Histochemistry of the developing notochord, perichordal sheath and vertebrae in Danforth’s short-tail (Sd) and normal C57BL/6 mice

Many studies have focused on the tissue interactions involved in chondrogenesis in an attempt to clarify the mechanisms underlying this process. The role of the notochord and neural tube in vertebral formation has been of particular interest, and the interactions between these structures and somatic mesenchyme during chondrogenesis have been examined extensively both in vivo and in vitro (see reviews by Holtzer, 1968; Lash, 1968a, b).

Danforth’s short-tail (Sd) mouse, which shows defective vertebral development in heterozygotes (Sd/ + ) and in homozygotes (Sd/Sd), provides a system in which these interactions can be examined. This mutant, first described by Dunn, Gluecksohn-Schoenheimer & Bryson (1940), exhibits a syndrome of defects, including anomalies of the axial skeleton, and of the urogenital and digestive systems (Gluecksohn-Schoenheimer, 1943). The disturbances of the axial skeleton include an abnormal articulation between the atlas and axis, reduction or absence of nuclei pulposi, smaller (split or indented) vertebrae, and a paucity of tail vertebrae (Theiler, 1951a, b, 1954; Grüneburg, 1953; Dürr, 1958). The last defect leads to the most conspicuous feature, a shortening and kinking of the tail.

The Sd mutant mouse is particularly interesting because of the controversy concerning the relationship of the notochord to the constellation of defects observed in this animal. Grüneburg (1958) suggested that the notochord is directly responsible for this syndrome, whereas Gluecksohn-Schoenheimer (1945) indicated that notochordal involvement occurs later and is secondary to degenerative changes in the mesenchyme of the tail. These uncertainties point to the need for further examination of the Sd mutant mouse during embryonic development.

Considerable evidence has accumulated indicating that extracellular matrix (ECM) materials (e.g. glycosaminoglycans such as chondroitin sulfates, and glycoproteins such as collagen) mediate vertebral chondrogenesis (Minor, 1973; Kosher & Lash, 1975). The presence of these substances or their precursors can be detected by various histochemical stains, including the periodic-acid Schiff (PAS) reagent and alcian blue. A positive PAS reaction that persists subsequent to diastase digestion suggests the presence of glycoproteins (LeBlond, Glegg & Eidinger, 1957), neutral polysaccharides (Kvist & Finnegan, 1970), or other substances containing neutral sugar residues. Neutral polysaccharides have been reported to be components of adult (Dische, Danilczenko & Zelmenis, 1958) and embryonic (Kvist & Finnegan, 1970) cartilage. In addition, the linkage regions connecting the precursors of chondroitin sulfates to core proteins contain neutral sugars (Rodén, 1970) that are potentially amenable to periodate oxidation. A positive alcian blue reaction is generally considered to detect glycosaminoglycans (acid mucopolysaccharides) (Yamada, 1963; Pearse, 1968). Thus, these stains are capable of detecting various components of the ECM or their precursors and were used as the main probes in this investigation.

The purpose of this study therefore has been to: (1) analyze and compare the histochemical staining properties of the normal and abnormal notochord, and (2) to ascertain the stage at which notochordal development in the abnormal embryo deviates from that in the normal embryo. Finally, particular attention will be focused on those aspects of vertebral development in the Sd mutant mouse which yield information on the role of the notochord in normal vertebral development.

The mutant embryos used in this study were obtained from matings of heterozygous (Sd/ + × Sd/ + ) adult mice. Litters from C57BL/6Sfd adult mice were used as controls to establish a morphological baseline. The embryos were timed by the vaginal plug method (vaginal plug = day zero) and staged according to Griineburg’s criteria (1943).

Embryos ranging in age from 9 to 14 days’ gestation were used in this study. After removal from the uterus, the embryos were placed in physiological saline and dissected free from the fetal membranes. The embryos were routinely fixed in either Bouin’s or Carnoy’s solution. The fixed specimens were dehydrated in graded ethanols, cleared in xylene or cedarwood oil and embedded in paraplast. Dur ing the final phases of embedding, the embryos were carefully oriented so that transverse or frontal sections could be obtained. Serial sections of 9- and 10-day embryos were cut at 7–8 μm and at 10 μm for 11–14 days’ gestation. The serial sections were then stained with alcian blue (Pearse, 1968), Mallory–Heidenhain stain for connective tissue (aniline blue), or periodic-acid Schiff reagent (Pearse, 1968) with or without a 20 min predigestion with saliva or 0-5% diastase to remove glycogen (Culling, 1963).

A total of 34 litters was obtained and 250 embryos were dissected out; only embryos of 12 days’ gestation and older could be identified as Sd/ + or Sd/Sd by external inspection alone. Microscopic examination was necessary to confirm the genotypes of younger embryos.

The distribution in 112 embryos from Sd/+ × Sd/+ matings was 24 +/ +, 68 Sd/ +, and 20 Sd/Sd. The expected ratio would be 28 +/ +, 56 Sd/ +, and 28 Sd/Sd. A similar deviation toward an excessive number of heterozygotes was observed by Dunn et al. (1940).

In the present study, most of the PAS-positive granules described in notochordal cells and many of those in mesenchymal cells, chondroblasts and chondrocytes (as well as the notochordal sheath) remain PAS-positive after diastase digestion in normal and abnormal embryos at all stages examined. No differences in notochordal or vertebral development were observed between C57BL/6Sfd embryos and the + / + embryos from the Sd line. Therefore, these two groups will be referred to collectively as ‘normal embryos’. The differences between Sd/ + and Sd/Sd embryos are a matter of degree; thus, these two categories will be grouped together as ‘abnormal embryos’. Most of the normal as well as the abnormal changes described below are first observed in the cervical or upper thoracic regions of the embryo, and later in more caudal and cranial regions. The development of the tail notochord parallels that of the trunk notochord, but lags behind it in time.

It should be noted that at 9 days’ gestation abnormal embryos could be distinguished from normal embryos on the basis of notochordal morphology and histochemistry. The remaining structures appeared normal, at least under the conditions of this study.

At 9 days’ gestation the notochord is a continuous rod of constant calibre extending from the hypophysis to almost the tip of the tail. By 12 days’ gestation the notochord in cranial regions has alternating narrower and wider segments while in caudal areas it still displays a constant calibre (Fig. 2a). This segmentation becomes more obvious in the 13-day embryo (Fig. 2c) and is exaggerated by 14 days (Fig. 2e).

The cytoplasm of notochordal cells displays a faint, diffuse PAS-positive reaction, which is more obvious in cranial regions of the embryos, and contains discrete PAS-positive granules (Fig. 1a). These PAS-positive granules gradually accumulate in numbers over days 10–11 (Fig. 3a, c) and by day 12 they are numerous (Fig. 4b, d). In 13-day embryos, the cells in intervertebral segments of the notochord are moderately vacuolated while those in vertebral regions remain non-vacuolated. At this time most notochordal cells are filled with PAS-positive non-granules (Fig. 7a). By 14 days, few notochordal cells remain in vertebral anlagen (Fig. 8a); however, vacuolation of notochordal cells is pronounced in intervertebral areas, and they display abundant PAS-positive granules. The PAS staining of these granules persists subsequent to treatment with diastase.

The notochord of abnormal embryos is a discontinuous structure even at 9 days of gestation (Fig. 1d), at which time it may be absent over distances up to 70 μm. In some instances, presumptive notochordal cells do not yet form a rodlike structure but are closely applied to the neural tube (Fig. 1c). As development continues, the notochord becomes increasingly fragmented. The notochordal remnants vary in size and shape, are often diffuse (Figs. 2d, 5b) and are always smaller in calibre and more irregular in outline than the normal notochord of a comparable age (Figs. 1b, 3d, 7b). In abnormals, notochordal fragments detach from the neural tube, but at any given stage less distance separates these two structures than in normal embryos. Even at later stages they remain abnormally close to the neural tube. On days 12–14, none of the notochordal fragments in abnormal embryos show signs of impending dilatation (Fig. 2b, d). In these older embryos, the remaining notochordal pieces often occupy an eccentric site in or lie entirely outside the anlagen of the vertebrae or disc (Figs. 6b, 1b). By 14 days, most sections of abnormal embryos lack notochordal fragments, and the few that remain generally occur in the sacral region.

In 9-day abnormal embryos notochordal cells display a faint PAS-positive reaction. However, most notochordal cells exhibit only a few PAS-positive granules and some lack granules altogether (Fig. 1b). By 10 days, many notochordal cells are devoid of PAS-reactive granules, and such granules are absent from virtually all notochordal cells of 11–14-day abnormal embryos (Figs. 2b, d, 3d, 6b, 1b). In addition, vacuolation does not occur in the notochordal cells remaining in intervertebral regions of day-13 (Figs. 2d, lb) or day-14 abnormal embryos.

A clear-cut notochordal sheath is obvious in 10-day embryos (Fig. 3 a), and is strongly PAS-positive by day 12 (Fig. 4c, d). In 13- and 14-day embryos, the sheath enclosing vertebral notochord is thick, while that around the notochord in intervertebral regions is attenuated. A sheath remains in all vertebral centra of 14-day embryos, attached to the surrounding cartilage, whether or not notochordal cells are present (Fig. 8a). In addition to being PAS-reactive, the sheath stains with alcian blue and aniline blue at all stages examined.

In 10-day abnormal embryos, the notochordal fragments are encircled by a thin, PAS-positive sheath (Fig. 3 b), which appears more or less normal. In sections where the notochord is absent, no sheath is seen. The notochordal sheath of abnormals does not undergo an increase in thickness as development continues (Figs. 3d, 5b). It is consistently much thinner than that observed in normal embryos of comparable ages. By 11 days, some notochordal fragments lack a sheath, and in 12- to 14-day abnormal embryos the remaining notochordal fragments commonly lack a sheath (Fig. 7b). When present, the sheath stains positively with the PAS-reagent, aniline blue and alcian blue.

In 9-day embryos the notochord is surrounded by sparse, unorganized mesenchyme, whose constituent cells are PAS-negative. At 10 days the mesenchyme around the notochord varies from loose to moderately dense, and lacks any specific pattern of organization (Fig. 3a). By 11 days a definite alignment of mesenchymal cells with respect to the notochord is obvious, with two or more rows of cells forming concentric rings around it (Fig. 3c), and PAS-positive granules occur in the cytoplasm of the mesenchymal cells closest to the notochord (Fig. 3c).

Further organization of mesenchymal cells is observed in embryos at 12 days’ gestation, at which time presumptive vertebral and intervertebral disc areas are readily recognized (Fig. 4). In regions where vertebrae will develop, the cells comprising the future centrum are separated from each other by considerable ground substance and contain PAS-positive granules (Fig. 4a, b), whereas the cells of the developing arches are densely packed, with little ground substance between them, and lack PAS-positive granules (Fig. 4a). By 13 days the vertebral models are enlarged and chondrified; their ground substance is moderately PAS-reactive, and more chondrocytes in the centrum contain PAS-positive granules (Fig. 6a). By 14 days, the ground substance is more intensely PAS-reactive, and nearly all chondrocytes are hypertrophied and contain abundant PAS-positive granules (Fig. 8a).

Intervertebral disc anlagen of 12-day embryos consist of the notochord surrounded by closely spaced rows of mesenchymal cells. PAS-positive granules are abundant in the cells of the concentric rows, but absent from the more peripheral cells (Fig. 4b, c). By 13 days the disc anlagen has enlarged and in contrast to the previous stage, most of its cells are devoid of PAS-stainable granules (Fig. 7a). Their cytoplasm, however, remains moderately PAS-positive. The cellular organization and staining of the disc anlagen at 14 days is similar to that at 13 days’ gestation.

The mesenchyme surrounding the notochords of 9- and 10-day abnormal embryos (Figs. 1b, 3b) resembles that described for normal specimens of these ages. When the notochord is absent, mesenchymal cells fill the site that this structure would normally occupy (Fig. 1 d). On day 11, these cells fail to be aligned in rows with respect to the notochord; instead, they remain randomly arranged (Fig. 3 d). In marked contrast to the normal situation, far fewer mesenchymal cells near the notochord contain PAS-positive granules.

As development proceeds the lack of organization in these mesenchymal derivatives persists. In the developing centra of 12-day embryos: (1) the cells nearest the notochord fail to tightly encircle it (Fig. 5 b), (2) the more peripheral cells of the precartilage model lack an ordered arrangement (Fig. 5a, b), (3) far less ground substance separates the chondroblasts, and fewer chondrocytes contain PAS-stained granules than in normal specimens (Fig. 5 b), and (4) fewer cells appear to comprise the anlagen. Often, when the notochord is absent the mesenchymal cells of the centrum assume a characteristic configuration, two side-by-side whorls (Fig. 5 c, d).

By 13 days, the vertebral model is chondrified, but is always much reduced in the dorsoventral dimension (Fig. 6 b), and its ground substance appears to be less PAS-positive than that in normal specimens. Of the chondrocytes in the centrum fewer have begun to hypertrophy and fewer contain PAS-positive granules than in normal embryos. In 14-day embryos, the centra are generally partially or completely bifurcated (Fig. 8b). The ground substance of these models is PAS-positive but the reaction is variable even within a given specimen, and few chondrocytes have hypertrophied.

Intervertebral disc anlagen also show disturbances of mesenchymal cell organization. The cells of the disc may form concentric rows but the rows are not precisely arranged (Fig. 5e,f), or they may be randomly placed, lacking any indication of order (Fig. 1b), or they may be arranged in side-by-side whorls. All arrangements are seen in any given specimen and no one type predominates over the others. In any of these situations, a notochordal fragment may or may not be present. As in vertebral regions fewer cells appear to form the disc primordia and fewer contain PAS-positive granules than in normal embryos.

The results presented here demonstrate that the histochemical properties of notochordal development in Sd mutant mice are clearly anomalous as early as the ninth day of gestation. At this stage, most notochordal cells of abnormal embryos are deficient in PAS-positive granules, whereas such granules are typically present in the notochords of normal specimens. Moreover, as embryogenesis continues, the notochordal sheath of Sd mutants fails to develop normally.

One of the striking findings of the present study, revealed by the use of the periodic-acid Schiff reagent, was the presence of discrete PAS-positive, diastaseresistant granules in the cytoplasm of notochordal cells. They are apparent in notochordal cells of normal mouse embryos at 9 days and subsequently increase in number until they become abundant. In earlier studies that employed the PAS reagent, similar PAS-stained granules were not described in notochordal cells (Leeson & Leeson, 1958; Leeson & Threadgold, 1960; Kvist & Finnegan, 1970).

The persistence of PAS-stainable material after diastase digestion is generally regarded as implying the presence of glycoproteins (LeBlond et al. 1957) or neutral polysaccharides (Kvist & Finnegan, 1970). Glycosaminoglycans such as chondroitin sulfates theoretically can react with the PAS reagent (Pearse, 1968) but such a reaction is unlikely when a routine PAS procedure is used (LeBlond et al. 1957; Zugibe, 1963), since more rigorous oxidation conditions are required to demonstrate these compounds (Scott & Dorling, 1969). The PAS-positive, diastase-resistant granules described in the current report accumulated in substantial numbers only in notochordal cells and those cells destined to become chondrocytes, making it tempting to speculate that they are involved in chondrogenesis.

Recently, considerable biochemical (Minor, 1973; Hay & Meier, 1974; Kosher & Lash, 1975) and morphological (Lauscher & Carlson, 1975; Kenney & Carlson, 1978) evidence has accumulated, strongly indicating that notochordal cells produce extracellular matrix (ECM) materials, glycosaminoglycans (chondroitin-4, and 6, sulfate, and heparan sulfate) and collagen. Moreover, these substances are produced at the time of somite chondrogenesis, and have been implicated as mediators of the interactions between the notochord and somitic mesoderm that bring about cartilage formation (Kosher & Lash, 1975; Hay & Meier, 1974). ECM components initially accumulate in the perichordal sheath, then appear to diffuse away from the notochord to become distributed among nearby mesenchymal cells (Ruggeri, 1972). If these proteoglycans are removed from the notochord by various digestion procedures, the ability of the notochord to support chondrogenesis is considerably impaired (Kosher & Lash, 1975).

Thus, the PAS-positive, diastase-resistant material seen as ‘granules’ in notochordal cells of the present study may reflect one or more of the substances produced by the notochord such as (1) collagen, a glycoprotein containing variable amounts of carbohydrate (Spiro, 1970; Dische, 1970), (2) the population of non-collagenous glycoproteins that typically occur in association with glycosaminoglycans (Rodén, 1970), or (3) precursors of glycosaminoglycans. Although it is not possible at this time to rule out the possibility that unrelated compounds are responsible for the observed staining, such an alternative seems less likely since the notochord appears to produce a limited number of substances (Kosher & Lash, 1975).

The decreased number or lack of notochordal granules in Sd mutant embryos appears to suggest that the secretory capacity of notochordal cells is in some way impaired. Our findings, however, do not rule out the unlikely possibility that the synthesis and turnover of these granules in abnormal embryos is accelerated, which would also lead to decreased numbers of granules in notochordal cells. The diminished ability of the young notochord in the present study to accumulate these granules also implies that notochordal cells of abnormal embryos are inherently defective, a proposal that gains support from our observation that these cells fail to vacuolate at the appropriate point in development. In chick and mouse embryos, the inability to vacuolate is characteristic of notochordal cells that fail to effectively promote chondrogenesis (Cooper, 1965). The decrease in PAS-positive granules in mesenchymal cells of disc regions and centra of abnormal embryos also suggests that the ability of the notochord to mediate chondrogenesis in Sd mutants may be faulty.

Our results tend to corroborate Grimeburg’s (1958) proposal of early involvement of the notochord in the Sd syndrome, as opposed to Gluecksohn-Schoenheimer’s (1945) contention that notochordal involvement is secondary to defects in the tail.

Of particular interest is the failure of somitic mesenchyme in the 11-day Sd mutant embryos to attain the typical organization observed in vertebral and intervertebral disc anlagen of normal specimens. Defects in the notochord or sheath, or both, may be responsible for the lack of cellular organization in mutants. Flower & Grobstein (1967) have indicated that although surface contact is not critical in the initiation of chondrogenesis, pre-existing surfaces may well have a role in later events in vertebral morphogenesis. Moreover, the properties of substrata or interfaces are reported to influence the morphogenetic movements of mesenchymal cells in embryonic and other systems (Weiss, 1961 ; Cohen & Hay, 1971 ; Toole et al. 1977). In view of these findings, it seems possible that in normal development the notochord-sheath complex contributes to vertebral formation by providing the appropriate surface about which mesenchymal cells initially align, especially since we have shown that when the notochord-sheath complex is demonstrably defective (in histochemical, and possibly chemical terms), the mesenchymal cells rarely show a definite orientation to the notochord (Figs. 3d, 4b). Although anomalies of mesenchymal cell organization in Sd mutants can be correlated with a dysfunctional notochord/ sheath, the possibility that mesenchymal cells themselves have limited capabilities cannot be excluded. This question hopefully will be answered by means of experimental manipulation involving reciprocal interchanges of notochord and mesenchymal cells between Sd and normal embryos.

Disturbances of axial organization as a consequence of notochordal anomalies occur in other mouse mutants, including brachyury, truncate, pintail and anury (see review by Grüneburg, 1963).

The axial defect common to the above mutants is reduction in number of tail vertebrae leading to shortening or absence of the tail; however, the involvement in these mice may be extensive, as in Sd, in which the entire vertebral column is involved, or localized, as in pintail, in which only axial structures of the tail are affected. Although the vertebral defects are often morphologically similar, the initial demonstrable abnormality of the notochord may be quite different. For example, the notochord of pintail is reduced in calibre, while that of truncate ends abruptly in the posterior region of the embryo, and that of anury is retained or incorporated into the hindgut or neural tube. On a cellular level, recent work suggests that brachyury (T) probably results from defects in cell surface components, which are expressed early in development and which may compromise the structural integrity of the notochord (Bennett, 1975). Differences in cell surfaces between normals and abnormals have also been implied by dissociation-reaggregation experiments using brachyury embryos (Yanagisawa & Fujimoto, 1978).

In conclusion, our histochemical results suggest that the development of normal vertebrae seems to be dependent, to a considerable extent, on the presence of a functional notochord. However, the possibility raised by Griineburg (1958) as to whether the abnormal Sd notochord, in turn, reflects improper development and regression of the primitive streak, remains to be explored.

The authors would like to thank Mr Charles O. Boyd for preparing the figures. This work was supported in part by Research Grant HD-09993 from the NICHD.