Zebrafish cdx1b regulates differentiation of various intestinal cell lineages

Authors

  • Yi-Hua Chen,

    1. Institute of Cellular and Organismic Biology (formerly Institute of Zoology), Academia Sinica, Nankang, Taipei, Taiwan, R.O.C.
    Search for more papers by this author
  • Yu-Fen Lu,

    1. Institute of Cellular and Organismic Biology (formerly Institute of Zoology), Academia Sinica, Nankang, Taipei, Taiwan, R.O.C.
    Search for more papers by this author
  • Ting-Yi Ko,

    1. Institute of Bioscience and Biotechnology, National Taiwan Ocean University, Keelung, Taiwan, R.O.C.
    Search for more papers by this author
  • Ming-Yuan Tsai,

    1. Institute of Bioscience and Biotechnology, National Taiwan Ocean University, Keelung, Taiwan, R.O.C.
    Current affiliation:
    1. Graduate Institute of Life Sciences, National Defense Medical Center, National Defense University, Neihu, Taipei, Taiwan 114, R.O.C.
    Search for more papers by this author
  • Che Yi Lin,

    1. Institute of Bioscience and Biotechnology, National Taiwan Ocean University, Keelung, Taiwan, R.O.C.
    Search for more papers by this author
  • Chia-Chi Lin,

    1. Institute of Cellular and Organismic Biology (formerly Institute of Zoology), Academia Sinica, Nankang, Taipei, Taiwan, R.O.C.
    Search for more papers by this author
  • Sheng-Ping L. Hwang

    Corresponding author
    1. Institute of Cellular and Organismic Biology (formerly Institute of Zoology), Academia Sinica, Nankang, Taipei, Taiwan, R.O.C.
    2. Institute of Bioscience and Biotechnology, National Taiwan Ocean University, Keelung, Taiwan, R.O.C.
    • Institute of Cellular and Organismic Biology (formerly Institute of Zoology), Academia Sinica, Nankang, Taipei, Taiwan 115, R.O.C.
    Search for more papers by this author

Abstract

Both antisense morpholino oligonucleotide (MO)-mediated knockdown and overexpression experiments were performed to analyze zebrafish cdx1b's function in intestinal cell differentiation. Substantial reductions in goblet cell numbers were detected in intestines of 102- and 120-hours post-fertilization (hpf) cdx1b MO-injected embryos (morphants) compared to cdx1b-4-base mismatched (4mm)-MO-injected and wild type embryos. A significant decrease in enteroendocrine cell numbers was also observed in intestines of 96-hpf cdx1b morphants. Furthermore, ectopic cdx1b expression caused notable increases in respective cell numbers of enteroendocrine and goblet cells in intestines of 96- and 98-hpf injected embryos. Decreased PepT1 expression was detected in enterocytes of intestines in cdx1b morphants from 80 to 102 hr of development. In addition, increased cell proliferation was detected in intestines of cdx1b morphants. Overall, our results suggest that zebrafish cdx1b plays important roles in regulating intestinal cell proliferation and the differentiation of various intestinal cell lineages. Developmental Dynamics 238:1021–1032, 2009. © 2009 Wiley-Liss, Inc.

INTRODUCTION

Mammalian intestines, which consist of the small intestine and large intestine, are essential for maintaining normal physiology and homeostasis. They are composed of an inner epithelium that is surrounded by connective tissue, smooth muscle, and enteric nerves. Small intestinal epithelium is organized into crypts of Lieberkuhn, a pocket-like invagination into the gut mucosa, and villi, finger-like projections extending into the intestinal lumen (Sancho et al.,2003). Crypts of Lieberkuhn contain proliferative stem cells that provide renewal of various differentiated intestinal cells. Villi contain three types of mature intestinal cells: enterocytes that function to absorb nutrients, goblet cells that secrete mucous as a protective barrier, and enteroendocrine cells that produce different gastrointestinal hormones. Paneth cells, a fourth type of mature epithelium cell, are located at the crypt base and secrete antibacterial peptides.

Cell fate decisions leading to the four types of differentiated epithelial cells have been shown to be regulated by different signaling pathways, which include Notch, BMP, and Wnt (Scoville et al.,2008). Hes1-deficient mice exhibited secretory cell expansion at the expense of enterocytes, indicating the importance of Notch signaling in the non-secretory lineage fate decision (Jensen et al.,2000). Villin-Cre BMPR1Afx/fx mutant mice exhibited defects in goblet, enteroendocrine, and Paneth cell maturation, indicating that BMP signaling is important in the maturation of these three secretory cell types (Auclair et al.,2007). Similarly, Wnt signaling was shown to play an important role in Paneth cell development (van Es et al.,2005).

In addition to paracrine factor signaling, several transcription factors have been implicated in secretory lineage decisions (Schonhoff et al.,2004). A basic helix-loop-helix protein, Math1, is required for the commitment of all secretory cell lineages, since loss of Math1 in mice caused depletion of goblet, enteroendocrine, and Paneth cells without affecting enterocytes (Yang et al.,2001). This result indicates that secretory cells arise from a common precursor that expresses Math1, while absorptive cells arise from a Math1-independent progenitor. Similarly, another bHLH protein, neurogenin 3 (NGN3), is also essential for the differentiation of endocrine cells in the intestinal epithelium. Overexpression of NGN3 via a villin promoter in mice produced increased numbers of cells expressing the pan-endocrine marker, chromogranin A, accompanied by decreased numbers of goblet cells in the intestines (Lopez-Diaz et al.,2007). That study indicated that neurogenin 3 can redirect the differentiation of bipotential secretory progenitors to an endocrine rather than a goblet cell fate. Approximately ten different endocrine lineages, producing a variety of hormones including somatostatin, serotonin, substance P, glucagon-like peptide, gastrin, ghrelin, GIP, secretin, cholecystokinin, and peptide YY, develop in the mouse intestine. NeuroD1/BETA2 is required for the expression of both secretin and cholecystokinin because NeuroD1/BETA2 null mice failed to develop endocrine cells that can produce these two hormones (Naya et al.,1997). Deletion of Pax6 in mice lacked endocrine cells that express the glucagon-like peptide, whereas Pax4 null mice failed to generate serotonin- and somatostatin-producing endocrine cells in the small intestine (Larsson et al.,1998; Hill et al.,1999).

Because of the availability of ENU mutant fish and reverse genetics, zebrafish have been adopted as a model organism for studying gastrointestinal development in recent years. Compared to mammalian intestines, zebrafish do not have a large intestine and lack the submucosal layer. There are no crypts of Lieberkuhn or Paneth cells in the intestinal epithelium. Thus, villi contain only three types of differentiated cells, namely, absorptive enterocytes, and goblet and enteroendocrine cells (Ng et al.,2005; Wallace et al.,2005). Although a study using both aei and mib mutants demonstrated the conserved role of Delta-Notch signaling in inhibiting the secretory cell fate in the zebrafish intestine (Crosnier et al.,2005), regulation of lineage fate decisions by other paracrine signaling pathways or transcription factors remains unclear. Previously, we reported that zebrafish cdx1b, a caudal-related homeobox gene, regulates foxa2 and gata5 expressions, which affect early endoderm formation (Cheng et al.,2008). In addition, cdx1b is an intestinal-specific gene due to its exclusive expression in the developing intestine after 28 hr of development. In this study, both antisense morpholino oligonucleotide (MO)-mediated knockdown and overexpression experiments were performed to analyze functions of cdx1b in intestinal cell proliferation and differentiation. Our results suggest that although cdx1b shares syntenic conservation with mammalian Cdx1, it functions like mammalian Cdx2 to regulate cell proliferation and the differentiation of enterocytes, and goblet and enteroendocrine cells in the zebrafish intestine.

RESULTS

Intestinal-Specific Expression of cdx1b and Abnormal Intestinal Development in cdx1b Morphants

cdx1b is an intestinal-specific gene because it is expressed in the developing intestine after 28 hr post-fertilization (hpf) (Cheng et al.,2008). Paraffin sectioning of 100-hpf embryos that were hybridized with cdx1b showed that cdx1b was uniformly expressed in all intestinal epithelial cells (Fig. 1). To investigate the function of cdx1b during zebrafish intestinal development, we conducted an antisense morpholino oligonucleotide (MO)-mediated knockdown experiment. A previously described cdx1b MO to block cdx1b translation and a control cdx1b-4mm MO were used in this study (Cheng et al.,2008). Roughly half of 72- and 96-hpf cdx1b morphants displayed a curled-down body axis (Fig. 2D,H). In addition, pronounced pericardial edema accompanied by a heart-looping defect was detected in 72- and 96-hpf cdx1b morphants compared to cdx1b-4mm-MO-injected and wild type embryos (Fig. 2A–H). Paraffin sectioning further demonstrated that epithelial cells appeared cuboidal with pleomorphic nuclei located at random positions with respect to the apical-basal axis in the intestinal bulb, mid-intestine, and posterior-intestine regions in 96-hpf cdx1b morphants compared to that of cdx1b-4mm-MO-injected embryos (Fig. 2I–P).

Figure 1.

cdx1b is expressed in the intestine. A: A 100-hr post-fertilization (hpf) embryo hybridized with a cdx1b antisense RNA probe. BE: Histological analyses of paraffin sagittal sections of a cdx1b antisense RNA-hybridized embryo. Scale bars = 100 μm.

Figure 2.

cdx1b antisense morpholino oligonucleotide (MO) knockdown analyses. AD: Wild type (A) and cdx1b MO-injected (B–D) 72 hr post-fertilization (hpf) embryos. EH: Wild-type (E) and cdx1b MO-injected (F–H) 96-hpf embryos. IL: Histological analyses of paraffin sagittal sections of a 96-hpf cdx1b-4mm-MO-injected embryo. Higher magnifications depicting the intestinal bulb (J), mid-intestine (K), and posterior intestine (L). MP: Histological analyses of paraffin sagittal sections of a 96-hpf cdx1b morphant. Higher magnifications depicting the intestinal bulb (N), mid-intestine (O), and posterior intestine (P). Arrows indicate locations of goblet cells in the mid-intestine (K,O). Scale bars = 100 μm.

Decreased Cell Numbers of Goblet and Enteroendocrine Cells Detected in Intestines of cdx1b Morphants

The zebrafish intestine contains three types of differentiated epithelial cells: enterocytes, and goblet and enteroendocrine cells. We used two different methods to evaluate goblet cell numbers in the intestine. We previously identified a homologue of Xenopus anterior gradient 2 (agr2) gene in zebrafish (Shih et al.,2007). Whole-mount in situ hybridization indicated that zebrafish agr2 is expressed in most organs that contain mucus-secreting cells including intestinal goblet cells. A significant reduction (75%) in agr2-expressing goblet cell number was observed in 102-hpf cdx1b morphants with or without the curled-down body axis compared to those of cdx1b-4mm-MO-injected and wild type embryos (Fig. 3A–H,U). We also used alcian blue staining to detect sulfated and carboxylated sialomucins in intestinal goblet cells. Similarly, substantial decreases in alcian blue–stained goblet cell numbers were detected in both 102- and 120-hpf cdx1b morphants compared to cdx1b-4mm-MO-injected and wild type embryos (Fig. 3I–L, Q–T). Approximately 60% reductions in the alcian blue–stained goblet cell number in the intestine of 102- and 120-hpf cdx1b morphants were observed (Fig. 3V).

Figure 3.

Inhibition of cdx1b function affects the development of goblet cells in the intestine. AD: Wild type (A), cdx1b-4mm MO-injected (B), and cdx1bMO-injected (C, D) 102-hr post-fertilization (hpf) embryos were hybridized with agr2 antisense RNA probes. EH: Wild type (E), cdx1b-4mm MO-injected (F), and cdx1b MO-injected (G, H) 102-hpf deyolked embryos that had been hybridized with agr2 antisense RNA probes. IL: Wild type (I), cdx1b-4mm MO-injected (J), and cdx1b MO-injected (K, L) 102-hpf embryos were stained with alcian blue. MP: Higher magnifications of alcian blue–stained paraffin sagittal sections corresponding to the mid-intestine (M, O) and posterior intestine (N, P) of 102-hpf cdx1b-4mm MO-injected (M, N) and cdx1b MO-injected (O, P) embryos. Arrows indicate locations of alcian blue–stained goblet cells in the intestine. QT: Wild type (Q), cdx1b-4mm MO-injected (R), and cdx1b MO-injected (S, T) 120-hpf embryos were stained with alcian blue. U: Comparison of agr2-expressing goblet cell numbers in the intestine of 102-hpf wild type (n = 28), cdx1b-4mm MO-injected (n = 28), and cdx1b MO-injected (n = 21) embryos. V: Comparison of alcian blue–stained goblet cell numbers in the intestine of wild type (n102 = 20; n120 = 21), cdx1b-4mm MO-injected (n102 = 23; n120 = 21), and cdx1b MO-injected (n102 = 22; n120 = 23) embryos from 102 (red bars) and 120 hpf (green bars). W: Comparison of the percentages of intestinal goblet cell numbers reveals decreased goblet cell numbers in cdx1b MO-injected (n = 7) embryos compared to wild type (n = 5) and cdx1b-4mm MO-injected (n = 5) embryos. Error bars indicate standard errors. Student's t-test was conducted to compare cdx1b MO-injected embryos with either wild type or cdx1b-4mm MO-injected embryos. *P < 0.001 in both comparisons. An, anus; gc, goblet cell; mc, mucus cell; ov, otic vesicle; ph, pharynx. Scale bars = 100 μm.

In our previous report, we showed that cdx1b regulates early endoderm formation by modulating downstream factors of Nodal signaling (Cheng et al.,2008). To ensure that reduction of intestinal goblet cell number in cdx1b morphants was not due to an early effect of cdx1b on the endodermal cell number, we performed alcian blue staining on paraffin sections of 102-hpf cdx1b morphants and cdx1b-4mm-MO-injected and wild type embryos (Fig. 3M–P). We counted the numbers of alcian blue–stained goblet cells and total cell numbers of mid- and posterior-intestine from respective embryos. Although approximately 20% reductions in total cell numbers in the mid- and posterior-intestine were detected in cdx1b morphants, a 55% reduction calculated as the percentage of goblet cell numbers in the intestine was still identified in 102-hpf cdx1b morphants compared to cdx1b-4mm-MO-injected and wild type embryos (Fig. 4W).

Figure 4.

Knockdown of cdx1b function affects the development of enteroendocrine cells in the intestine. A, B: cdx1b-4mm MO-injected (A) and cdx1b MO-injected (B) 96-hr post-fertilization (hpf) deyolked embryos hybridized with nkx2.2a antisense RNA probes. C: Comparison of nkx2.2a-expressing enteroendocrine cell numbers in cdx1b-4mm MO-injected (n = 21) and cdx1b MO-injected (n = 21) embryos. DH: Wild type (D), cdx1b-4mm MO-injected (E, F), and cdx1b MO-injected (G, H) 96-hpf deyolked embryos hybridized with glucagon antisense RNA probes. I: Comparison of glucagon-expressing enteroendocrine cell numbers in wild type (n = 15), cdx1b-4mm MO-injected (n = 31), and cdx1b MO-injected (n = 29) embryos. Error bars indicate standard errors. Student's t-test was conducted to compare cdx1b MO-injected embryos with either wild type or cdx1b-4mm MO-injected embryos. *P < 0.001 in both comparisons. Ac, pancreatic α cells; ee, enteroendocrine cells; p, pancreas. Scale bars = 100 μm.

We also used two different probes to evaluate enteroendocrine cell numbers in the intestine. Zebrafish nkx2.2a is expressed in intestinal enteroendocrine cells in addition to being expressed in the brain, ventral neural tube, and pancreas (Ng et al.,2005). Nkx2.2a-expressing enteroendocrine cells were distributed throughout the entire intestinal tract of 96-hpf cdx1b-4mm-MO-injected embryos (Fig. 4A). In cdx1b morphants, a substantial decline (63 %) in the nkx2.2a-expressing enteroendocrine cell number was observed (Fig. 4A–C). Intestinal enteroendocrine cells produce different hormones, including glucagon, which is also synthesized in pancreatic α cells. Glucagon-expressing enteroendocrine cells were mainly distributed in the posterior intestinal bulb and mid-intestine of 96-hpf wild type embryos (Fig. 4D). A significant decline (73%) in glucagon-expressing enteroendocrine cell numbers was detected in 96-hpf cdx1b morphants compared to cdx1b-4mm-MO-injected and wild type embryos (Fig. 4D–I). Taken together, development of both goblet and enteroendocrine secretory cells in the intestine was affected in cdx1b morphants.

Decreased Expression of PepT1 Detected in Enterocytes in the Intestine of cdx1b Morphants

The intestinal absorption of di- and tri-peptides generally occurs via the oligopeptide transporter, PepT1, while the intestinal fatty acid–binding protein is responsible for intracellular transport of long-chain fatty acids (Verri et al.,2003; Her et al.,2004). Both PepT1 and IFABP have been used as terminal differentiation markers for intestinal enterocytes. Previously, we demonstrated that decreased expression levels and domains of IFABP were detected in a majority of 72-hpf cdx1b morphants (Cheng et al.,2008). Similarly, substantial reductions in both the expression level and domain of PepT1 in the intestinal bulb were identified in a majority of 80-hpf cdx1b morphants compared to cdx1b-4mm-MO-injected and wild type embryos (Fig. 5A–C). Although reduced PepT1 expression level and domain in the intestinal bulb were not easily distinguished in respective 96- and 102-hpf cdx1b morphants compared to control embryos, quantitative PCR results confirmed decreased PepT1 expression levels in respective cdx1b morphants from 80 to 102 hr of development (Fig. 5). Overall, these results demonstrated that the development of enterocytes in the intestine was affected in cdx1b morphants.

Figure 5.

Knockdown of cdx1b function affects PepT1 expression in enterocytes. AC: Wild type (A), cdx1b-4mm MO-injected (B), and cdx1b MO-injected (C) 80-hr post-fertilization (hpf) embryos hybridized with PepT1 antisense RNA probes. DF: Wild type (D), cdx1b-4mm MO-injected (E), and cdx1b MO-injected (F) 96-hpf embryos hybridized with PepT1 antisense RNA probes. G: Q-PCR indicated decreased PepT1 expression in cdx1b morphants compared to wild type and cdx1b-4mm MO-injected embryos from 80 to 102 hr of development. Error bars indicate standard errors. Results from duplicated experiments are shown. Significant differences between ΔΔCt values based on Student's t-test were used to indicate different gene expressions expressed as fold changes. Student's t-test was conducted to compare cdx1b MO-injected embryos with either wild type or cdx1b-4mm MO-injected embryos.*P < 0.05 in both comparisons. H, I: cdx1b-4mm MO-injected (H) and cdx1b MO-injected (I) 102-hpf embryos hybridized with PepT1 antisense RNA probes. Ib, intestinal bulb. Scale bars = 100 μm.

Intestinal Epithelial Cell Morphology Affected in cdx1b Morphants

Since an abnormal appearance of intestinal epithelium and decreased numbers of both enteroendocrine and goblet cells were detected in cdx1b morphants, we used electron microscopy to better define the ultrastructures of enterocytes and goblet cells. In 80-hpf cdx1b morphants, enterocytes were less elongated, and apical microvilli were shorter and less numerous in the intestinal bulb and mid-intestine compared to those of wild type embryos (Fig. 6A–D). No goblet cells with mucous granules were detected in the mid-intestines of 80-hpf cdx1b morphants compared to wild type embryos (Fig. 6B,D). In 96-hpf cdx1b morphants, enterocytes remained cuboidal shaped with shorter but numerous apical microvilli in the intestinal bulb and mid-intestine compared to wild type embryos (Fig. 6E–H). The majority of goblet cells were pre-goblet cells with a small number of mucous granules in the mid-intestines of 96-hpf cdx1b morphants (Fig. 6H). Some of these were clustered together and a similar spatial distribution pattern of goblet cells could be visualized in both 102- and 120-hpf cdx1b morphants by both alcian blue staining and agr2 RNA hybridization (Fig. 3).

Figure 6.

Transmission electron microscopy (TEM) revealed abnormal ultrastructural morphology in the intestine of cdx1b morphants. AD: EM sagittal sections of the intestinal bulb (A, C) and mid-intestine (B, D) of 80-hr post-fertilization (hpf) wild type (A, B) and cdx1b MO-injected (C, D) embryos. EH: EM sagittal sections of the intestinal bulb (E, G) and mid-intestine (F, H) of 96-hpf wild type (E, F) and cdx1b MO-injected embryos (G, H). Arrows indicate the locations of goblet cells in the mid-intestine. Scale bars =10 μm.

Ectopic cdx1b Expression Increased the Respective Cell Numbers of Goblet and Enteroendocrine Cells in the Intestine of Injected Embryos

To further confirm the decreases in the respective cell numbers of goblet and enteroendocrine cells in the intestine of cdx1b morphants, we also overexpressed cdx1b by either injecting cdx1b mRNA or Tol2-CMV-cdx1b DNA into 1–2-cell zygotes. An approximately 1.6-fold increase in agr2-expressing goblet cell numbers was detected in 98-hpf cdx1b ectopically expressed embryos compared to lacZ-injected or wild type embryos (Fig. 7A–D, I). A similar increase (1.4-fold) in alcian blue–stained goblet cell numbers was identified in 98-hpf Tol2-CMV-cdx1b-injected embryos compared to Tol2-CMV-GFP-injected or wild type embryos (Fig. 7E–H,J). In addition, secreted mucins could easily be detected in the lumen of the intestine of Tol2-CMV-cdx1b-injected embryos (Fig. 7G). In 96-hpf cdx1b ectopically expressed embryos, an increase (1.2-fold) in the glucagon-expressing enteroendocrine cell numbers was also identified compared to Tol2-CMV-GFP-injected or wild type embryos (Fig. 8). Overall, these results indicate that overexpression of cdx1b can induce the formation of extra goblet and glucagon-expressing enteroendocrine cells in the intestine of injected embryos.

Figure 7.

Overexpression of cdx1b increased goblet cell numbers in the intestine. AD: Wild type (A), lacZ mRNA-injected (B), and cdx1b mRNA-injected (C, D) 98-hr post-fertilization (hpf) deyolked embryos hybridized with agr2 antisense RNA probes. EH: Wild type (E), CMV-GFP-injected (F), and CMV-cdx1b-injected (G, H) 98-hpf deyolked embryos stained with alcian blue. The arrowhead in G indicates secreted mucin in the lumen. I: Comparison of agr2-expressing goblet cell numbers in the intestine of lacZ (n = 20) and cdx1b (n = 20) ectopically expressed embryos. J: Comparison of alcian blue–stained goblet cell numbers in the intestine of GFP (n = 20) and cdx1b (n = 20) ectopically expressed embryos. Error bars indicate standard errors. *P < 0.001. Gc, goblet cell. Scale bars =100 μm.

Figure 8.

Effects of ectopic cdx1b expression on the development of enteroendocrine cells in the intestine. AE: Wild type (A), CMV-GFP-injected (B, C), and CMV-cdx1b-injected (D, E) 96-hr post-fertilization (hpf) deyolked embryos hybridized with glucagon antisense RNA probes. F: Comparison of glucagon-expressing enteroendocrine cell numbers between CMV-GFP-injected (n = 29) and CMV-cdx1b-injected (n = 27) embryos. Error bars indicate standard errors. *P < 0.001. Ac, pancreatic α cells; ee, enteroendocrine cells. Scale bars = 100 μm.

Increased BrdU-Labeled S Phase Cell Population Detected in cdx1b Morphants

To investigate whether cdx1b may play a role in modulating cell proliferation in the intestine, BrdU labeling and immunocytochemistry using anti-p-Histone H3 antiserum were conducted to analyze respective cell populations in either the S or M phase of the cell cycle. Since a high proliferation rate of intestinal cells was reported to occur between 34 and 74 hr of development in zebrafish, we fixed both 72-hpf cdx1b morphants and cdx1b-4mm MO-injected embryos for comparison (Ng et al.,2005; Wallace et al.,2005). Significant increased percentages (18 and 22%) of BrdU-labeled S phase cells were respectively detected in the intestinal bulb and mid-intestine of 72-hpf cdx1b morphants compared to cdx1b-4mm MO-injected embryos (Fig. 9A–F, M). However, similar percentages of p-Histone H3-stained M phase cells were detected in the entire intestines of both 72-hpf cdx1b morphants and cdx1b-4mm MO-injected embryos (Fig. 9G–L,N). A TUNEL assay was conducted to analyze the occurrence of apoptosis in the intestine, and results indicated that very low numbers of apoptotic cells could be detected in the entire intestines of both 72-hpf cdx1b morphants and cdx1b-4mm MO-injected embryos (Fig. 10). Taken together, these results suggest that cdx1b negatively regulates intestinal cell proliferation.

Figure 9.

Cell proliferation affected in cdx1b morphants. AF: BrdU-labeled paraffin transverse sections corresponding to the intestinal bulb (A, D), mid-intestine (B, E), and posterior-intestine (C, F) of respective 72-hr post-fertilization (hpf) cdx1b-4mm-MO-injected (A–C) and cdx1b MO-injected (D–F) embryos. Arrows indicate BrdU-labeled cells. GL: p-Histone H3-stained cryostat transverse sections corresponding to intestinal bulb (G, J), mid-intestine (H, K), and posterior-intestine (I, L) of respective 72-hpf cdx1b-4mm-MO-injected (G–I) and cdx1b MO-injected (J–L) embryos. Arrows indicate p-Histone H3-stained cells. M: Comparison of the percentages of BrdU-labeled cells in the intestinal bulb, mid-intestine, and posterior-intestine of cdx1b-4mm-MO-injected (n=8) and cdx1b MO-injected (n=9) embryos. N: Comparison of the percentages of p-Histone H3-stained cells in the intestinal bulb, mid-intestine, and posterior-intestine of cdx1b-4mm-MO-injected (n = 5) and cdx1b MO-injected (n = 6) embryos. Total cell number counted for intestinal bulb, mid-intestine, and posterior intestine of both 72-hpf cdx1b -4mm-MO-injected and cdx1b MO-injected embryos are shown. Error bars indicate standard errors. *P < 0.001. Ns, not significant. Scale bars = 100 μm.

Figure 10.

TUNEL assay revealed a low level of apoptosis in cdx1b morphants. AI: Paraffin sagittal sections corresponding to the intestinal bulb (A, D, G), mid-intestine (B, E, H), and posterior-intestine (C, F, I) of respective 72-hr post-fertilization (hpf) wild type (A–C), cdx1b-4mm-MO-injected (D–F), and cdx1b MO-injected (G–I) embryos treated with TUNEL reactions using FITC-dUTP (green). Nuclei were stained with DAPI (blue). Embryos in A–C and G–I are oriented with the anterior to the left, while those in D–F are oriented with the anterior to the right. Scale bars = 100 μm.

DISCUSSION

Both mammalian Cdx1 and Cdx2 have been implicated in intestinal development (Guo et al.,2004). Intestinal metaplasia consisting of alkaline phosphatase-producing absorptive cells, and goblet and enteroendocrine cells were detected in transgenic mice expressing Cdx2 in parietal cells using the H+/K+-ATPase promoter (Mutoh et al.,2002,2005). Another line of Foxa3/Cdx2 transgenic mice also revealed the presence of intestinal goblet cells in the gastric mucosa (Silberg et al.,2002). Those studies demonstrated that Cdx2 regulates the differentiation of different intestinal cell lineages. However, Vil-Cdx1 transgenic mice showed no alteration in the intestinal architecture, cell proliferation, or cell type specification, indicating that only Cdx2 plays important roles in controlling intestinal development (Beck,2004; Bonhomme et al.,2008).

Although zebrafish cdx1b was named based on syntenic conservation with mammalian Cdx1, results from this study suggest that zebrafish cdx1b functions like mammalian Cdx2 in the intestinal cell differentiation. Similar to those Cdx2 transgenic mice studies, induction of extra enteroendocrine and goblet cells was detected in both 96- and 98-hpf cdx1b ectopically expressed embryos (Figs. 7, 8). Moreover, we identified significant reductions in the respective numbers of goblet and enteroendocrine cells as well as decreased enterocyte PepT1 expression in the intestine of cdx1b morphants, which were not revealed in studies from either heterozygote Cdx2+/− mutant mice or Cdx2-null mutant chimeric mice (Figs. 3–5) (Chawengsaksophak et al.,1997; Beck et al.,2003). A similar result was revealed in a recent study using a cdx1b splice-blocking MO for knockdown experiments alone (Flores et al., 2008). Ultrastructural analyses by electron microscopy further demonstrated that the cell morphology of enterocytes and goblet cells was affected in cdx1b morphants (Fig. 6). These results support that zebrafish cdx1b like mammalian Cdx2 functions in regulating the differentiation of various intestinal cell lineages.

Because Notch signaling was implicated in intestinal cell fate decisions between secretory and non-secretory lineages and Notch 3 was identified as one of the downstream targets of mouse Cdx2 by a microarray analysis, we also investigated whether Notch signaling is affected in cdx1b morphants (Uesaka et al.,2004; Stanger et al.,2005). Similar expression levels and patterns of delta D, notch 4/6, and her6 in the intestine were detected in both 72-hpf cdx1b morphants and wild type embryos (Supp. Fig. S1, which is available online). Therefore, the defects in the differentiation of various intestinal cell lineages observed in cdx1b morphants were not due to alteration of Notch signaling activity.

Math1, a basic helix-loop-helix transcription factor, was shown to be required for the commitment of all secretory cell lineages, since loss of Math1 in mice caused depletion of goblet, enteroendocrine, and Paneth cells without affecting enterocytes (Yang et al.,2001). Mammalian Cdx2 was implicated in activation of Math1 expression in intestinal epithelial cells (Mutoh et al.,2006). We tried to investigate whether expression levels of putative Math1/atonal-1 homologs were affected in cdx1b morphants; however, we failed to localize their expressions in the intestine of respective embryos. At present, we do not know whether a true ortholog of the Math1/atonal-1 gene exists in zebrafish. Therefore, how cdx1b regulates the development of both secretory cells in zebrafish remains to be further studied.

Polyps with stomach heteroplasia were found in midgut tissues of heterozygote Cdx2+/− mutant mice and Cdx2-null mutant chimeric mice (Chawengsaksophak et al.,1997; Beck et al.,2003). In the intestine of cdx1b morphants, we detected no trans-differentiation of the intestinal epithelium to stomach gland–like structures by electron microscopic analyses (Fig. 6). The vertebrate transcription factor, Sox2, is specifically expressed in the epithelia of the esophagus and stomach, and the expression of zebrafish sox2 was detected in the pharynx and esophagus (Muncan et al.,2007). We compared expression patterns of sox2 in 72-hpf cdx1b morphant and wild type embryos and detected no ectopic expression of sox2 in cdx1b morphants (Supp. Fig. S2). These results indicate that stomach heteroplasia did not occur in cdx1b morphants.

In addition to regulating intestinal cell differentiation, mammalian Cdx2 was implicated as a tumor-suppressor gene (Bonhomme et al.,2003; Guo et al.,2004). Conditional expression of Cdx2 in IEC-6 cells resulted in arrest of cell proliferation (Suh and Traber,1996). Mammalian Cdx2 was shown to directly upregulate transcription of P21/WAF1/CIP1, a cyclin-dependent kinase inhibitor that blocks cell cycle G1-S transition, and the expression of MOK, a mitogen-activated kinase involved in growth arrest and differentiation in intestinal epithelium (Bai et al.,2003; Uesaka and Kageyama,2004). Compared to cdx1b-4mm MO-injected embryos, an increased percentage of BrdU-labeled S phase cells and a very low number of apoptotic cells were detected in the intestine of cdx1b morphants (Figs. 9, 10). This may have resulted from reduced P21 expression when cdx1b protein synthesis was knocked down, thus promoting the entrance of cells into the S phase. Overall, similar to mammalian Cdx2, cdx1b negatively regulates cell cycle progression at the G1- to -S transition. This result is consistent with a recent study that demonstrated the inhibition of intestinal cell proliferation by cdx1b (Flores et al., 2008). However, aberrant increase in proliferative cells were detected in the mid- and posterior intestines of 6 dpf cdx1b splicing MO-injectred embryos in Flores et al. (2008) while we detected significant increase of cell proliferation mainly in the intestinal bulb and mid-intestines of 72 hpf cdx1b morphants (Fig. 9). Such discrepancy may be due to the observation that embryonic intestinal regions differ in their proliferative capabilities as development proceeds (Ng et al., 2005).

In conclusion, our loss- and gain-of-function studies demonstrated that zebrafish cdx1b functions like mammalian Cdx2 in regulating intestinal cell proliferation and the differentiation of various intestinal cell lineages. The absence of stomach heteroplasia in the intestine of cdx1b morphants may have been due to zebrafish not forming a stomach; thus, there is no default pathway allowing the trans-differentiation of intestinal epithelial cells into stomach epithelium with a cdx1b deficiency.

EXPERIMENTAL PROCEDURES

Zebrafish Maintenance and Collection of Embryos

Adult zebrafish (Danio rerio) were raised at the zebrafish facility of the Institute of Cellular and Organismic Biology, Academia Sinica, Taipei, Taiwan. The fish were maintained in 20-L aquariums supplied with filtered fresh water and aeration under a photoperiod of 14 hr of light and 10 hr of dark as described in The Zebrafish Book (Westerfield,1995). Different developmental stages were determined based on previously described morphological criteria (Kimmel et al.,1995).

Tol2-CMV-cdx1b Construction, Morpholino Oligonucleotide, cdx1b mRNA and Tol2-CMV-cdx1b Plasmid Injection

Tol2-CMV-cdx1b was constructed by polymerase chain reaction (PCR) amplification for the cytomegalovirus (CMV) enhancer/promoter-cdx1b coding region and directional cloning into Xho I and Cla I-digested pT2KXIGΔin plasmid (Kawakami et al.,2004; Kotani et al.,2006; Urasaki et al.,2006). Two morpholino oligonucleotides (MO), the cdx1b MO comprising sequences complementary to the AUG translational start site and the 21 bases in the 5′ UTR region, and the cdx1b-4mm MO containing the same nucleotide sequences as cdx1b MO except for four mismatched sequences, were used (Cheng et al.,2008). They were dissolved in Danieau solution (58 mM NaCl, 0.7 mM KCl, 0.4 mM MgSO4, 0.6 mM Ca(NO3)2, and 5 mM Hepes; pH 7.6) at a 1-mM stock concentration. The stock solution was diluted to a working concentration of 0.4 mM with phenol red solution, and 2.3 nl was microinjected (to a final concentration of 7.5 ng or 0.92 pmol) into the cytoplasm of 1- or 2-cell-stage embryos. Capped cdx1b and lacZ mRNAs were synthesized using a T7 or SP6 mMESSAGE mMACHINE Kit (Ambion). Microinjection was performed using a Nanoject II automatic injector (Drummond). To ectopically express cdx1b, cdx1b mRNA (40 pg) was injected into the cytoplasm of 1–2-cell zygotes. LacZ mRNA (40 pg) was injected into the cytoplasm of 1–2-cell zygotes as the control. To inject Tol2-CMV-cdx1b, fertilized eggs were injected with 2.3 nl of a mixture containing 25 ng/μl of circular plasmid DNA of Tol2-CMV-cdx1b and 25 ng/μl of the transposase mRNA. Tol2-CMV-GFP was co-injected with transposase mRNA under the same conditions as Tol2-CMV-cdx1b was injected as the control.

Whole-Mount In Situ Hybridization

Whole-mount in situ hybridization was performed on embryos treated with 0.003% phenylthiocarbamide using digoxigenin-labeled antisense RNA probes and alkaline phosphatase-conjugated anti-digoxigenin antibodies as described previously (Peng et al.,2002). Various templates were linearized, and antisense RNA probes were generated as follows: agr2 (Bam HI/T7), cdx1b (Hind III/T3), glucagon (Nco I/SP6), nkx2.2a (Bam HI/T7), PepT1 (Nco I/SP6), and sox2 (Not I/T7).

Histologic Methods and Photography

Paraffin sectioning and hematoxylin and eosin (H&E) staining were conducted based on standard procedures. Cryostat sectioning of whole-mount in situ embryos was performed as described in Westerfield (1995). Images of embryos from whole-mount in situ hybridization, alcian blue staining, p-Histone H3 and BrdU labeling, TUNEL assay, and paraffin sectioning were taken using an RT color digital camera (SPOT) on a Zeiss Axioplan 2 microscope equipped with DIC or FITC mode. The obtained images were used to count goblet and enteroendocrine cell numbers as well as p-Histone H3- and BrdU-labeled cell percentages in the intestine using the Image-Pro Plus program (Media Cybernetics). Respective percentages of BrdU-labeled and p-Histone H3-stained cells were calculated by counting either BrdU-labeled or p-Histone H3-stained cell numbers and the respective total intestinal epithelial cell numbers in 10 sections from the intestinal bulb, mid-intestine, and posterior intestine of both 72-hpf cdx1b morphants and cdx1b-4mm MO-injected embryos.

Alcian Blue Staining

For whole-mount alcian blue staining, embryos were fixed in 4% paraformaldehyde overnight. After rinsing with PBS several times, embryos were rinsed with acid alcohol (70% ethanol and 0.37% HCl). Embryos were then stained with 0.1% alcian blue 8GX in acid alcohol for 3.5 hr at room temperature and destained in acid alcohol overnight. After rehydration in PBS, embryos were digested with 1% trypsin for 15–60 min at room temperature. Embryos were washed with PBS several times and stored in 70% glycerol. For alcian blue staining of paraffin sections, paraffin sections were first dewaxed and rehydrated. After rinsing with acid alcohol, sections were incubated with 0.1% alcian blue 8GX in acid alcohol for 30 min and then washed with distilled water. Sections were then stained with hematoxylin according to standard procedures.

Immunocytochemistry and TUNEL Assay

Antigen retrieval was performed on cryostat sections in 10 mM citrate buffer (pH 6.0) and 0.05% Tween 20 at 98–120°C for 10 min using an RHS-1 Microwave Vacuum Histoprocessor (Milestone) and then cooled in PBST for 10 min. After sections (8–10 μm) were incubated with blocking solution (0.1% Tween 20, 1% bovine serum albumin [BSA], and 2% goat serum in PBST) for 30 min at room temperature, anti-p-Histone H3 antiserum (1:200; Santa Cruz) was added and incubated overnight at 4°C. Sections were washed with PBST and incubated with a peroxidase-conjugated goat anti-rabbit IgG antibody (1:200; Jackson) for 2 hr at room temperature. The antibodies were visualized with 3,3'-diaminobenzidine tetrahydrochloride (DAB, Vector) and counterstained with hematoxylin. For BrdU labeling, 71 hr post-fertilization (hpf) embryos were incubated in 10 mM BrdU for 10 min at room temperature. After rinsing with egg water several times, embryos were incubated at 28°C for 1 hr before fixation with 4% paraformaldehyde overnight at 4°C. Following dehydration, paraffin embedding and sectioning were performed according to standard protocols. After sections (5 μm) were incubated with blocking solution overnight at 4°C, an anti-BrdU antibody (0.0025 mg/ml; Becton Dickinson) was added and incubated for 2 hr at room temperature. Sections were washed with PBST and incubated with a peroxidase-conjugated goat anti-mouse IgG antibody (1:200; Jackson) for 2 hr at room temperature. The antibodies were visualized with DAB and counterstained with hematoxylin. For the TUNEL assay, paraffin sections were dewaxed and rehydrated. They were treated with 10 μg/ml proteinase K for 15 min at room temperature before labeling of DNA breaks by terminal deoxynucleotidyl transferase with fluorescein-dUTP according to the protocol provided by the manufacturer (Roche Applied Bioscience). Sections were mounted in medium containing DAPI for counterstaining (Vector).

Electron Microscopy

Embryos were fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 2 hr at 4°C. After washing with 0.1 M cacodylate buffer, embryos were postfixed with 1% osmium tetroxide in 0.1 M cacodylate buffer, dehydrated, and embedded in Spurr's resin. Sections (70–80 nm) were examined using an electron microscope (Hitachi), and digital images were collected using an AMT XR40 CCD camera (Advanced Microscopy Technique).

Real-Time Quantitative (Q) Reverse-Transcription (RT)-PCR

DNase-I-treated total RNA (1 μg) was used to generate first-strand cDNA using random hexamers and ImpromII reverse transcriptase (Promega) at 25°C for 10 min and 48°C for 1 hr. Q-PCRs were initiated by adding cDNAs into 1× SYBR green PCR master mix (Applied Biosystems) containing 300 nM of the respective forward and reverse primers for PepT1 or β-actin. The reactions were then conducted in a GeneAmp 7000 sequence detection system (Applied Biosystems). PCR conditions were set to 95°C for 10 min for 1 cycle; and 95°C for 15 s and 60°C for 1 min for 40 cycles. The primer pair for PepT1 was F-AACACAAACATCAAGCAAACC and R-AACTACCAACCCTCAAGCCC. The primer pair for β-actin was F-CCATTGGCAATGAGAGGTTCAG and R-TGA- TGGAGTTGAAAGTGGTCTCG.

The cycle threshold (CT) at which the fluorescence intensity exceeded a preset threshold was used to calculate the respective expression levels of PepT1 and β-actin. To present the expression level, the comparative Ct method was used. ΔCt = avg. CtPepT1 - avg. Ctβ-actin; Δ ΔCt = ΔCtcdx1bmo - ΔCtwt; stdevΔ ΔCt = √(stdev of β-actin)2 + (stdev of PepT1)2; Fold-change = 2 (-Δ ΔCt); S.D.fold change = (ln2) (stdevΔΔCt)(2 (-Δ ΔCt)).

Acknowledgements

We thank Dr. K. Kawakami for providing pT2KXIGΔin and pCS-TP plasmids. The authors also thank Dr. Cheng-Chen Huang for his thought-provoking discussion, Ms. Chun-Shiu Wu for her technical assistance, and Ms. Wun-Hsiang Yu for her maintenance of the fish room. The authors thank the Core Facility of the Institute of Cellular and Organismic Biology, Academia Sinica, for assistance with transmission electron microscopy analyses and DNA sequencing.

Ancillary