchased from The Jackson Laboratory. Timed-pregnancies were established by housing two female mice with one male mouse for a period of two hours. GD0 was defined as the beginning of the breeding period in which a copulation plug was found. Pregnant dams were administered two 25% ethanol (v/v in lactated Ringer’s solution) dosages of 2.9 g/kg by ip injection four hrs apart beginning at GD7 or GD8.5. Blood ethanol concentrations resulting from this exposure paradigm have already been described
MRM was performed on a 9.4 T vertical bore magnet interfaced to a GE console running Epic 12.4X (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×25 mm long solenoid radiofrequency coil using a radiofrequency refocused spin echo sequence (TR = 50 ms, TE = 5.2 ms, field of view = 20×10×10 mm, matrix size = 512×256×256), resulting in isotropic spatial resolution of 39 µm. A novel acquisition strategy that amplifies the high-frequency 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
Forebrain and pituitary regions were manually segmented using ITK-Snap (Version 2.1.4) as previously described
Linear facial measurements were obtained from 3D head reconstructions using netfabb Studio Basic (Version 4.7). Volumetric brain measurements were obtained with ITK-Snap.
DSM was performed as previously described
Linear brain measurements were obtained from MRM sections using ImageJ (Version 1.43;
Following MRM imaging, fetuses fixed in Bouin's solution were transferred to 70% ethanol for at least two weeks. After paraffin embedding, 10 µm sections were produced and stained with hematoxylin and eosin by standard protocols.
Multivariate analyses of variance (MANOVAs) were used to determine significant linear and volumetric group differences. Significant between-subject effects were followed by Student Newman Keuls posthoc tests when appropriate. To explore the relationship between regional brain volumes and facial linear measurements, Pearson’s correlation coefficients were utilized. An alpha value of 0.05 was maintained for all analyses.
For each treatment group, the number of litters and total fetuses produced is listed in
(GIF)
(TIF)
(TIF)
(GIF)
(GIF)
(TIF)
(TIF)
(WMV)
The authors thank Drs. Michael Charness and Arthur Aylsworth for thoughtful discussion. Magnetic Resonance Microscopy was performed at the Duke Center for In Vivo Microscopy, an NIH/NCRR/NIBIB National Biotechnology Research Center. Scanning electron microscopy was performed at the Chapel Hill Analytical and Nanofabrication Laboratory (CHANL), Institute of Advanced Materials, University of North Carolina. This study was done in conjunction with the Collaborative Initiative on Fetal Alcohol Spectrum Disorders (CIFASD). Additional information about the CIFASD can be found at