Animals

Sexually mature adult X. laevis were purchased from an Aquatic Animal Supply (Misato, Japan). Tadpoles were obtained by the induction of breeding. For breeding, chorionic gonadotropin (hCG-Mochida, Mochida Pharmaceutical, Tokyo, Japan) was injected subcutaneously into adult X. laevis. Tadpoles were raised in a conventional culture room under light- (12-h/12-h light/dark cycle) and temperature- (22°C) controlled conditions. Green juice, including kale and green barley, was fed to the tadpoles at the pre-metamorphic stages, and fish food pellets were added after metamorphosis. All animal experiments in this study were conducted in accordance with the Regulations for Animal Experimentation of Waseda University, Tokyo, Japan. Larvae were staged based on the definition of Nieuwkoop and Faber (Nieuwkoop and Faber, 1967).

Histology

Tadpoles were anesthetized with 1 mg/L of ethyl 3-aminobenzoate methanesulfonate (MS-222; Sigma-Aldrich, St Louis, MO, USA) in ice-cold water. The livers were immediately collected from anesthetized tadpoles. Tissue samples were fixed for 24 hours at 4°C in Bouin's solution (Sigma-Aldrich, St Louis, MO, USA), dehydrated through a graded series of ethanol, embedded in Paraplast (Kendall-Tyco Healthcare, Mansfield, MA, USA) and cut into 4-µm thick sections. The sections were stained with hematoxylin-eosin solution (Wako Pure Chemical Industries, Osaka, Japan) and mounted with Bioleit (Okenshoji, Tokyo, Japan).

Cytology

Dulbecco's modified phosphate-buffered saline without Mg²⁺ and Ca²⁺ ions (dDPBS (-)) was used to avoid blood coagulation. Cytocentrifuged preparations were made with 2 × 10⁵ cells diluted in 0.8 × dDPBS (-), containing 50% fetal calf serum (FCS) (AusGeneX, Brisbane, Australia). The cell suspension was centrifuged at 1,100 × g for four minutes onto glass slides. After air drying, preparations were fixed for 5 minutes in methanol, incubated in 0.1% o-dianisidine (Wako Pure Chemical Industries, Osaka, Japan) in methanol, incubated 30% hydrogen peroxide for 1.5 minutes, washed in running water, and counterstained with Giemsa solution (Okui et al., 2013).

Cell culture

The isolation of hematopoietic progenitor cells from the liver was performed as previously described (Nogawa-Kosaka et al., 2011). Briefly, livers were dissected and placed in 0.8 × dDPBS (-) or 0.8 × alpha minimum essential medium (α-MEM) (Invitrogen, Carlsbad, CA, USA). Organs were washed, minced, and drawn in and out of a syringe through a 27 gauge needle, and treated with a hemolysis solution (150 mM NH₄Cl, 14 mM NaHCO₃, 0.12 mM EDTA-2Na) to remove most erythrocytes. To prepare a single-cell suspension, the cell suspension was further filtered through a 40-µm nylon cell strainer (BD Biosciences, San Jose, CA, USA). Cells (1.5 × 10⁴) were cultured in 0.8 × α-MEM including 10% heat-inactivated FCS, streptomycin (Invitrogen, Carlsbad, CA, USA), penicillin (Invitrogen, Carlsbad, CA, USA), and 1% adult X. laevis serum in 96-well plates. Epo produced by Escherichia coli was added to culture medium. The assay for cell proliferation was conducted in the presence of 100 ng/ml Epo in accordance with the protocol of a previous report (Nogawa-Kosaka et al., 2011). Cells on each day were stained with Trypan Blue solution and counted with a hemocytometer.

Table 1.

Sequences of primers for RT-PCR.

Reverse transcription polymerase chain reaction (RT-PCR)

Total RNA was extracted from cells on each culturing day and whole liver using TriPure Isolation Reagent (Roche Applied Science, Indianapolis, IN, USA) according to manufacturer's instructions. Total RNA was reverse-transcribed into cDNA using ReverTra Ace (Toyobo, Osaka, Japan). Fragments of ribosomal protein l13a (rpl13a) and glyceraldehyde-3-phosphate dehydrogenase (gapdh) were amplified as an internal standard by reference to a previous study (Nogawa-Kosaka et al., 2010). The primer sequences were listed in Table 1. For genes and proteins of X. laevis, we followed the notation method declared in Xenbase ( http://www.xenbase.org/gene/static/geneNomenclature.jsp).

Statistical analysis

All experiments were performed in triplicate. Error bars in figures show standard errors. Student's t-test and Dunnett's test were used when appropriate. Results which p-values were under 0.05 were considered to be significant.

Liver structure at the metamorphic stage

At NF56, the length of the body reached approximately 30 mm under our breeding conditions. Forelimb and hind limb were formed clearly, and elimination of the caudal portion started at NF60 (middle metamorphic stage). At the climax of metamorphosis (NF66), caudal elimination was complete, and the body morphology transformed to that of adult frogs (Fig. 1A). In the metamorphic stages, from NF56 to NF66, changes occurred in both the external morphology and the internal structure. The lung could be observed at NF56, and tissue arrangement of the intestine, which was in a spiral configuration at NF56, changed to a more complicated structure at NF66. Sternum formation began at NF60. From NF56 to NF60, larval liver increased in size and turned red (Fig. 1B).

The structure of the liver changed greatly during metamorphosis. The number of lobes varied from three at NF56 to two at NF66. HE-stained liver sections at NF49, 56, 60, and 66 clearly demonstrated the morphological changes in hepatocytes and the increase in sinusoid diameter (Fig. 1C). At NF49, the structure of the hepatic cell cord was unclear, and the diameter of the sinusoid was wider than that at NF56, 60, and 66. It is known that the pigment cells, melanomacrophages, appear in adult liver (Wiechmann et al., 2003). Melanomacrophages were not observed in the liver at NF49 and NF56, and they began to appear at NF60 (Fig. 1C). The staining property also changed during the development, and the area of sinusoid decreased from NF56 (Fig. 1D). The size of hepatocytes increased significantly at NF66 (Fig. 1E). We then examined erythroid progenitors in the liver during the metamorphic period and evaluated their response to Epo.

Fig. 1.

(A) Comparison of liver structure at different metamorphic stages; NF49, 56, 60, and 66 are shown. Scale bars represent 10 mm. (B) Internal organs at stages NF49, 56, 60, and 66 from the ventral direction are shown. The directions of bodies are same as that seen in (A). Arrows indicate the liver. (C) Paraffin-embedded liver sections (4-µm thick-ness) at NF49, 56, 60, and 66, stained with HE solution. Yellow arrows show melanomac-rophages. Scale bars represent 50 µm. Si; sinusoid. (D) Ratios of sinusoidal area in one visual field in NF49, 56, 60, 66 livers. Values represent means (n = 3), and error bars rep-resent standard errors. (E) The area of a hepatocyte of NF49, 56, 60, 66 are shown. Val-ues represent means (n = 3), and error bars represent standard errors. ***; P < 0.001, Dunnett's test (vs. NF49).

Fig. 2.

Difference in cell proliferation after hemolysis processing. (A) Whole liver cells of NF56 were cultured in the presence of Epo (100 ng/ml). Values represent means (n = 3), and error bars represent standard errors. Circles show 100 ng/ml Epo. The dashed line indicates the control group. *; P < 0.05, **; P < 0.01, Student's t-test (vs. control). (B) Liquid culture of liver cells at NF56 with Epo. Cells were treated with hemolysis solution. The sequential cell counts of viable cells are shown. Values represent means (n = 3), and error bars represent standard errors. Solid lines indicate the number of cells, which are cultured with Epo (100 ng/ml), and the dashed line indicates the control. *; P < 0.05, **; P < 0.01, ***; P < 0.001, Student's t-test (vs. control). N.D.; not detected.

Fig. 3.

Morphology of the liver cells cultured with Epo. (A) Micrographs of liquid cul-tured cells at day 0 (left column) and day 8 control and Epo (middle and left column, respectively) are shown. Scale bars represent 50 µm. (B, C) Liquid-cultured cells of NF56 (B) and NF66 (C) livers were treated with o-dianisidine and Giemsa solution on day 0, day 5, and day 8. Red arrows indicate blast cells, and blue arrows indicate the o-dianisidine+ cells. Scale bars represent 50 µm. (D) Morphological typing of cultured cells. Cells were treated with o-dianisidine solution, and stained with Giemsa solution for counter staining. The o-dianisidine+ cells, blast cells, and other cells were defined.

Reactivity of pre- and post-metamorphic liver cells to Epo

When whole liver cells of NF56 larvae were cultured in the presence of Epo, the survival rates of the cells increased from that of the control, but the cell counts did not increase (Fig. 2A). Since it was possible that mature erythrocytes express Epor to avoid apoptosis due to Epo-Epor signaling (Aizawa et al., 2005; Tamura et al., 2015), we removed mature erythrocytes by hemolysis prior to liquid culture. This also suppressed the absorption of unbound Epo to Epor in the culture. After removal of mature erythrocytes by hemolysis, the number of proliferating cells increased (Fig. 2B).

Previous studies have shown that erythropoiesis takes place in the liver of tadpoles; therefore, we examined whether erythroid progenitors that responded to Epo resided in the larval liver. When larval hepatic cells at NF56 or NF66 were cultured in the presence of Epo, the cells proliferated, and several types of erythrocytic cells emerged in the respective cultures at day 8 (Fig. 3A). To characterize the types of erythrocytic cells induced by stimulation by Epo, we stained the cultured cells on each day with o-dianisidine to classify them based on hemoglobinization. A number of blast cells and pale brown cells that appeared on day 5 were not validated as erythrocytic cells. Subsequently, mature erythrocytes emerged abundantly at day 8 from liver cells at both NF56 and NF66 (Fig. 3B and 3C). Next, viable blast cells and mature erythrocytes from liver cells at NF56 or NF66 were counted on each culture day. We classified cultured cells into three cell subpopulations based on the size of nuclei and the results of o-dianisidine and Giemsa staining: (1) mature erythrocytes robustly stained with o-dianisidine (brown cytoplasm), (2) blast cells with large nuclei (blue or very pale brown cytoplasm stained with Giemsa and o-dianisidine), and (3) other cells with lobulated nuclei (negative to o-dianisidine-staining) (Fig. 3D).

The viable cells increased in the cultures of the liver cells of NF56 and NF66 (Fig. 4A, B). In the culture of the cells derived from the NF56 liver, half maximal of the number of viable cells was attained at 5 days and the maximal was attained at day 7 after the start of culture (Fig. 4A). At NF66, the cell number did not reach the maximum plateau until day 8 (Fig. 4B). The difference in the pattern of cell proliferation between NF56 and NF66 was that the NF56 cells increased in a sigmoid pattern, while NF66 cells expanded in a linear pattern. Dead cell counts for NF56 and NF66 were similar to those for the control (data not shown).

Fig. 4.

Liquid culture of liver cells at NF56 and NF66 with Epo. (A, B) Sequential cell counts of viable cells from NF56 (A) and NF66 (B) are shown. Values represent means (n = 3, * P < 0.05, ** P < 0.01), and error bars represent standard errors. Solid lines indicate the number of cells which are cultured with Epo, and dashed lines are control. (C, D) The sequential cell counts of blast cells and o-dianisidine+ cells from NF56 (C) and NF66 (D) liver cells are shown. Values represent means (n = 3, *, # P < 0.05, **, ## P < 0.01 vs. con-trol, Student's t-test), and error bars represent standard errors. Dash lines indicate the number of blast cells, and solid lines indicate the number of o-dianisidine+ cells. (E, F) The sequential change of cell components in the culture. Blast cells (shaded area), o-dianisidine+ cells (dotted area), and other cells (blank area) were counted and represented in stacked bar graphs. Values represent means (n = 3), and error bars represent standard errors.

Blast cell numbers at both NF56 and NF66 started to increase at day 3, peaked at day 5 (NF56) and day 4 (NF66), and then decreased at day 7 (Fig. 4C, D). The number of o-dianisidine+ cells decreased by day 5 for NF56, while a slight rise was observed at day 2 for NF66 (Fig. 4C, D). However, a rapid increase was observed at day 7 (NF56) and day 6 (NF66). Additionally, at day 8, the number of o-dianisidine+ cells continued to increase for NF66, while it decreased for NF56. In the light of cell proportion in the culture of both NF56- and NF66-derived cells, blast cells increased at first, and o-dianisidine+ cells increased thereafter, then blast cells decreased (Fig. 4E, F).

Gene expression in pre- and post-metamorphic tissues

In adult X. laevis, the liver is the predominant hematopoietic organ (Hadji-Azimi et al., 1987; Nogawa-Kosaka et al., 2011). The whole liver at NF56 and NF66 expressed epo, epor, gata2, and gata1(Fig. 5A). To characterize the cells responding to Epo, the expression of various genes at each day was analyzed by RT-PCR (Fig. 5B, C). We evaluated the mRNA expression of Epor and Gata, which participate in erythropoiesis in mammals and fish. We first examined the Gata-binding regions on the X. laevisglobin gene (data not shown). In the NF56 and NF66 liver cells cultured in the presence of Epo, the expression of both larval and adult globin were up-regulated at days 6–8, consistent with the growth pattern of o-dianisidine+ cells. Interestingly, the expression of larval or adult globin gene increased predominantly in NF56- and NF66-derived cells, respectively. The wavy change of globin expression was detected in the culture (Fig. 5B, C). Both larval and adult globin were expressed initially in the cultures NF56 and NF66 liver cells, and their expression levels gradually declined to the nadir at around day 4, and then the expression levels increased again at day 6. The expression of epor reached a peak in both liver cultures (NF56 and NF66) as erythrocytic blast cells expressing eporemerged. Similar to globin expression, the expression of epor was detected at day 0, decreased to the nadir at days 1–2, and then increased to reach a peak at around day 3. The expression of gata2 was detected at days 3–4 (NF56 and NF66). Theexpression of gata1 peaked at day 6 in the NF56 liver cells, while that in NF66 liver cells peaked at day 5. The peak of gata1 expression was reached about 2 days after that of gata2. Albumin expression was observed in the early culture days, and it showed a sequential decline, indicating that exogenous Epo could not maintain hepatocyte culture under this condition. The changes in gene expression over the course of erythroid development are shown in Fig. 6.

Fig. 5.

Gene expressions of epo, epor, gata2, and gata1 in whole liver of NF56 and NF66 were shown (A). Gapdh was used as an internal control. Gene expressions of liquid cultured liver cells at NF56 (B) and NF66 (C). Sequential gene expressions of liq-uid-cultured cells at NF56 and NF66 are shown for (B) and (C), respectively. Rpl13A was used as an internal control.

Fig. 6.

Conceptual diagram of erythropoiesis with morphology and gene expression in X. laevis. Gata2 expressed highly in days 3–4. After a day from the peak of gata2 expression, gata1, and epor expressed highly, and differentiated into erythroblasts like mammals. Furthermore, erythroblasts accumulated larval or adult hemoglobin over about two days and differentiated into mature erythrocytes.