Our first case study examining the degree to which cell division is required for posterior growth is the branchiopod crustacean Thamnocephalus platyurus. Thamnocephalus hatch as free-swimming larvae with three pairs of head appendages and an undifferentiated trunk. Sequential segment addition and progressive differentiation gradually produce the adult morphology of eleven limb-bearing thoracic segments and eight abdominal segments, the first two of which are fused to form the genital region (Figure 3.2; Linder, 1941; Anderson, 1967; Freyer, 1983; Mpller et ah, 2003). Precise staging of developmental cohorts allowed for quantification of the changing dimensions of the growth zone as well as cell cycling behaviors (Constantinou et ah, 2020).
Growth Zone Mitoses Are Few with More Extensive Cell Cycling in the Segmented Region
In Thamnocephalus, the initial growth zone has about 40% of the total tissue needed to add post-hatching trunk segments (14). During segment addition, the growth zone shrinks. Without mitosis, the initial growth zone would be quickly depleted, used up after adding four segments, despite the fact that the size of new segment anlage also decreases over development. However, mitosis does occur in the growth zone and, since the total cell number in the growth zone is quite large, we estimate that the population of cells in the initial hatchling growth zone need only divide approximately 1.5 times to produce the tissue necessary to make all segments (Constantinou et al., 2020). Consistent with this estimate, direct counts of mitosis (using Hoechst staining or an anti-phosphorylated histone 3 [pH3], which marks cells with mitotic chromosome condensation), show surprisingly little mitosis in the growth zone: less than 5% of the cells are in mitosis at any given stage (Constantinou et al., 2020). This shows that not only are there very few mitotic cells in the posterior but, given the relatively large numbers of cells in the growth zone, not much division may be required.
What then accounts for the remaining growth? When larvae are exposed to a nucleotide analog (5-ethynyl-2'-deoxyuridine, EdU, which is incorporated into cells during S-phase) for 30 minutes, few EdU positive cells are found in the growth zone; instead many more are detected in the regions anterior to the growth zone, or after segment specification. Thus, it is growth in regions of the larvae anterior to the growth zone—the already specified segments—and not the growth zone itself that accounts for most of the elongation (Constantinou et al., 2020).
FIGURE 3.2 Sequential addition of segments from the posterior region in larval fairy shrimp. Thamnocephalus platyurus. A. Outlined drawings of successive one-hour time points indicating the addition of segments (as measured by Engrailed/Invected stripes) beginning at hatching to 4 hours post-hatching (graphics by P. Storrer). B. DAPI-stained larvae at hatching (top) and 4 hours post-hatching (bottom), and C. SEM of older larva showing in more detail segments beginning to undergo morphogenesis.
Growth Is Regulated through Both Synchronization of S-Phase Domains and Slowing the Rate of S-Phase
In Drosophila, discrete clusters of cells specified for common cell fates undergo mitosis together, resulting in repeatable, progressive patterns of mitosis (Foe, 1989). We, and others, had anticipated finding similar mitotic domains of some sort in sequentially segmenting arthropods, but as of yet, have failed to find domains as intricate as the 25 distinct mitotic domains described in Drosophila (Nagy et al., 1994; Rosenberg et ah, 2014) or failed to find them at all (Brown et ah, 1994; Liu and Kaufman, 2009). Interestingly, EdU incorporation in Thamnocephalus larvae unexpectedly revealed discrete S-phase domains, demonstrating a spatial coordination in cell cycling not captured by examining mitosis (Figure 3.3). The cells in the most posterior growth zone are variably in S-phase. Anterior to this, two stable
FIGURE 3.3 Spatially discrete domains of EdU incorporation in the posterior of Thamnocephalus larvae indicate coordinated cell cycling. A. An early larval stage showing band of EdU cells in the newest specified segment (flanked by En/In stripes, red arrows). This band is followed by a region in the anterior GZ of with very few EdU incorporating cells and then a region in the posterior GZ with scattered EdU incorporating cells. B. Similar larvae showing that the EdU band (green) excludes cells in M-phase (pH3 staining, pink). (After Constantinou et al., 2020.)
S-phase domains are associated with segmentation: a band of cells not undergoing S-phase in the anterior growth zone and a synchronized band of cells undergoing S-phase in the most recently specified segment. The stable domains indicate a synchronized transition into S-phase in the newly specified segments. In addition, during a 30 minute exposure to EdU, cells in the growth zone undergoing DNA replication incorporate less EdU than replicating cells in the segmented region of the larvae. This suggests that the rate of cell division in the posterior is regulated, at least in part, by controlling the rate of DNA synthesis, in this case by a slower the cell cycle.
Wnt/Caudal Cene Expression Are Linked to the Temporal Control of S-Phase Strikingly, the S-phase domains correlate with expression domains of specific Wnt paralogs (Figure 3.4): WntA mRNA expression is restricted to the anterior growth zone and Wnt4 mRNA is expressed only in the posterior growth zone (Constantinou et al., 2016). The boundary between WntA and Wnt4 corresponds to the posterior boundary of no EdU incorporation (Constantinou et al., 2020). The broad posterior expression domain of Wnt4 unfortunately does not provide clues to understand which 5% of the growth zone cells are given the cue to enter either S- or M-phase. Additionally, the anterior GZ boundary where cells transition to synchronized S-phase corresponds to the anterior boundary of caudal expression. Wnt signaling functions via caudal to integrate patterning, metabolism, and cell division in the posterior growth zone in other arthropods (Oberhofer et al., 2014; reviewed in Williams and Nagy, 2017). These data suggest distinct roles for Wnt paralogs in regulating the cell cycle within the growth zone.
FIGURE 3.4 Expression of segment patterning genes in the growth zone of Thamnocephalus coincides with domains of cell cycling. A. caudal expression extends throughout the GZ to the border of the newest segment and abuts the band of EdU incorporating cells. B. WntA extends in a gradient through the anterior GZ, while C. Wnt4 extends in a gradient throughout the posterior GZ. In all the photos, the red arrowhead marks the boundary between the anterior GZ and the newest segment. (From Constantinou et al., 2020.)
Thus, the emerging model of the growth zone from Thamnocephalus, is scattered, infrequent cell division in the posterior growth zone; rare cell division in the anterior growth zone; and high numbers of cell division in newly formed segments (Figure 3.7A). Actual mitoses in the growth zone are limited. Not only are cells of the growth zone not generating a large amount of additional tissue, they are also cycling more slowly than cells in the already specified segments. Furthermore, the cell cycle in the growth zone is highly regulated both temporally and spatially. By the time cells reach the anterior growth zone, S-phase is completed (cells rarely incorporate EdU in the anterior growth zone); immediately afterward, they synchronize to cycle together in the newly specified segment (all cells are in S-phase; Constantinou et al., 2020). These cell cycle domains correlate with expression of specific Wnt paralogs. This coincident mapping of cell cycling and gene expression domains is consistent with the idea of a posterior signaling region where cells are maintained in a multipo- tent state until they transit anteriorly and, under new regulatory signals, move toward segment specification.