Protocol for the Development and characterization of craniofacial mouse models

Newly discovered craniofacial deviants are evaluated for their potential value as new models using a standard set of genetic and phenotypic tests. If the craniofacial dysmorphology is heritable, amply penetrant and consistent, we establish a colony and make the new model available. The following overview describes our process and offers a model for others who wish to characterize craniofacial mutations in their own laboratories.

Deviant search

As part of The Jackson Laboratory quality assurance program for its production colonies, trained animal care technicians observe mice within an inbred strain.  When abnormal mice with overt anatomical or behavioral abnormalities are identified, these mice and their parents and siblings are sent to the Deviant Search Program. Mice are evaluated for the uniqueness of their abnormality and then offered to staff members interested in working on the new deviant.  Members of our group routinely attend Deviant Search, and the majority of new models we develop come from this program; other deviants are donated from individual research colonies.

Heritability testing and mode of inheritance

Not all apparent deviants pass their unique trait on to subsequent generations. Our first test with each new mutant is to discover whether the trait is heritable and, if so, its mode of inheritance.

To determine definitively whether the mutation follows a dominant or recessive inheritance pattern, we mate the affected mutant mouse to an unrelated wild type mouse with the same or similar inbred background.  Unrelated wild types are critical because mating to a relative may confound the results if an unaffected sibling is a carrier for the mutation.  Mating to wild-type controls from the same or a closely related inbred strain also reduces the chance of introducing modifying genes from another strain that may alter the phenotypic outcome.  We have also found that most phenotypes are more penetrant on the background in which they arose than others, indicating modifying gene(s) may interact with the causative gene.

If mutants appear in the first generation (F1), we can conclude that the new mutation is dominant or semi-dominant. If mutants appear in the second generation (F2), and not in the first, then the mutation is recessive. If the appearance of the F1 and F2 affected mice differs, the mutation is semi-dominant. As a rule of thumb, we screen 28 F1 and F2 offspring.


Many craniofacial mutants display incomplete phenotypic penetrance wherein the severity of the phenotype varies among mutants. Heritable phenotypes with low penetrance are significantly more difficult and costly to maintain and map; therefore, we typically do not pursue strain development when the phenotype penetrance is less than 20% of expected Mendelian ratios.


Many newly discovered deviant phenotypes resemble those of previously identified mutations. When this happens, a complementation or allele test is done to test for a remutation of a known gene. We may also utilize the Mouse Genome Informatics (MGI) database to search for established strains with known mutations resulting in a similar phenotype to our deviant. If an established strain is available, we order a mouse to do a complementation test. 

Genetic analysis: molecular mapping

Mutations resulting in unique phenotypes are genetically mapped to establish the chromosomal location of the causative gene. In the past, we relied on high-resolution mapping to narrow the genetic interval to a manageable size for candidate gene analysis. With the advent of sequence capture and high-throughput sequencing (HTPS) techniques, this level of resolution is no longer required. Genetic mapping to rough chromosomal position still provides a number of useful advantages however, such as a reduced computational burden, fewer variants to validate, and greater confidence in variant causality. To identify causative mutations we use a combination of genetic mapping and HTPS. We have found that mapping a gene at least to a chromosome greatly facilitates the analyses of HTPS. Our studies use a combination of 1-10 Mb interval-specific, gene-specific, and whole-exome approaches to identify a wide spectrum of mutation types across diverse genetic backgrounds.

Genetic analysis: exome capture and sequencing

When linkage is established and a broad chromosomal location identified, we employ whole-exome capture and HTPS to identify potential causative variants. Whole DNA exomes from mutant samples are captured using an in-solution, hybridization-based probe pool developed in our group in collaboration with Roche-Nimblegen. The content of the probe pool is defined by the unified mouse gene catalog, which, excluding UTR sequences, olfactory receptors and pseudogenes, encompasses approximately 50 Mb of genomic sequence. Our preliminary exome data indicate high capture sensitivity and specificity; >96.7% of the targeted bases are covered with just one lane of 75 bp paired-end on the Illumina GAIIx.

Our primary sequencing approach is to sequence whole exomes from enriched mutant DNA samples and to multiplex where possible (Fairfield, et al., 2011). An additional advantage of the paired end sequencing approach is that it provides positional information that is critical for the identification of spontaneous mutations that are due to genome rearrangements (larger insertions or deletions).

Genetic analysis: analysis and validation

All raw sequence data analysis, including read mapping and SNP/mutation calling, are performed by the Computational Science service at JAX, using Galaxy sequence analysis tools. Multiple candidate variations are detected in each strain, but most are eliminated upon validation. For validation, each candidate mutation is PCR amplified from up to 10 other individuals within the same mutant pedigree. Each PCR amplimer is subjected to Sanger sequencing. In the majority of the cases, non-mutagenic variants will not segregate with the phenotype but bona fide mutations will. Validation is not attempted until sufficient sequencing coverage has been obtained, as indicated by comparison of computational analysis of parameters like '% target bases covered' and by comparison of the variant profiles obtained to the Sanger whole genome sequencing data.

Colony maintenance

Craniofacial Resource mice are housed in 51 square inch polycarbonate boxes, on bedding composed of sterilized shavings of Northern White Pine, under 14:10 hour light/dark cycles. A diet of autoclaved NIH 31 (6% fat diet, Ca:P of 1.15:0.85, 19% protein, vitamin and mineral fortified; Purina Mills International, Richmond, Ind.) and water acidified with HCl to achieve a pH of 2.8-3.2 (which prevents bacterial growth) are freely available. Mouse colony maintenance and use is reviewed and approved by The Jackson Laboratory Institutional Animal Care and Use Committee and is in accordance with The National Institutes of Health guidelines for the care and use of animals in research.

Phenotypic analyses: hearing

We first test two mutant and controls of each sex are first gross hearing abnormalities via click box testing (RIKEN Japan Mouse Clinic, 2011).  If abnormalities are apparent additional animals are submitted for comprehensive hearing testing via auditory brainstem response (ABR) by the phenotyping group at JAX.

Phenotypic analyses: vision

Our collaborator, Dr. Bo Chang, screens mutant and control mice for eye abnormalities first by clinical examination. If an irregular eye phenotype is suspected, further screening procedures are done by slit lamp examination or ERG.


Three mutants and one control mouse of each sex are examined for anatomical lesions at 12 weeks of age by our pathologist. If a lesion is found, more mice and appropriate sectioning and staining are done to confirm the histopathological finding.  Three mutants of each sex are used to ensure that lesions are penetrant enough to warrant inclusion in the list of reported phenotypes.

Standard histology protocol

For fixation of tissues, mice are deeply anesthetized with tribromoethanol (avertin) until they no longer display a withdrawal reflex in the hind limbs and then perfused intracardially with Bouin's fixative following a flush of the vasculature with saline solution. After soaking in Bouin's for one week to demineralize bones, tissues are dissected. Six segments of spine with axial muscles and spinal cord in situ, representing cervical, thoracic and lumbar spinal segments, are dissected. The brain is removed and sliced into 6 cross sectional pieces at the levels of olfactory lobes, frontal cortex, striatum, thalamus, midbrain, rostral and caudal medulla with cerebellum. Midsagittal slices of hind leg through the knees are prepared. Slices of basal skull through the pituitary and inner ears are taken. Both eyes, salivary glands and submandibular lymph node, trachea plus thyroid and sometimes parathyroid are removed and cassetted. A longitudinal slice of skin from the back is removed. The thymus, slices of lung, and a longitudinal slice of heart are cassetted. Similarly, slices of liver through gall bladder, kidney with adrenal attached, pancreas and spleen are prepared. The stomach is sliced longitudinally to include both squamous and glandular portions. Loops of small intestine from 3 levels and slices of large intestine and cecum are removed, as are slices of urinary bladder. The whole uterus and attached ovaries is taken. In males, testes are sliced longitudinally. The accessory male organs including seminal vesicles, coagulating gland, and prostate are removed en block. Altogether in most cases, all tissues fit into a total of 10 cassettes. The cassettes are processed in an automatic tissue processor to dehydrate tissues, which are then embedded in paraffin. Six micron sections are cut and stained with hematoxylin and eosin (H&E). Sections of brain and spinal cord in vertebral bones are also stained with luxol fast blue (LFB) for myelin and cresylecht violet (CV) for cellular detail.

Skeletal morphometry and bone density

PIXImus densitometry

PIXImus scans (PIXImus, LUNAR, Madison, Wis.), which provide skeletal and body composition data such as bone mineral density (BMD, g/cm2), bone mineral content (BMC, g/cm2), body mass (g), lean mass (g), fat mass (g), and % fat mass, are completed on groups of 6 male and 6 female 12-week-old mutant and control mice. The skulls and bodies are scanned separately to provide independent data on skull BMD and BMC and body BMD and BMC. The PIXImus small animal densitometer (DEXA) has a resolution of 0.18 x 0.18 mm pixels and is equipped with software version 1.46. The PIXImus is calibrated routinely with a phantom using known values, and a quality assurance test is performed daily. The variability in precision for measuring total body BMD is less than 1% and approximately 1.5% for specialized regions such as the skull. The correlation between PIXImus BMD measurements of 614 lumbar vertebrae compared to peripheral quantitative computerized tomography (pQCT) measurements was found to be significant (p<0.001; r=.704) (Donahue, 1999).

Faxitron X-rays

We obtain x-rays at 5X magnification of the pelvis and femurs and at 3X magnification of the body of two male and female mutants and controls at 12 weeks of age using a Faxitron MX20 cabinet X-ray (Faxitron X-Ray Corp., Wheeling, Ill., USA) and Kodak Min-R 2000 mammography film (Eastman Kodak Co., Windsor, Col., USA). We then analyze the images to identify any skeletal phenotypes.

Skeletal preps

In some cases whole skeletons of mutant and control mice are cleared in 1% KOH, stained with alizarin red, stored in glycerol (Green, 1952), and then evaluated for skeletal malformations. Malformations found may indicate that the craniofacial phenotype is part of a greater syndrome.

Craniofacial morphometry: skull preparation

Skulls of 6 male and 6 female mutants and controls are collected at 12 weeks of age, prepared by incomplete maceration in potassium hydroxide, stained with alizarin red, and stored in undiluted glycerol (Green, 1952). During the collection process, right ear pinnae are measured with digital hand calipers (Stoelting, Wood Dale, IL, USA).

Craniofacial morphometry: Digital caliper skull measurements

We use seven measurements taken with hand-held digital calipers to define skull morphology.  These measures have a high degree of accuracy and precision and discriminate differences between mutant and control skull characteristics. Our linear measures have been added to the illustration below, which was taken from a publication by Dr. Joan Richtsmeier, who characterized craniofacial differences in mouse models of Down syndrome using three-dimensional anatomical landmarks (Richtsmeier, 2000).

Developmental phenotyping: Body growth and tail length

Many craniofacial models also display an overt difference in body size compared to wild-type controls.  We measure body weight (g) and tail length (cm) at two, four, six, eight and twelve weeks of age in six mutants and wild types of each sex to quantify these differences.

Developmental phenotyping: embryos

Strains with dominant craniofacial mutations are tested to determine if a more severe phenotype, possibly embryonic lethal, results when the mutation is made homozygous. Heterozygous mice are intercrossed, and litters are scored for expected number of pups and clinical appearance, and are typed for markers at the established interval. If no homozygotes are found, timed matings between F1 animals are set up, and embryos are harvested at E18.5, E14.5 and E10.5. If homozygotes are not identified, the investigation is no longer pursued. If validated homozygotes are found, they are examined for phenotypic abnormalities and the results are presented in web paper format (described below).

Strain preservation

Most strains are archived via frozen sperm; others require cryopreservation of embryos, depending on the type of mutation and the genetic background of the mutant strain.

Data analysis

Skull and PIXImus measurements, body growth, and tail length curves are analyzed; statistical differences are considered significant when p < 0.05.  We previously evaluated data using StatView 4.5 software (Abaccus Cary, NC) for Macintosh computers. Currently our group employs JMP11 software (SAS Institute Inc. Cary, N.C.), available for Macintosh and PC.


Donahue LR, Rosen CJ, Beamer WG. (1999) Comparison of Bone Mineral Content and Bone Mineral Density in C57BL/6J and C3H/HeJ Female Mice by pQCR (Stratec XCT 960M) and DEXA (PIXImus). Thirteenth International Mouse Genome Conference. Philadelphia, PA.

Fairfield H, Gilbert GJ, Barter M, Corrigan RR, Curtain M, Ding Y, D'Ascenzo M, Gerhardt DJ, He C, Huang W, Richmond T, Rowe L, Probst FJ, Bergstrom DE, Murray SA, Bult C, Richardson J, Kile BT, Gut I, Hager J, Sigurdsson S, Mauceli E, Di Palma F, Lindblad-Toh K, Cunningham ML, Cox TC, Justice MJ, Spector MS, Lowe SW, Albert T, Donahue LR, Jeddeloh J, Shendure J, Reinholdt LG. 2011. Mutation discovery in mice by whole exome sequencing. Genome Biology 12: R86.

Green, MC. (1952) A rapid method for clearing and staining specimens for the demonstration of bone. The Ohio Journal of Scince 52(1):31-33. January 1952.

Manly KF, Cudmore RH, Jr., Meer JM. (2001) Map Manager QTX, cross-platform software for genetic mapping. Mamm. Genome 12:930-932.

RIKEN Japan Mouse Clinic. (2011, 7 15). Modified-SHIRPA v4. Retrieved 5 14, 2014, from Riken BRC:

Richtsmeier JT, Baxter, LL, Reeves, RH. (2000) Parallels of craniofacial maldevelopment in Down syndrome and Ts65Dn mice. Dev. Dyn. Feb;217(2):137-45.

Truett GE, Heeger P, Mynatt RL, Truett AA, Walker JA, Warman ML. (2000) Preparation of PCR-quality mouse genomic DNA with hot sodium hydroxide and Tris (HoSHOT). Biotechniques 29;52-54.