An analysis of lateralized neural crest marker expression across development in the Mexican tetra, Astyanax mexicanus

Abstract

The biological basis of lateralized cranial aberrations can be rooted in early asymmetric patterning of developmental tissues. However, precisely how development impacts natural cranial asymmetries remains incompletely understood. Here, we examined embryonic patterning of the cranial neural crest at two phases of embryonic development in a natural animal system with two morphotypes: cave-dwelling and surface-dwelling fish. Surface fish are highly symmetric with respect to cranial form at adulthood, however adult cavefish harbor diverse cranial asymmetries. To examine if lateralized aberrations of the developing neural crest underpin these asymmetries, we used an automated technique to quantify the area and expression level of cranial neural crest markers on the left and right sides of the embryonic head. We examined the expression of marker genes encoding both structural proteins and transcription factors at two key stages of development: 36 hpf (∼mid-migration of the neural crest) and 72 hpf (∼early differentiation of neural crest derivatives). Interestingly, our results revealed asymmetric biases at both phases of development in both morphotypes, however consistent lateral biases were less common in surface fish as development progressed. Additionally, this work provides the information on neural crest development, based on whole-mount expression patterns of 19 genes, between stage-matched cave and surface morphs. Further, this study revealed 'asymmetric' noise as a likely normative component of early neural crest development in natural Astyanax fish. Mature cranial asymmetries in cave morphs may arise from persistence of asymmetric processes during development, or as a function of asymmetric processes occurring later in the life history.

Keywords: cranial neural crest; developmental patterning; facial bones; intramembranous bones; osteocranium.

FIGURE 1

A direct method for quantifying gene expression area and level on the developing lateral head. Using the open-source software program ImageJ, we defined the area of gene expression, and quantified the level of expression within that area using a semi-automated approach. The lateral sides of the embryo [(A), right; (B), left] were imaged following in situ hybridization for one of 19 probes. A region of the head was delimited in each image [white dotted box in (A,B)]. The limit of expression for the probe (in this example, col9a3) was automatically outlined in yellow (C,D) using ImageJ, which capitalized on the purple hue of the substrate used in our whole-mount in situ hybridization protocol. Note that the purple hue was calibrated (based on H, S, B values, see Methods) consistently applied across all specimens (cave and surface). From this region, we calculated the area (µm²) and relative expression of the purple substrate (scale 0-255), as a proxy for expression level. We then determined the larger value (in this case the right side, represented as an orange dot) for both area and expression. A summary of all individual, and averaged aggregate data (n = 4) is presented in Figure 2. Scale = 200 µm.

FIGURE 2

Asymmetry measures for gene expression area across the left-right axis reveals more lateralization at early stages in both cave and surface morphs. Each of 19 gene markers involved in diverse stages of neural crest development are arrayed for two morphotypes [cave at (A), surface at (B)] across two stages of development (36 hpf and 72 hpf). The first four circles in each row represent the individual polarities (“I”) for lateralization in quantified expression area (right = orange, left = purple). The final circle is the polarity for the aggregate mean value across n = 4 specimens (“A”). Those genes demonstrating consistency in asymmetric polarity across all four individuals and the aggregate are indicated (dashed red rectangle, red arrow). Both cave and surface fish showed fewer examples of consistent asymmetry at 72 hpf compared to 36 hpf. Overall polarities were mixed, with no obvious bias towards the right or left for any gene, with the exception of sox10 (red asterisk, see Figure 4).

FIGURE 3

Asymmetry measures for gene expression level across the left-right axis reveals diverse patterns of lateralization in cave and surface morphs. 19 neural crest genes are arrayed for cavefish (A) and surface fish (B) at 36 hpf and 72 hpf. Individual (“I”) and aggregate (“A”) polarities reveal more examples of consistent asymmetry at later stages for cavefish, but more examples at earlier stages for surface fish. Five genes demonstrate consistent asymmetric polarity in cavefish (dashed red rectangle, red arrow), but the polarities were mixed (i.e., three leftward, two rightward). Similarly, four genes show consistent asymmetric polarity in surface fish (dashed red rectangle, red arrow), but again polarities were mixed (i.e., one leftward, three rightward). Overall, this analysis reveals common left-right variation between morphs. Note that the gene sox10 harbors consistent leftward bias in both expression area (Figure 2) and expression level (red asterisk).

FIGURE 4

Neural crest gene expression is comparable between embryonic Astyanax cave and surface morphs despite substantial differences in adult phenotype. From the panel of 19 genes, only sox10 showed consistent asymmetric expression at the same stage of development (36 hpf) for both expression area and level [purple rectangle, (A’)]. Sox10 expression is shown for the right (A,C) and left (A’,C’) sides of cavefish at 36 (A,A’) and 72 hpf (C,C’), respectively. Surface fish sox10 expression is shown for the right (B,D) and left (B’,D’) lateral sides at 36 (B,B’) and 72 hpf (D,D’). Relative expression level is depicted by color intensity, comparing cave and surface fish at both stages of development for eight genes encoding structural proteins [blue, (E)] and eleven genes encoding transcription factors [green, (F)]. Note two genes show significant differences between morphotypes, with surface fish showing less expression, including phf20a at 72 hpf, and tfap2a at 36 hpf. Scale in (A–D’) = 200 µm.