Early development of the thymus in Xenopus laevis

Abstract

Background

Although Xenopus laevis has been a model of choice for comparative and developmental studies of the immune system, little is known about organogenesis of the thymus, a primary lymphoid organ in vertebrates. Here we examined the expression of three transcription factors that have been functionally associated with pharyngeal gland development, gcm2, hoxa3 and foxn1, and evaluated the neural crest contribution to thymus development.

Results

In most species Hoxa3 is expressed in the third pharyngeal pouch endoderm where it directs thymus formation. In Xenopus, the thymus primordium is derived from the second pharyngeal pouch endoderm, which is hoxa3-negative, suggesting that a different mechanism regulates thymus formation in frogs. Unlike other species foxn1 is not detected in the epithelium of the pharyngeal pouch in Xenopus, rather, its expression is initiated as thymic epithelial cell starts to differentiate and express MHC class II molecules. Using transplantation experiments we show that while neural crest cells populate the thymus primordia, they are not required for the specification and initial development of this organ or for T cell differentiation in frogs.

Conclusions

These studies provide novel information on early thymus development in Xenopus, and highlight a number of features that distinguish Xenopus from other organisms.

Introduction

Located in the thoracic cavity of most vertebrates the thymus is a primary lymphoid organ. It provides a unique microenvironment for T lymphocytes maturation an essential step in the development of the adaptive immune system (Miller, 1961). Failure to organize or to maintain a proper thymic structure can have serious health consequences ranging from immunodeficiency to autoimmunity.

The thymus is one of four glands, along with the parathyroid, thyroid and ultimobranchial bodies, that are derived from the foregut endoderm in the pharyngeal region. The pharyngeal apparatus is a transient structure composed of a series of bulges that form on either side of the developing head: the pharyngeal arches. Each pharyngeal arches is separated externally by ectoderm-lined pharyngeal clefts and internally by endoderm-lined pharyngeal pouches. Each pouch/cleft gives rise to a specific set of organs, although there is variation between species (reviewed in Grevellec and Tucker, 2010). In most organisms thymus development is tightly linked to formation of the parathyroid glands (reviewed in Blackburn and Manley, 2004; Rodewald, 2008; Gordon and Manley, 2011). The two organs develop from a single primordium arising bilaterally from an outpocketing of the third pharyngeal pouch endoderm. The epithelial cells of the common primordium are initially indistinguishable, and patterning of the third pharyngeal pouch endoderm into distinct domains becomes apparent with the expression of two transcription factors, Foxn1 and Gcm2, that have restricted and complementary expression patterns in the prospective thymus and parathyroid domains, respectively (Gordon et al., 2001). Foxn1-expressing cells will form the thymus epithelial cells (TEC), the major component of the thymic stroma required for all stages of thymocyte development and differentiation. Lineage and functional analyses in mice have shown unambiguously that TEC from both the cortex and the medulla are exclusively derived from the pharyngeal endoderm (Gordon et al., 2004).

Each thymic primordium is surrounded by mesenchyme derived from the neural crest (NC) that migrates from the posterior hindbrain into the pharyngeal region. In chicken and mouse these NC cells contribute to the connective tissue of the thymic capsule and become associated with the vasculature of the fetal and adult thymus (Le Lievre and Le Douarin, 1975; Le Douarin and Jotereau, 1975; Bockman and Kirby, 1984; Jiang et al., 2000, Foster et al., 2008; Muller et al., 2008). The development of a mature thymus depends on interactions between the NC-derived mesenchyme and the pharyngeal endoderm-derived TEC, NC cells stimulating proliferation and maturation of the epithelial compartment of the thymus (Itoi et al., 2007). Among the factors implicated are Fgf7 and Fgf10 produced by the NC mesenchyme, which promote both TEC proliferation and differentiation (Suniara et al., 2000; Revest et al., 2001; Dooley et al., 2007). The extent to which the NC participates in early thymus development is not as well described. In chicken embryos, NC ablation results in small thymi with delayed development and poor function (Bockman and Kirby, 1984). In the classic mouse mutants Splotch, carrying a deletion of Pax3, the embryos have a dramatically reduced number of NC cells in the vicinity of the third pharyngeal pouch. However, the thymus and parathyroid primordia are normally specified (Griffith et al., 2009). A more recent study looking at compound mutant embryos homozygous for a NC-specific Foxd3 mutation and heterozygous for Pax3 reported severe thymus hypoplasia (Nelms et al., 2011). Moreover, DiGeorge syndrome patients have a range of congenital defects involving structures influenced by the cardiac NC, including cardiac deficiencies and craniofacial abnormalities but they also show thymus hypoplasia leading to T cell deficiency and increased susceptibility to infections (reviewed in Wurdak et al., 2006).

The African clawed frog, Xenopus laevis, has been a model of choice for comparative and developmental studies of the immune system (reviewed in Du Pasquier et al., 1989; Robert and Ohta, 2009; Robert and Cohen, 2011). Its adaptive immune system in the adults is remarkably similar to that of mammals, and Xenopus has been an especially valuable system to analyze T cell ontogeny and explore self-tolerance (reviewed in Du Pasquier et al., 1989; Robert and Ohta, 2009). While a great deal has been learned about the development and maturation of the immune system in Xenopus, less is known about early thymus development. A few reports published in the late sixties briefly described thymus development in Xenopus (Nieuwkoop and Faber, 1967; Manning and Horton, 1969), however these studies have provided only limited visual information on thymus organogenesis. Here we present a histological and molecular characterization of the emergence and development of the thymic rudiments in Xenopus. In particular, we examined the expression of three transcription factors that have been associated with pharyngeal gland development, gcm2, hoxa3 and foxn1, and highlight some of the more unique features of these genes as they relate to thymus development in frogs. Moreover, using transplantation experiments we analyzed the NC contribution to early thymus development.

Results

Anatomy of the developing thymus

The thymus primordia were first morphologically recognizable at around stage 47 as two spherical masses on each side of the hindbrain. They were located dorsal-laterally between the developing inner ear and the eye (Fig 1A–F). This is the stage at which the thymus rudiment is initially colonized by lymphoid precursors, which express the cell surface marker, RC47 (Du Pasquier and Flajnik, 1990). At stage 50 the thymi had increased in size and they retained their dorsolateral position (Fig 1G–J), and expressed p63 (Fig 1K–L), a factor essential for the proliferative capacity of the epithelial stem cells of the thymus (Senoo et al., 2007). By stage 55 the thymus rudiments were extensively colonized by NC-derived pigment cells that reside in the cortex (Fig 1M–N), and the rudiments had shifted into a more superficial position, underneath the skin (Fig 1O–P). In a transverse section the cortex and medulla of the thymus were clearly visible (Fig 1P–Q). While morphologically recognizable at this stage, the architecture of the cortex and medulla can be identified as early as stage 48/49 with the differential expression of MHC class II molecules by TEC (Du Pasquier and Flajnik, 1990) and CTX, a cell surface marker, by cortical thymocytes (Robert and Cohen, 1998).

Figure 1. Developmental anatomy of the thymus.

(A–F) The thymus primordia are recognizable at around stage 47 (arrows). (C) Transverse section of a stage 47 tadpole, dorsal towards the top. The thymus primordia (arrows) are located lateral and ventral to the hindbrain (hb), between the eye and the developing inner ear (ie), as seen on a longitudinal section (D–F). (F) Higher magnification view of the right thymic primordium shown in panel (E). Longitudinal section of a stage 47 tadpole, anterior towards the top. (G–L) At stage 50 the thymus primordia have increased in size but retained their dorsolateral position (arrows). (G, I) Phase contrast. (H–J) Hematoxylin and eosin staining of the same section. (K–L) Expression of p63, an epithelial stem cell factor. Panels (I–J) are higher magnification view of panels (G–H). Panel (L) is a higher magnification view of panel (K). Transverse sections, dorsal towards the top. (M–N) At stage 55 the thymus has been extensively colonized by the NC-derived pigment cells (arrows). Panel (N) is a higher magnification view of the thymus shown in panel (M). Dorsal views, anterior toward the top. (O–Q) On a transverse section the cortex (cx) and medulla (md) of the thymus are now clearly distinguishable. The dashed circle outlines the medulla in panel (Q). Panels (P–Q) are higher magnification views of panel (O). Scale bars in (A and D) = 500 μm; (B, C, E, G, H, N and P) = 200 μm; (F, I, J, K and Q) = 100 μm; (M) = 2 mm; (O) = 1 mm.

Histological analysis of thymus organogenesis

We next analyzed initial thymus development in Xenopus using serial sections prepared from tadpoles at stage 41 through stage 45 (Fig 2). The transverse sections were made at a level posterior to the eye and anterior to the developing inner ear. Dorsal bilateral outpocketings of the second pharyngeal pouch endoderm were first visible at stage 41 (Fig 2). As development proceeds the thymus primordia became more rounded in shape but still connected to the pharyngeal pouch (stage 42 to stage 44; Fig 2). The connection with the epithelium of the pharyngeal pouch was completely lost by stage 45 (Fig 2). These findings are largely consistent with previous studies (Nieuwkoop and Faber, 1967; Manning and Horton, 1969).

Figure 2. Early development of the thymus primordia.

Histological sections illustrating the formation of the thymus primordium (arrows) from the pharyngeal pouch epithelium (arrowheads) in Xenopus tadpoles at stage 41, stage 42, stage 43, stage 44 and stage 45. For each stage both sides of the pharyngeal region are shown (lateral panels). The central panels are lower magnification views, composite picture of both sides. Dorsal bilateral outpocketings of the pharyngeal epithelium are first visible at stage 41. The thymus primordium is completely separated from the pharyngeal pouch by stage 45. Transverse sections through the hindbrain region at a level posterior to the eye and anterior to the developing inner ear, dorsal to top. The scale bars in the lateral and middle panels represent 100 μm and 200 μm, respectively.

Expression of gcm2, hoxa3 and foxn1 during thymus development

The three transcription factors Gcm2, Hoxa3 and Foxn1 have been functionally associated with pharyngeal gland development in mammals (reviewed in Manley and Condie, 2010). Here we analyzed the expression of these three genes as they relate to Xenopus thymus development.

gcm2

The vertebrate homologue of Drosophila gcm2 (glial cells missing 2) is exclusively expressed in the parathyroid glands and in the pharyngeal pouches from which they derive: the third and fourth pouches in chicken (Okabe and Graham, 2004; Neves et al., 2011; Grevellec et al., 2011), and the third pharyngeal pouch in mouse (Gordon et al., 2001). In fish that lack parathyroid glands, gcm2 is expressed in the developing gills (Hogan et al., 2004; Okabe and Graham, 2004).

Purely aquatic amphibians lack parathyroid glands, and in Xenopus the parathyroid glands develop during metamorphosis (Shapiro, 1933; Srivastav et al., 1995). By whole mount in situ hybridization Xenopus gcm2 was first detected at stage 31 in two vertical stripes in the pharyngeal region, with stronger expression in the most posterior stripe (Fig 3A–B). At stage 32 gcm2 was also weakly expressed in a third stripe posterior to the first two (Fig 3C–F). In situ hybridization on longitudinal sections of stage 32 and stage 34 embryos indicated that these stripes corresponded to gcm2 expression in the endoderm (pouch) and ectoderm (cleft) components of the second, third and fourth pharyngeal arches (Fig 3J–M). Later in development gcm2 expression appeared confined to the developing gills (Fig 3 G–I), as seen in teleost fish (Hogan et al., 2004; Okabe and Graham, 2004). We also assessed whether gcm2 was expressed in the parathyroid glands after metamorphosis. In 3-months old post-metamorphic froglets, parathyroid glands were identified in the thoracic cavity based on their position relative to the carotid arteries, as seen on transverse sections (Fig 3N–O; Shapiro, 1933; Coleman, 1969). We performed real time RT-PCR on dissected tissues from 3-months old froglets and found that gcm2 was specifically expressed in the parathyroid glands and was virtually undetectable in other tissues such as brain, heart and thymus (Fig 3P). This result indicates that gcm2 expression in the parathyroid is conserved across many species.

Figure 3. Developmental expression of gcm2.

(A–B) gcm2 was first detected at stage 31 by whole mount in situ hybridization in two longitudinal stripes in the pharyngeal region. At stage 32 (C–D), and stage 34 (E–F), gcm2 was now expressed in three distinct stripes. At stage 40 (G–H) and by stage 41 (I), gcm2 expression appeared to be confined to the developing gills (arrows). (A–I) Lateral views, dorsal towards the top, anterior to the left. Panels (B, D, F, H) are higher magnification views of the embryos shown in panels (A, C, E, G). In situ hybridization on longitudinal sections of stage 32 (J–K) and stage 34 (L–M) embryos confirmed the expression of gcm2 in the endoderm and ectoderm components of the second, third and fourth pharyngeal arches, with stronger expression in the third arch. For (J–K) and (L–M) the level of the section is indicated by a white line in panels (D) and (F), respectively. In panels (J–M) anterior towards the top. (N–O) Hematoxylin and eosin staining of a section through the thoracic region of a 3-months old (90 days) froglet. The parathyroid (arrow) was located in the vicinity of the aortic arch arteries. Panel (O) is a higher magnification view of the parathyroid shown in panel (N). aa, aortic arch artery; ca, carotid artery; pa, pulmonary artery; ph, pharynx; pp, pharyngeal pouches; II–IV, pharyngeal arches. In each panel the stage/age is indicated in the lower left corner. The scale bars in (A, C, E and G) = 500 μm; (B, D, F, H and I) = 250 μm; (J, L and N) = 200 μm; (K, M and O) = 100 μm. (P) In 3-months old froglets Gcm2 was specifically detected in the parathyroid (Pa) by real time RT-PCR, while in other tissues, including brain (Br), heart (He) and thymus (Th), gcm2 was essentially undetectable.

hoxa3

In mice the homeobox containing gene, Hoxa3 is expressed in the third and fourth pharyngeal pouches endoderm and in the NC-derived mesenchyme of the third and fourth pharyngeal arches (Hunt et al., 1991; Manley and Capecchi, 1995; 1998). HOXA3 is essential for pharyngeal glands development as Hoxa3 mutant mouse embryos lack both the thymus and parathyroid (Manley and Capecchi, 1995).

In Xenopus embryos, hoxa3 was expressed at the early neurula stage in the developing spinal cord and migrating NC cells (Fig 4A–B). hoxa3 showed an anterior boundary of expression corresponding to rhombomere 5 (McNulty et al, 2005). At stage 32 (Fig 4E–F) and stage 35 (Fig 4I–J) the anterior boundary of expression was maintained with hoxa3-positive NC cells populating the third pharyngeal arch. In situ hybridization on section at these stages showed that hoxa3 was expressed in the mesenchyme of the third and fourth pharyngeal arches, as well as in the endoderm-derived epithelium of the third and fourth pharyngeal pouches (Fig 4G–H, K–L). By contrast the population of NC cells that migrated into the second pharyngeal arch did not express hoxa3 (Fig 4G–H, K–L). At stage 41 the thymus primordia were exclusively surrounded by hoxa3-negative NC cells (Fig 4M–P), confirming that the thymus in Xenopus is exclusively derived from the second pharyngeal pouch.

Figure 4. Developmental expression of hoxa3.

(A–C) hoxa3 expression at stage 22 (A–B) and at stage 24 (C–D) was restricted to the spinal cord with an anterior boundary of expression corresponding to rhombomere 5. hoxa3 was also expressed in the migrating NC cells. At stage 32 (E–H) and stage 34 (I–L) NC cells had colonized the third and fourth pharyngeal arches (III–IV) where they still expressed hoxa3. At this stage hoxa3 was also detected in the pharyngeal endoderm epithelium of the arches, but was excluded from the mesoderm core. The level of the sections is indicated by a black line in panel (F, J). At stage 41 (M–P), the primordium of the thymus emerging from the second pharyngeal pouch was surrounded by hoxa3-negative NC cells (arrows). For whole mount in situ hybridization, anterior to left, dorsal to top except for panels (A, C), which are dorsal view. Panels (G, H, K, L) are longitudinal sections, anterior towards the top. Panels (O–P) are transverse sections, dorsal to top. ph, pharynx; pp, pharyngeal pouches; II–IV, pharyngeal arches. In each panel the stage is indicated in the lower left corner. The scale bars in (A–E, I and M) = 500 μm; (F, J and N) = 250 μm; (G, K and O) = 200 μm; (H, L and P) = 100 μm.

foxn1

The winged helix/forkhead transcription factor Foxn1 (formerly called whn/Hfh11; Kaestner et al., 2000) is highly expressed in the thymus in mouse, birds and fish (Nehls et al., 1996; Neves et al., 2011; Gordon et al., 2001 Schorpp et al., 2002). FOXN1 is not required for thymus specification (Nehls et al., 1996), however it is essential later for TEC differentiation in mice (Blackburn et al., 1996).

A PCR product corresponding to Xenopus laevis foxn1 was amplified from stage 55 thymus cDNA. Sequence analysis (Fig 5A; Table 1) indicated that at the amino acid level, the partial sequence of Xenopus laevis Foxn1 shared 53% identity with human FOXN1 (Schorpp et al., 1997), 53% identity with mouse FOXN1 (Nehls et al., 1996), 57% identity with chicken Foxn1 (Neves et al., 2011) and 42% with zebrafish Foxn1 (Schorpp et al., 2002). Xenopus laevis and Xenopus tropicalis Foxn1 sequences showed a very similar identity to that of other species (Table 1). The forkhead DNA binding domain shows 100% conservation between human, mouse, chicken and Xenopus (Fig 5A). While in most species Foxn1 is expressed early in the pharyngeal pouch endoderm, in Xenopus foxn1 was first detected at stage 47, when the thymus primordia had already detached from the pharyngeal pouch. Whole-mount in situ hybridization shows that foxn1 expression was restricted to the thymus (Fig 5B). This expression was confirmed by in situ hybridization on sections (Fig 5C–D). At stage 48 the thymus primordia had increased in size and foxn1 expression appeared to be stronger (Fig 5E–F). foxn1 can be detected up to stage 57 (the last stage examined in this study), however at this stage the expression was more diffuse and appeared scattered throughout the entire organ (Fig 5G), presumably due to the expansion of the lymphoid cells in the thymus (Du Pasquier and Weiss, 1973; Robert and Ohta, 2009).

Figure 5. Sequence and developmental expression of foxn1.

(A) The deduced amino acid sequences of Foxn1 from human, mouse, chicken and zebrafish were aligned using ClustalW2 and compared to the partial amino acid sequence of Xenopus laevis Foxn1. The highly conserved forkhead domain is underlined. (B) By whole mount in situ hybridization foxn1 was first detected at stage 47 (arrows) lateral to the hindbrain (hb). In situ hybridization on transverse sections of stage 47 (C–D) and stage 48 (E–F) tadpoles. foxn1 was completely restricted to the thymus primordium (arrows). foxn1 expression in the thymus persisted at least up to stage 57 (G), at which stage foxn1 was weakly detected throughout the entire organ. Panels (D, F) are higher magnification views of the thymus primordium shown in panels (C, E), respectively. In each panel the stage is indicated in the lower left corner. The scale bar in (B) = 500 μm; (C, E and G) = 200 μm; (D and F) = 200 μm.

Table 1.

Sequence homology comparison of Foxn1 in different species

	X. laevis	X. tropicalis	Human	Mouse	Chicken	Zebrafish
X. laevis		91%	53%	53%	57%	42%
X. tropicalis			49%	50%	51%	40%
Human				86%	54%	48%
Mouse					55%	40%
Chicken						43%
Zebrafish

The temporal expression of hoxa3, gcm2 and foxn1 genes was confirmed by RT-PCR (Fig 6). These results are consistent with our in situ hybridization data and further demonstrate the lack of foxn1 expression in the thymus primordium prior to its separation from the pharynx.

Figure 6. RT-PCR analysis of the developmental expression of hoxa3, gcm2 and foxn1.

The stages are according to Nieuwkoop and Faber (1967). Th, thymus from stage 55 tadpoles. Ornithine decarboxylase (odc) is shown as a loading control.

NC contribution to thymus development

The cardiac NC is a subdivision of the cranial NC that makes an important contribution to the pharyngeal glands (Le Lievre and Le Douarin, 1975; Bockman and Kirby, 1984). To evaluate whether NC cells contribute to thymus development in Xenopus, we performed transplantation experiments. NC grafts isolated from stage 17 embryos injected with mRNA encoding Red Fluorescent Protein (RFP) were transplanted onto the same region of stage-matched unlabeled embryos (Fig 7A; Lee and Saint-Jeannet, 2011). At stage 26, the RFP-labeled NC grafts displayed a normal pattern of cranial NC migration in the host embryo, and cells derived from the RFP-labeled NC graft contributed to the hyoid (pharyngeal arch 2), anterior (pharyngeal arch 3) and posterior (pharyngeal arch 4–5) branchial streams of the NC, but not to the mandibular (pharyngeal arch 1) stream of NC (Fig 7B–E), as previously described (Lee and Saint-Jeannet, 2011).

Figure 7. NC contribution to the thymus.

(A) Experimental design to analyze the NC contribution to the thymus. Two-cell stage embryos were injected in the animal pole with mRNA encoding Red Fluorescent Protein (RFP). At stage 17 the RFP-labeled cardiac NC was transplanted onto the same region of an unlabeled host embryo (NC graft). (B–E) RFP-labeled NC graft at stage 26. Lateral views of the same embryo. Dorsal towards the top, anterior to the right (B–C) and left (DE). Cells derived from the RFP-labeled NC graft were confined to the three most posterior streams of cranial NC, as previously described (Lee and Saint-Jeannet, 2011). At stage 41 (F–O) cells derived from the RFP-labeled NC graft (J–O) formed most of the head mesenchyme, surrounding the thymus primordial (arrows). A fraction of the RFP-labeled NC cells also express hoxa3, however these cells remained confined to a region posterior to the thymus primordia (G, I, M, J, N). Panels (H, L) are higher magnification of panel (F). Panels (I, M) are higher magnification of panel (G). Longitudinal sections, anterior towards the top. (K) and (O) are overlay of panels (H–J) and (L–N), respectively. Insets in panels (K, O) are higher magnification views of the thymus primordia (arrows). (P–Y) At stage 47 the thymus primordia can be identified by Foxn1 expression (Q, S, W). A small number of RFP-labeled NC cells have colonized the thymus (T, U, X, Y). Panels (R–U) are higher magnification views of the thymus primordium (arrows) in panels (P–Q). Panels (R–U) and (V–Y) show two representative tadpoles form two independent experiments. (U) and (Y) are overlay of panels (R–T) and (V–X), respectively. The scale bars in (B–G, P and Q) = 200 μm; (H–O and R–Y) = 50 μm.

The progeny of the RFP-labeled NC cells was analyzed in transplanted embryos at stage 41 and stage 47 (n=16; three independent experiments). On longitudinal sections through the pharyngeal region of stage 41 embryos, RFP-positive cells formed most of the mesenchyme surrounding the thymus primordium, which remained free of NC cells at this stage (Fig 7 F–O). In situ hybridization for hoxa3, which has an anterior boundary of expression in the third pharyngeal arch mesenchyme, indicated that the RFP-labeled NC cells surrounding the thymus primordia were hoxa3-negative, consistent with its derivation from the second pharyngeal pouch. In all cases examined (n=10), transverse sections of stage 47 tadpoles showed that RFP-labeled NC cells populated the thymus primordial visualized by foxn1 expression (Fig 7P–Y).

The NC is not required for initial thymus development

We next performed NC ablation experiments to evaluate the role of the NC-derived mesenchyme, surrounding the pharyngeal pouch, during initial thymus development. A segment of the NC, corresponding to prospective cardiac NC, was ablated at stage 17 (Fig 8A, B, E; Lee and Saint-Jeannet, 2011). The successful ablation of the NC was assessed by expression of the NC marker sox10 at stage 25, when the NC cells populated the pharyngeal arches in control embryos (Fig 8C–D, F–G). In all NC-ablated tadpoles examined at stage 47 (n=13, in three independent experiments) the bilateral thymus primordia formed in their proper location and expressed foxn1 gene (Fig 8H–M). These results indicate that NC cells-derived signals are not required for the specification and the initial organogenesis of the thymus primordium in Xenopus. To further evaluate the expression levels of foxn1 in control and manipulated embryos we performed real time RT-PCR on thymi isolated from stage 48 tadpoles. We normalized foxn1 expression levels to the expression of the TEC-specific marker, keratin 8 (Klug et al., 2002), to account for any changes in TEC number between control and NC-ablated samples. We found that foxn1 expression in the NC-ablated thymi was not statistically different from that of control thymi, indicating that NC cells are not required for foxn1 expression in developing TEC (Fig 8N). Next we determined whether NC cells were needed for the function of TEC to support T cell development. To address this issue we analyzed the expression of the recombination-activating gene 1 (rag1) in dissected thymi from stage 48 tadpoles following NC ablation (Greenhalgh et al., 1993). Rag1 is specifically expressed in lymphocytes as they initiate the assembly of the T cell receptor genes, and is therefore a good marker to assess intrathymic T cell development (reviewed in Schatz and Swanson, 2011). By real time RT-PCR we found that rag1 expression in the thymi of NC ablated tadpoles (n=9 tadpoles), was similar to that of control thymi (Fig 8O), suggesting that intrathymic T cell development was not impaired by NC ablation.

Figure 8. NC cells are dispensable for thymus development.

(A) Diagram illustrating cardiac NC ablation at stage 17. (B–G) sox10 expression in control and NC-ablated embryos at stage 17 (B and E), and stage 25 (C–D and F–G). The red brackets in control (B) and NC-ablated (E) embryos indicate the position of cardiac NC territory ablated at stage 17. NC ablation resulted in the bilateral loss of the three most posterior NC streams at stage 25 (F–G) as compared to control (red brackets in C and D). In NC-ablated embryos sox10 expression in the otic vesicle is unaffected (arrows). (B and E) are dorsal views, anterior to top. (C and D; control) and (F and G; NC-ablated) are showing both sides of the same embryo. Dorsal to top, anterior to right (C and F) or to left (D and G). (H–M) Transverse sections (dorsal to top) through two stage 47 tadpoles (H–J and K–M) following NC ablation at stage 17. foxn1 expression was detected in the thymus primordia of both tadpoles. Panels (H, J) and (K, M) are higher magnification views of the thymi shown in (I) and (L), respectively. (I, L) The arrows point to the thymus primordia expressing foxn1. The scale bars in (B–G) = 500 μm; (H, J, K and M) = 50 μm; (I and L) = 200 μm. (N–O) The thymi of control (n=9) and NC-ablated (n=9) tadpoles were dissected at stage 48 and analyzed by real time RT-PCR for foxn1 (N) and rag1 (O) expression. foxn1 and rag1 expression were normalized to the expression of the TEC-specific marker, keratin 8. Statistical significance was determined using the T-test. The values are presented as mean SEM.

Discussion

In this study we analyzed the development of the thymus in Xenopus laevis using histological and molecular tools. In agreement with earlier studies (Nieuwkoop and Faber, 1967; Manning and Horton, 1969), we found that the thymus rudiment arises from the second pharyngeal pouch endoderm at around stage 41, and that by stage 45 the rudiment separates from the pharynx to become a fully individualized paired organ. This was documented by the expression of hoxa3, a gene with an anterior boundary of expression in the third pharyngeal arch. While NC cells migrating into the third pharyngeal arch express Hoxa3, the population of NC cells that travels rostral to the otic vesicle into the second pharyngeal arch remains hoxa3-negative. In Xenopus the thymus primordium is exclusively surrounded by hoxa3-negative NC cells consistent with its derivation from the second pharyngeal pouch. This contrasts with mouse, chicken and zebrafish embryos in which the thymus primordium originates from the third and/or fourth pharyngeal pouch, and is surrounded by Hoxa3-positive NC cells (Hunt et al., 1991; Manley and Capecchi, 1995; 1998; Saldivar et al., 1996; Hogan et al., 2004; Chen et al., 2010; see Table 2). While in all species examined so far the thymus originates from pharyngeal pouches, its precise embryological origin and the number of thymus glands in the adult vary greatly. For example sharks develop a thymic primordium in every pouch except the first one, resulting in five bilateral thymi. In reptiles, the thymus arises from the second and third pharyngeal pouches, and in chicken the thymus primordia are located in the third and fourth pharyngeal pouches and give rise to seven thymic lobes (reviewed in Rodewald et al., 2008; Ge and Zhao, 2012). These differences in the origin of the thymus presumably reflect the specialized characteristics of individual pharyngeal arches acquired during evolution (Grevellec and Tucker, 2010).

Table 2.

Summary of the developmental expression of Gcm2, Hoxa3 and Foxn1 in different species.

Genes	Species	Early expression	Late expression	References
Gcm2	mouse	3 ppe	parathyroid rudiment	Kim et al., 1998; Gordon et al., 2001
	bird	2, 3 & 4 ppe	parathyroid rudiment	Okabe and Graham, 2004; Neves et al., 2011; Grevellec et al., 2011
	frog	2, 3 & 4 ppe+pce	gills	This study
	fish	2, 3, 4, 5 & 6 ppe+pce*	gills	Okabe and Graham, 2004; Hogan et al., 2004
Hoxa3	mouse	3 & 4 pam+ppe	thymus rudiment	Hunt et al., 1991; Manley and Capecchi, 1995; 1998
	bird	3 & 4 pam+ppe	nd	Saldivar et al., 1996
	frog	3 & 4 pam+ppe	thymus rudiment	This study
	fish	3 & 4 pam+ppe+pce*	nd	Hogan et al., 2004; Chen et al., 2010
Foxn1	mouse	3 ppe	thymus rudiment	Gordon et al., 2001
	bird	3 & 4 ppe	thymus rudiment	Neves et al., 2011
	frog	-	thymus rudiment	This study
	fish	3 ppe	thymus rudiment	Schorpp et al., 2002

pam, pharyngeal arch mesenchyme; pce, pharyngeal cleft ectoderm; ppe, pharyngeal pouch endoderm. -, not detected in the pharyngeal apparatus; nd, not determined;

^*pce expression has been reported in one study (Hogan et al., 2004).

In most species Hoxa3 is expressed in the NC mesenchyme of the pharyngeal arches as well as in the pharyngeal pouches endoderm (see Table 2). Hoxa3 mutant mouse embryos lack both the thymus and parathyroid, however the amount of NC cells migrating in the pharyngeal arches is essentially normal in the mutants (Manley and Capecchi, 1995), suggesting that it is Hoxa3 expression in the pharyngeal pouch endoderm that is critical for thymus formation. This was recently confirmed by a NC-specific deletion of Hoxa3 in mice, which did not affect specification of the thymus and parathyroid, rather both organs failed to properly sever their connection with the pharynx and therefore were ectopically positioned (Chen et al., 2010). Because thymus organogenesis in Hoxa3 null mutants fails after formation of the third pharyngeal pouch and prior to organ specification, it has been proposed that HOXA3 functions to establish the identity of the third pharyngeal pouch to form the thymus and parathyroid territories (Blackburn and Manley, 2004; Manley and Condie, 2010). The fact that the thymus primodium arises from the second pharyngeal pouch endoderm in Xenopus, a tissue that is hoxa3-negative, clearly indicates that thymus organogenesis can occur independently of Hoxa3, raising the question of the precise function of HOXA3 during thymus development in mammals. If HOXA3 is indeed providing positional identity to the third pouch compatible with thymus formation, it is possible that this function may have been co-opted by another Hox gene product in frogs, presumably Hoxa2, which has an anterior boundary of expression in the second pharyngeal arch (Pasqualetti et al., 2000).

In non-aquatic vertebrates the thymus and parathyroid glands are derived from a common primordium and their segregation into two distinct domains can be first visualized by the expression of two transcription factors Foxn1 and Gcm2, which are largely thymus- and parathyroid-specific, respectively. Gcm2 is critically required for parathyroid gland formation. It is expressed in the pharyngeal pouches and the forming parathyroid glands (see Table 2), and Gcm2 mutant mice fail to form parathyroid glands (Günther et al., 2000; Liu et al., 2007). While fish and purely aquatic amphibians lack parathyroid glands, Xenopus develops parathyroid glands during metamorphosis (Shapiro, 1933; Srivastav et al., 1995). In Xenopus embryos, gcm2 is initially detected in the second, third and fourth pharyngeal pouches and clefts and then its expression becomes restricted to the gills. This is consistent with previous reports in teleost fish showing that gcm2 was expressed within the pharyngeal apparatus and internal gill buds (Hogan et al., 2004; Okabe and Graham, 2004; see Table 2). Interestingly, parathyroid hormone encoding genes were also detected in the gills, suggesting that the tetrapod parathyroid glands and the gills of fish are evolutionarily related and may play a similar role in controlling calcium levels (Okabe and Graham, 2004). Because in post-metamorphic froglets gcm2 is specifically expressed in the parathyroid glands, it is likely that Gcm2 in Xenopus may carry a similar function to that of higher vertebrates (Liu et al., 2007).

The transcription factor Foxn1 is the gene mutated in the nude mice (Nehls et al., 1994; Blackburn et al., 1996). While Foxn1 is not required for initial thymus specification and organogenesis (Nehls et al., 1996), it is one of the earliest genes expressed in the thymus primordium. Foxn1 is first detected in the presumptive thymic epithelium of the third pharyngeal pouch at 72 hpf in zebrafish, and at E11.25 in the mice (Gordon et al., 2001; Schorpp et al., 2002). In birds it is detected in the dorsal-anterior region of the third and fourth pharyngeal pouch endoderm as early as at E5 (Neves et al., 2011). In these species Foxn1 is detected in the pharyngeal pouch endoderm well before the thymus primordium detaches from the pharynx and starts to differentiate (Gordon et al., 2001; Schorpp et al., 2002; Neves et al., 2011; see Table 2). Interestingly, in Xenopus, foxn1 was not detected until the thymus primordium separated from the pharynx at around stage 47. This is approximately the stage at which the TEC starts to differentiate and express MHC class II molecules, and when RC47-positive leukocytes colonize the thymus (Du Pasquier and Flajnick, 1990). This suggests that initial TEC development in Xenopus is Foxn1-independent.

The NC contributions to thymus development have been somewhat controversial. Initial NC ablation experiments in chicken embryos suggest that thymus organogenesis depends on signals derived from the NC cells surrounding the third and fourth pharyngeal pouches (Bockman and Kirby, 1984). In the mouse mutants Splotch, that have a dramatically reduced number of NC cells in the third pharyngeal pouch, the thymus and parathyroid primordia are normally specified as determined by the expression of Gcm2 and Foxn1 (Griffith et al., 2009). In these mutants the boundary between the two glands primordia is shifted such that the presumptive thymic region is expanded at the expense of the parathyroid resulting in a larger fetal thymus (Griffith et al., 2009). A more recent study suggests that thymus formation in mouse embryos is NC-dependent (Nelms et al., 2011). Compound mutant embryos homozygous for a NC-specific Foxd3 mutation and heterozygous for Pax3, which lack NC cells caudal to the first pharyngeal arch, show severe thymus hypoplasia (Nelms et al., 2011). It is unclear why these two mouse models have somewhat different outcomes with respect to thymus development. Our findings are consistent with the former observations since NC ablation did not affect specification and initial organogenesis of the thymus. Moreover, the thymus primordium expressed Foxn1 in the absence of NC cells, suggesting that NC-mesenchyme signals are not required for these processes. We also found that ablation of NC cells did not alter the expression of rag1, an essential factor for thymic T cell development. Altogether these results indicate that the NC cells are largely dispensable for early thymus development and differentiation in Xenopus.

Although the mechanisms regulating thymus development have been extensively studied in the last decade (Reviewed in Blackburn and Manley, 2004; Hollander et al., 2006; Rodewald, 2008; Gordon and Manley, 2011), less is known about the factors that initiate the specification of the thymus primordium. Because Xenopus is an excellent model to analyze cell fate choices during embryogenesis, the work presented here provides the basis for future investigations of thymus development in vertebrates.

Experimental Procedures

Embryos, injections, tissues labeling and transplantation

Xenopus laevis embryos were staged according to Nieuwkoop and Faber (1967) and raised in 0.1X NAM (Normal Amphibian Medium; Slack and Forman, 1980). Monomeric Red Fluorescent Protein (RFP) mRNAs were synthesized in vitro using the Message Machine kit (Ambion, Austin, TX). Labeling and transplantation of the NC explants were performed as previously described (Lee and Saint-Jeannet, 2011). Briefly, embryos were injected into the animal pole at the 2-cell stage with 1 ng of RFP mRNA. NC labeled-explants, corresponding to prospective cardiac NC, were dissected at stage 17 in 1X NAM supplemented with 50 μg/ml of gentamycin using a 25-gauge syringe needle. Grafts were then transplanted onto the same region of unlabeled stage-matched sibling embryos. Grafted embryos were allowed to heal for 30 minutes in the dissection medium, before being transferred for long-term culture into 0.1X NAM. Embryos were observed under an epifluorescence microscope (Nikon, Eclipse E800).

Isolation of Xenopus laevis foxn1 and gcm2

Xenopus laevis foxn1 was amplified by PCR from stage 55 thymus cDNA generated by the First-Strand cDNA Synthesis Kit (GE Healthcare). The primers (F: 5′-TTTCATCCGTATAAGAGG-3′; and R: 5′-ATCTGTAAGAGAAGGATT-3′) were designed based on the sequence of Xenopus tropicalis foxn1 (accession # XM_002937144). These primers were located at position 478 and 1764 of Xenopus tropicalis sequence. The PCR conditions were as follow: denaturation 94 °C (1 min), annealing at 60 °C (1 min), and extension at 72 °C (1 min.) for 35 cycles. A 1287-bp PCR product was purified, ligated into pGEMTeasy (Promega), and sequenced. The partial sequence of Xenopus laevis foxn1, that covers 129 amino acids, has been deposited into GeneBank (accession # JQ182306). Xenopus laevis gcm2 was amplified by PCR from stage 35 cDNA using primers (F: 5′-TCCAAAGATGCCTCAGGACT-3′; and R: 5′-TATATCCGTGCATTGGAGCA-3′) designed based on the published sequence (accession # AB175676; Okabe and Graham, 2004). The PCR condition were as follow: denaturation 94 °C (1 min), annealing at 55 °C (1 min), and extension at 72 °C (1 min) for 35 cycles. A 1100-bp PCR product was purified, ligated into pGEMTeasy (Promega), and sequenced. Xenopus laevis hoxa3 cDNA (NP_001080293) was purchased from Open Biosystems (Thermo Scientific, USA).

RT-PCR analysis

Total RNAs were extracted from embryos or dissected tissues using an RNeasy micro RNA isolation kit (Qiagen). The RNA samples were treated with RNase-free DNase I before RT-PCR. The reactions were performed using the One Step RT-PCR kit (Qiagen) according to the manufacturer’s instructions using the following primer sets: hoxa3 (F: 5′-TCCCTACCAAGGAGCAAATG-3′; R: 5′-GCAGGCAGGTCGATGATATT-3′), gcm2 (F: 5′-GATGCCTCAGGACTTCAAGC-3′; R: 5′-CATGGCCCAACCACTTAGAT-3′), foxn1 (F: 5′-ATCCGTGAGAAAAAGCATGG-3′; R: 5′-TAGTGGGGGAGTTCACAAGG-3′), rag1 (F: 5′-GAGGTGGCATTGCCTAATGT-3′; R: 5′-TTCCCTCTGTTTCCTCCTCA-3′); keratin 8 (F: 5′-AGGCTGCAGTCCGAAATAGA-3′; R: 5′-CTCCTTCCAGCAGTTTCCTG-3′) and odc (F: 5′-ACATGGCATTCTCCCTGAAG-3′; R: 5′-TGGTCCCAAGGCT AAAGTTG-3′). The PCR conditions were as follow: denaturation 94°C (30 sec), annealing at 55°C (30 sec), and extension at 72°C (30 sec) for 30 cycles. For real-time RT-PCR the reaction was performed using the QuantiTect SYBR Green RT-PCR kit (QIAGEN) on a LightCycler (Roche Diagnostics). The reaction mixture consisted of 10 μl of QuantiTect SYBR Green RT-PCR Master Mix, 500 nM forward and reverse primers, 0.2 μl of RT, and 60 ng of template RNA in a total volume of 20 μl. The cycling conditions were as follow: denaturation at 95°C (15 sec), annealing at 55°C (20 sec), and extension at 72°C (15 sec). By optimizing primers and reaction conditions, a single specific product was amplified as confirmed by melting curve analysis. Each reaction included a control without template and a standard curve of serial dilutions (in 10-fold increments) of test RNAs. In each case, ornithine decarboxylase (odc) was used as an internal reference (data not shown). Each bar on the histograms has been normalized to the level of odc.

In situ hybridization

Digoxygenin (DIG)-labeled antisense RNA probes (Genius Kit, Roche) were transcribed from linearized plasmids using the indicated RNA polymerase: pGEMT-foxn1 (SacII, SP6 polymerase), pGEMT-gcm2 (SalI, T7 polymerase) and pSport6-hoxa3 (SalI, T7 polymerase). For each gene sense probes were used as negative control (not shown). Whole-mount in situ hybridization was performed as previously described (Harland, 1991). For in situ hybridization on sections, embryos were fixed in 4% paraformaldehyde in phosphate buffer saline (PBS; Gibco) for 1 hour, embedded in Paraplast+ and 12 μm sections hybridized with the appropriate DIG-labeled probes as described (Henry et al. 1996). Because RFP can no longer be visualized following this procedure, sections of RFP-labeled embryos were first individually photographed and then processed for in situ hybridization. Sections were then briefly counter stained with eosin.

Histology and immunohistochemistry

Embryos were fixed in 4% paraformaldehyde in PBS for 1 hour. After dehydration through a graded series of ethyl-alcohol, embryos were embedded in Paraplast+. Transverse serial sections (12 μm) were performed on an Olympus rotary microtome and stained with Hematoxylin and Eosin. For immunohistochemistry, after rehydration in PBS, 10 μm sections were boiled for 20 min in sodium citrate buffer (10 mM sodium citrate, 0.05% Tween 20, pH 6.0) to retrieve antigens. The sections were then blocked in PBST (0.1% Triton X-100 in PBS, containing 0.2% bovine serum albumin and 5% inactivated lamb serum) and incubated successively with mouse anti-p63 (Santa Cruz, clone 4A4; Senoo et al., 2007) antibody overnight at 4°C (1:50 dilution), and an anti-mouse IgG-FITC conjugated secondary antibody (Jackson ImmunoResearch; 1:100 dilution) for 1 hour at room temperature. All antibodies were diluted in PBST.

Highlights.

• The thymus primordia arise from the second pharyngeal pouch.

• Thymus formation is independent of hoxa3 expression in the pouch endoderm.

• foxn1 expression correlates with the initiation of thymic epithelial differentiation.

• Neural crest cells are not required for thymus specification or T cell differentiation.

KEY POINTS.

• Asthma prevalence and severity are greater in women than in men, and a large body of evidence suggests that this is related to estrogen.

• Estrogen receptors are found on numerous immunoregulatory cells and estrogen’s actions skew the immune response toward allergy.

• Estrogen may act directly to create deleterious effects in asthma, or may affect other intermediaries to exert effects on lung mechanics and airway inflammation.

• Environmental estrogens (xenoestrogens) may enhance development of atopic disorders like asthma.