Since extraembryonic tissues supply nutrients to the developing embryo, we tested the possibility that huntingtin deficiency may perturb this function by performing RT-PCR analysis to examine the expression of a panel of 'nutritive' genes in E7.5 wild-type and Hdh^ex4/5/Hdh^ex4/5 huntingtin deficient embryos. Consistent with a previous report 16, no obvious differences were found in the expression of "nutritive" genes (Hnf4, Afp, Tfn, ApoAI, Apo-AIV, and ApoB) or genes involved in yolk sac hematopoiesis or vasculogenesis (Ttr, Rbp, Flt1, Flk1, Tal1, Rbtn2, GATA1) (data not shown), suggesting that huntingtin is not essential for the proper expression of genes required for the nutritive function of the extraembryonic tissues.

To investigate huntingtin's developmental activities, we then analyzed the expression of genes which pattern the early embryo or mark morphogenic landmarks in wild-type and Hdh^ex4/5/Hdh^ex4/5 embryos by whole mount and section in situ hybridization. The dissections confirmed previous morphologic data at E7.0–7.5 that all Hdh^ex4/5/Hdh^ex4/5 homozygotes exhibit abnormal morphology, including shortened primitive streak and a lack of morphological head folds or node 1013. The results of in situ hybridization analysis also confirmed that all three germ layers and extraembryonic tissue are formed in huntingtin deficient embryos.

Otx2, normally expressed in the anterior neuroectoderm and anterior visceral endoderm 19, is expressed in mutant embryos at E7.5 (Fig. 1A,B) although the expression domain appears reduced. Similarly, Hesx1 expression is grossly normal in mutant embryos, with expression localized to the AVE and neuroectoderm (Fig. 1C–F, 20), although the expression domain also appears reduced. These results indicate appropriate specification and movement of anterior visceral endoderm (AVE) cells from the distal tip and suggest that neuroectoderm is induced in the mutant embryos.

To examine definitive endoderm formation, the expression of Hnf3β (FoxA2) in mutant and wild-type embryos was analyzed. In wild-type embryos, Hnf3β expression is confined to the node and anterior definitive endoderm (Fig. 1G,I21). Mutant embryos exhibit Hnf3β-reactive definitive endoderm over the disorganized anterior streak region and proceeding rostrally around the distal tip (Fig. 1H,J). In both normal and mutant embryos, the AVE exhibits little Hnf3β expression. Therefore, huntingtin deficiency does not greatly affect Hnf3β regulation or the reorganization of the visceral endoderm.

The lack of a morphological node and presence of a shortened streak, together with reduced neuroectoderm and lack of headfolds, suggest that anterior streak formation may be impaired in huntingtin deficient embryos. To investigate this possibility, we examined mesoderm formation in mutant embryos. Mesoderm is specified in the mutant embryos, as marked by the expression of T (Brachyury) and Evx1 (Fig. 2A–F). However, close inspection of the data reveals abnormal patterning within this tissue and its derivatives.T, normally expressed in the primitive streak, node and axial head process/notochord mesoderm 22, is detected in the shortened streak and axial mesoderm in Hdh^ex4/5/Hdh^ex4/5 embryos, extending rostrally from a region of weakly positive cells (Fig. 2A,B). T expression appears weaker, however, in the anterior streak, corresponding to cells that will give rise to axial mesoderm (Fig. 2D). T is also ectopically expressed in mutant extraembryonic mesoderm at the anterior embryonic junction and along the chorion (Fig. 2B,D). Similarly, Evx1, normally expressed in primitive streak mesoderm at E7.5 with highest levels in proximal cells 23, is expressed in the proximal shortened streak but is also aberrantly expressed throughout the extraembryonic mesoderm, allantois and chorion (Fig. 2E,F). Extraembryonic mesoderm, derived from the proximal streak, does not normally express T or Evx1 in wild-type embryos 22. Therefore, the inappropriate expression of T and Evx1, the shortened primitive streak, and the absence of a morphological node, all suggest that the anterior primitive streak is deficient in the mutant embryos.

The anterior streak generates paraxial mesoderm. Therefore we examined paraxial mesoderm formation in wild-type and mutant embryos, revealing deficits in mesoderm patterning. Starting at E.7.5, Wnt3A is expressed in the primitive streak in cells fated to become paraxial mesoderm. In huntingtin deficient mutants, Wnt3a is induced in the proximal streak (Fig. 2G,H), confirming stage appropriate posterior development, in contrast to the absence of anterior head folds. However, expression of Wnt3a is noticeably reduced in Hdh^ex4/5/Hdh^ex4/5 embryos, suggesting a defect in paraxial mesoderm development (Fig. 2H). Reduced expression of Tbx6 in the mesoderm lateral to the primitive streak in mutant embryos confirms this interpretation (Fig. 2I,J). Furthermore, in mutant embryos at E7.5, the expression of Dll1 in the distal streak region and in only a narrow swath of cells located laterally confirms the paucity of paraxial mesoderm (Fig. 2K,L, 24). These results strongly suggest that anterior primitive streak formation is impaired, resulting in reduced axial and paraxial mesoderm formation and impaired neural development.

To elucidate the apparent patterning deficits, we next analyzed signaling molecules that are required for early patterning. Nodal, a member of the Tgfβ family of secreted molecules is required for the formation and maintenance of the primitive streak and induction of the AVE 252627. Nodal is normally expressed throughout the epiblast and overlying visceral endoderm at early post implantation stages 28, but later becomes restricted to the posterior of the embryo to the site of primitive streak with asymmetrical visceral endoderm expression marking the left-right axis. By E7.5, Nodal expression is restricted to the node. Nodal expression was assessed in Hdh^ex4/5/Hdh^ex4/5 embryos heterozygous for the Ndl lac Z allele 2829. Notably, heterozygous loss of nodal does not alter the Hdh^ex4/5/Hdh^ex4/5 phenotype, as determined by morphology of Hdh^ex4/5/Hdh^ex4/5:Ndllacz/Ndl+ embryos compared with Hdh^ex4/5/Hdh^ex4/5 embryos (data not shown). In contrast to wild-type embryos, which exhibit tight restriction of Nodal.LacZ expression to the node, Hdh^ex4/5/Hdh^ex4/5:Ndllacz/Ndl+ embryos express Nodal.LacZ throughout the endoderm overlying the epiblast, with higher levels in the posterior in an asymmetric pattern (Fig. 3A–D). The lack of tight restriction of nodal signal is consistent with a failure to form an organized node structure.

Fgf8 signaling is also essential for normal gastrulation in the mouse embryo. Fgf8 is required for cell migration away from the primitive streak 30. Expressed just prior to streak formation in the posterior epiblast and visceral endoderm, Fgf8 is restricted to the streak mesoderm at E7.5 in a decreasing proximal-distal gradient and is downregulated in cells shortly after they exit the streak (Fig. 3E,G). In Hdh^ex4/5/Hdh^ex4/5 embryos, Fgf8 expression is strongly expressed in the posterior region in the primitive streak and ectopically in the endoderm overlying the entire epiblast (Fig. 3F,H). However, streak derivatives appear to migrate normally as evidenced by the proper anterior expression of markers such as Otx2, Hnf3β and Hesx1 anteriorly (Fig. 1). Therefore, mutant embryos exhibit normal migration of streak derivatives but display impaired Fgf8 repression in mutant endoderm.

Hdh^ex4/5/Hdh^ex4/5 embryos also fail to restrict the expression of goosecoid (Gsc). Normally, Gsc is initially expressed in the visceral endoderm and proximal, posterior streak where the primitive streak will form prior to gastrulation. As the primitive streak forms and extends, Gsc is expressed in the distal streak, the node, and the axial mesoderm extending anteriorly from the node (Fig. 3I,K, 3132). However, in the mutant Hdh^ex4/5/Hdh^ex4/5 embryos, high levels of Gsc expression remain unrestricted in the endoderm overlying the entire embryo and ectopically in cells adjacent to the ectoplacental cone (Fig. 3J,L). These results suggest that, in contrast to proper Hnf3β regulation, Gsc remains inappropriately activated in mutant visceral and definitive endoderm, implicating huntingtin in the proper restriction of this homeodomain transcription factor.

Previous studies of chimeric embryos suggest that huntingtin is required only in the extraembryonic tissue for proper development 16. Signals from the extraembryonic tissue are critical for the induction of embryonic signals and for patterning the epiblast. Consequently, we examined extraembryonic development in huntingtin deficient embryos. Hnf4 is a transcription factor expressed in the primitive endoderm as soon as this tissue becomes distinct and is a key regulator of visceral endoderm secreted factors such as alphafetoprotein, apolipoproteins, and transferrin. Inactivation of Hnf4 results in impaired gastrulation 3334. At E7.5, Hnf4 is expressed in the columnar visceral endoderm cells at the extraembryonic-ectoderm junction (Fig. 4A, 33). In Hdh^ex4/5/Hdh^ex4/5 embryos, consistent with normal primitive and visceral endoderm differentiation, Hnf4 expression appears normal, although the signal is stronger in mutant embryos compared to wild-type embryos (Fig. 4B). Similarly, Pem, a transcription factor expressed in proximal visceral endoderm and ectoplacental cone in wild-type embryos at E7.5, also is expressed in these tissues in the mutant embryos (Fig. 4C,D35). However, Pem expressing visceral endoderm hangs over the anterior of the mutant embryos, revealing abnormal location despite grossly normal differentiation.

Signals from the extraembryonic tissues, including the anterior visceral endoderm and extraembryonic ectoderm are required for proper formation and patterning of the epiblast 17. Bmp4 is a signaling molecule that is first expressed uniformly throughout the extraembryonic ectoderm and subsequently is localized to a ring of extraembryonic ectoderm adjacent to the epiblast (Fig. 4E, 36). A key factor in regulating the formation of the node and primitive streak, Bmp4 is required for patterning the embryo along the proximodistal axis 37383940. In the absence of huntingtin, Bmp4 expression is properly maintained in the Hdh^ex4/5/Hdh^ex4/5 extraembryonic ectoderm but is also expressed throughout the extraembryonic ectoderm (Fig. 4F) in a pattern that is similar to early Bmp4 expression rather than being restricted to a ring of extraembryonic ectoderm as seen in the wild-type embryos To assess Bmp4 signaling from the extraembryonic ectoderm, we evaluated primordial germ cells (PGCs), which require Bmp4 for their induction 37. PGCs can first be detected at E7.0 and subsequently underlie the posterior portion of the primitive streak. Whole mount staining of E7.5 mutant and wild-type embryos for alkaline phosphatase activity reveals that PCGs form in Hdh^ex4/5/Hdh^ex4/5 embryos, suggesting that Bmp4 signaling is functional in the absence of huntingtin (Fig. 4G,H).

The anterior visceral endoderm (AVE) is also an extraembryonic source of signals that are critical for early patterning. Wnt and nodal antagonists, Dkk1 (mdkk-1) and Lefty1 respectively, are expressed in the AVE and are important in limiting the posteriorization of the anterior embryo by restricting Nodal and Wnt signaling 414243. In Hdh^ex4/5/Hdh^ex4/5 embryos, both Dkk1 (Fig. 4I,J) and Lefty1 (Fig. 4K,L) are expressed normally in the AVE as compared with wild-type embryos. However, Dkk-1 levels appear to be slightly increased in Hdh^ex4/5/Hdh^ex4/5 embryos, although the pattern of Dkk-1 expression remains unchanged and this increase may just reflect the same amount of expression in a smaller area. Therefore, the ectopic expression of Nodal (Fig. 3A–D) and the decreased Wnt3a expression (Fig. 2H) in mutant embryos do not appear to be result of changes in the expression pattern of Lefty1 or Dkk1.

Despite normal AVE formation and neuroectoderm induction, head folds do not form in Hdh^ex4/5/Hdh^ex4/5 embryos. Therefore, to determine whether mutant embryos are inherently capable of forming head folds, embryos harvested at stage E7.5 were allowed to progress in rich culture medium in vitro for 24 hours. Wild-type embryos continued to develop head folds, somites and hearts (Fig. 4M,N). In contrast, mutant stage 7.5 embryos did not develop headfolds, hearts or somites, although these embryos continued to live (Fig. 4O). These results strongly suggest that in the absence of huntingtin, embryos are unable to undergo organogenesis, even if they continue to live past E7.5 in a nutrient rich environment.

The Hdh^ex4/5 mice carrying a pGKneo insertion/replacement inactivating mutation in the mouse HD gene homologue have been described previously 10. The experiments were conducted in accordance with an IACUC approved protocol, through the MGH Subcommittee on Animal Research. Mutant Hdh^ex4/5/Hdh^ex4/5 and normal littermates were obtained in timed pregnancies from mating of Hdh^ex4/5/Hdh⁺ heterozygotes, genotyped by PCR assay, as described 10. The day of plug was taken to be E0.5. Embryos that were morphologically normal were pooled separately from morphologically mutant embryos for analysis. Nodal expression was determined in embryos from matings of Hdh^ex4/5/Hdh⁺; Ndll^acZ/Ndl⁺ compound heterozygotes genotyped by PCR assay as described in 29.

After dissection in PBS, embryos were fixed overnight in 4% paraformaldehyde at 4°C. For sections, decidua fixed in 4% paraformaldehyde, were embedded in paraffin and sectioned at 7 microns. RNA in situ hybridizations were performed as described previously 55. Nodal.lacZ expression was assessed by β-galactosidase staining as reported 29, on embryos post fixed in 4% paraformaldehyde. Embryos were mounted in 80% glycerol before being photographed.

The huntingtin deficient phenotype is fully penetrant at each of the stages that were assessed 10. Three to six embryos were evaluated for each marker, with every embryo exhibiting the same mutant phenotype in each case.

After dissections, embryos were fixed in 4% paraformaldehyde briefly and washed and stored in 1 × PBS/0.1% TX-100 at 4°C. Embryos were washed once with Tris-Maleate Buffer (25 mM Tris-Maleate, pH = 9.0, 0.8 mM MgCl₂) and were subsequently incubated in alkaline phosphatase staining solution (25 mM Tris-Maleate, pH = 9.0, 0.8 mM MgCl₂, 0.4 mg/ml alpha-naphthyl phosphate, 1 mg/ml Fast-Red). Stained embryos were washed in 1 × PBS/0.1% TX-100.

Embryos were dissected at E7.5 and washed in DMEM. Embryos were then cultured individually in 1 ml of culture media (75% immediately centrifuged rat serum and 25% DMEM 56) for 24 hours while rotating in a 37°C incubator in 5% CO₂. Embryos were then fixed in 4% paraformaldehyde for analysis.