Regulated cell polarity defines long bone morphology

We are interested in how tissue architecture contributes to embryonic growth. As a model system we use the growth plate cartilage, a tissue that resides at the ends of long bones and that directly determines growth rate in the embryo and in early post-natal life. In the growth plate, Resting progenitor cells are progressively recruited to a proliferative phase in which cells undergo up to six rounds of amplifying cell divisions. Proliferative chondrocytes then initiate cell cycle exit becoming prehypertrophic chondrocytes before ceasing cell division and swelling as hypertrophic chondrocytes. It is the process of hypertrophy that determines the growth rate of long bones. The transition to proliferative chondrocyte involves a morphological change in which the round, disorganized resting chondrocytes flatten and arrange in columns like stacks of coins. We asked what is the physiological role of column formation in the growth plate cartilage. As a first step, we sought to determine the molecular mechanisms that regulate column formation. Previous work by Dodds (1930) suggested that columns are formed by clonal expansion of progenitor cells
through a process involving daughter cell rearrangement following cell division. We tested this model using the modern methods of lineage analysis of cells labeled with replication defective avian retroviruses and of three-dimensional imaging of labeled cells to calculate the angle of the spindle relative to the long axis of the cartilage. Our work confirmed Dodd’s model and suggested that proliferative chondrocytes have polarity that regulates spindle orientation and either the same or different polarity pathway regulates cell intercalation.
We next sought to determine if known signaling pathways regulated polarity in proliferative chondrocytes. We found that disrupting frizzed signaling in chick long bones via retrovirus-mediated expression of full-length or dominant-negative frizzled receptors specifically affected the alignment and orientation of proliferative chondrocyes. These effects were phenocopied by altered signaling through the noncanonical frizzled regulated polarity pathways (by Vangl2 or dominant-negative Rock) by not by interfering with the canonical β-catenin regulated pathway (data not shown). One potential caveat is if expression of the receptors interferes globally with the signaling network known to be required for maintenance of the growth plate. However, injection of low titer retrovirus produces small patches of infected cells that display misorientation even when surrounded by phenotypically wildtype cells. These data initially suggested that both proliferative chondrocyte cell polarity and column formation, requires cell autonomous noncanonical frizzled signaling. However, we discovered through the analysis of Piga mutant mice that these two processes were mechanistically separable. Piga mutant mice lack gpi-anchored cell surface proteins on chondrocytes. We showed that Piga mutant chondrocytes undergo normal maturation, and have normal cell proliferation and cell death profiles. Yet, the cells fail to flatten or arrange in columns. These defects are not the result of improper orientation of the division plane, but rather lineage analysis demonstrates that Piga mutant proliferative chondrocytes fail to intercalate following cell division. Thus, there are different molecular requirements for orienting the division plane and cell
intercalation.

Why are columns required? One common aspect of the chick and mouse phenotypes is decreased longitudinal growth and concomitant increased lateral growth. This phenotype is similar to defects in convergent extension cell movements that promote tissue elongation during embryogenesis. Therefore, we propose that column formation serves to orient the vectors of long bone growth by generating a higher cell density in the longitudinal axis than the lateral axis and thereby maximizing elongation following cell hypertrophy.

No relevant conflicts of interest declared.