Influence of epithelial-mesenchymal interaction on the viability of facial mesenchyme in vitro

Endogenous bone morphogenetic proteins regulate outgrowth and epithelial survival during avian lip fusion

Our expression studies of bone morphogenetic proteins (BMPs) and Noggin (a BMP antagonist) in the embryonic chicken face suggested that BMP signals were important for closure of the upper lip or primary palate. We noted that Noggin expression was restricted to the frontonasal mass epithelium but was reduced at the corners of the frontonasal mass (globular processes) just prior to fusion with the adjacent maxillary prominences. We therefore performed gain- and loss-of-function experiments to determine the role of BMPs in lip formation. Noggin treatment led to reduced proliferation and outgrowth of the frontonasal mass and maxillary prominences and ultimately to the deletion of the maxillary and palatine bones. The temporary block in BMP signalling in the mesenchyme also promoted epithelial survival. Noggin treatment also upregulated expression of endogenous BMPs, therefore we investigated whether increasing BMP levels would lead to the same phenotype. A BMP2 bead was implanted into the globular process and a similar phenotype to that produced by Noggin resulted. However, instead of a decrease in proliferation, defects were caused by increased programmed cell death, first in the epithelium and then in the mesenchyme. Programmed cell death was induced primarily in the lateral frontonasal mass with very little cell death medial to the bead. The asymmetric cell death pattern was correlated with a rapid induction of Noggin in the same embryos, with transcripts complementary to the regions with increased cell death. We have demonstrated a requirement for endogenous BMP in the proliferation of facial mesenchyme and that mesenchymal signals promote either survival or thinning of the epithelium. We furthermore demonstrated in vivo that BMP homeostasis is regulated by increasing expression of ligand or antagonist and that such mechanisms may help to protect the embryo from changes in growth factor levels during development or after exposure to teratogens.

Effect of fibroblast growth factors on outgrowth of facial mesenchyme

The ectoderm is required for outgrowth of facial prominences and facial ectoderm from all facial prominences is interchangeable. Signals provided by the ectoderm may include members of the fibroblast growth factor family (FGF). In order to test whether FGFs could replace facial ectoderm and promote outgrowth, stage 24 frontonasal mass or mandibular mesenchyme was grafted to a host chick limb and a bead soaked in FGF-2 or FGF-4 was placed on top of the mesenchyme. Following 7 days of incubation, the amount of outgrowth was quantified by measuring the rods of cartilage that formed from the grafts. FGF-2 and FGF-4 stimulated an increase in length of cartilage rods in mandibular grafts compared to mandibular mesenchyme grafted without ectoderm (P < 0.05). FGF-4 stimulated a small increase in length of frontonasal mass mesenchyme (P < 0.05) and both FGFs increased the frequency of egg tooth formation in frontonasal mass mesenchyme compared to frontonasal mass mesenchyme grafted without ectoderm. FGFs can partially but not completely replace facial ectoderm since homotypic recombinations of frontonasal mass and mandibular tissues were significantly longer than mesenchyme grafts treated with FGF-soaked beads (P < 0.05). The addition of a second FGF-soaked bead did not significantly increase the length of the frontonasal mass or the mandibular mesenchyme. We have determined that FGF-2 protein is expressed in facial ectoderm and could be an endogenous signal for outgrowth. In contrast, FGF-8 transcripts are not expressed in the ectoderm covering the areas of the face that were grafted; thus, it is less likely that FGF-8 is required for outgrowth. Our results indicate that FGFs are part of an endogenous signaling pathway involved in distal outgrowth and chondrogenesis of the facial prominences.

Chicken transcription factor AP-2: cloning, expression and its role in outgrowth of facial prominences and limb buds

Embryonic facial development in chick embryos involves a sequential activation of genes that control differential growth and patterning of the beak. In the present study we isolate one such gene, the transcription factor, AP-2, that is known to be expressed in the face of mouse embryos. The protein sequence of chick AP-2alpha is 94% homologous to human and mouse AP-2. Wholemount in situ hybridization with a probe for chick AP-2 identifies expression from primitive streak stages up to stage 28. The most striking expression patterns in the head are during neural crest cell migration when AP-2 transcripts follow closely the tracts previously mapped for neural crest cells. Later, expression in the facial mesenchyme is strongest in the frontonasal mass and lateral nasal prominences and is downregulated in the maxillary and mandibular prominences. Once limb buds are visible, high expression is seen in the distal mesenchyme but not in the apical ectodermal ridge. The expression patterns of AP-2 in stage 20 embryos suggested that the gene may be important in "budding out" of facial prominences and limb buds. We implanted beads soaked in retinoic acid in the right nasal pit of stage 20 embryos resulting in a specific inhibition of outgrowth of the frontonasal mass and lateral nasal prominences. AP-2 expression was completely down-regulated in the lateral nasal within 8 hr of bead application. In addition, the normal up-regulation of AP-2 in the frontonasal mass did not occur following retinoic-acid treatment. There was an increase in programmed cell death around the right nasal pit that accompanied the down-regulation of AP-2. Prominences whose morphogenesis were not affected by retinoic acid did not have altered expression patterns. We removed the apical ectodermal ridge in stage 20 limb buds and found that AP-2 expression was partially downregulated 4 hr following ridge removal and completely downregulated 8 hr following stripping. Application of an FGF-4 soaked bead to the apex of the limb bud maintained AP-2 expression. Thus AP-2 is involved in outgrowth and could be regulated by factors such as FGFs that are present in the ectoderm of both the face and limb.

Distribution of p21ras during primary palate formation of non-cleft and cleft strains of mice

Cleft lip, with or without cleft palate, is one of the most common defects in craniofacial formation. The primary palatogenesis of mice is similar to that of humans and spontaneous cleft lip is associated with genotype in both mice and humans. To investigate the temporal and spatial expression of ras genes in cleft (A/WySn) and non-cleft strains of mice (BALB/cBy), a broad spectrum ras antibody was used. Positive staining was found in ectodermal, mesenchymal, and neuroepithelial cells of facial prominences before the primary palate formation stage (10 d 20 hr) in both strains. During the primary palate formation stage (11 d 20 hr), positive staining was found in the ectodermal and mesenchymal cells of the facial prominences of the non-cleft strain but not in those of the cleft strain. These results suggest ras genes may play a role in the primary palatogenesis of mice. Cleft lip could be associated with the deficiency of ras gene expression during primary palate formation of mice.

Enhancement of avian mandibular chondrogenesis in vitro in the absence of epithelium

The roles of mandibular epithelium in chondrogenesis and growth of mandibular mesenchyme were examined in organ cultures. Epithelium and mesenchyme were separated from the mandibular arches of chick embryos at stages before and after the onset of chondrogenesis in vivo (stages 18-28). Isochronic and heterochronic tissue recombinations were prepared. Removal of the mandibular epithelium resulted in reduced growth of the explants and enhanced chondrogenesis, resulting in increased levels of mRNAs for type II collagen and aggrecan. The presence of mandibular epithelium promoted cell division in loosely arranged undifferentiated tissue from the mandibular mesenchyme and resulted in increased levels of type I collagen mRNA. Enhanced chondrogenesis was also observed in the mesenchyme isolated with basement membrane and isolated mesenchyme grown within Matrigel. These findings suggest that mandibular epithelium has mitogenic and chondrogenic-inhibitory effects on the underlying mesenchyme that are stage independent. Furthermore, the chondrogenic-inhibitory effect of mandibular epithelium on the underlying mesenchymal cells is not mediated by basement membrane.

Modulation of gap junction-mediated intercellular communication in embryonic chick mesenchyme during tissue remodeling in vitro

Gap junction-mediated intercellular communication was analyzed in a model system in which tissue necrosis and remodeling could be modulated. This in vitro system, previously used for analysis of epithelial-mesenchymal tissue interaction, was modified to permit analysis of the presence and extent of intercellular communication by monitoring intercellular transfer of the microinjected fluorescent dye, Lucifer Yellow. Light and transmission electronmicroscopy were employed to correlate the presence and degree of gap junctional communication (coupling) with tissue morphology. Digital image analysis was used to determine cell density and mitotic indices within the outgrowths of explants. Our results indicated that cell communication in outgrowths adjacent to necrotic foci within an explant was minimal or absent. Cell-coupling in outgrowths adjacent to a compartment of viable mesenchyme was significantly higher - equivalent to unseparated control cultures. A time-course study demonstrated correlation of increased levels of cell-coupling in outgrowths with the level of tissue remodeling within an explant. Our conclusions from these studies are that embryonic mesenchymal cell populations may be selectively uncoupled as a result of alterations in the microenvironment produced by a proximate impaired cell population. It is proposed that endogenous factors in the microenvironment ("wound signals"), emanating from impaired cell populations, regulate gap junction-mediated intercellular communication in adjacent viable tissue. Normal, unimpaired populations of cells surrounding an area of injury are thereby isolated from the effects of a potentially toxic environment. This could serve as a protective function in development and may represent, in a more general sense, part of the repertoire of events associated with tissue repair and remodeling.

Serotonin and morphogenesis. Transient expression of serotonin uptake and binding protein during craniofacial morphogenesis in the mouse

This study describes the timecourse of expression of low-affinity serotonin uptake sites in the developing craniofacial region of the mouse embryo. Whole mouse embryos were incubated in the presence of various serotonergic compounds followed by immunocytochemical localization of serotonin (5-HT) and its binding protein. In the gestational day 9 embryo (3-5 somites), 5-HT uptake was observed in the myocardium of the heart, the visceral yolk sac and foregut. A specific and transient pattern of 5-HT uptake was observed in the hindbrain neuroepithelium from day 9.5-11, where it was localized in rhombomeres 2-5 in the day 9.5 embryo. By day 10, when rhombomeres were no longer evident, uptake was present in the dorso-lateral neuroepithelium surrounding the fourth ventricle (rhombic lip; cerebellar anlage). Uptake of 5-HT was initially observed in the surface epithelium of the craniofacial region at day 10 (20-25 somites) and was greatly increased at day 11. The invaginating lens, nasal placode epithelium and otocyst also took up 5-HT at day 11. During these stages a 45 kD serotonin-binding protein (SBP) was expressed in craniofacial mesenchyme, and became progressively restricted to regions subjacent to epithelial uptake sites. These staining patterns were shown to be specific for 5-HT and SBP by their absence in embryos stained using preabsorbed antisera. The timecourse of these patterns are correlated with critical events in craniofacial morphogenesis including (1) onset of inductive epithelial-mesenchymal interactions, (2) invagination and fusion of placodal structures, (3) presence of rhombomeres, and (4) regions of low proliferative activity.

Serotonin as a regulator of craniofacial morphogenesis: site specific malformations following exposure to serotonin uptake inhibitors

During craniofacial development in the mouse embryo (days 9-12 of gestation; plug day = day 1), transient expression of serotonin (5-HT) uptake in epithelial structures of this region correlates with critical morphogenetic events (Lauder et al., '88; Shuey, '91; Shuey et al., '89, '92). The purpose of the present investigation was to assess the possible functional significance of these uptake sites by examination of patterns of dysmorphology following exposure of embryos to selective 5-HT uptake inhibitors. Exposure of mouse embryos in whole embryo culture to sertraline, at a concentration (10 microM) which produced no evidence of general embryotoxicity, caused craniofacial malformations consistent with direct action at 5-HT uptake sites. Two other 5-HT uptake inhibitors, fluoxetine and amitriptyline, produced similar defects. The critical period of sertraline exposure occurred on days 10-11. The observed craniofacial defects were associated with decreased proliferation and extensive cell death in mesenchyme located 5-6 cell layers deep from the overlying epithelium. In contrast, the subepithelial mesenchymal layers showed normal or elevated levels of proliferation. From these results it appears that inhibition of 5-HT uptake into craniofacial epithelia may produce developmental defects by interference with serotonergic regulation of epithelial-mesenchymal interactions important for normal craniofacial morphogenesis.

Insulin-like growth factor I and II and insulin-like growth factor binding protein-2 RNAs are expressed in adjacent tissues within rat embryonic and fetal limbs

Since the rapid proliferation of cells in a directed manner is a necessary component of limb formation, the distribution of locally produced mitogenic molecules within the developing limb is of considerable interest. We have used in situ hybridization to localize transcripts for both insulin-like growth factor binding protein-2 (IGFBP-2) and its ligands, the insulin-like growth factors I and II (IGF-I and IGF-II), within limb buds of rat embryos 10-16 days after conception (equivalent to stages 1-12 of mouse limb morphogenesis, Wanek et al, 1989. J. Exp. Zool. 249, 41-49). The mRNA for IGFBP-2 is very abundant in an anterior-posterior strip of ectoderm along the distal edge of the limb bud (the progenitor of the apical ectodermal ridge or AER) from as early as limb stage 1 (Embryonic Day 10) and is much less abundant in the rest of the limb ectoderm. A high level of IGFBP-2 expression continues to characterize the AER following its definitive appearance (stage 3) and throughout its existence (until stage 7). This is a period of rapid outgrowth during which the rate of mesodermal cell division is highest in cells nearest to the AER. The AER is known to have mitogenic activity in vitro and to direct limb outgrowth in vivo, but, until recently, few putative molecular correlates of these activities have been detected. The transcripts for IGF-I and IGF-II are also present at high abundance in developing limbs, especially in mesodermally derived cells. IGF-I mRNA is abundant in presumptive limb mesoderm from the beginning of limb outgrowth (just before stage 1), but is very low or undetectable in much of the rest of the embryo, while IGF-II mRNA becomes very abundant in limb mesoderm at stage 2. The distribution in limbs of both IGF-I and IGF-II mRNA changes dramatically during outgrowth and differentiation, so that their expression characterizes complementary populations of cells by stage 11. Taken together, these data suggest that IGFs and the IGF binding proteins, which may modulate IGF action, contribute to limb outgrowth and patterning.

Influence of epithelial-mesenchymal interaction on the viability of facial mesenchyme. II: Synthesis of basement-membrane components during tissue recombination

The presence of basement-membrane components during tissue separation procedures was determined employing monoclonal antibodies to laminin and type IV collagen. In addition, the reconstitution of basement-membrane components and the formation of the basement-membrane were examined in isolated epithelium and mesenchyme and in tissue recombination. Epithelium and mesenchyme of maxillary processes of chick embryos were separated by a variety of protocols, including those employed in a prior study (Saber et al: Anat. Rec. 225:56-66, 1989). Results indicated that the protocol previously employed did not remove basement-membrane components after enzymatic tissue separation. A revised protocol in which the basement-membrane components (i.e., laminin and type IV collagen) were removed from isolated tissues prior to recombination revealed that a developmental compartment and a gradient of cell viability, comparable in size and dimensions to that observed in the study of Saber et al. (ibid.) was present in the mesenchyme of recombined explants. Type IV collagen and laminin, therefore, do not appear to be required initially during tissue recombination in order for subsequent growth-sustaining effects to be expressed. Additional studies revealed, however, that synthesis of basement-membrane components occurred not only in isolated tissues but was altered markedly by tissue recombination. Culture of isolated tissues demonstrated induction of laminin synthesis in separated epithelium by 24 hours and induction of collagen synthesis in isolated mesenchyme by 24 hours. Recombination of epithelium and mesenchyme, however, resulted in rapid induction of laminin synthesis within 1 hour. Recombination of epithelium and mesenchyme after 24 hours resulted in the presence of laminin not only in epithelium but in mesenchyme as well. Both tissues were required for basement-membrane formation which appeared to be fully reconstituted by 24 hours in culture. These observations indicate that recombination in culture alters the pattern of synthetic activity of these basement-membrane components. These can be characterized as "early" (temporal) and "late" spatial) responses by the recombined tissues.

Epithelial growth by rat vibrissae follicles in vitro requires mesenchymal contact via native extracellular matrix

An in vitro assay utilizing the rat vibrissa anagen follicle as a model for studying the epithelial-mesenchymal interactions (EMI) in hair growth is described. Through selective disruption of the epithelial-mesenchymal interface, we investigate whether the specialized extracellular matrix (ECM) of the dermal papilla and basement membrane zone (BMZ) serves a crucial function in hair follicle EMI. Epithelial bulbs incubated intact within their follicular sheaths incorporate thymidine primarily into cells of the hair matrix and outer root sheath, as shown by autoradiography. However, after removal of its mesenchymal associations (dermal papilla and extrabulbar connective tissue), the epithelial bulb showed no incorporation. Neither externally added collagen (type I or IV) nor the basement membrane components in Matrigel could substitute for the growth supporting influence of native surrounding stroma. Mechanical separation of the bulb from the dermal papilla in the basement membrane zone inhibited thymidine incorporation by the epithelium even though mesenchyme was still in close proximity. Enzymatic digestion of the dermal papilla ECM and the basal lamina by Dispase, a fibronectinase and type IV collagenase, also inhibited bulb growth without evidence of cytotoxicity. These experiments suggest that direct epithelial to mesenchymal contact is required for the support of follicular epithelial growth in vitro and that specific ECM components, possibly fibronectin and/or type IV collagen, rather than diffusable factors alone, play a crucial role in the mechanism of hair follicle EMI. The in vitro system described here provides an alternative to developmental EMI models and may serve as a valuable tool for studying EMI in the adult mammalian organism.