Vertebrates exhibit prominent left-right asymmetries in both the arrangement of internal organs relative to the midline and in the anatomical differences that are associated with individual organs (reviewed by Levin,2005). Defects in establishing the left-right body plan during embryogenesis result in abnormal laterality phenotypes that are typically classified as either situs inversus or heterotaxy (also termed “situs ambiguus”). Situs inversus is a condition in which all organ left-right asymmetries are mirror-image-reversed. Heterotaxy is a condition defined as neither situs solitus (normal) nor situs inversus, in which some organs are left-right-reversed and others are not, resulting in a wide range of gross anatomical phenotypes. Individuals afflicted with either situs inversus or heterotaxy are at significant risk for harboring congenital heart defects (reviewed by Bowers et al.,1996), an association that is not surprising given the many morphological and molecular left-right asymmetries that must become established during normal heart development (reviewed by Ramsdell,2005).
Using an innovative technique to trace left-right cell lineages in the Xenopus embryo, we recently found that left-right cardiomyocyte lineage composition differed between different regions of the heart with respect to its anteroposterior axis (data not shown). Moreover, we found that allocation of left-right cardiomyocyte lineages was altered in embryos with experimentally induced laterality defects, indicating that cardiac left-right cell lineage composition is regulated by the left-right body axis. One region of the heart that was frequently affected in laterality mutant embryos was the atrioventricular (AV) canal. Because the AV canal is a site for formation of cushion mesenchyme, a progenitor tissue for valve and septal structures in the mature heart (reviewed by Eisenberg and Markwald,1995), we hypothesized that the high incidence of valvuloseptal defects found in individuals with laterality disease might be causatively linked with abnormal left-right patterning of the AV canal region of the heart.
In the present study, we addressed this possibility by analyzing AV cushion tissue in left-right lineage-labeled Xenopus hearts. Unlike results obtained for the AV myocardium, in which left- vs. right-side-derived cells do not mix across the cardiac midline, we find that the mesenchyme present in the AV cushions is derived from left- and right-side lineages that are mixed together in both the superior and inferior AV cushions. Moreover, we find that the relative left-right origins as well as amount of AV cushion cells typically are altered in embryos with laterality defects induced by misexpression of constitutively active activin-like kinase-4 (CA-ALK4), a type I TGF-β receptor that we previously have shown to be necessary and sufficient for normal left-right axis determination in Xenopus (Chen et al.,2004). These results demonstrate that similar to cardiomyocytes, allocation of left-right mesenchyme cell lineages also is regulated by the left-right body axis. We propose that left-right cell lineage composition in the AV canal is critical for normal valvuloseptal morphogenesis and that left-right lineage abnormalities in this region might be related to the onset of valvuloseptal defects that are characteristically associated with laterality disease.
MATERIALS AND METHODS
Left-Right Cell Lineage Labeling and Induction of Laterality Defects
Egg-laying in adult pigmented Xenopus laevis females was induced by injection of pregnant mare serum gonadotropin and human chorionic gonadotropins in accordance with a standard protocol approved by the university's Institutional Animal Care and Use Committee. Eggs were fertilized, dejellied with 2.5% cysteine, and staged according to Nieuwkoop and Faber (1967). Oregon Green-conjugated dextran (10,000 MW; lysine-fixable) and Alexa 647-conjugated dextran (10,000 MW; Molecular Probes) were each diluted to 2.5 mg/ml in RNAse-free water. The two left blastomeres of a four-cell stage embryo were each pressure-injected in the marginal zone with 2 nl of Oregon Green-dextran and the two right blastomeres were injected with 2 nl of Alexa 647-dextran. Embryos were collected at stages 45–46 and processed as described below.
GpppG-capped RNA was transcribed with the mMessage mMachine kit (Ambion) using 300 ng of linearized CA-ALK4 cDNA (Chang et al.,1997) as template and 500 pg of RNA was pressure-injected into left or right ventrolateral vegetal cells of 16-cell stage embryos as described previously (Chen et al.,2004). After reaching stages 45–46, embryos were anesthetized with 0.01% benzocaine and scored for dorsoanterior phenotype using the dorsoanterior index (DAI) (Kao and Elinson,1988) and LR phenotype using orientation of the heart, gut, and gallbladder as indicators of body situs (Ramsdell and Yost,1999; Chen et al.,2004). Embryos exhibiting normal orientation of these organs were classified as situs solitus (normal) and embryos exhibiting reversed orientation of all of these organs were classified as situs inversus. Embryos exhibiting any combination of normal and reversed organ orientations were classified as heterotaxic. Embryos exhibiting an abnormal DAI or other gross morphological defects were not used for lineage analysis.
Histology, Confocal Imaging, and Cell Counting
Embryos were fixed in MEMFA, washed five times with PBS, and stored at 4°C until further use (Sive et al.,2000). For processing into paraffin, embryos were dehydrated in a graded ethanol series, cleared with Histoclear, and infiltrated and embedded with Paraplast Plus (Fisherbrand). The paraffin blocks were serially sectioned at 10 microns and deparaffinized with Histoclear. Slides were coverslipped with DPX mounting medium prior to viewing.
Images of embryo sections were collected using a Leica TCS SP2 AOBS Confocal System mounted onto a Leica DM RE-7 upright microscope. Prior to imaging the heart, embryos were prescreened at low (5–10×) magnification to ascertain that the dextran lineage labels were appropriately targeted to one side only of the embryo. Only embryos showing distinct left-right hemilabeling were used for further analysis. Hearts were typically viewed using a 40× oil Plan APO objective (NA = 1.25) using a 2× zoom factor, generating 1,024 × 1,024 pixel images at 80× magnification. The dextran fluorophores were excited with 488 nm (Argon laser) for Oregon Green or 633 nm (HeNe laser) for Alexa 647. Each tissue section was imaged at a depth giving maximum emission signal in both channels and pseudocolored green for Oregon Green and red for Alexa 647. Green and red mesenchyme cells present in the AV cushions were counted by visual inspection of images collected from alternating sections encompassing the entire AV canal region of each heart (typically 4–6 sections total were collected per heart). Hearts from wild-type embryos (n = 3), situs inversus embryos (n = 3), heterotaxic embryos (n = 4), and embryos exhibiting normal body situs following ectopic CA-ALK4 expression (n = 6) were examined; however, phenotypic information was keyed separately from specimen number until cell counts were completed.
The total count of red and green cells in the inferior and superior AV cushions of wild-type embryos was compared using a two-tailed t-test with α = 0.05. There were three wild-type cases used in that analysis, and the small difference in cell number for the two cushions did not reach statistical significance. In another similar t-test, the ratio of red-to-green cells was calculated for these same cushions in the three wild-type hearts. Although the average ratio was 1 in the superior AV cushion (SAVC) and 2.2 in the inferior AV cushion (IAVC), the t-test value of 1.7 was not enough to indicate a reliable difference, even if one used a less conservative one-tailed test assumption.
Oregon green-conjugated dextran and Alexa 647-conjugated dextrans were used to trace left-side and right-side cell lineages in wild-type (i.e., control) embryos, which appear as green and red cells, respectively. Prior to heart tube looping, the myocardium overlying the IAVC is located on the left side of the heart and the myocardium overlying the SAVC is located on the right. As looping proceeds, the IAVC region becomes positioned at the inner curvature of the heart and the SAVC region at the outer curvature of the heart (Fig. 1). We found that consistent with their initial left- and right-side positions in the heart prior to looping, the tissue overlying the IAVC was marked by green, and the tissue overlying the SAVC was marked by red (Fig. 2A). This segregation of left- vs. right-side lineages, which are predominantly cardiomyocytes, was observed in the AV canal of all wild-type embryos used in this study. When mesenchyme cells present in each AV cushion were examined, no apparent differences either in the total number of cells present or in the relative contributions of red vs. green cells were observed (Fig. 3A and B). These observations were confirmed by performing cell counts in three wild-type hearts yielding on average 20 ± 3.8 (SEM) cells in the SAVC and 16 ± 3.6 cells in the IAVC (t = 0.77; P = NS). The ratio of right- to left-derived cell lineages was 0.99 ± 0.15 in the SAVC and 2.2 ± 0.7 in the IAVC (t = 1.69; P = NS). Thus, in the wild-type Xenopus heart, the relative contribution of left- vs. right-side cell lineages is the same in both cushions of the AV canal, and unlike in other vertebrates, e.g., mouse (Webb et al.,1996) or chick (Moreno-Rodriguez et al.,1997), the SAVC and IAVC in Xenopus do not exhibit left-right asymmetry in the total amount of cells present.
To determine whether left-right cell lineage composition in the AV cushions is regulated by the left-right body axis, we next performed cell lineage analyses in embryos with experimentally induced laterality defects. This was accomplished by ectopic expression of CA-ALK4, a type I TGF-β receptor that we previously found to cause heterotaxy and situs inversus when targeted to the right side of the early Xenopus blastula (Chen et al.,2004). Of the hearts examined in three CA-ALK4-induced situs inversus embryos (embryos 4–6), none was found to be completely normal with respect to the IAVC. Consistent with the mirror-image-reversed position of the cushions that occurs in situs inversus, the tissue overlying the IAVC (i.e., at the outer curvature) and SAVC (i.e., at the inner curvature) regions of embryo 4 (Fig. 2B) was red and green, respectively, which is the mirror-image opposite of that seen in wild-type embryos. The relative contributions of right vs. left cell lineages in both cushions differed from the 1:1 ratios seen in wild-type embryos, with a predominance of left-side-derived cells present in the SAVC and a predominance of right-side-derived cells present in the IAVC (Fig. 3A and B). In contrast to these results, embryo 5 showed reduced cushion formation in the SAVC, although the relative proportion of red vs. green cells was similar to that observed in wild-type hearts (Figs. 2C and 3A). The relative contributions of left- and right-side-derived cells in the IAVC differed from that of wild-type hearts, with a right-side predominance observed in the mutant heart (Fig. 3B). Interestingly, the AV myocardium of the heart was not strictly comprised of single left or right cell lineages; that is, a scattering of red cells was seen in the otherwise mostly green SAVC region and a few green cells were observed in the IAVC region (Fig. 2C). The other embryo in this group (embryo 6) showed normal SAVC tissue formation, but lacked left-side-derived cells in the IAVC (Figs. 2D and 3A and B). When the left-right composition of the myocardium was examined, it was found that the SAVC myocardium was comprised exclusively of green (left-side) cells, but the IAVC myocardium was comprised of red (right-side) cells with a few scattered green cells mixed in (Fig. 2D). Thus, although the overall body situs of embryos in this group was identical, including the orientation of heart looping, the hearts of these embryos showed differences not only compared to wild-type embryos, but also compared to one another when analyzed at the histological level.
As with situs inversus embryos, embryos exhibiting CA-ALK4-induced heterotaxy (embryos 7–10) were found to differ not only from wild-type embryos, but also from one another within this phenotypic group. The amount of mesenchyme was reduced in the SAVC region in two of four embryos (embryos 7 and 9; Fig. 2E and G), although the relative contribution of left and right lineages was normal (Fig. 3A and B). Both parameters of SAVC formation were normal for embryo 8 (Fig. 2F), but the greater right-side lineage contribution seen in the IAVC was different from the 1:1 ratio seen in both wild-type and in other embryos in the heterotaxy group (Fig. 3B). Different results were observed for embryo 10 (Fig. 2H), which showed a left-side predominant shift in mesenchyme lineage in the SAVC, and reduced left-side lineage composition noted for the IAVC (Fig. 3A and B). It was additionally noted that the differences observed in the amount of cushion mesenchyme formed or the relative left-right lineage contributions did not correspond to the orientation of heart looping; hearts of embryos 7 and 9 had d-loops (normal) and hearts of embryos 8 and 10 had l-loops (reversed). Myocardial left-right cell composition was normal in embryos 8 and 10 (Fig. 2F and H), but embryos 7 and 9 showed a mix of red and green cells in both the SAVC and IAVC regions (Fig. 2E and G). Thus, in all heterotaxic embryos with abnormal AV myocardial cell composition, differences in either the amount and/or relative left-right lineage contributions of the underlying cushion mesenchyme were found. However, the corollary was not observed; that is, not all differences in the amount and/or relative left-right lineage contributions of cushion mesenchyme were accompanied by abnormalities in left-right myocyte lineage composition.
The final group of embryos examined was those that exhibited normal body situs (situs solitus) despite ectopic expression of CA-ALK4 RNA. As previously reported (Chen et al.,2004), when CA-ALK4 RNA is expressed on the right side of the embryo, a minority of embryos fails to develop overt laterality defects, as assessed by the orientation of heart looping, gut looping and rotation, and the position of the gall bladder relative to the embryonic midline. Embryos in this category (embryos 11 and 12) showed normal SAVC formation with respect to the amount and relative left-right lineage contributions of cells present (Figs. 2I and J and 3A and B). However, one embryo (embryo 11) showed not only fewer cells in the IAVC (Fig. 2I), but also a complete lack of right-side-derived cells in this cushion (Fig. 3B). No abnormalities in left-right myocyte lineage composition were noted for the IAVC regions in these embryos, although the SAVC myocardium in embryo 11 contained green cells mixed in with a majority of red cells (Fig. 2I). Similarly, embryos in which CA-ALK4 RNA had been expressed on the left side showed altered left-right AV myocyte composition in some, but not all, embryos in this group. The SAVC myocardium of embryo 14 (Fig. 2K) contained mostly green cells, and although the amount of mesenchyme formation in the SAVC appeared elevated, the left-right lineage composition was normal. The SAVC myocardium of embryo 16 also showed abnormal left-right lineage composition, with some green cells mixed in within the mostly red myocardium (Fig. 2L). Abnormal left-right lineage composition was noted for mesenchyme in this cushion, which showed a left-side predominance (Fig. 3A). The myocyte composition of the IAVC appeared normal, and the amount and left-right lineage proportions of IAVC mesenchyme formation appeared normal in this embryo (Fig. 3B). The other embryos in this group showed normal amounts and relative left-right contributions of cells in both the SAVC and IAVC (Fig. 3A and B) as well as normal left-right cell composition in the overlying myocardium (not shown). Thus, as was found for embryos in the situs inversus and heterotaxy groups, phenotypic heterogeneity was present between embryos within the same experimental situs solitus groups, despite the overall outward appearance of identical (and normal) body situs.
Despite advances in defining many of the genes and signaling molecules that drive left-right axis determination in the vertebrate, relatively little is known about how organ primordia ultimately respond to left-right positional information to generate anatomical asymmetries. Results obtained in this and a previous study (data not shown) suggest that left-right lineage composition of the heart is an important parameter that is regulated by the left-right body axis during organogenesis. In approximately half of laterality mutant embryos examined, defects in the relative proportions of left and right cell lineages were observed for the AV cushions, and in many (but not all) instances, defects in the overall amount of mesenchyme formation also were observed. There did not appear to be a correlation between the presence or absence of defects in the SAVC vs. the IAVC, i.e., each cushion appeared to be equally affected, although not necessarily within the same embryo. Intriguingly, no two embryos within a given category of abnormal laterality phenotype exhibited the same defects in either AV mesenchyme cell count or AV mesenchyme left-right lineage, demonstrating that extreme heterogeneity exists in the cellular composition in the AV canal of embryos with situs inversus and heterotaxy. This observation is consistent with findings in our earlier study in which other areas of the heart (e.g., atria, outflow tract) appeared to be stochastically affected among the different phenotypic groups (data not shown). Thus, one important conclusion that can be drawn from these results is that body situs classification is not predictive of whether AV canal abnormalities will be present, or even which aspects of cushion tissue formation will be abnormal in affected hearts.
In addition to wild-type, situs inversus, and heterotaxic embryos, embryos that failed to develop abnormal body situs following misexpression of ALK4, a well-characterized laterality gene (Chen et al.,2004), were examined. Somewhat surprisingly, despite normal gross body situs, altered left-right cell lineage contributions of AV mesenchyme were observed in some embryos in this category, as were defects in the overall amount of cushion tissue in others. One interpretation of such data is that ALK4, the laterality gene manipulated in the experimental embryos, can affect cardiogenesis independently from its role in left-right axis determination. Alternatively, it is possible that the AV mesenchyme defects, particularly the left-right lineage abnormalities, represent subtle laterality defects that can nevertheless occur in embryos with otherwise normal body situs. Future study to distinguish between these two possibilities will be important because if the latter is correct, then this could account for many of the seemingly isolated congenital heart defects that are present in families with heritable forms of laterality disease (Morelli et al.,2001).
It should be noted that in wild-type hearts, neither the overall amount nor the relative proportion of left- and right-side-derived cells significantly differed between the SAVC and the IAVC. This is in contrast to results obtained in mouse (Webb et al.,1996) and chick (Moreno-Rodriguez et al.,1997), in which the SAVC contains far fewer mesenchyme cells than the IAVC. Although the reason for this difference between frog and other vertebrates is not clear, one possibility is that it might reflect the variation in septation and valve formation necessitated by differences in three- vs. four-chamber hearts. As others and we have proposed for septation of the common atrium (Gormley and Nascone-Yoder,2003), the significance of differences in left- and right-side-derived cell lineages in the heart may be related to their ability to become differentiated into particular tissues. Thus, in vertebrates in which AV cushion mesenchyme cells contribute to multiple structures (e.g., mitral valve, tricuspid valve, interatrial and interventricular septum) (reviewed by Markwald et al.,1998), asymmetries in left-right cell lineage composition may be necessary for normal morphogenesis. By comparison, the AV cushions in the univentricular frog heart form a relatively simple valve present in the common atrium (Mohun et al.,2000), a process that may not require left-right cell lineage asymmetries in cushion tissue itself. However, this is not to imply that asymmetry in cell lineages is not important in the AV canal of the frog heart; as shown in the present study, the myocardium overlying the SAVC and IAVC regions of the Xenopus heart exhibits distinct differences in left-right cell lineage composition. In all cases in which abnormalities in AV myocardial cell lineages were detected, defects in either AV mesenchyme amount or left-right derivation were seen. Because a major source of AV cushion mesenchyme is formed via a myocardially induced epithelial-mesenchymal transformation of endocardial cells (reviewed by Eisenberg and Markwald,1995), this suggests that left-right lineage could play a role in the ability of the AV myocardium to modulate this process. Interestingly, it also was noted that there were some embryos that exhibited AV mesenchyme defects (either amount or lineage) even in the presence of normally allocated left-right AV myocyte lineages. There are at least two possible (and not necessarily mutually exclusive) interpretations of this observation that warrant further study. First, it is possible that there are sources of AV mesenchyme (other than the endocardium) that are targets of the laterality pathway. Second, it is possible that AV endocardial cells are themselves direct targets of left-right axial signaling. Because it is known that there is at least one, if not more, additional sources of AV mesenchyme other than the endocardium (Manner et al.,1993; Manner,1999; Gittenberger-de Groot et al.,2000; Perez-Pomares et al.,2002), it will be interesting to learn whether findings made in higher vertebrates can be extended to Xenopus. If so, the ability to uniquely and reliably perform left-right cardiac lineage analyses in combination with the ability to consistently elicit different types of laterality defects will position Xenopus as an important and valuable vertebrate model for study of many aspects of left-right cardiogenesis.
The authors thank Austin Ramsdell for technical assistance with the confocal image analysis. J.J., a participant in the 2005 MUSC Summer Undergraduate Research Program, was supported by T35 HL07769 (to Dr. Cynthia Wright). This work was supported by HL73270 (to A.F.R.) and an SC COBRE for Cardiovascular Disease P20-RR-1634 (to Dr. Roger R. Markwald).