Programming Pluripotent Precursor Cells Derived from Xenopus Embryos to Generate Specific Tissues and Organs

Xenopus embryos provide a rich source of pluripotent cells that can be differentiated into functional organs. Since the molecular principles of vertebrate organogenesis appear to be conserved between Xenopus and mammals, this system can provide useful guidelines for the directional manipulation of human embryonic stem cells. Pluripotent Xenopus cells can be easily isolated from the animal pole of blastula stage Xenopus embryos. These so called “animal cap” cells represent prospective ectodermal cells, but give rise to endodermal, mesodermal and neuro-ectodermal derivatives if treated with the appropriate factors. These factors include evolutionary conserved modulators of the key developmental signal transduction pathways that can be supplied either by mRNA microinjection or direct application of recombinant proteins. This relatively simple system has added to our understanding of pancreas, liver, kidney, eye and heart development. In particular, recent studies have used animal cap cells to generate ectopic eyes and hearts, setting the stage for future work aimed at programming pluripotent cells for regenerative medicine.


Introduction
Amphibians have traditionally been used as model systems for vertebrate embryogenesis. They are particularly suited for the study of early embryonic patterning and induction events. In the 1920s, Hans Spemann and Hilde Mangold demonstrated that transplantation of a small piece of dorsal mesoderm (the "Spemann organizer") into the ventral epidermis of a recipient embryo results in the formation of a complete secondary body axis [1]. This epochal work inspired generations of developmental biologists. However, the molecular control of this phenomenon has only been determined in the last two decades [2].
During the late 1960s, Peter Nieuwkoop and colleagues discovered that ectodermal cells can differentiate into mesodermal tissue if combined with prospective endodermal cells of the vegetal pole [3]. These so-called animal cap cells (ACC) can be isolated from the blastocoel roof of a blastula stage Xenopus embryo. If ACC are cultured without further manipulation they develop into "atypical epidermis". Combination of ACC with vegetal cells generates dorsal, intermediate or ventral mesodermal tissue, depending on whether the vegetal fragment was taken from the dorsal or ventral side of the embryo [4]. Subsequent work demonstrated that ACC can also differentiate into neural or endodermal tissue. Neural induction can be achieved by inhibition of TGFß-signaling, thereby defining the neural fate as a default state [5], or by application of FGF [6]. Finally, ACC can also be turned into endoderm: In a cell-autonomous manner, VegT, a vegetally localizing, maternal T-box transcription factor, or its downstream transcription factor Mixer, can induce the expression of endodermal marker genes [7][8][9]. All of these studies using manipulated ACC were primarily based on the analysis of the activation of germ layer-/tissue-specific marker gene expression, as analyzed by either RT-PCR or whole mount in situ hybridization based techniques.
Taking the use of the ACC system a step further, various higher order organotypic structures were generated employing a two-step protocol. First, ACC were programmed using different concentrations/combinations of regulatory molecules as outlined above. While the programming protocols utilized in these studies were mostly empirical, others studies have aimed to recapitulate the genetic program that controls in vivo organogenesis. Second, programmed ACC were transplanted into Xenopus embryos where they generated organ specific cell types or functional organs (see also the following paragraphs on individual organ systems, and for a recent review [10]). Striking examples are the generation of beating hearts or functional eyes from Xenopus ACC (see also below). Thus, the discovery of the ACC as a source for pluripotent precursor cells, which can be directed to differentiate into a broad variety of specialized cell types, has paved the road for their use in the generation of ectopic, functional organs in vivo.

The Experimental Basis
The pluripotent cells of the blastocoel roof, the ACC, are organized in several ectodermal cell layers, normally fated to develop into epidermis. They can be cultivated in simple saline solution for up to one or two weeks [11], corresponding to tadpole stages of development, when the primary organs have been formed. Manipulation of the corresponding explants can be performed in various ways: a) ACC can be exposed to extracellular signaling molecules like growth factors or chemical agents affecting signal transduction. Recombinant growth factors can be applied for short intervals or throughout the incubation time. Alternatively, growth factors can be produced by the ACC themselves. To this end, fertilized embryos are injected with mRNA or DNA coding for the respective growth factor prior to dissection of the ACC. b) Intracellularly active proteins, such as transcription factors, signal transduction proteins, etc. can also be supplied by microinjection of the corresponding mRNAs. As a variation of this procedure, microinjected explants can be sandwiched with naïve ACC, allowing for the distinction of cell-autonomous and non-cell-autonomous effects. c) The developmental potential of specific tissues can be analyzed by heterotopic "Sandwich" assays combining ACC with tissue explants from the same or different embryonic sources. d) Manipulated ACC can be transplanted into recipient embryos to analyze if they can form functional ectopic organs.
These different experimental protocols have not only been applied to the analysis of tissue/organ formation, but quite extensively also to problems relating to signal transduction mechanisms, cell adhesion phenomena and more. In addition to the activities employed in the pioneering studies mentioned above, various other factors, including secreted modulators and transcription factors, have successfully been used to direct ACC development. As outlined below, activin and retinoic acid, as well as the BMP-inhibitor noggin, have proven particularly useful to generate a broad range of tissue types from ACC. Furthermore, transcription factors can not only be used to generate a germ layer specific ground state (such as VegT for the endoderm), but combinations of transcription factors can drive organ specific development with high efficiency and specificity. An example of this is the generation of functional eyes by the eye field transcription factors. These and further examples of use of the ACC to induce different organs will be summarized in the following chapter.

Pancreas and Liver Development
The vertebrate pancreas and liver are organs that play a key role in energy metabolism and homeostasis. They both originate from the anterior foregut endoderm, where the pancreas derives from a dorsal and a ventral anlage. The liver and ventral pancreas originate from a common population of precursor cells. High levels of FGF and BMP signaling from the adjacent mesodermal tissue induce the liver specific gene program and suppress pancreas specification. Additionally, retinoic acid (RA) signaling has a conserved, essential role for pancreas development (for a recent review, see [12]). The first success in the generation of pancreatic tissue from ACC came from the use of a two-step protocol, where the explants were first treated with activin and after its removal with RA [13]. The pancreatic tissue generated contained both endocrine (insulin and glucagon positive cells) as well as exocrine (acinus-like structures) components. In an attempt to more closely recapitulate early developmental control of pancreas formation, ACC were subsequently programmed with VegT, the key maternal transcription factor for formation of the endodermal germ layer. VegT alone was sufficient to induce the onset of a liver specific gene program, but addition of RA to the system proved essential and sufficient to induce both endocrine and exocrine pancreas-specific gene activity ( Figure 1A, B). Additional application of a BMP-inhibitor was found to inhibit liver gene expression, thereby increasing the yield of pancreatic cells, which is consistent with the in vivo signaling situation [14][15][16] ( Figure 1A, B). Indeed, at the appropriate concentration of RA, the majority of cells exhibited either endo-or exocrine pancreatic character ( Figure 1C). Using Xenopus embryo microinjection, Afelik et al. [17] were able to demonstrate that ectopic expression of a combination of only two transcription factors, namely Pdx1 and Ptf1a, is sufficient to transform posterior endoderm into pancreas. Attempts to circumvent the need for RA in respect to pancreas induction in ACC, by applying VegT, Pdx1 and Ptf1a, were unsuccessful so far [18].  Due to the obvious medical interest in the generation of insulin-producing pancreatic ß-cells in vitro, numerous attempts using mammalian embryonic stem cells have been described in the literature (for recent reviews see [19,20]). Protocols for the generation of ß-cells from mammalian ES cells or from induced pluripotent stem cells (iPSCs), designed with the idea of recapitulating the relevant regulatory mechanisms as they operate during embryogenesis, have incorporated elements of the experiments using Xenopus ACC. In addition to other agents, including small molecules, activin and RA were successfully employed in the generation of pancreatic tissue from murine ES cells in vitro [21].

Kidney Development
The pronephros constitutes the primary secretory organ of Xenopus embryos. It is a derivative of the intermediate mesoderm and requires inducing signals to specify the pronephric anlagen during early gastrula stages of development. At tailbud stage, the developing pronephros becomes morphologically distinguishable and turns into a functional secretory organ (reviewed in [22]).
It was again Asashima and colleagues who first reported that pronephric structures can be generated from ACC in vitro. Histological analysis showed that ACC treated with a combination of activin and RA developed into pronephric tubules at a high frequency [23]. With the advent of pronephros-specific antibodies and molecular probes for pronephros-specific transcription activity, these initial observations were confirmed by several additional studies [24][25][26]. Strikingly, transplantation of these programmed ACC into Xenopus embryos, which had undergone bilateral pronephrectomy, suggests that they develop into functional secretory units. Although pronephrectomized embryos exhibited edema formation at high frequency and died after 7-9 days of surgery, transplanted animals formed less edema and had an increased survival rate [27].
A recent, elegant study, using a "sandwich" explant system including programmed ACC, identified a prime candidate signaling molecule for pronephros induction [28]. ACC were programmed with Wnt11 or Wnt11b by means of mRNA microinjection and used to form conjugate explants with intermediate mesoderm dissected from early gastrula stage Xenopus embryos. The programmed ACC or the intermediate mesoderm cultured in isolation did not reveal any significant indication of pronephros formation; however, their combination ("Holtfreter sandwich") resulted in the formation of pronephric structures at high frequency. This procedure takes in vitro organ formation a step further by mimicking aspects of the often very complex tissue interactions.

Heart Development
The development of the heart is highly conserved among vertebrates. In Xenopus, heart development is initiated at the onset of gastrulation and requires an interplay of signals from the dorsal lip and the underlying endoderm. Initially, two patches of the mesoderm on the dorsal side of the embryo are specified towards cardiac fate. During gastrulation, these heart progenitors move dorso-anteriorly and continue to migrate ventrally at neurula stages. After meeting at the ventral midline, they fuse and form the linear heart tube, which then undergoes the looping and remodeling processes of cardiac morphogenesis.
Analyzing the precise molecular events that determine cardiac fate and ultimately formation of the heart is a topic of intense ongoing research. Xenopus provides a unique combination of straightforward gain-and loss-of-function analyses with a remarkable ability to heal after microsurgery, which has enabled Xenopus to become one of the favorite model systems for studying cardiogenesis [29,30]. In a landmark study, Horst Grunz exploited these advantages and generated beating heart muscle tissue in vitro by treating dorsal lip explants with suramin, a polyanionic compound [31]. Transplantation of these suramin-treated explants into the posterior trunk area of early tailbud embryos led to the formation of ectopic beating heart structures [32]. These experiments set the stage for Xenopus explants as valuable tools for the molecular analysis of cardiac induction and differentiation.
In particular, ACC have proven to be a powerful and resourceful system to analyze the regulation of cardiogenesis. Using ACC, numerous transcription factors like GATA4, members of the Sox family or secreted signaling molecules like Wnt11 have been shown to induce cardiac marker expression [33][34][35]. The frequent use of ACC for the analysis of cardiogenesis started with the observation that activin-treatment of newt ACC induced a beating heart-like phenotype [36]. Asashima and colleagues succeeded in developing a method that induced cardiomyocytes at a high frequency [37] (Figure 2). ACC were dissociated by the removal of calcium and reassociated using a calcium-and activin-containing medium (Figure 2A). Reaggregated cells began beating at the same time as heart beating is initiated in embryos. In order to analyze if the in vitro-generated myocardial cells were functional, activin-treated ACC reaggregates were implanted into the posterior abdomen of neurula stage control embryos. Ectopic beating hearts were generated ( Figure 2B-D), which consisted of at least two chambers and were involved in the systemic circulation. Since the first usage of ACC by Asashima and colleagues, this assay system has been extensively used as well as further modified. It has been demonstrated that injection of femtomolar concentrations of activin-encoding mRNA can replace the dissociation step [38]. Further, conjugation of ACC with anterior endoderm has also been shown to lead to cardiac induction, thus allowing analysis of the molecular contribution of endoderm to this process. Using this strategy, Samuel et al. [39] defined the time windows where FGF, Nodal and Wnt/ß-catenin signaling affect cardiogenesis. Their data demonstrate that FGF and Nodal signaling are required for cardiogenesis, while Wnt/ß-catenin signaling antagonizes cardiac differentiation. Thus, ACC proved to be a valuable tool for characterizing the molecular events involved in cardiac specification. Further use of this system for research may contribute to an understanding of congenital heart disorders and to developing therapies for cardiac disease.

Eye Development
The vertebrate eye detects incoming light and converts it into electrical signals; this strikingly complex structure poses particular challenges for regenerative medicine. Not only are eye tissues, like the retina, composed of an intricate array of different cell types, but embryonic eye development is also achieved by a sequential induction process that is difficult to recapitulate in vitro. Asahima and colleagues succeeded in inducing eye development in vitro at a high frequency by combining animal caps with dorsal lip and lateral marginal zone cells [40]. The in vitro generated eyes were morphologically similar to normal eyes and extended and connected to the optic tectum if transplanted into host embryos. Furthermore, if the in vitro induced eyes were transplanted into tadpole embryos, where both endogenous eyes had been removed, the resulting frogs could adapt their skin color to the color intensity of their background, indicating that the eyes were functional. These experiments showed that pluripotent ACC can be used to reproduce eye development in vitro and set the stage for the identification of the molecular factors involved in this process. The generation of retinal precursor cells is of particular interest, as it could provide an opportunity to treat retinal injuries or degeneration. Recently, factors like the BMP inhibitor noggin, or a set of genes commonly called the eye field transcription factors, have been shown to be able to drive pluripotent ACC towards a retinal fate. Lan et al. [41] showed that high concentrations of noggin, a secreted molecule that generates neural tissue by inhibition of BMP signaling, is sufficient to elicit retinal fates in animal caps. Noggin-injected animal caps cultured to tadpole stages at which embryonic retinogenesis is almost complete, expressed terminal differentiation markers of specific retinal cell types. Interestingly, the cell distribution was reminiscent of the retina lamination suggesting that noggin-injected animal caps display a basic retinal patterning. Furthermore, if the presumptive eye field of an embryo was replaced with an isochronic animal cap expressing a high concentration of noggin, the transplanted ACC formed a complete eye with a lens. Moreover, light-induced c-fos expression and electrophysiological characterization of the single-rod photoreceptors strongly suggest the functionality of the animal-cap-derived eye.
An inspiring recent study demonstrated that ACC overexpressing a combination of seven transcription factors, the so-called eye field transcription factors (EFTF) and the neural patterning gene otx2, differentiate into all retinal cell classes and form a neuronal network necessary for vision [42]. Using an extended microarray analysis Viczian et al. showed that EFTF-expressing pluripotent ACC share a common transcriptional profile with the eye field. EFTF-expressing cells gave rise to ectopic eyes when they were transplanted into the flanks of a control embryo ( Figure 3A-D), indicating that they are determined to a retinal fate. However, electrophysiological characterization was not possible, because these eyes were small and could not connect to their normal tectal targets in this ectopic location. Thus, the authors grafted EFTF-expressing animal caps into the eye field of host embryos ( Figure 3E-J). These transplanted ACC generated all seven classes of differentiated retinal cells observed in the normal retina and showed a stereotypical retinal cell morphology, suggesting that these eyes are functionally equivalent to normal eyes. This was supported by the finding that electroretinograms of eyes, exclusively derived from EFTF-expressing ACC, were similar to those of eyes from the unoperated side of the same animal. The authors even went one step further and used phototropic behavior tests to confirm eye functionality. Interestingly, tadpoles with induced eyes showed the same vision-dependent preference towards a white background as normal embryos, confirming that these eyes originating from manipulated ACC were indeed functional.
Taken together, these data show that pluripotent ACC can form diverse retinal cell classes and even functional eyes, if treated with the appropriate combination of factors. Thus, these studies open the way to characterize the role of various intrinsic and extrinsic factors and their combinations for their ability to direct pluripotent ACC to a retinal fate. Future challenges will be to determine how to maintain cultures of retinal precursor cells and to analyze if they could be used for retinal repair of damaged or diseased eye tissue.

Conclusions
The Xenopus ACC system has already a long and successful tradition, but it is expected that it will further benefit from recent genetic and genomic advances. The use of mutant and transgenic Xenopus lines, which are increasingly available, remain to be exploited in this context. For example, the generation of transgenes carrying inducible versions of transcription factors would allow temporal control of their activation. Thus, more refined, multistep differentiation protocols could be developed. Once Xenopus lines with mutations in disease-associated genes are available, in vitro organ generation from such animals may be suitable for drug development. Moreover, advanced technologies that are based upon genomic sequence information, like microarray analysis and proteomics, are now readily available. Organ-specific tissues generated from manipulated ACC making use of wild-type as well as mutant Xenopus lines hold the promise of the discovery of novel genes and a more complete definition of the gene networks that govern organ development in vertebrates.
Several of the experiments outlined here were designed with the strategy to recapitulate key regulatory steps in organ development as they are known from embryological studies, others were equally successful using a more empirical approach. The most impressive example for the latter scenario is provided by the pioneering studies of the Asashima group, where different concentrations of RA in combination with activin were found to be sufficient to generate a whole variety of different organs and tissues [10]. These include pancreatic tissue, where RA has a conserved, essential signaling function during vertebrate embryogenesis. While this may also be the case for the development of the kidney [43], this is not clear yet for other organs. The use of combinations of eye field specific transcription factors is certainly the most rational strategy of mimicking the highly specialized situation characteristic of a given group of organ precursor cells. However, it remains to be shown if this strategy can also be applied to other types of organ precursor cells.
Xenopus ACC define a source of pluripotent vertebrate precursor cells; their versatility is comparable to other populations of pluripotent precursor cells, such as mammalian IPS cells. A hope that is connected to the IPS system is that these pluripotent cells could be used to engineer patient-specific tissue or even organs for regenerative medicine. In general, the underlying molecular principles in vertebrate organogenesis appear to be highly conserved. Therefore, protocols established in the very robust-and therefore easy to handle-ACC system can serve as guidelines for the directional manipulation of mammalian embryonic stem cells.
Thus, looking back on half a century of exciting research with Xenopus animal cap explants makes us optimistic for another 50 years of groundbreaking insights into the secrets of organ development.