Cephalopod lentigenic cell differentiation and early anterior segment heterogeneity
The anterior of the cephalopod eye, or the anterior segment, is composed primarily of lens generating cells (lentigenic cells) [40,41,42]. Lentigenic cells are arranged circumferentially around the developing lens and extend long cellular processes, fusing into plates to form the lens (Fig. 1A) [40, 41, 43,44,45]. We identified the first evidence of differentiated lentigenic cells starting at late stage 21, using a previously described nuclear morphology, unique to one of the three lentigenic cell types (LC2) (Figs. 1B and 2A) [43, 44, 46]. The number of LC2 cells continues to grow until reaching pre-hatching stage (stage 29). We performed staged in situ hybridization for a homolog of DpS-Crystallin, the most abundant family of proteins in the cephalopod lens [47, 48] (Supplemental Figure 1). The first evidence of expression correlates with changes in nuclear morphology at stage 21 (Fig. 1C).
We sought to understand the molecular heterogeneity of cells in the early developing anterior segment, of which nothing is currently known. Using previously published candidates and RNA-seq data, we performed in situ hybridization screens at stage 23 to identify unique cell populations [46, 50]. We find DpSix3/6 at stage 23 expressed in the anterior segment in the distal cells that make a central cup (cc), as well as a marginal population of cells in the most proximal tissue (pm) (Fig. 2B”, Supplemental Figure 2, Supplemental Figure 3). The proximal central cells lacking DpSix3/6 expression correspond to the LC2 population (Fig. 2A”, B”). Asymmetry along the animal anterior-posterior axis in the eye is also apparent, with enrichment on the anterior side of the animal (Fig. 2B”). We also find the gene DpLhx1/5, expressed in a distal-marginal population of cells in the anterior segment (dm), and excluded from the distal central cup cells (cc) (Fig. 2C”, Supplemental Figure 2, Supplemental Figure 3). Together these genes show distinct populations of cells present early in development and provide a helpful molecular map of the anterior segment tissue at this time point: central cup cells (cc), LC2 cells (lc2), proximal-marginal cells (pm), and distal-marginal cells (dm) (Fig. 2).
Proximal-distal limb patterning genes in the anterior segment of the cephalopod
To assess whether genes involved in appendage patterning may be required for cephalopod lens development, we identified and performed in situ hybridization for the genes Dlx, Meis, Pbx, and Dac at stages 21 and 23 (Fig. 2, Supplemental Figure 2, Supplemental Figure 3). All genes were clearly expressed in the developing anterior segment and lentigenic cells with the exception of DpDac (Fig. 2E–G, Supplemental Figure 2C-2J’, Supplemental Figure 3). We find DpSP6-9a and DpDlx have overlapping expression, in the central cup cells (cc) and all proximal cells (LC2 and pm) (Fig. 2D–E”, Supplemental Figure 3). DpMeis and DpPbx are both broadly expressed in the anterior segment during lens development, with DpPbx excluded from the LC2 cells (Fig. 2F”, G”, Supplemental Figure 3).
It is known that the transcription factor aristaless is necessary for the most distal tip of the Drosophila limb in the limb program [9]. The evolutionary relationship of Prd-like homologs (Arx/Aristaless, Alx/Aristaless-like, Rx/Retinal Homeobox, and Hbn/Homeobrain) is ambiguous across species [51]. We identified three candidate Prd-like genes in D. pealeii and performed in situ hybridization for all three homologs, DpHbn, DpPrdl-1, and DpPrdl-2 (Supplemental Figure 2K, L) [46]. DpHbn is expressed in the anterior segment in the distal central cup cells (cc) while DpPrdl-1 and DpPrdl-2 are excluded from the eye (Fig. 2H” and Supplemental Figure 2C, C’, K and L, Supplemental Figure 3). DpHbn’s central, distal expression recapitulates aristaless expression in the developing Drosophila limb.
Our data show that the majority of the proximodistal patterning genes in the developing limb, including SP6-9, Dlx, Meis, Pbx, as well as the Prd-like homolog, Hbn, show expression in concentric and overlapping cell populations surrounding the developing lens in the squid (Fig. 2). This pattern of expression is similar to the bullseye-like pattern of expression of these genes in the developing Drosophila limb imaginal disc and suggests a co-option of this regulatory program for a new function: patterning the cephalopod anterior segment and lens [14].
Canonical Wnt signaling genes expressed during anterior segment development
The duplication of SP6-9 in cephalopods may provide a substrate for the evolution of cis-regulation, resulting in novel expression of the limb patterning program in the cephalopod lens. In Drosophila appendage outgrowth, active Wnt signaling is upstream of the expression of SP6-9 [52, 53]. To assess whether Wnt may be acting upstream in the cephalopod anterior segment or whether novel regulatory mechanisms may be at play, we performed in situ hybridization for members of the Wnt signaling pathway at stage 21 and stage 23 (Fig. 3, Supplemental Figure 4). We were interested in identifying cells in the anterior segment or in adjacent tissue that may be a source of the Wnt morphogen. We performed in situ hybridization for seven Wnt homologs, with most Wnt genes expressed in the retina (Fig. 3A’, C’, and D–G). DpWnt8, DpWnt11, and DpProtostome-specific Wnt show the most robust retinal expression (Fig. 3A’, F, and G), and DpWnt7 is the only Wnt expressed in the anterior segment (Fig. 3C). DpWnt6 showed no evidence of expression in the developing eye (data not shown). These data support the hypothesis that Wnt signals emanating from the anterior segment or neighboring tissues could regulate anterior segment development.
To identify cells with potential active Wnt signaling, we analyzed the expression of Fz genes, which encode a family of Wnt receptors. We find that DpFz receptors are expressed broadly throughout the embryo. A subset of these (e.g. DpFz1/2/7, DpFz4, and DpFz5/8) are expressed in a subset of cells in the anterior segment, while others, like DpFz9/10, are excluded from the anterior segment (Fig. 3H–K, Supplemental Figure 4). On close examination, we find that DpFz5/8 is excluded asymmetrically in the anterior segment and may be important for anterior-posterior patterning (Fig. 3J’, J”, Supplemental Figure 4D). DpFz1/2/7 is excluded from the distal-marginal cells and central cup cells and interestingly, the central cup cells lacking DpFz1/2/7 are those that express all the limb patterning program genes (Fig. 3K’, K”, Supplemental Figure 4D). These data suggested active Wnt signaling may be important in the cephalopod anterior segment.
Ectopic Wnt activation leads to the loss of the lens
To assess the hypothesis that Wnt signaling is playing a regulatory role in anterior segment development, we utilized well-characterized pharmacological compounds that act as agonists and antagonists of the Wnt pathway [54,55,56,57]. We empirically determined a working concentration of LiCl (0.15 M), CHIR99021 (250 μm), and Quercetin (50 μM). We bathed embryos in the compound or vehicle control for 24 h at stage 21, the onset of lentigenic cell differentiation, and immediately fixed thereafter. Embryos were sectioned and assessed for phenotypes. Stage 21 control embryos show a thickened anterior segment, identifiable lentigenic cells, and small lens primordia (Fig. 3L). LiCl-treated stage 21 embryos show a complete absence of lens formation: no anterior segment thickening, lentigenic cells, or lens tissue. These data suggest that ectopic Wnt pathway activation inhibits lens and anterior segment development (Fig. 3L’, Supplemental Figure 5A). CHIR99021 treatment showed similar phenotypes (Supplemental Figure 5A). We assessed LiCl treated and control animals for cell death and find little difference between control and treated eyes suggesting that toxicity is unlikely the reason for these phenotypic changes (Supplemental Figure 5B). Wnt antagonist treatments (Quercetin) starting at stage 21 show lens development unaffected (Supplemental Figure 5C).
We were interested in the consequence of activating or inhibiting the Wnt pathway on lens development after the beginning of lentigenic cell differentiation. We performed the same 24-h LiCl exposure at stage 23 and find the lens smaller and the anterior segment less thick than control animals, but lentigenic cells and lens tissue remain identifiable. This suggests that ectopic Wnt signaling does not impact cell identity in differentiated lentigenic cells (Fig. 3M, M’). In Quercetin-treated animals starting at stage 23, the anterior segment shows minor organizational defects, but lens development appears unaffected (Supplemental Figure 5C).
The lack of lens growth in stage 21 treated animals may be a result of an imposed delay in lens formation or it may be a result of the loss of lens potential. To differentiate between these possibilities we allowed treated animals to recover. We bathed experimental and control embryos, at both stages 21 and 23, for 24 h, washed out the solution, and allowed animals to develop for an additional 48 h. LiCl-treated stage 21 embryos never recover a lens (Fig. 3N, 3N’) while LiCl treated stage 23 embryos do form a small but morphologically abnormal lens (Fig. 3O, O’). This abnormal lens is larger than the lens found in animals immediately fixed after treatment, suggesting that existing lentigenic cells at stage 23 continue to contribute to lens formation and growth. However, because the stage 23 treated lens is markedly smaller than the control, it suggests that further lentigenic cell differentiation is lost in treated animals. These data suggest that ectopic Wnt signaling leads to the disruption of lens potential and the lack of proper lentigenic cell differentiation.
Despite the remarkable loss of the lens as a consequence of ectopic Wnt signaling, these data do not clearly distinguish between the loss of lentigenic cell fate or proper cell function, such as the growth of the cellular processes that form the lens. To assess if lentigenic cell fate is lost, we performed in situ hybridization experiments for DpS-Crystallin on LiCl-treated animals. We saw two types of expression phenotypes, either a significant decrease (type I) or a complete loss (type II) in DpS-Crystallin expression as compared to control (Fig. 4P, P’, and P”, Supplemental Figure 6). We find all DpS-Crystallin expression exclusively dorsal to the site of lens formation suggesting that these cells may differentiate first. These data show that ectopic Wnt signaling results in the loss of lentigenic cell fate and that our treatment may have interrupted a dorsal-to-ventral wave of differentiation in some embryos (Fig. 4A). In addition, we assessed other anterior segment markers, including DpSix3/6 and DpLhx1/5, and these genes show a consistent loss of expression in the most severe phenotypes, (Supplemental Figure 6A-6C).
Limb patterning program regulatory evolution
To address if Wnt signaling is upstream of the limb patterning program, we performed in situ hybridization of limb transcription factors after LiCl treatment (Fig. 3Q–S, Supplementary Figure 6A-6C). Similar to DpS-Crystallin expression, we again see a mild reduction (Type I) or loss and severe reduction (Type II) in region of expression. Our milder phenotypes, again, show a dorsal asymmetry, which can be most easily seen in DpSP6-9A, DpDlx, and DpHbn (Fig. 3Q, Q’, Q”; R, R’, R”; and S, S’, S”). Changes are also visible but less obvious in DpPbx and DpMeis expression, with DpPbx only showing a mild phenotype (Supplemental Figure 6A-6C). These data are consistent with the placement of Wnt signaling upstream of the limb patterning program in a negative regulatory role.