Major malformations of the human eye, including microphthalmia and anophthalmia, are examples of phenotypes that recur in families yet often show no clear Mendelian inheritance pattern. Defining loci by mapping is therefore rarely feasible. Using a candidate-gene approach, we have identified heterozygous coding-region changes in the homeobox gene OTX2 in eight families with ocular malformations. The expression pattern of OTX2 in human embryos is consistent with the eye phenotypes observed in the patients, which range from bilateral anophthalmia to retinal defects resembling Leber congenital amaurosis and pigmentary retinopathy. Magnetic resonance imaging scans revealed defects of the optic nerve, optic chiasm, and, in some cases, brain. In two families, the mutations appear to have occurred de novo in severely affected offspring, and, in two other families, the mutations have been inherited from a gonosomal mosaic parent. Data from these four families support a simple model in which OTX2 heterozygous loss-of-function mutations cause ocular malformations. Four additional families display complex inheritance patterns, suggesting that OTX2 mutations alone may not lead to consistent phenotypes. The high incidence of mosaicism and the reduced penetrance have implications for genetic counseling.

We have isolated mammalian homologues of the Drosophila dachshund gene. Two domains of high conservation, one of which contains an α-helical, coiled-coil motif, show similarity to the Ski family of genes. We therefore propose that Dachshund belongs to a superfamily including these genes. Mouse Dachshund (Dach) is expressed in the eye and limb, structures affected by the Drosophila loss-of-function mutant, and rib primordia, CNS and genital eminence. Pax6 and Dach show overlapping but non-identical expression patterns. Dach expression is unaffected in smalleye mouse brain, indicating that Pax6 is not directly activating Dach. In Drosophila eye development dachshund is a component of an interacting network of proteins. Genes homologous to many of these exist in mammals; Dach joins this expanding group.

Mutation or deletion of the PAX6 gene underlies many cases of aniridia. Three lines of evidence now converge to implicate PAX6 more widely in anterior segment malformations including Peters' anomaly. First, a child with Peters' anomaly is deleted for one copy of PAX6. Second, affected members of a family with dominantly inherited anterior segment malformations, including Peters' anomaly are heterozygous for an R26G mutation in the PAX6 paired box. Third, a proportion of Sey/+ Smalleye mice, heterozygous for a nonsense mutation in murine Pax-6, have an ocular phenotype resembling Peters' anomaly. We therefore propose that a variety of anterior segment anomalies may be associated with PAX6 mutations.

Twenty-seven chemicals were tested for their mutagenic potential in the L5178Y tk+/tk- mouse lymphoma cell forward mutation assay using procedures based upon those described by McGregor et al. (McGregor DB, Martin R, Cattanach P, Edwards I, McBride D, Caspary WJ (1987): Environ Mol Mutagen 9:143-160). Cultures were exposed to the chemicals for 4 hr, then cultured for 2 days before plating in soft agar with or without trifluorothymidine (TFT), 3 micrograms/ml. The chemicals were tested at least twice. Statistically significant responses were obtained with acid orange 10, aniline, benzaldehyde, o-chloroaniline, chlorodibromomethane, cytembena, 1,2-dibromo-4-(1,2-dibromomethyl) cyclohexane, dieldrin, lithocholic acid, oxytetracycline, phenazopyridine HCl, 1-phenyl-3-methyl-5-pyrazolone, sodium diethyldithiocarbamate, solvent yellow 14, tetraethylthiuram disulfide (disulfiram), 2,4-toluene diisocyanate, and 2,6-toluene diisocyanate. Apart from phenazopyridine HCl, acid orange 10, and solvent yellow 14, rat liver S9 mix was not a requirement for the mutagenic activity of these compounds. Chemical not identified as mutagens were N-4-acetylaminofluorene, chlorpheniramine maleate, chloropropamide, 1,4-dioxane, endrin, ethylene glycol, iron dextran, methapyrilene, sodium(2-ethylhexyl)alcohol

A new protocol for testing vapors and gases in the L5178Y mouse lymphoma assay is presented. Four chemicals, propylene, 1,2‐propylene oxide, 1,3‐butadiene, and vinylidene chloride, were tested for their mutagenic potential. Cultures were exposed to the chemicals, which were delivered as vapors or gases, for 4 hr, then cultured for 2 days before plating in soft agar with or without trifluorothymidine (TFT), 3 μg/ml. Each chemical was tested at least twice. Significant responses were obtained with 1,2‐propylene oxide and vinylidene chloride, but neither cytotoxicity nor mutagenicity was induced by 1,3‐butadiene; propylene could not be classified as either mutagenic or non‐mutagenic in the assay. Rat liver S9 mix was not a requirement for the mutagenic activity of 1,2‐propylene oxide, whereas the liver preparation markedly enhanced both the cytotoxicity and mutagenicity of vinylidene chloride.

Forty-one chemicals were tested for their abilities to induce trifluorothymidine resistance in L5178Y mouse lymphoma (MOLY) cells. These chemicals were included in the National Toxicology Program's evaluation of four in vitro short-term toxicity assays for predicting carcinogenicity in the rodent bioassay. Of the 41 chemicals examined for this report, 8 were equivocal in the rodent bioassay, and 7 were questionable in- the MOLY assay. If these chemicals are eliminated from an analysis of concordance, the remaining 26 chemicals lead to a concordance of 69% with a sensitivity of 71%. The specificity could not be determined because only two non-carcinogens were detected.