Gradient expectations: revisiting Charles Manning Child’s theory of metabolic regionalization in developmental patterning and regeneration

Abstract

Charles Manning Child introduced one of several early models to explain how an organism can both establish and reestablish positional identity during embryogenesis and regeneration. In his gradient theory model, tissues along an axis exhibit graded levels of metabolic activity, demonstrated through their differential susceptibility to metabolic inhibitors. While Child’s work was difficult to place in a mechanistic framework in his own time, technological advances and recent discoveries in both embryos and regenerating organisms make his early work on redox signaling as a positional cue newly pertinent.

Graphical Abstract

Graphical Abstract 1: Model of gradient identification via susceptibility. A target animal or tissue, such as a planaria, is exposed to a targeted inhibitor. Here, potassium cyanide, which inhibits mitochondrial respiration, is schematized. Necrosis is assayed periodically. Tissues which die early (gray) are considered to have greater activity of the targeted process. In planaria, inhibition of respiration first leads to death of anterior structures and, after more time has passed, posterior structures. This allows the conclusion that there are 2 gradients for mitochondrial respiration, a stronger anterior based gradient and a weaker posterior based one.

Introduction

In the first years of the 20th century, Charles Manning Child, then a junior faculty member at the University of Chicago, was among several prominent developmental biologists captivated by the puzzle of morphogenesis. At the forefront of his attention was the remarkable capacity for certain invertebrates, having lost the structures at one end of their primary axis, to reliably regenerate only those missing structures. Child inferred that the fidelity of regeneration arose because tissues along an axis exhibited a graded property. If a portion of the tissue was removed, the gradient remained and could be re-calibrated to reestablish the value of the missing structure. Through a series of careful experiments, Child and his contemporaries found that the anterior- and posterior-most tissues of animals from planaria to hydrozoans were especially sensitive to metabolic inhibition. He concluded that at least one aspect of axial gradation was metabolic activity¹.

In his own time, Child’s elucidation of the gradient theory and its articulation in diverse animals was regarded as an important addition to developmental biology. He was a successful mentor and a prolific scholar, writing several books in addition to his numerous articles. Many of the latter were published in the journal Physiological Zoology, which Child established, and which continues to this day as Physiological and Biochemical Zoology. He was an active contributor to the ongoing debate regarding preformation (the idea that the fertilized egg contains within it all the determinants of its eventual form, which simply unfold as the animal develops) versus epigenesis (the idea that an animal’s form emerges gradually, in layers, from the unformed egg or zygote, and is vulnerable to external influences)². This debate put Child somewhat at odds with other leaders in developmental biology and genetics, most notably Thomas Hunt Morgan.

Child, like Morgan, was struck by the parallels between the embryonic development of a structure and its regeneration⁷. Indeed, Child considered regeneration to be a subtype of development¹. Despite their shared fundamental interest in the process of regeneration, mechanistic conceptions of development, and therefore regeneration, differed substantially between Child and Morgan. By providing a mechanistic explanation for chromosome theory and Mendelian inheritance, Morgan was able to bring these ideas, which he himself had regarded skeptically as too preformationist, into a contemporary cell biological framework³. Morgan’s new theory of genetic inheritance quickly gained popularity and independent experimental support, bringing it to the fore. But Child continued to find strict preformation understandably difficult to reconcile with his own work demonstrating the plasticity of tissues to external cues. Morgan’s mechanistic model of gametic inheritance also did not encompass the non-gametic forms of reproduction that many of Child’s models utilized, such as binary fission or budding. He also was very much aware of discoveries such as the principles of induction (articulated by Ethel Browne and Hilde Mangold), which showed clearly that cells could be induced to take on different fates through the influence of their neighbors. These musings formed the basis of some of his later writings, particularly in his book Patterns and Problems in Development¹. By the mid-20th century, Child’s conception of graded tissue identity had fallen into the background, as the mechanisms of genetic inheritance became more and more clearly articulated⁴. But, as new concepts such as morphogen gradients and epigenetic inheritance have been added to our routine conception of cell identity, Child’s work no longer seems difficult to reconcile with Morgan’s. Child’s theory of physiological gradients, a term adopted by Child and his contemporaries to be more inclusive of not only metabolic gradients, but also those of cellular and environmental properties^5,6, has been given new breath in recent years as tools to characterize and functionally test metabolic activity have improved. Present day researchers have begun to elucidate how various physiological gradients instruct cell differentiation and tissue patterning. Considering Child’s work in light of recent findings in both regeneration and development, here we draw attention to major advances in genetic and metabolic assays that have influenced how we can study cell differentiation and regionalization of tissues.

Charles Manning Child

Charles was born in Ypsilanti Michigan in 1869⁴. His mother, Mary Elizabeth Manning, was the daughter of a physician and his father, Charles Chauncey Child, was a descendant of a long line of shipbuilders⁸. Young Charles Manning Child was raised in Higganum, Connecticut, where the family’s shipbuilding business had been based. As a child, Charles was interested in reading and nature; he collected and studied plants and minerals using his mother’s microscope, which she had acquired from the highly influential natural historian Louis Agassiz⁹. Child’s interests in natural science developed through his high school years and he studied both chemistry and zoology following his enrollment at Wesleyan University in 1886, eventually settling on zoology. He received a bachelor’s and a master’s degree from Wesleyan in 1890 and 1892, respectively. Upon leaving Wesleyan, he traveled to Germany, where he completed his PhD degree with Rudolf Leuckart in 1894⁹. His scholarly work during these early years in Germany focused on several aspects of invertebrate nervous system development and morphology. In 1895 he returned to the United States, becoming part of the zoological staff of the new University of Chicago, having been recruited by Charles Otis Whitman. He served first as a zoological assistant, then associate, from 1895–1898, and taught in the Embryology course at Woods Hole during the summers of 1895 and 1896. Child was therefore a contemporary of several highly influential members of the Woods Hole Embryology faculty, including Whitman, Edward Conklin, and Frank Lillie, who were all interested in articulating cell lineages in order to understand the emergence of animal form⁹. Child also took on this problem, establishing early embryonic lineages in the annelid worm Arenicola¹⁰. By 1905, he was an assistant professor at U. Chicago and had turned his attention to regeneration in coelenterates, hydrozoans, and planaria, publishing several dozen articles from 1900 to 1915, and two books: Individuality in Organisms and Senescence and Rejuvenescence, both published in 1915.

Susceptibility: early identification of metabolic gradients

Child’s work on metabolic gradients began in earnest in 1915. He made use of a general paradigm, which he termed susceptibility, in which he would first inhibit a metabolic activity and then assess how quickly tissues succumbed to necrosis after that inhibition¹¹. Tissues which died more rapidly after treatment with the inhibitor were considered to be more dependent on the underlying process, and therefore to have more of that activity. By observing a single tissue or organism over time, regional differences in activity could be deduced (see graphical abstract). In one specific series of experiments, Child would take whole organisms, or remove segments of tissue from different axial positions, and place them in a solution of potassium cyanide, which was known at that time to be an inhibitor of oxidative cellular respiration. Child and his colleagues found that the susceptibility of tissues to cyanide existed as a gradient that correlated with its axial position. The most susceptible axial positions were at the anterior and posterior poles, gradually declining towards the mid-region (see graphical abstract)⁵. Further, some tissues such as the heart and eyes were much more susceptible to cyanide toxicity, suggesting they were more oxidatively active, as has been confirmed by recent studies in retinal pigment epithelium and cardiomyocytes, proving how mechanistically fruitful this early assay was^12–15.

Oxygen availability impacts differentiation and cell fate specification

Due to the availability of mitochondrial toxins, most of the gradient studies in Child’s era focused on oxygen gradients, identifying primarily anterior biased gradients as well as weaker secondary posterior gradients of activity in embryos^13,16. Initial studies by Child showing the importance of oxygen gradients were largely restricted to embryos that develop externally due to the nature of his experimental toolkit. Today, the use of dyes such as EF5 and pimonidazole, which form adducts with thiol groups specifically in hypoxic cells, now allow identification of regional differences in oxygen availability in tissues and enable us to define oxygen gradients in organisms that develop in utero^17,18. In placental mammals, embryos develop in a naturally hypoxic environment; that is, cells in the embryo have graded oxygen availability up to less than half the amount of oxygen in adult tissues^19,20. in vitro culture of rat embryos under varying oxygen conditions has shown that hypoxic conditions induce microcephaly and overall reduction in growth. The reduction in head size observed suggests that normal anterior development requires oxygen, in line with Child’s finding that there are anterior requirements for oxidative metabolism in invertebrates and Hyman’s similar observations in teleosts, while a reduced overall size of the embryo may suggest other regional requirements for oxygen, but these have not been thoroughly classified thus far^13,19. Hyperoxic conditions result in failure to close the neural tube and form somites, indicating that regulation of oxygen availability is crucial for developmental organization across multiple cell lineages and developmental axes¹⁹. These tools for visualization of oxygen levels have greatly improved our understanding of oxygen availability in healthy embryonic tissues and are critical to interpreting perturbations to oxygen homeostasis in development.

The importance of oxygen availability to differentiation has also been extensively studied in the context of cell culture of numerous cell lines. Embryonic stem cells (ESCs) remain quiescent when incubated under hypoxic conditions but begin to proliferate as oxygen availability increases²¹. Varying oxygen levels also has dramatic effects on cell fate even along a single trajectory. ESCs cultured in hypoxia are less likely to adopt the neural stem cell (NSC) lineage, though further differentiation of NSCs to neurons is actually promoted under low oxygen conditions^22,23. Because the partial pressure of oxygen varies across tissues, it is difficult to generalize principles of oxygen availability on differentiation, though modulation of oxygen concentration has become a useful experimental variable for manipulating progenitor differentiation and proliferation^21,24. The use of cell culture allows more regulated control of oxygen availability in cellular monolayers and has provided unique insight into regulation of cell fates by local oxygen signals which could translate to better understanding of differentiation in vivo.

Physiological gradients in vertebrates are required for development and regeneration

Metabolic pathways have also recently re-emerged as developmental regulators. One major advancement in this vein is the identification of posterior biased glycolytic gradients, which have been described in both mammalian and chick embryogenesis^25,26. The central observation of this work is that transcriptional gradients for glycolytic enzymes exist in the presomitic mesoderm (PSM). Notably, rate-regulating enzymes for glycolysis, such as phosphofructokinase (pfkp) and pyruvate kinase (pkm), are enriched in the posterior extreme of mouse and chick embryos and gradually decline in expression in the anterior PSM and somites^25,26. To characterize the function of these transcriptional gradients, mass spectrometry and pharmacological perturbations were leveraged. In mice, ¹³C-isotope tracing found that the posterior PSM was indeed heavily glycolytic relative to the anterior PSM as evidenced by increase in labeled glycolytic intermediates²⁵. Unlabeled mass spectrometry comparing somites to anterior and posterior PSM also revealed a greater amount of glycolytic intermediates in the posterior PSM, in line with transcript based readouts²⁶. In both systems, the posterior region was shown to have significant increases in lactate production, indicative of a Warburg-like metabolic state favoring aerobic glycolysis. Upon addition of 2-deoxyglucose, a non-hydrolysable glucose mimic, or substitution of glucose for its downstream metabolic pyruvate, somitogenesis and elongation of the primary axis were disrupted^25,26. These studies were able to show that the early reactions of glycolysis are required for PSM elongation and segmentation.

More recently, carbon metabolism has been explored in the context of regeneration using zebrafish larval tails as a model. Quickly after injury, the regenerating tail blastema takes up a large degree of glucose and the mitochondria in this structure exhibit a fragmented phenotype, indicating that the cells in this structure are undergoing aerobic glycolysis²⁷. This injury-induced metabolic shift is required for proper activation of TGF® signaling and blastema formation, highlighting a critical role in signal transduction and regeneration. This causal relationship between metabolic flux and developmental growth and patterning is a logical progression of Child’s initial theories, merging advances in transcriptomic, pharmacological, and metabolic analysis from the last century to test the importance of various physiological gradients in development and regeneration.

Oxygen regulation is critical for regeneration

Not only have metabolic pathways been shown to be critical in regeneration, but so too have the oxidative gradients initially described by Child. A major interest in tissue regeneration has been the role of injury induced reactive oxygen signals (ROS) in facilitating wound response^28–33. In most of these regenerative contexts, failure to induce ROS leads to reduced or failed regeneration and can increase scar tissue formation. As more attention has been paid to transcriptional regulation in regeneration, we now have more insight into how oxygen metabolism in the form of ROS facilitate regeneration. Much as Child’s early work in planaria showed increased activity of mitochondrial oxidative phosphorylation following injury, Pirotte and colleagues found that ROS accumulate at wound sites and that this ROS accumulation in required for proper regeneration of both anterior and posterior aspects of planaria³¹. Specifically, these injury-induced ROS are required for activation of differentiation programs in neoblasts and activation of neuronal gene signatures, suggesting that ROS are important for initiating transcriptional programs in regeneration. Similarly, in Xenopus ROS have been shown to be upstream of FGF signaling and in zebrafish they are required upstream of Hedgehog signaling, both pathways well-established regulators of cell fate and patterning^30,32. This body of work is actively growing and not only indicates that activation of physiological gradients is required for responding to injury stresses but suggests that these gradients could facilitate tissue patterning in regeneration as they do in development.

Recent technological advances in oxygen recording have enabled more careful observation of oxygen flux following injury. A study by Ferreira et al. using Xenopus laevis tapdoles, which are capable of whole tail regeneration, adapted an optrode system to measure changes in oxygen flux during regeneration²⁸. Following injury, there is a large influx of oxygen into the freshly amputated tissue and this influx is sustained even following wound epidermis formation, or closure of the open wound. Prior studies looking in the hours immediately following injury had also shown an increase in ROS at the wound site, highlighting the critical role of O₂ metabolism in regeneration³⁰. This study further identified a critical hypoxic microenvironment, generated by the ROS produced by increased oxygen flux, in regenerative animals but not in non-regenerative developmental stages, suggesting that this regionalized hypoxia is instructive in regeneration. While the local hypoxic niche is non-intuitively a result of increased oxygen flux into the regenerating tail, these results are in direct agreement with Child’s findings that regenerating tissues have increased oxygen demand based on increased susceptibility to mitochondrial toxins.

Conclusion

While Child’s work has become something of a historical footnote, recent work has highlighted previously understudied intersections between metabolic and transcriptional perspectives on biology that make his early insights into redox signaling newly relevant. As our understanding of and ability to study gene regulation deepens, it becomes clear that cell environment and behavior are critical components of transcriptional output. For developing and regenerating tissues, this reinforces that physiological contexts, including access to oxygen or availability of metabolites, are as important inputs to the establishment and maintenance of cell identity as the inherited genome.

The proliferation of new tools to query oxygen tension, redox state, and glycolytic activity make it an exciting time to pursue these variables in the development and regeneration of axial structures and specific organs. One could readily imagine directly revisiting Child’s gradient theory with these tools to address questions he was not able to fully resolve. Would we find, for example, that the highly oxidative tissues at the axial termini of tubularia are less glycolytic, due to their reliance on oxidative phosphorylation, thus representing a counterpoint to the highly glycolytic PSM of amniotes? Or would we find that these regions are just highly metabolically active in general, with elevated levels of both glycolysis and oxygen consumption? Similarly, now that we have learned that ROS are important signals to initiate regeneration in planaria and vertebrates, does the intensity and functional requirement for these signals parallel the axial gradients of oxidative activity that Child’s susceptibility hypothesis proposed^30–32? The experimental tractability of multiple models in which embryogenesis, regeneration, and homeostatic modulation of tissue identity can all be queried put present day researchers in the enviable position of being able to articulate conserved principles for these processes as well as those that are unique to particular organisms^34–37. As regional and tissue specific differences in these factors become easier to study, we are eager to learn how Child’s early theories of axial gradients integrate with our current understanding of developmental and regenerative patterning.

Abbreviations

ESC

embryonic stem cell

NSC

neural stem cell

PSM

presomitic mesoderm