ABSTRACT
Chlorcyclizine (CHLR) enhances the degradation of hyaluronate (HA) into smaller molecular weight pieces with no effect on its synthesis. Administration of CHLR to pregnant CD-I mice on gestational days 10·5, 11·5 and 12·5 results in 100 % cleft palate in the fetuses. The caudal two thirds of the palatal shelves are reduced in size and unable to reorient in vitro, while anterior shelf regions are relatively unaffected. Alcian blue staining combined with specific enzymic digestion was used to identify HA in sections of CHLR-treated shelves. With the aid of computer-assisted image subtraction the patterns of HA distribution across the tissue section were objectively identified. Anterior, posterior and presumptive soft palatal shelf regions were examined at gestational days 13·25, 13·5, 13·75 and 14·5. Acquisition of a normal pattern of HA distribution was delayed by about 24 h, as compared to untreated specimens in all three shelf regions. The posterior and soft regions, comprising the caudal two thirds of the shelf, also showed pronounced shape change. These regions only displayed normal curvature of the nasal surface when a normal pattern of HA distribution was attained. These results suggest that, for the caudal two thirds of the palatal shelf, normal shape and the ability to remodel are linked to the molecular configuration of HA and to a specific pattern of HA distribution.
Introduction
HA is the predominant glycosaminoglycan (GAG) component of the secondary palatal shelves around the time of shelf reorientation (Pratt, Goggins, Wilk & King, 1973). Chlorcyclizine, an antihistaminic benzhydrylpiperazine compound, enhances the degradation of hyaluronate (HA) and chondroitin sulphates with little effect on their synthesis and no appreciable effect on DNA synthesis (Wilk, King & Pratt, 1978). Administration to pregnant mice results in cleft palate in their offspring. The caudal two thirds of the palatal shelves of drug-treated fetuses are reduced in size and unable to reorient in vitro, while anterior shelf regions are relatively unaffected (Brinkley & Vickerman, 1982). Chlorcyclizine appears to impair the functional role of HA in remodelling the caudal two thirds of the palatal shelf.
Utilizing Alcian blue staining, specific enzymic digestion and a computer-assisted method of image subtraction, we have identified characteristic regionspecific patterns of HA distribution in mouse palatal shelves prior to shelf reorientation which are uniform following reorientation (Brinkley & Morris-Wiman, 1987). Our results suggest the HA-rich extracellular matrix is being hydrated, thus moving stain-binding sites further apart. These observations also support the hypothesis that expansion of an HA-rich extracellular matrix is involved in the remodelling of the caudal two thirds of the palatal shelf.
CHLR-induced reductions in the size of HA molecules could alter their ability to complex with collagens and other GAGs, which might make them more labile within the extracellular matrix. Thus, they might assume an abnormal distribution within the shelf and in that way interfere with shelf remodelling. The present study was undertaken to determine the patterns of distribution of chlorcycli-zine-degraded HA over the time course of expected in vivo palatal closure.
Materials and methods
Specific details of animal care and husbandry, histological techniques and image analysis are given in the preceding paper (Brinkley & Morris-Wiman, 1987). Only information pertaining specifically to these studies and a brief review of the computer-assisted method will be given here.
CHLR administration
Sequential doses (250 mg kg-1) of chlorcyclizine hydrochloride (Burroughs-Wellcome) were administered to pregnant CD-I mice by gavage on gestational days 10·5, 11·5 and 12·5, as this treatment regimen results in 100% cleft palate in the offspring (Brinkley & Vickerman, 1982). Each dose was delivered in a total volume of 0·5 ml sterile tap water.
Specimens
CHLR-treated pregnant dams that were gestational day 13·25,13·5,13·75 or 14·5 were killed by cervical dislocation. The fetuses were delivered and staged according to their crown-rump length and a morphological rating based on the developmental state of fore- and hindlimb, ears, eyelids and hair follicles (Walker & Crain, 1960) as previously described (Brinkley & Bookstein, 1986). The groups will be referred to by gestational age. These gestational times and developmental stages correspond to 30, 24 and 18 h prior to normal shelf reorientation, and after shelf reorientation and initial adhesion, respectively. Four to six specimens from a minimum of three litters were examined for each gestational age and shelf region.
Image acquisition and analysis
Serial 7 μm coronal sections of palatal shelves, which were fixed in phosphate-buffered formalin containing 2% (w/v) tannic acid (Mallinckrodt, Paris, KY) to precipitate the GAGs (Singley & Solursh, 1980), were cut, mounted on slides, deparaffinized in xylene and rehydrated through a graded ethanol series. Alternate sections from the anterior, posterior and soft shelf regions were either digested with Streptomyces hyaluronidase to specifically remove HA (Ohya & Kaneko, 1970), or incubated in buffer alone as a control.
Following staining with Alcian blue and counterstaining with nuclear fast red, videoimages were taken with the aid of a Newvicon television camera mounted on a Leitz Diaplan microscope. The videoimages were converted to 512×512×8 bit digital images by an International Imaging Systems image processor (San Jose, California, USA) linked to a MassComp MC-500 computer (Westford, Massachusetts, USA). Two different images, a matrix image and a cell image, were made of both digested and control tissue sections using narrow band-width filters at or near the maximum absorbance for the respective dye products. The cell image was subtracted from the matrix image of each pair. The resulting matrix-only image of the digested section was then subtracted from the matrix-only image of the control section. The result is a difference picture of the distribution of what was enzymically removed, the HA. Next these final images were thresholded to remove background grey values, and colour-coded to facilitate visual comparisons. Details of this methodology are given in the preceding paper (Brinkley & Morris-Wiman, 1987).
Results
The size and shape of all shelf regions were about 24 h delayed as compared to normal shelves (Brinkley & Morris-Wiman, 1987).
Anterior (Fig. 1)
Hyaluronate distribution in anterior shelves over the time course of expected shelf reorientation, days 13·25 (A), 13·5 (B), 13·75 (C) and 14·5 (D). The maxillary regions imp adjacent to the shelves themselves are HA-rich, while the shelves themselves are not. By day 14·5 (D), a pattern of HA-distribution similar to that seen 24h earlier in untreated specimens is now observed. Grey values used for colour coding: grey, 1–42; blue, 43–84; green, 85–126; yellow, 127–168; red, 169–210; white, 211–255. M, medial; N, nasal; O, oral. Bars, 10 μm.
Hyaluronate distribution in anterior shelves over the time course of expected shelf reorientation, days 13·25 (A), 13·5 (B), 13·75 (C) and 14·5 (D). The maxillary regions imp adjacent to the shelves themselves are HA-rich, while the shelves themselves are not. By day 14·5 (D), a pattern of HA-distribution similar to that seen 24h earlier in untreated specimens is now observed. Grey values used for colour coding: grey, 1–42; blue, 43–84; green, 85–126; yellow, 127–168; red, 169–210; white, 211–255. M, medial; N, nasal; O, oral. Bars, 10 μm.
Day-13·25 (Fig. 1A) and -13·5 (Fig. 1B) shelves showed little HA. By day 13·75 (Fig. 1C) HA was observed in the medial part of the shelf and by day 14·5 (Fig. 1D) HA was diffusely distributed throughout the entire shelf except for a midoral, HA-poor region. At all four ages, there were considerable amounts of HA in the adjacent maxillary regions.
Posterior (Fig. 2)
Hyaluronate distribution in posterior shelves over the time course of expected shelf reorientation, days 13·25 (A), 13·5 (B), 13·75 (C) and 14·5 (D). Uniform patterns of distribution are observed for about 24 h following the last maternal administration of CHLR (day 13·25 (A) –13·5 (B)). A pattern similar to that seen 24 h earlier was noted at day 14·5 (D). As in untreated specimens, the maxillary regions adjacent to the shelves were always HA-poor. Labels and grey values assigned to each colour are as described for Fig. 1. tg, toothgerm. Bars, 10μm.
Hyaluronate distribution in posterior shelves over the time course of expected shelf reorientation, days 13·25 (A), 13·5 (B), 13·75 (C) and 14·5 (D). Uniform patterns of distribution are observed for about 24 h following the last maternal administration of CHLR (day 13·25 (A) –13·5 (B)). A pattern similar to that seen 24 h earlier was noted at day 14·5 (D). As in untreated specimens, the maxillary regions adjacent to the shelves were always HA-poor. Labels and grey values assigned to each colour are as described for Fig. 1. tg, toothgerm. Bars, 10μm.
Day-13·25 (Fig. 2A) shelves showed homogeneous distribution of HA, with a sizeable, HA-poor maxillary region immediately adjacent to the shelf. At day 13·5 (Fig. 2B), the overall pattern was still observed, but the HA was more dispersed. 6 h later at day 13·75 (Fig. 2C), two regions of increased HA density could be discerned, one medial, the other lateral-oral. At all three ages, the nasal surface of the shelves was markedly straight with little or no curvature. By day 14·5 (Fig. 2D), the increased HA density in the medial and lateral-oral regions was even more pronounced, as was the paucity of HA in the adjacent maxillary region. A normal nasal curvature is now visible.
Soft (Fig. 3)
Hyaluronate distribution in soft shelves over the time course of expected shelf reorientation, days 13·25(A), 13 5(B), 13·75(C) and 14·5(D). As with the other shelf regions the acquisition of a normal pattern of HA distribution was delayed about 24 h from that seen in untreated specimens. The superior nasal surface of the shelves also did not display a normal curvature until that time. As seen in the posterior, the adjacent maxillary regions (mx) were HA-poor. Labels and grey values assigned to each colour are as described for Fig. 1. tg, toothgerm. Bars, 10μm.
Hyaluronate distribution in soft shelves over the time course of expected shelf reorientation, days 13·25(A), 13 5(B), 13·75(C) and 14·5(D). As with the other shelf regions the acquisition of a normal pattern of HA distribution was delayed about 24 h from that seen in untreated specimens. The superior nasal surface of the shelves also did not display a normal curvature until that time. As seen in the posterior, the adjacent maxillary regions (mx) were HA-poor. Labels and grey values assigned to each colour are as described for Fig. 1. tg, toothgerm. Bars, 10μm.
Although HA staining was distributed evenly throughout most of the shelf at day 13·25 (Fig. 3A), a lateral-oral area of increased intensity was often noted. Considerably more HA was visible by day 13·5 (Fig. 3B), with somewhat more noted in the naso-medial one fourth to one third of the shelf and no change in staining in the midoral region. This pattern was even more pronounced at day 13·75 (Fig. 3C). As was observed in the posterior region, the nasal surface of shelves of all three ages was noticeably lacking in curvature. By day 14·5 (Fig. 3D), HA was distributed throughout the body of the shelf with somewhat more intense areas visible in the naso-medial and lateral-oral regions. A peripheral, midoral HA-poor region was still visible.
Discussion
The advantages and limitations of the computer-assisted method employed here are discussed fully in the preceding paper. The major advantage is that it is an objective method of identifying the distribution of HA while ensuring it is not obscured in any way by artefactual cellular staining. The major limitation is that the patterns identified by this method are conservative depictions of the location and density of HA.
The altered patterns of HA distribution observed in CHLR-treated shelves could be attributed to decreased synthesis, increased removal because smaller molecular weight HA might be degraded faster, altered staining properties of the molecules themselves or changes in the ability of the HA molecules to complex with other extracellular matrix components which in turn could alter their topographical distribution (Welsh, Rees, Morris & Madden, 1980). Wilk, King & Pratt (1978) showed that CHLR did not affect the synthesis of GAGs or the synthesis or distribution of lysosomal enzymes. Staining properties should not be affected, as the smaller molecular weight HA molecules produced after CHLR treatment are still well above the threshold for detection .with Alcian blue (Turner & Cowman, 1985). Another possibility is that the smaller HA molecules now have more staining sites available, and therefore greater staining intensity might be expected. No significant increases in HA staining of CHLR-treated as compared to normal palatal shelves (Brinkley & Morris-Wiman, 1987) were observed in the present study.
The interactions of smaller HA molecules with each other and other GAGs and collagen in the extracellular matrix should produce a gel-fibre matrix with greatly altered biophysical properties (Comper & Laurent, 1978; Welsh et al. 1980). Such things as the topological stability of the HA molecules in the gel, as well as the hydrophilic characteristics of the gel, are only two such properties that could be affected. The former would alter the pattern of HA distribution, while the latter would affect the function of the gel-fibre matrix. It seems possible, if not probable, that the altered patterns of HA distribution observed in the present study at days 13·25–13·75, as well as the inability of the CHLR-treated posterior and soft shelves to reorient (Brinkley & Vickerman, 1982), are the result of these phenomena.
The CHLR-treated shelves observed in the present study were harvested from 18 to 48 h after the last maternal CHLR treatment. Treated shelves were delayed about 24 h in the acquisition of a normal pattern of HA distribution. Previous work has shown that gestational day-13·5, -14·25 and -14·5 fetuses of CHLR-treated mothers are significantly smaller in crown-rump length and are less-morphologically advanced. However, this retardation is completely overcome by day 15·25 (Brinkley & Vickerman, 1982). The same study demonstrated that CHLR treatment also directly affects the shelves themselves. The anterior regions of day-13·5 CHLR-treated shelves were able to reorient in vitro, whereas the posterior and soft regions were not able to reorient until day 14·25. Present and past results suggest that the ability of the posterior and soft regions to remodel successfully is closely linked to the acquisition of a normal pattern of HA distribution. In contrast, the ability of the anterior shelf to reorient does not seem linked to HA configuration or distribution.
The effects of CHLR treatment on shelf shape also support these suggestions. Despite reduced size at day 13·25, anterior shelves display normal curvatures of the nasal surface. In contrast, the nasal surfaces of the posterior and soft shelf regions are abnormally flat. One possible explanation for these findings is that, in the anterior shelf region, shape and the ability to reorient are derived from an inherent horizontal position built into its structure by the arrangement of both epithelial and mesenchymal cells as well as constituents of the extracellular matrix other than HA. In contrast, the shape and remodelling ability of posterior and soft shelf regions are affected by the distribution and functional state of HA.
CHLR-induced changes in the HA molecules could alter how cells interact with one another via changes in composition of the palatal epithelial basement membrane as well as the mesenchymal extracellular matrix. GAGs in the basement membrane are known to play an essential role in directing and maintaining changes in epithelial cell morphology in developing systems (Bernfield, Cohen & Banerjee, 1973; Spooner & Faubion, 1980). Epithelial cell packing is denser around CHLR-treated as compared to normal shelves (unpublished observation). We have also previously shown that CHLR affects mesenchymal cell-basal lamina relationships in this region (Brinkley & Morris-Wiman, 1986). Present and past results (Brinkley & Morris-Wiman, 1984; Brinkley, 1984) suggest that hydration of a patterned HA-rich extracellular matrix directed by local changes in epithelial cell density underlies the remodelling of the caudal two thirds to three quarters of the palatal shelves. Normal spatial and temporal distribution of native HA and, possibly, other GAG molecules both in the extracellular matrix and the basement membrane may be required for these events to take place.
ACKNOWLEDGEMENTS
We wish to acknowledge the excellent technical assistance of Susan Tachna. This work was supported by NIH-NIDR grants DE02774 and K04-00104 to L.L.B.