Development of the Thymus Requires Signaling Through the Fibroblast Growth Factor Receptor R2-IIIb

Jean-Michel Revest, Ravinder K. Suniara, Karen Kerr, John J. T. Owen and Clive Dickson
J Immunol August 15, 2001, 167 (4) 1954-1961; DOI: https://doi.org/10.4049/jimmunol.167.4.1954

Abstract

Mice deficient for fibroblast growth factor (Fgf)R2-IIIb show a block in thymic growth after embryonic day 12.5, a stage that just precedes its detection in thymic epithelial cells. Fgf7 and Fgf10, the main ligands for FgfR2-IIIb, are expressed in the mesenchyme surrounding the thymic epithelial primordium, and Fgf10-deficient mice also exhibit impaired thymic growth. Hence, Fgf signaling is essential for thymic epithelial proliferation. In addition to the proliferative block, most thymic epithelial cells fail to progress from an immature cytokeratin 5-positive to a cytokeratin 5-negative phenotype. Nevertheless, sufficient epithelial cell differentiation occurs in the severely hypoplastic thymus to allow the development of CD4/CD8-double-positive thymocytes and a very small number of single-positive thymocytes expressing TCRs.

The thymus is the site of T cell development, where through a series of interactions with the thymic epithelium, lymphoid precursors mature and are exported to the periphery. In the mouse, the thymus primordium emerges around embryonic day (E)3 10.5 as an epithelial bud encompassing endoderm from the third pharyngeal pouch and possibly ectoderm of the third branchial cleft (1, 2, 3). After E11, the bilaterally formed thymic primordia with accompanying neural crest-derived mesenchyme separate from the pharynx and migrate to form a bilobed organ located above the heart (3). The process of thymic lymphopoiesis starts around E11.5–E12.5, following migration of lymphoid stem cells into the epithelial primordium, and is concomitant with organ migration and rapid epithelial cell proliferation (4, 5). Subsequently, these stem cells proliferate, rearrange their Ag receptor genes, and then differentiate into functionally mature T cells. Moreover, differentiation of the thymic epithelium and T cell development are interdependent (6, 7). Organ explant cultures have demonstrated that the E12 mesenchyme surrounding the thymic epithelial rudiment is necessary for thymic development (8). From E12.5 to birth, a high proliferation rate results in the thymic mass markedly expanding in all dimensions with the formation of numerous lobules. The late maturation of the thymus is marked by the development of the first Hassall’s bodies.

The development of the thymus has been shown to require the expression of a number of genes, including members of the Hox and Pax gene transcription factor families. Mice deficient for Hoxa3 are athymic and have several other pharyngeal organ abnormalities (9, 10). Its expression in both the neural crest mesenchyme and pharyngeal endoderm suggests that Hoxa3 could affect both these cell lineages. Pax1 and Pax9 are expressed in an overlapping pattern in the pharyngeal endodermal cells and continue throughout fetal development. Mice deficient for Pax1 have a hypoplastic thymus and show abnormal thymocyte maturation, whereas Pax9-mutant mice are athymic (11, 12).

Fibroblast growth factors (Fgfs) are a large family of intercellular signaling molecules and mediate their biological responses by binding and activating high affinity cell surface receptors (FgfR) with intrinsic tyrosine kinase activity (reviewed in Refs. 13, 14). Fgfs have been implicated in a number of important cell activities including differentiation, proliferation, and migration (reviewed in Refs. 15, 16). The Fgfs signal through four Fgf receptor genes (FgfR1–FgfR4), but alternative splicing increases FgfR diversity. FgfR2 can generate two isoforms, FgfR2-IIIb and FgfR2-IIIc, with different Fgf-binding specificity and different sites of expression (17). FgfR2-IIIb is located in many types of epithelia and is activated by four known ligands (Fgf1, Fgf3, Fgf7, and Fgf10) that are synthesized predominantly in mesenchyme (17, 18, 19, 20, 21, 22). Gene targeting studies on FgfR2, which encompass both receptor isoforms, have given varying results with peri-implantation lethality at E4.5–E5.5 (23) to early embryonic lethality of embryos showing no limb buds and a defective placenta (24, 25). More recently, FgfR2-IIIb isoform-specific abrogations have shown this isoform is essential for epithelial-mesenchymal interactions during the development of several organs (26, 27, 28, 29, 30).

Mice deficient for FgfR2-IIIb survive in utero to term but cannot survive after birth because they have no lungs. These mice also show a severely hypoplastic thymus of comparable size to the normal E12 thymus. Here, we show that Fgf signaling is essential for thymic epithelial cell proliferation, and although very low numbers of thymocytes are generated, differentiation occurs to the CD4/CD8-double-positive and -single-positive stages.

Materials and Methods

Gene targeting, mouse strains, and genotyping

Gene targeting of isoform IIIb of FgfR2 has been realized by introducing an internal ribosomal entry site (IRES)-lacZ gene into exon IIIb as previously described (30). Briefly, stop codons in different reading frames were introduced upstream of the IRES. A neomycin-resistant gene, driven by the HSV-tk promoter and flanked by LoxP sites from plasmid pL2Neo (a gift of H. Gu, Institute for Genetics, University of Cologne, Cologne, Germany; Ref. 31), was inserted into the intron between exons IIIa and IIIb. The resulting construct was electroporated into 129P2/OlaHsd embryonic stem cells (Fig. 1⇓B). Neomycin-resistant clones were screened by both Southern blotting and PCR. Five different lines of mice were used (2G1, 3G9, 2D9, 5A7, and 2G1 Cre). The 2G1 Cre line was obtained by deletion of the neomycin-resistant gene from the 2G1 line by injecting eggs derived from a 2G1-heterozygous male crossed with a wild-type Bl6 female with 10 ng/ml of pMC-Cre (kindly provided by H. Gu; Ref. 31). No differences in phenotype between the various lines were observed. The mice were genotyped as previously described (30). Fgf10−/− mouse embryos were kindly provided by D. Ornitz (Washington University Medical School, St. Louis, MO; Ref. 32).

FIGURE 1.
FIGURE 1.

Disruption of the IIIb isoform of FgfR2 affects normal development of the thymus. A, Schematic diagram showing the FgfR2 receptor, which is composed of three Ig-like loops (Ig-I–Ig-III), a transmembrane domain (TM), and a cytoplasmic TK domain (TK1–TK2). B, Schematic depiction of the targeting construct used to disrupt FgfR2-IIIb. Translation termination codons in all three reading frames and an IRES-lacZ gene were used to specifically disrupt exon IIIb. A neomycin-resistant gene was inserted into the intron between exons IIIa and IIIb and flanked by loxP sites to allow its excision by Cre recombinase after germline transmission. C and F, Phenotype of mice heterozygous or homozygous for the disrupted FgfR2-IIIb allele at E18.5. Homozygous mutant mice show an absence of limbs, eyelids, a tail dorsally curved, and abnormal skin (F). Thymuses of wild-type mice and mice deficient for isoform IIIb of FgfR2 were compared at E18.5. D, The two wild-type thymus lobes just above the heart surrounded by the lungs. E, A magnification of D. G, The thymus, indicated by the arrow, just above the heart in FgfR2-IIIb−/−lacZ mice. Note an absence of lungs in FgfR2-IIIb−/−lacZ mice (G). H, A magnification of G showing a thymus highly reduced in size compared with the wild-type thymus. H, Heart; rL, right lung; lL, left lung; and T, thymus.

In situ hybridization

Radioactive in situ hybridization was performed on dewaxed tissue sections essentially as described previously (33). A mouse FgfR2-IIIc probe (IIIc) was kindly provided by P. Kettunen (University of Bergen, Bergen, Norway) (33). The lacZ probe and the pan-FgfR2 probe from the tyrosine kinase (TK) domain were previously described (30). Fgf10 and Fgf7 probes correspond to full-length cDNA cloned, respectively, in pBluescript KS and pGEM3 (Stratagene, La Jolla, CA, and Promega, Madison, WI, respectively). Plasmids used to generate antisense Pax1 and Pax9 were kindly provided by R. Balling (Max Planck Institute for Immunobiology, Freiburg, Germany) and I. Thesleff (University of Helsinki, Helsinki, Finland), respectively. Plasmids were linearized and antisense single-stranded RNA probes generated with the appropriate RNA polymerase.

Histology

Tissues were fixed overnight in neutral-buffered formalin, dehydrated through ethanol, embedded in paraffin, and 5-μm sections were cut for staining in H&E.

Immunohistochemistry

The 8-μm sections were cut, and when necessary, Ags were unmasked by prior microwaving for 10 min in 0.01 M trisodium citrate (pH 6). For detection of proliferating cells, a standard indirect immunoperoxidase protocol was used using a rabbit polyclonal Ab, NCL-Ki67p (NovoCastra, Newcastle, U.K.), and revealed by a secondary Ab, biotinylated swine anti-rabbit IgG (DAKO, Ely, U.K.) followed by streptavidin-peroxidase (DAKO) and 3,3′-diaminobenzidine (Sigma, St. Louis, MO) visualization. The cells were then counterstained with hematoxylin. Apoptotic cells were detected by the incorporation of TUNEL using the ApopTag Plus in situ apoptosis detection kit (Oncor, Gaithersburg, MD) according to the manufacturer’s recommended protocol.

For cytokeratin 5 detection, thymic lobes were dissected from E17 embryos, embedded in OCT compound (TissueTek; Sakara Finctek Europe BV) and snap-frozen in liquid N2. Next, 5-μm cryosections were cut onto multispot slides and allowed to air-dry for 1 h before being fixed in cold methanol for 30 min. Sections were then triple-labeled for 45 min for Thy1.2-biotin, cytokeratin 5, and pan-cytokeratin, followed by two 5-min washes in PBS. These were then revealed using extravidin-tetramethylrhodamine isothiocyanate, anti-rabbit Ig-7-amino-4-methylcoumarin-3-acetic acid, and anti-mouse Ig-FITC, respectively, for 30 min. After further washes, sections were mounted in SlowFade Light Antifade reagent in glycerol buffer (Molecular Probes, Eugene, OR). These were then photographed using a Zeiss Axioplan epifluorescence microscope (Zeiss, Oberkochen, Germany; ×40 objective). The following reagents were used: biotinylated rat anti-Thy1.2 (clone 53-2.1; BD PharMingen, San Diego, CA), rabbit anti-keratin 5 (clone AF138; Covance; Cambridge Bioscience, Cambridge, U.K.), mouse anti-pan cytokeratin (clone C-11; Sigma), extravidin-tetramethylrhodamine isothiocyanate (Sigma), goat anti-rabbit Ig-7-amino-4-methylcoumarin-3-acetic acid (DAKO), and goat anti-mouse Ig-FITC (Caltag, South San Francisco, CA).

Flow cytometric analysis

Thymic lobes were dissected from E17- and E18-heterozygous and -homozygous mutant embryos and teased using fine cataract knives to yield a cell suspension. Cells were counted, spun down, and the supernatant removed before triple labeling for CD4, CD8, and TCRαβ for 45 min. After washing, cells were fixed in 1% paraformaldehyde. Samples were subsequently analyzed on a BD LSR flow cytometer (BD Biosciences, San Jose, CA). Dead cells and debris were excluded by selective scatter gating. For flow cytometric analysis, the following Abs were used: PE-conjugated anti-CD4 (clone RM4-5; BD PharMingen), allophycocyanin-conjugated anti-CD8 (clone 53-6.7; PharMingen) and FITC-conjugated antiαβ (clone HS7-597; Sigma).

Results

Mice deficient for FgfR2-IIIb have a severely hypoplastic thymus

The generation and characterization of mice deficient for FgfR2-IIIb has been described previously (30). In brief, translational termination codons in three reading frames were placed close to the start of exon IIIb, followed by an IRES-lacZ gene (Fig. 1⇑, A and B). These changes result in a truncated form of FgfR2-IIIb that lacks part of the ligand binding domain, the transmembrane element, and the TK domain, whereas expression of the alternatively spliced isoform, FgfR2-IIIc, remains intact.

The thymus is one of several organs showing developmental abnormality in FgfR2-IIIb−/−lacZ mice (Fig. 1⇑, C and F). The dissected thoracic regions of E18.5 fetuses show that homozygous mutant mice have a profoundly smaller thymus, as well as no significant lung development, although the heart, which lies immediately posterior to the thymus, appears of normal size (Fig. 1⇑, D, E, G, and H). To characterize thymus development in more detail, consecutive sagittal sections from E12–E17.5 fetuses were stained and examined for histological differences (Fig. 2⇓). By E12 of gestation, the mouse thymus consists of a region of epithelium surrounded by a capsule, and the whole is embedded in a thick layer of mesenchyme. At this stage, there were no discernible histological differences in the thymus of wild-type and FgfR2-IIIb−/−lacZ mice (Fig. 2⇓, A and B). However, from E12.5 to E17.5, the thymus of heterozygous and wild-type mice grew progressively larger (Fig. 2⇓, C, E, and G), whereas the thymus of homozygous mutant mice remained the same size (Fig. 2⇓, D, F, and H). These data suggest that FgfR2-IIIb signaling is essential for normal thymus development, and the block in thymic development occurs between E12 and E12.5.

FIGURE 2.
FIGURE 2.

Temporal development of the thymus in FgfR2-IIIb−/−lacZ mice. Sagittal sections of H&E-stained E12 (A and B), E12.5 (C and D), E13.5 (E and F), and E17.5 (G and H) fetuses from heterozygous (A, C, E, and G) and null (B, D, F, and H) mice. Although there is no discernible difference in size at E12, by E12.5, the thymus of the wild-type mouse is larger and continues to enlarge, whereas the thymus of the FgfR2-IIIb−/−lacZ mouse remains at the size of E12 embryos. Arrowheads indicate the thymus. H, Heart. A–D, Bar = 0.1 mm; E–H, bar = 0.5 mm.

Expression of FgfR2-IIIb, Fgf7, and Fgf10 in the thymus

The mutant FgfR2 allele contains an IRES-lacZ cassette in the disrupted IIIb exon, although β-galactosidase activity is below detectable levels in heterozygous or homozygous embryos (30). Consequently, expression of FgfR2-IIIb RNA was assessed by radioactive in situ hybridization on E13.5 heterozygote and homozygote thymus using a IIIb isoform-specific probe (lacZ) or a pan-specific FgfR2 TK probe, which detects both FgfR2-IIIb and FgfR2-IIIc (Fig. 3⇓, A–D). The lacZ and TK probe staining were very similar, establishing the presence of FgfR2-IIIb RNA in the thymic epithelium and suggesting no discernible expression of FgfR2-IIIc in the surrounding mesenchyme. Although encoding for a defective receptor, FgfR2-IIIb RNA expression was also detected in the null mice using both the TK (Fig. 3⇓, E and F) and lacZ (Fig. 3⇓, G and H) probes. As a negative control for lacZ staining, the probe was shown to give no signal on sections from wild-type embryos (data not shown). The finding that FgfR2-IIIb is expressed by the thymic epithelial cells is consistent with other FgfR2 expression studies that show that this receptor is expressed specifically in several epithelial cell lineages (18, 19). FgfR2-IIIb was not detected at E12 (data not shown), but was present by E13.5 in the thymic epithelial compartment in both normal and mutant mice. Expression of FgfR2-IIIb continued in the thymic cortical region throughout fetal life, although the progressive infiltration of lymphoid and mesenchymal cells at later stages of development makes assignment of FgfR2-IIIb to specific cell lineages ambiguous. Moreover, FgfR2-IIIb expression was not detected in the surrounding mesenchyme or on the infiltrating lymphoid cells in the mesenchyme. These data suggest a strong spatio-temporal correlation between the loss of FgfR2-IIIb expression in the epithelium and the onset of thymic hypoplasia.

FIGURE 3.
FIGURE 3.

Expression of FgfR2-IIIb, Fgf7, and Fgf10 RNA in mutant mice. Radioactive in situ hybridization on E13.5 embryo sections was used to assess the expression pattern of FgfR2-IIIb in the thymus. An antisense probe against FgfR2 TK domain (A, B, E, and F) and a FgfR2-IIIb isoform-specific probe (lacZ) (C, D, G, and H) was compared on the thymus of heterozygous mice (A–D) or homozygous (E–H) mice. A strong positive signal for TK and lacZ expression is shown over the stromal compartment of the thymus, indicating expression of FgfR2-IIIb. Note that the cartilage primordium of the clavicle is also positive for FgfR2-IIIb expression (A–D). Expression of Fgf7 (I and J) and Fgf10 (K and L) RNA in the mesenchymal cells that surround the thymus stroma of heterozygous mice. Cla, Cartilage primordium of the clavicle; and ca, carotid artery. Bar = 0.1 mm.

FgfR2-IIIb is activated by three main ligands, Fgf3, Fgf7, and Fgf10. An in situ hybridization analysis at E13.5 showed no Fgf3 expression (data not shown), but Fgf7 and Fgf10 were both detected in the surrounding mesenchyme (Fig. 3⇑, I–L), suggesting that they may induce thymic epithelial proliferation. However, mice deficient for Fgf7 are viable and fertile, and no thymic phenotype has been reported (34). Meanwhile, very recent data showed that mice deficient for Fgf10 have a small thymus (35). We have confirmed this observation with an analysis of similar Fgf10-deficient fetuses (Fig. 4⇓). Embryos were assessed at E12.5 and E14.5 by standard histological staining procedures, and a significant reduction in the size of the thymus was observed (Fig. 4⇓, B and D). As seen in the FgfR2-IIIb−/−lacZ mice, the parathyroid appeared to be unaffected (Fig. 4⇓, C and D). A comparison between the thymus of stage-matched Fgf10−/− and FgfR2-IIIb−/−lacZ mice showed that those from FgfR2-IIIb−/−lacZ mice were significantly smaller, suggesting than Fgf7 might be able to provide some compensatory signaling through FgfR2-IIIb in Fgf10-null mice. These findings support the idea that Fgf10 is the major ligand activating FgfR2-IIIb in the thymic epithelium, and this signaling is necessary for normal thymic development.

FIGURE 4.
FIGURE 4.

Targeted disruption of Fgf10 affects normal development of the thymus. Sagittal H&E-stained sections through E12.5 (A and B) and E14.5 (C and D) fetuses of wild-type (A and C) and Fgf10−/− mice (B and D). The smaller size of the thymus is already apparent at E12.5 of Fgf10-deficient mice (A and B). The position of the parathyroid (P) in (C and D) provides a comparison of the size reduction of the thymus at E14.5 in Fgf10-null mice. Bar = 0.1 mm.

Expression of Pax1 and Pax9 are unaffected in FgfR2-IIIb−/−lacZ mice

As an assessment of thymic endoderm differentiation into the epithelial primordium, we examined the expression of two markers of thymic epithelium, the transcription factors Pax1 and Pax9. Moreover, mice deficient for Pax1 also show a reduced thymic size, similar phenotypically to FgfR2-IIIb−/−lacZ mice, as well as quantitative alterations in T cell maturation (11). Although structurally similar and showing a similar thymus-related expression pattern to Pax1, mice deficient for Pax9 are athymic (12). To determine whether these transcription factor genes have an altered thymic expression pattern in the FgfR2-IIIb-deficient mouse, they were analyzed using in situ hybridization at E13.5 both in the heterozygous and homozygous mutant mice. Pax1 and Pax9 were both strongly expressed by the thymic epithelial compartment of heterozygous mice (Fig. 5⇓, A–D). Similarly, in FgfR2-IIIb-deficient mice, expression of both Pax1 and Pax9 occurs normally and at about the expected levels, consistent with a role upstream of FgfR2-IIIb receptor signaling (Fig. 5⇓, E–H).

FIGURE 5.
FIGURE 5.

Expression of Pax1 and Pax9 in the thymus of heterozygous and FgfR2-IIIb−/−lacZ mice. Radioactive in situ hybridization on sagittal sections of E13.5 embryos was used to assess the expression pattern of Pax1 and Pax9 in the thymus. Pax1 (A, B, E, and F) and Pax9 (C, D, G, and H) are expressed in the stromal compartment of the thymus in heterozygous (A–D) and FgfR2-IIIb−/−lacZ mice (E–H). The parathyroid (C and D) also expresses Pax9. Expression of Pax1 and Pax9 is unaffected in FgfR2-IIIb−/−lacZ mice, suggesting that these genes act upstream of FgfR2-IIIb signaling. P, Parathyroid; ca, cartilage primordium of the clavicle; and sa, subclavian artery. Bar = 0.1 mm.

Loss of FgfR2-IIIb or Fgf10 compromises thymic epithelial proliferation

To investigate the mechanism behind the reduced size of the thymus, the levels of apoptosis and cell proliferation were analyzed. Sections were stained using the TUNEL method for the detection of apoptosis and for Ki67 Ag to measure proliferating cells. The TUNEL assay revealed a few apoptotic cells in the thymus of both normal and homozygous mutant mice. As a positive control, other tissues in the sections were stained for the expected level of apoptosis (Fig. 6⇓, A and B). Staining with Ki67 showed that nearly all the stromal cells in the normal thymus stained positive, indicating a high level of proliferation, consistent with its increase in size (Fig. 6⇓C). However, a significant decrease in staining of thymic stromal cells in the homozygous mutant was apparent, indicative of a proliferation block (Fig. 6⇓D). A similar decrease in cell proliferation was observed in mice deficient for Fgf10 (Fig. 6⇓F), whereas the heterozygous mice displayed normal proliferation, indicated by the extensive brown immunoperoxidase staining (Fig. 6⇓E). Because most of the thymic stroma at E13.5 is composed of epithelial cells, this suggests that Fgf10 is the main mesenchymal inducer for the thymic epithelial cell proliferation. However, it is also likely that the reduced proliferation of epithelial cells has a secondary effect on the proliferation of the infiltrating lymphoid cells.

FIGURE 6.
FIGURE 6.

Detection of proliferation and apoptosis in thymus of FgfR2-IIIb−/−lacZ and Fgf10−/− mice. Sagittal sections from E13.5 (A–D) and E14.5 (E and F) fetuses were used to detect proliferating and apoptotic cells by immunohistochemistry. Apoptotic cells from thymus of FgfR2-IIIb−/−lacZ mice were detected using a TUNEL assay (A and B). No detectable difference in the level of apoptosis was seen between the thymus of heterozygous and homozygous mutant mice (A and B). Arrowheads indicate the position of apoptotic cells. Proliferating cells were detected with an Ab against Ki67 in heterozygous (C and E), FgfR2-IIIb−/−lacZ (D), and Fgf10−/− mice (F). A high rate of proliferation was detected in the heterozygous thymus as indicated by the brown immunoperoxidase staining of the cells (C and E), whereas many fewer cells were shown to be proliferating in FgfR2-IIIb−/−lacZ (D) and Fgf10 −/− mice (F). Bar = 0.1 mm.

T cell maturation occurs despite a block of thymic growth

As a preliminary assessment of thymic development, frozen sections of the thymus from E17 heterozygous and homozygous mutant mice were stained with a pan-cytokeratin Ab to identify the epithelium, cytokeratin 5 as a marker of early epithelial cell differentiation, and Thy1.2 for thymocytes. Although the thymus fails to grow in FgfR2-IIIb−/−lacZ mice, many of the epithelial cells remain positive for cytokeratin 5 (Fig. 7⇓, A and B), suggesting that they remain in a precursor form and may not therefore be competent to promote T cell maturation (7). Surprisingly, flow cytometry using cells teased from E17–E18 lobes of homozygous mutant mice showed that T cell maturation is able to proceed to the CD4/CD8-single-positive stage (Fig. 7⇓C). Because the thymuses from FgfR2-IIIb−/−lacZ mice are extremely small, cells from E17 and E18 were pooled for the analyses; in heterozygous mice at E17, there were ∼3 × 106 cells/lobe and, at E18, ∼6 × 106 cells/lobe, but from FgfR2-IIIb−/−lacZ mice, only 15,500 cells/lobe were obtained. With regard to TCR expression, a small proportion of TCR-positive cells was found in the thymus of FgfR2-IIIb−/−lacZ mice. Analysis of αβTCR expression showed ∼4 and 7% positive in the heterozygous and homozygous mutant mice, respectively. Within the αβTCR population from FgfR2-IIIb−/−lacZ mice, 8% were CD4-single-positive, and 24% were CD8-single-positive (data not shown). Similar distributions were found in thymocytes from the heterozygous mice. These findings clearly demonstrate that differentiation to the double-positive stage of thymocyte development occurs in FgfR2-IIIb−/−lacZ mice. A similar proportion of single-positive cells was also found in both heterozygous and homozygous mutant mice. However, the total yield of thymocytes was considerably reduced from the severely hypoplastic thymus. Consequently, we cannot be certain that positive selection occurs normally, although we detect low numbers of single-positive cells that are αβTCR positive. Further analysis of the T cell development, which occurs in neonatal mice, was not possible because these mice die at birth due to agenesis of the lungs.

FIGURE 7.
FIGURE 7.

Immunohistochemistry and FACSanalyses on heterozygous and FgfR2-IIIb−/−lacZ mice. Immunohistochemistry showing sections stained with Abs to Thy1.2, pan-cytokeratin, and keratin 5 on E17 heterozygous (A) and FgfR2-IIIb−/−lacZ (B) mice. In both sections, Thy1.2-positive cells are red, pan-cytokeratin-positive cells are blue, and keratin 5-positive cells are green. Although Thy1.2-positive cells are present in both sections, a greater proportion of cytokeratin-expressing cells (blue) are keratin 5 positive (green) in the mutant thymus compared with the heterozygote thymus. C, FACS analyses of CD4/CD8-positive cells isolated from E17 and E18 heterozygous and pooled E17 and E18 FgfR2-IIIb−/−lacZ mice. The yields of cells were as follows: E18 heterozygotes, 6 × 106 cells/lobe; E17 heterozygotes, 3 × 106 cells/lobe; but for E18/E17 mutant mice, 15,500 cells/lobe. Note that homozygous mutant mice yield CD4- and CD8-single-positive cells.

Discussion

The mouse thymic primordium is formed around E10.5 as an epithelial bud from the third pharyngeal pouch and mesenchyme that differentiates from neural crest cells (1, 2, 3). At about E11.5, the epithelial primordium is first colonized by T cell progenitors that migrate in from blood vessels in the surrounding mesenchyme. At E12, the mesenchyme surrounds, but has not penetrated, the epithelium. Removal of the mesenchyme at this critical phase prevented the development of the epithelium in vitro (36). Subsequently, Shinohara and Honjo showed that the mesenchyme of the early E12 thymus influences MHC class II expression in the epithelium, but they did not study the effects on lymphopoiesis (37, 38). At E12, thymic explants show a limited lymphoid developmental capacity, which can be enhanced by addition of fibroblasts (39, 40).

In a recent study, E12 thymic lobes with intact mesenchyme were shown to generate all T cell populations in vitro, but lobes from which mesenchyme has been removed show poor lymphopoiesis (8). At E13, mesenchyme was shown to start migrating into the epithelial primordium where it can continue to interact with epithelial cells as well as thymocytes (8). Previous studies have reported expression of FgfR2-IIIb in the late gestation thymus (18, 19), but here we show expression of the FgfR2-IIIb receptor on the epithelial primordium of the E13.5 thymus, a stage when only very few lymphoid or mesenchymal cells have infiltrated the epithelium (41). Furthermore, the ligands for FgfR2-IIIb, Fgf7 and Fgf10, are present in the surrounding mesenchyme, indicative of paracrine growth regulation. Consistent with this suggestion, FgfR2 expression was not detected in surrounding thymic mesenchyme, which also contains infiltrating lymphoid cells, nor were Fgf7 and Fgf10 detected in the epithelial compartment. Significantly, the block in thymic growth in FgfR2-IIIb-deficient mice correlates with the stage of development when receptor expression is first detected. Mice deficient for Fgf10, the major ligand for FgfR2-IIIb, also have a similar but less severe hypoplasia (Figs. 4⇑ and 6⇑) (35), providing strong circumstantial evidence that Fgf10 in the mesenchyme acts as the paracrine inducer of stromal epithelial cell proliferation.

Formation and migration of the thymic primordium appears to be normal, suggesting that the role of the Hox3 paralogs is unaffected (42). This view is supported by expression in the hypoplastic thymus of the transcription factors Pax1 and Pax9, which play important roles at this early stage of thymic morphogenesis (11, 12). Moreover, as T cell infiltration and differentiation of the thymus occurs in the FgfR2-IIIb−/−lacZ mice, the function of the whn gene responsible for the athymic nude mouse mutation also appears unaffected (43). Therefore, we would conclude that FgfR2-IIIb most likely functions downstream of the transcription factor cascade involved in thymus morphogenesis, which includes Hox3 paralogs, Pax1/Pax9, and whn, although this does not exclude a continued requirement for these transcription factors (reviewed in Ref. 44).

Thymic epithelial cells express distinct cytokeratins that reflect their proliferative potential and maturational status (7). Most cells express cytokeratins 8 and 18, but a distinct subset also express cytokeratin 5. This subset contains precursors that generate the major population of cytokeratin 5-negative cortical epithelial cells (7, 45). Our studies show that, in the hypoplastic thymus of mice deficient for FgfR2-IIIb, many of the epithelial cells fail to progress from the cytokeratin 5/cytokeratin 18-positive stage (Fig. 7⇑A and data not shown). A similar block in maturation of thymic epithelial cells has been reported in mice expressing a human CD3ε transgene (7), but in this case, the effect has been attributed to a failure to generate thymocytes that are required for epithelial maturation (46). In our studies we show that, although the epithelial phenotype is similar to that above, the thymus of FgfR2-IIIb−/−lacZ mice is capable of supporting thymocyte development. However, we cannot exclude that the small number of cytokeratin 5-negative cells are both necessary and sufficient to facilitate the observed thymocyte maturation. As well as generating CD4/CD8-double- and -single-positive cells, there is also a small proportion of TCR αβ cells in the thymus of mutant mice that includes a small subset of the single-positive populations. Although thymocytes transit the double-negative 1-double-negative 4 pathway, we cannot exclude changes in the CD25:CD44 ratios during the earlier highly proliferative phase of T cell expansion, because in line with the small size of the thymus, thymocyte numbers are severely reduced.

In summary, signaling through the FgfR2-IIIb receptor on thymic epithelial cells is crucial for thymic development. Mesenchyme, which initially surrounds the epithelium, produces Fgf7 and Fgf10 capable of activating this receptor, although the genetic evidence suggests the latter is the major ligand. While in the limb, FgfR2-IIIb signaling appears to be crucial for cell survival (30), and in the thymus it is required for proliferation of the thymic epithelium after E12.5. The lack of epithelial cell proliferation and the associated retention of primarily cytokeratin 5/cytokeratin 18-positive cells does not block the maturation of thymocytes to the CD4/CD8-double-positive stage, although the yield of cells is severely reduced. It is more difficult to be sure that normal positive selection occurs because of the gestational age and low cell numbers. It seems likely that the effect on thymocyte yield is secondary to the failure of epithelial proliferation, but we cannot exclude the possibility that there are direct effects on thymocytes themselves. These findings suggest that different paracrine signals may be required for growth and differentiation of the thymic epithelium. As mesenchyme also infiltrates the thymic stroma, these cells might have additional direct effects on thymocyte development, because they express an extracellular matrix able to activate thymocyte integrins (8). In this context, it will be interesting to investigate whether Fgfs can wholly substitute for the effect of removal of mesenchyme on lymphopoiesis in cultures of E12 thymus.

Acknowledgments

We thank Dr. David Ornitz for generously providing Fgf-10−/− mouse embryos and Dr. Mike Owen and Prof. Eric Jenkinson for suggestions and critically reading the manuscript. We also thank George Elia for help with producing numerous histological sections and Peter Hagger and Janet MacDonald for looking after the mice.

Footnotes

  • ↵1 Current address: Laboratoire de Psychobiologie des Comportements Adaptatifs, Institut National de la Santé et de la Recherche Médicale Unité 259, Université de Bordeaux II, Domaine de Carreire, Bordeaux, France

  • ↵2 Address correspondence and reprint requests to Dr. Clive Dickson, Imperial Cancer Research Fund, 44 Lincoln’s Inn Fields, London WC2A 3PX, U.K. E-mail address: dickson{at}icrf.icnet.uk

  • 3 Abbreviations used in this paper: E, embryonic day; Fgf, fibroblast growth factor; IRES, internal ribosomal entry site; TK, tyrosine kinase.

  • Received April 30, 2001.
  • Accepted June 18, 2001.

References

  1. ↵
    Cordier, A., S. Haumont. 1980. Development of thymus, parathyroids and ultim-branchial bodies in NMR1 and nude mice. Am. J. Anat. 157: 227
    OpenUrlCrossRefPubMed
  2. ↵
    van Ewijk, W.. 1991. T-cell differentiation is influenced by thymic microenvironments. Annu. Rev. Immunol. 23: 2614
    OpenUrl
  3. ↵
    Anderson, G., N. Moore, J. Owen. 1996. Cellular interactions in thymocyte development. Annu. Rev. Immunol. 14: 73
    OpenUrlCrossRefPubMed
  4. ↵
    Le Douarin, N., F. Jotereau. 1975. Tracing cells of the avian thymus through embryonic life in interspecific chimeras. J. Exp. Med. 142: 17
    OpenUrlAbstract/FREE Full Text
  5. ↵
    Moore, M., J. Owen. 1967. Experimental studies on the development of the thymus. J. Exp. Med. 126: 715
    OpenUrlAbstract
  6. ↵
    Ritter, M., R. Boyd. 1993. Development in the thymus: it takes two to tango. Immunol. Today 14: 462
    OpenUrlCrossRefPubMed
  7. ↵
    Klug, D., C. Carter, E. Crouch, D. Roop, C. Conti, E. Richie. 1998. Interdependence of cortical thymic epithelial cell differentiation and T-lineage commitment. Proc. Natl. Acad. Sci. USA 95: 11822
    OpenUrlAbstract/FREE Full Text
  8. ↵
    Suniara, R., E. Jenkinson, J. Owen. 2000. An essential role for thymic mesenchyme in early T cell development. J. Exp. Med. 191: 1051
    OpenUrlAbstract/FREE Full Text
  9. ↵
    Manley, N., M. Capecchi. 1995. The role of Hox-3 in mouse thymus and thyroid development. Development 121: 1989
    OpenUrlAbstract
  10. ↵
    Su, D., N. Manley. 2000. Hoxa3 and Pax1 transcription factors regulate the ability of fetal thymic epithelial cells to promote thymocyte development. J. Immunol. 164: 5753
    OpenUrlAbstract/FREE Full Text
  11. ↵
    Wallin, J., H. Eibel, A. Neubuser, J. Wilting, H. Koseki, R. Balling. 1996. Pax1 is expressed during development of the thymus epithelium and is required for normal T-cell maturation. Development 122: 23
    OpenUrlAbstract
  12. ↵
    Peters, H., A. Neubuser, K. Kratochwil, R. Balling. 1998. Pax9-deficient mice lack pharyngeal pouch derivatives and teeth and exhibit craniofacial and limb abnormalities. Genes Dev. 12: 2735
    OpenUrlAbstract/FREE Full Text
  13. ↵
    Johnson, D., L. Williams. 1993. Structural and functional diversity in the FGF receptor multigene family. Adv. Cancer Res. 60: 1
    OpenUrlPubMed
  14. ↵
    McKeehan, W. L., F. Wang, M. Kan. 1998. The heparan-sulfate fibroblast growth-factor family–diversity of structure and function. Prog. Nucleic Acid Res. Mol. Biol. 59: 135
    OpenUrlPubMed
  15. ↵
    Burgess, W. H., T. Maciag. 1989. The heparin binding (fibroblast) growth factor family proteins. Annu. Rev. Biochem. 58: 575
    OpenUrlCrossRefPubMed
  16. ↵
    Basilico, C., D. Moscatelli. 1992. The FGF family of growth-factors and oncogenes. Adv. Cancer Res. 59: 115
    OpenUrlCrossRefPubMed
  17. ↵
    Ornitz, D. M., J. S. Xu, J. S. Colvin, D. G. McEwen, C. A. MacArthur, F. Coulier, G. X. Gao, M. Goldfarb. 1996. Receptor specificity of the fibroblast growth-factor family. J. Biol. Chem. 271: 15292
    OpenUrlAbstract/FREE Full Text
  18. ↵
    Finch, P., G. Cunha, J. Rubin, J. Wong, D. Ron. 1995. Pattern of keratinocyte growth factor and keratinocyte growth factor receptor expression during mouse fetal development suggests a role mediating morphogenetic mesenchymal-epithelial interactions. Dev. Dyn. 203: 223
    OpenUrlCrossRefPubMed
  19. ↵
    Orr-Urtreger, A., M. Bedford, T. Burakova, E. Arman, Y. Zimmer, A. Yayon, D. Givol, P. Lonai. 1993. Developmental localization of the splicing alternatives of fibroblast growth-factor receptor-2 (FGFR2). Dev. Biol. 158: 475
    OpenUrlCrossRefPubMed
  20. ↵
    Peters, K., S. Werner, G. Chen, L. Williams. 1992. Two FGF receptor genes are differentially expressed in epithelial and mesenchymal tissues during limb formation and organogenesis in the mouse. Development 114: 233
    OpenUrlAbstract
  21. ↵
    Yamasaki, M., A. Miyake, S. Tagashira, N. Itoh. 1996. Structure and expression of the rat messenger-RNA encoding a novel member of the fibroblast growth-factor family. J. Biol. Chem. 271: 15918
    OpenUrlAbstract/FREE Full Text
  22. ↵
    Mason, I., F. Fuller-Pace, R. Smith, C. Dickson. 1994. FGF-7 (keratinocyte growth factor) expression during mouse development suggests roles in myogenesis, forebrain regionalisation and epithelial-mesenchymal interactions. Mech. Dev. 45: 15
    OpenUrlCrossRefPubMed
  23. ↵
    Arman, E., R. Haffner-Krausz, Y. Chen, J. K. Heath, P. Lonai. 1998. Targeted disruption of fibroblast growth-factor (FGF) receptor-2 suggests a role for FGF signaling in pregastrulation mammalian development. Proc. Natl. Acad. Sci. USA 95: 5082
    OpenUrlAbstract/FREE Full Text
  24. ↵
    Xu, X. L., M. Weinstein, C. L. Li, M. Naski, R. I. Cohen, D. M. Ornitz, P. Leder, C. X. Deng. 1998. Fibroblast-growth-factor-receptor-2 (FGFR2)-mediated reciprocal regulation loop between FGF8 and FGF10 is essential for limb induction. Development 125: 753
    OpenUrlAbstract
  25. ↵
    Arman, E., R. Haffner-Krausz, M. Gorivodsky, P. Lonai. 1999. Fgfr2 is required for limb outgrowth and lung-branching morphogenesis. Proc. Natl. Acad. Sci. USA 96: 11895
    OpenUrlAbstract/FREE Full Text
  26. ↵
    Werner, S., H. Smola, X. Liao, M. Longaker, T. Krieg, P. Hofschneider, L. Williams. 1994. The function of KGF in morphogenesis of epithelium and reepithelialization of wounds. Science 266: 819
    OpenUrlAbstract/FREE Full Text
  27. ↵
    Peters, K., S. Werner, X. Liao, S. Wert, J. Whitsett, L. Williams. 1994. Targeted expression of a dominant negative FGF receptor blocks branching morphogenesis and epithelial differentiation of the mouse lung. EMBO J. 13: 3296
    OpenUrlPubMed
  28. ↵
    Celli, G., W. J. Larochelle, S. MacKem, R. Sharp, G. Merlino. 1998. Soluble dominant-negative receptor uncovers essential roles for fibroblast growth-factors in multiorgan induction and patterning. EMBO J. 17: 1642
    OpenUrlAbstract
  29. ↵
    DeMoerlooze, L., B. Spencer-Dene, J. M. Revest, M. Hajihosseini, I. Rosewell, C. Dickson. 2000. An important role for the IIIb isoform of fibroblast growth factor receptor 2 (FGFR2) in mesenchymal-epithelial signalling during mouse organogenesis. Development 127: 483
    OpenUrlAbstract
  30. ↵
    Revest, J.-M., B. Spencer-Dene, K. Kerr, L. De Moerlooze, I. Rosewell, C. Dickson. 2001. Fibroblast growth factor receptor 2-IIIb acts upstream of Shh and Fgf4 and is required for limb bud maintenance but not for the induction of Fgf8, Fgf10, Msx1 or Bmp4. Dev. Biol. 231: 47
    OpenUrlCrossRefPubMed
  31. ↵
    Gu, H., Y.-R. Zou, K. Rajewsky. 1993. Independent control of immunoglobulin switch recombination at individual switch regions evidenced through Cre-loxP-mediated gene targeting. Cell 73: 1155
    OpenUrlCrossRefPubMed
  32. ↵
    Min, H., D. Danilenko, S. Scully, B. Bolon, B. Ring, J. Tarpley, M. DeRose, W. Simonet. 1998. Fgf-10 is required for both limb and lung development and exhibits striking functional similarity to Drosophila branchless. Genes Dev. 12: 3156
    OpenUrlAbstract/FREE Full Text
  33. ↵
    Kettunen, P., I. Karavanova, I. Thesleff. 1998. Responsiveness of developing dental tissues to fibroblast growth factors: expression of splicing alternatives of FGFR1, -2, -3, and of FGFR4; and stimulation of cell proliferation by FGF-2, -4, -8, and -9. Dev. Genet. 22: 374
    OpenUrlCrossRefPubMed
  34. ↵
    Guo, L., L. Degenstein, E. Fuchs. 1996. Keratinocyte growth factor is required for hair development but not for wound healing. Genes Dev. 10: 165
    OpenUrlAbstract/FREE Full Text
  35. ↵
    Ohuchi. 2000. FGF10 acts as a major ligand for FGF receptor 2 IIIb in mouse multi-organ development. Biochem. Biophys. Res. Commun. 277: 643
    OpenUrlCrossRefPubMed
  36. ↵
    Auerbach, R.. 1960. Morphogenetic interactions in the development of the mouse thymus. Dev. Biol. 2: 271
    OpenUrlCrossRefPubMed
  37. ↵
    Shinohara, T., T. Honjo. 1996. Epidermal growth factor can replace thymic mesenchyme in induction of embryonic thymus morphogenesis in vitro. Eur. J. Immunol. 26: 747
    OpenUrlCrossRefPubMed
  38. ↵
    Shinohara, T., T. Honjo. 1997. Studies in vitro on the mechanism of the epithelial/mesenchymal interaction in the early fetal thymus. Eur. J. Immunol. 27: 522
    OpenUrlCrossRefPubMed
  39. ↵
    Itoi, M., T. Amagai. 1998. Inductive role of fibroblastic cell lines in development of the mouse thymus anlage in organ culture. Cell. Immunol. 183: 32
    OpenUrlCrossRefPubMed
  40. ↵
    Amagai, T., M. Itoi, Y. Kondo. 1995. Limited development capacity of the earliest embryonic murine thymus. Eur. J. Immunol. 25: 757
    OpenUrlCrossRefPubMed
  41. ↵
    Suniara, R., E. Jenkinson, J. Owen. 1999. Studies on the phenotype of migrant thymic stem cells. Eur. J. Immunol. 29: 75
    OpenUrlCrossRefPubMed
  42. ↵
    Manley, N., M. Capecchi. 1998. Hox group 3 paralogs regulate the development and migration of the thymus, thyroid and parathyroid glands. Dev. Biol. 1951..
  43. ↵
    Nehls, M., B. Kyewski, M. Messerle, R. Waldschütz, K. Schuddekopf, A. Smith, T. Boehm. 1996. Two genetically separable steps in the differentiation of thymic epithelium. Science 272: 886
    OpenUrlAbstract
  44. ↵
    Manley, N.. 2000. Thymus organogenesis and molecular mechanisms of thymic epithelial cell differentiation. Semin. Immunol. 12: 421
    OpenUrlCrossRefPubMed
  45. ↵
    Klug, D., E. Crouch, C. Carter, L. Coghlan, C. Conti, E. Richie. 2000. Transgenic expression of cyclin D1 in thymic epithelial precursors promotes epithelial and T cell development. J. Immunol. 164: 1881
    OpenUrlAbstract/FREE Full Text
  46. ↵
    van Ewijk, W., G. Hollander, C. Terhorst, B. Wang. 2000. Stepwise development of thymic microenvironments in vivo is regulated by thymocyte subsets. Development 127: 1583
    OpenUrlAbstract
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools

Jump to section