NT-4 protein is localized in neuronal cells in the brain stem as well as the dorsal root ganglion of embryonic and adult rats


Address correspondence and reprint requests to Ritsuko Katoh-Semba, Department of Perinatology, Institute for Developmental Research, Aichi Human Service Center, Kasugai, Aichi 480–03, Japan. E-mail: katohsemba@inst-hsc.pref.aichi.jp


We have newly established a sensitive, two-site enzyme immunoassay system for neurotrophin-4 (NT-4) and investigated its tissue distribution in the rat nervous system. The minimal limit of detection of the assay is 0.3 pg/0.2 mL of assay mixture. Concentrations of NT-4 were found to be extremely low in all brain regions, irrespective of the animal age, the highest level being found in the brain stem of 40-day-old rats, at 0.12 ng/g wet weight. NT-4 levels in young adult rats were significantly lower in the thalamus and higher in the olfactory bulb, neocortex, hypothalamus and brain stem than respective levels in 1-week-old rats. NT-4 immunoreactivity was strong in large neurons of the red nucleus and pontine reticular nucleus as well as the locus coeruleus, and moderate in cells in the mesencephalic trigeminal nucleus and interstitial nucleus of the medial longitudinal fasciculus. In the rat embryo, stong staining of NT-4 was detected in cells of regions corresponding to the midbrain/pons from E11.5 through E15.5. The intensity was decreased after E13.5 when the cytoplasm of cells in the medulla oblongata, fibers of the cerebellar primordium, and both cells and fibers of the dorsal root ganglion were also stained. Concentrations of NT-4 were detected in regions including the hindbrain and the dorsal root ganglion. Immunoblotting of NT-4-immunoreactive proteins extracted from these two regions revealed a band corresponding to mature NT-4 with a molecular mass of ∼14 kDa. Kainic acid and another glutamte agonist, (+/–)-α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid did not affect NT-4 levels in the hippocampus. The present results show NT-4 to be localized in very limited brain cells and fibers from the embyonic period through to the young adult, suggesting specific roles in brain functions.

Abbreviations used

(+/–)-α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid hydrobromide


brain-derived neurotrophic factor


bovine serum albumin


3,3′-diaminobenzidine tetrahydrochloride




dorsal root ganglion




kainic acid


lateral geniculate nucleus


nerve growth factor






phosphate-buffered saline


sodium dodecyl sulfate


tyrosine kinase

Neurotrophin-4 (NT-4) is the fourth member of the nerve growth factor family, its gene encoding a precursor protein of 236 amino acids which is processed into a 123 amino acid mature NT-4 form (Hallböök et al. 1991; Hallböök 1999). NT-4 is structurally and functionally related to other neurotrophins, nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), and neurotrophin-3 (NT-3), which share approximately 50% homology in amino acid sequences (Ip et al. 1992) and are known to have individual functions in limited regions and at specific ages. It has been shown that NT-4 selectively promotes survival of cultured striatal neurons and midbrain dopaminergic neurons, protects midbrain dopaminergic neurons from toxic damage, and rescues lateral geniculate neurons from dystrophic effects of monocular visual deprivation (Ardelt et al. 1994; Hynes et al. 1994; Riddle et al. 1995; Alexi and Hefti 1996; Lingor et al. 2000). It has been also demonstrated that NT-4 elicits growth of neurites from the dorsal and nodose ganglia, but not from the sympathetic ganglia, similar to BDNF. Both neurotrophins, in fact, act on cells through the tyrosine kinase B (TrkB) receptor (Berkemeier et al. 1991). However, they also have differences: for example, NT-4, but not BDNF, induces neuritic growth from PC12 cells through TrkA (Berkemeier et al. 1991). NT-4 mRNA is widely expressed during embryonic development, while BDNF mRNA is increased during post-natal development of the brain (Timmusk et al. 1993). BDNF levels are rapidly and potently upregulated by epileptiform activity in the hippocampus and cerebral cortex, while NT-4 mRNA levels are strongly regulated by neuromuscular activity in skeletal muscles (Funakoshi et al. 1995).

The expression of NT-4 mRNA seems to be widely distributed and developmentally regulated in the brain and peripheral tissues during embryonic and post-natal development, like neurotrophin receptor expression (Timmusk et al. 1993; Escandón et al. 1994; Hafidi 1999). In rats, almost all peripheral tissues express NT-4 mRNA and high levels are found in the prostate, testis, and ovary (Ip et al. 1992; Timmusk et al. 1993). In contrast, brain mRNA levels are extremely low, irrespective of the age, although they appear to be slightly higher on embryonic day 13.5 (E13.5) than at other ages. Moreover, levels of NT-4 mRNA in the brain stem demonstrate a small increase during post-natal development. It is reported that lack of the NT-4 gene causes a loss of sensory, but not motor neurons, along with selective structural and chemical deficits in sympathetic ganglia (Conover et al. 1995; Liu et al. 1995; Roosen et al. 2001; Stucky et al. 2002). It has also been shown that infusion of NT-4 into the cat primary visual cortex inhibits formation of the ocular dominance column, and blocks the effect of monocular deprivation on binocular responsiveness of cortical neurons in kittens (Cabelli et al. 1995; Gillespie et al. 2000). Thus, NT-4 is considered to play specific roles in nervous tissues even though its mRNA levels are very low (brain: testis at 1 day of age = 1 : 75). However, the localization of NT-4 protein in brains has remained unclear. Therefore we have prepared NT-4-specific antibodies useful for immunocytochemistry and established a specific enzyme immunoassay system. In the present communication, we report on the distribution of NT-4 protein in nervous tissues of embryonic and young adult rats.

Materials and methods

Animals and surgery

Sprague–Dawley rats at various ages were used for the experiments. Noon of the day on which a vaginal plug was detected was designated as embryonic day 0.5 (E0.5). Some rats underwent operations under anesthesia to insert polyethylene tubing into the left lateral ventricle with fixation to the surface of the skull with dental cement (1.5 mm lateral to the midline; 0.8 mm posterior to bregma; 3.5 mm ventral to the cortical surface). All the experiments with animals were carried out according to the guidelines for animal experimentation at the Institute for Developmental Research.

Preparation and purification of antibodies

Polyclonal antibodies against recombinant human NT-4/5 (Peprotech Inc., Rocky Hill, NJ, USA) and its carboxy-terminal decapeptide (CTLLSRTGRA)–hemocyanin complex conjugated using N-succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate were raised by immunizing rabbits (Katoh-Semba et al. 1997). IgG specific for NT-4/5 (pAb rNT-4) was purified from the antiserum on a column of Sepharose 4B coupled with the antigen (NT-4 from Peprotech Inc.) and some of the purified pAb rNT-4 was treated with pepsin to obtain F(ab′)2 fragments as described previously (Katoh-Semba et al. 1989). pAb rNT-4 and its F(ab′)2 fragments were used for immunohistochemistry and the NT-4 assay, respectively. Specific IgGs against its carboxy-terminus (pAb C10) were purified by the same procedure, except for coupling of recombinant rat NT-4 with the his-tag prepared in our laboratory and were used for the western blotting analysis. Some purified pAb C10 was also degraded to F(ab′)2 fragments and then reduced with 2-mercaptoethylamine to obtain Fab′ fragments, followed by conjugation with β-d-galactosidase (EC from Escherichia coli, as described previously (Katoh-Semba et al. 1989). Labeled Fab′ fragments were subjected to enzyme immunoassay. Polyclonal antibodies against recombinant rat NT-4 with the his-tag were raised by immunizing goats (pAb gNT-4) and were affinity-purified using recombinant human NT-4/5 from Peprotech Inc. The resultant purified pAb gNT-4 was used for the western blotting analysis. pAb rNT-4 and pAb gNT-4 immunoreacted with recombinant human NT-4/5 obtained from three companies. However, pAb C10 only reacted with recombinant human NT-4 kindly supplied by Genentech Inc. (San Francisco, CA, USA), and not with recombinant human NT-4/5 from Peprotech Inc. or Promega Inc (Madison, WI, USA). Affinity-purified antibodies did not react with NGF, BDNF or NT-3 (Fig. 1a).

Figure 1.

Analysis of the specificity of polyclonal antibodies against NT-4 and the sensitivity of the immunoassay system. (a) Western blotting analysis. Five-nanogram aliquots of each neurotrophin were loaded in the respective lanes. Blotted membranes were incubated with pAB rNT-4 or pAb C10 at a concentration of 0.2 µg/mL, overnight. std, pre-stained molecular marker protein; lower band in the lane of NT-4, mature NT-4 with a molecular mass of ∼14 kDa; upper band, NT-4 dimer. (b) β-Galactosidase activity as a function of the concentration of individual neurotrophins. The activity is expressed as the intensity of fluorescence of 4-methylumbelliferone produced during a 20-min incubation at 30°C with 4-methylumbelliferyl-β-galactoside as the substrate. A fluorescence intensity of 1000 is equivalent to 1 µm 4-methylumbelliferone. ●, Recombinant human NT-4 provided from Genentech Inc.; ○, purified submandibular mouse β-NGF; ▵, recombinant human BDNF; x, recombinant human NT-3. (c) Note the linear relationship between the wet weight of the testis (mg) used for the assay and the estimated NT-4 (pg).

Drug treatments

Kainic acid (KA) and (+/–)-α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid hydrobromide (AMPA) were intraventricularly administered at a dose of 0.05 µg in 5 µL of phosphate-buffered saline (PBS)/rat and 2 µg in 2 µL of dimethylsulfoxide (DMSO)/rat, respectively. Control rats received the same volumes of PBS or DMSO.

Extraction of BDNF and NT-4 from tissues

Testes and 10 regions of brains from rats at various post-natal ages were dissected out, weighed immediately and frozen at − 80°C. Regions containing the dorsal root ganglion (DRG) and corresponding to the hindbrain (HB) were also dissected from fetuses on embryonic day 13.5 (E13.5). Hippocampi and entorhinal cortices from KA- or AMPA-treated rats were removed 8 h after injection. Dissected tissues were homogenized with 10 volumes of 100 mm phosphate buffer containing 1 mm EDTA, 2 m guanidine hydrochloride (pH 7.2), and three protease inhibitors (10 mmN-ethylmaleimide, 0.36 mm pepstatin, and 1 mm phenylmethylsulfonyl fluoride). Homogenates were sonicated and centrifuged at 46 000 g for 30 min at 4°C and the resultant supernatants were subjected to measurement of NT-4 amounts.

Enzyme immunoassay

NT-4 was assayed by the same method as for BDNF (Katoh-Semba et al. 1997). Briefly, standard recombinant human NT-4 (Genentech Inc.) or tissue extracts were bound to polystyrene beads bearing immobilized F(ab′)2 fragments of pAb rNT-4 in 0.2 mL of 60 mm phosphate buffer containing 0.3 m NaCl, 1 mm MgCl2, 0.1% bovine serum albumin (BSA), 0.5% protease-treated gelatin, 0.1% NaN3 (buffer B) and 1 m guanidine hydrochloride. After incubation at 30°C, with shaking, for 5 h, each bead was washed two times with 1 mL of cold 10 mm phosphate buffer (pH 7.0) containing 0.1 m NaCl, 1 mm MgCl2, 0.1% BSA, and 0.1% NaN3 (buffer A) and was transferred into a new test tube with 0.1 mL of buffer A containing 2% skimmed milk and 1 mU of β-d-galactosidase-labeled Fab′ fragments prepared from the purified pAb C10, followed by incubation at 4°C for 16 h or more. Then beads were washed two times with cold buffer A and the bound galactosidase activity was assayed with 4-methylumbelliferyl-β-d-galactoside as the substrate. The detection limit of the present assay was 0.3 pg per tube and other neurotrophins did not cross-react (Fig. 1b). The assay for NT-4 was carried out within the range defined by a linear relationship between the amount of extract used and the activity of bound-β-d-galactosidase. The activity increased linearly with increasing wet weight of tissue up to 10 mg for the testis (Fig. 1c). The recoveries of standard recombinant human NT-4 added to homogenates of the testis and hippocampus were 89.7 ± 9.7 (SD) and 80.1 ± 9.4 (n = 5), respectively.

Analysis of proteins immunoreactive with antibodies against NT-4

Proteins that immunoreacted in our assay system were investigated by an immunoblotting method as described previously (Katoh-Semba et al. 2001). In brief, NT-4 protein was extracted from testes on post-natal days 7, 14, and 40, or from the DRG or the HB pooled from 92 embryonic rats at E13.5 as described above in the section on extraction of NT-4 from tissues. The extracts were mixed with the same volume of 2× buffer B. Seven and 15 beads with immobilized pAb gNT-4 were added to the mixtures for the testes and embryonic regions, respectively, and shaken in the mixture at 4°C for 16 h or more. The same volume of buffer B containing 1 m guanidine hydrochloride was shaken with seven or 15 beads at the same time in the presence or absence of recombinant human NT-4. Beads were washed twice with 10 mL of cold buffer A and incubated at room temperature for ≥ 30 min in 10 mL of buffer A containing 2% skimmed milk, followed by washing once with buffer A and three times with 10 mL of 0.1 m phosphate buffer (pH 7.0). Proteins bound to beads were dissociated with 0.4 mL of a 0.4% solution of sodium dodecyl sulfate (SDS) in distilled water and the solution was evaporated to dryness. The resultant powder was dissolved in 40 µL of Laemmli's buffer without additional SDS, heated at 100°C for 3 min and subjected to SDS–polyacrylamide gel electrophoresis (PAGE). Immunoblotting was carried out as described earlier (Katoh-Semba et al. 1997). Blots were blocked with buffer B supplemented with 5% BSA, and then they were incubated in buffer B with pAb C10 (0.2 µg/mL) followed by peroxidase-labeled anti-rabbit IgG in 20 mm Tris–HCl buffer (pH 7.5) containing 0.5 m NaCl (Tris-buffered saline; 1 : 5000) as described previously (Katoh-Semba et al. 1997, 2001). In the embryonic case, after incubation with pAb C10 (0.02 µg/mL), blots were incubated with goat antiserum to rabbit IgG in Tris-buffered saline containing 2% skimmed milk (1: 5000) and then peroxidase-labeled rabbit antiperoxidase in Tris-buffered saline (1 : 5000). The membranes were reacted with the chemiluminescence reagent (Perkin Elmer Life Science Inc., Boston, MA, USA) and used to expose X-ray film.

Immunocytochemical analysis

Paraffin sections for immunocytochemical analysis were prepared as described previously (Katoh-Semba et al. 2001). In brief, 40-day-old rats were anesthetized with ether and perfused via the heart with 30 mm piperazine-1,4-bis (2-ethansulfonate) buffer (pH 7.2) containing 2% paraformaldehyde, 0.5% glutaraldehyde, 10% DMSO, and 8% sucrose after an initial perfusion with 0.9% NaCl. Brains were dissected out and post-fixed in the same solution at room temperature for about 4 h and at 4°C, overnight. Fixed brains were incubated at 4°C with 0.1 m phosphate buffer (pH 7.4) containing 8% sucrose for 24 h after one wash with the same buffer, and then dehydrated through a graded alcohol series (each step, 1–2 h). Rat fetuses were removed at various embryonic days, rinsed twice with Hank's solution, and immersed in Bouin's fixative overnight. Then, fixed fetuses were dehydrated with 95% alcohol overnight without washing and 100% alcohol for 4 h twice. Dehydrated brains and fetuses were then embedded in paraffin after immersing in 100% xylene, a mixture of xylene–paraffin (1 : 1), and then 100% paraffin with appropriate temperatures and time periods. Sagittal sections were cut at a 7-µm thickness for brains and at an 8-µm thickness for fetuses.

For immunohistochemical staining, sections were deparaffinized in 100% xylene, and hydrated through a graded alcohol series. Then they were treated for 20 min at room temperature with 2% H2O2 in Tris-buffered saline. Treated sections were washed once with Tris-buffered saline and then incubated with blocking reagent composed of Tris-buffered saline supplemented with 4% goat serum and 2% BSA for 1 h or more in a humid container at room temperature. Sections after one wash in Tris-buffered saline were treated with the first antibody, affinity-purified pAb rNT-4 (1 and 2 µg/mL in Tris-buffered saline containing 1% BSA for brains and fetuses, respectively) for appropriate time periods at 4°C. Control sections were incubated with non-specific rabbit IgG at the same concentration. After three washes, sections were incubated at room temperature for 2 h with biotinylated anti-rabbit IgG (Vector Laboratories, Burlingame, CA, USA) in Tris-buffered saline containing 1% BSA as the second antibody (1 : 200). Then, avidin–peroxidase-labeled biotin complex (Vector Laboratories) in Tris-buffered saline containing 0.1% BSA (1 : 100) was allowed to bind to the second antibody (1 h). The peroxidase activity was visualized by reaction with diaminobenzidine at a concentration of 0.2 mg/mL of 20 mm Tris–HCl (pH 7.5) containing 0.006% H2O2.

Other methods

Before protein determination, guanidine hydrochloride was removed from the tissue homogenates used for the NT-4 assay. Twenty volumes of 95% alcohol were added to each homogenate and the mixture was centrifuged at 9000 g for 20 min at 4°C. The resultant precipitates were washed three times with alcohol and then dissolved in the original volumes of 0.1 m phosphate buffer, pH 7.0 containing 2% SDS. Amounts of protein were measured using a micro BCA protein assay reagent kit (Pierce, Rockford, IL, USA) with BSA as a standard. All data are expressed as mean ± SD values. Differences between brain regions were examined for statistical significance using anova (followed by a post-hoc test).


Reliability of our assay system

Levels of NT-4 mRNA in the brain have been reported to be lower than those in peripheral tissues (Timmusk et al. 1993). Before determination of NT-4 in the nervous tissues of rats, we measured changes with post-natal age in the concentration of NT-4 in the testis, where its mRNA is highly expressed, to confirm the reliability of our methodology. When NT-4 in the testis was expressed as amounts per g of wet weight, mean values decreased with post-natal age (Fig. 2a), in accordance with data for NT-4 mRNA. The developmental change in average amounts per mg of protein was also similar to the pattern for values per wet weight (Fig. 2b). Western blot analysis showed the testis protein to have a molecular mass of ∼14 kDa, corresponding to the mature form of NT-4 (Fig. 2c). However, its precursor form was not stained, differing from the BDNF case (Katoh-Semba et al. 2001).

Figure 2.

Developmental changes in levels of NT-4 in the testis. (a and b) Concentrations of NT-4 in the testis estimated by the immunoassay system specific for NT-4. Data are mean ± SD (bars) values from five, six, and six animals on post-natal days 7, 14, and 40, respectively, and are expressed based on g wet weight (a) or mg protein (b). (c) Western blots of testis proteins immunoreacted with NT-4-specific antibodies. Extracts were prepared from 150, 500, and 500 mg wet weight testis at 7, 14, and 40 post-natal days of age, respectively, as described in Methods. NT-4-immunoreactive proteins were concentrated using beads immobilized with pAb gNT-4 and dissociated from the beads in SDS solution. Blots were incubated with pAb C10 after blocking and then with peroxidase-labeled anti-rabbit IgG. Lane 1, recombinant human NT-4 (5 ng); lane 2, recombinant human NT-4 (5 ng) recovered from beads; lanes 3–5, testis proteins from rats on days 7 (lane 3), 14 (lane 4), and 40 (lane 5); lane 6, proteins recovered from beads after they were incubated with the homogenizing buffer. Only mature NT-4 with the molecular mass of ∼14 kDa was observed in the testis.

Distribution in the rat brain

Figure 3 summarizes data for the concentrations of NT-4 in the 10 brain regions of 7-day-old and 40-day-old rats. Appreciable levels of NT-4 were found in all regions at both ages (Fig. 3b). However, as shown in Fig. 3(a), mean levels at 40 days of age, when expressed as g of wet weight, were extremely low compared to those for other neurotrophins in the rat brain (Katoh-Semba et al. 1998). The concentration in the hippocampus was about one-hundredth of the mean value for BDNF in the same region (see Table 1) and, even in the brain stem with the highest level, the average concentration was only ∼0.120 ng/g wet weight. Differences between brain regions at this age were significant (F9,40 = 5.597, p < 0.001). Significantly high and low levels of NT-4 were observed in the brain stem and the thalamus, respectively (p < 0.005). In 7-day-old rats, the olfactory bulb showed a low level of NT-4 (data not shown; p < 0.001; refer to Fig. 3b, gray column). When the concentration was expressed as values per mg of protein, there was a significant difference between 7- and 40-day-old rats (F1,80 = 27.71, p < 0.001). Levels in the eight brain regions of young adult rats, except for the cerebellum and thalamus, demonstrated a tendency for increase with post-natal age. Significantly high levels of NT-4 were found in the olfactory bulb, neocortex, hypothalamus and brain stem (p < 0.01 for the olfactory bulb, neocortex, and brain stem; p < 0.05 for the hypothalamus). In contrast, in the cerebellum and thalamus, the average concentrations tend to decrease with post-natal age (p < 0.05 for the thalamus).

Figure 3.

Concentrations of NT-4 in various regions of the rat brain. Data are mean ± SD (bars) values from five rats based on g wet weight (a) and mg of protein (b). (a) Values obtained from 40-day-old rats; #p < 0.005. (b) Values obtained from 7-day-old (gray columns) and 40-day-old rats (closed columns); *p < 0.05 and **p < 0.01, respectively.

Table 1.  Effects of an intraventricular injection of KA or AMPA on levels of BDNF and NT-4 in the hippocampus and entorhinal cortex of 40-day-old rats
 nNeurotrophins (ng/g wet weight)
  1. AMPA (2 µg/rat) or kainic acid (0.05 µg/rat) was intraventricularly injected and after 8 h, the hippocampus and entorhinal cortex were dissected. BDNF and NT-4 were, respectively, determined using the right and left regions from the same animals. Values are means ± SD of results from five to seven animals. *p < 0.01 as compared to the controls.

 Controls5 7.11 ± 0.320.074 ± 0.008
 KA750.02 ± 14.84*0.076 ± 0.009
 AMPA619.17 ± 1.56*0.086 ± 0.009
Entorhinal cortex
 Controls5 2.51 ± 0.610.078 ± 0.011
 KA718.95 ± 19.42*0.072 ± 0.013
 AMPA6 2.06 ± 0.410.094 ± 0.005

Sagittal sections of brains from 40-day-old rats were stained with pAb rNT-4. Figure 4(a) is an illustration of an adult rat brain cut laterally at 0.9 mm. NT-4 immunoreactivity was found in several nuclei such as the pontine reticular nucleus, the red nucleus (R), and interstitial nucleus of the medial longitudinal fasciculus greater part (IMLFG). Figure 4(b) shows a stained section corresponding to the part enclosed by the square in Fig. 4(a). NT-4-immunoreactive cells were found in the red nucleus (R) and the pontine reticular nucleus oral part (PnO) of the sagittal section. Relatively large cells, probably neurons, in the two nuclei were strongly stained with the antibodies (Figs 4d and e for the pontine reticular nucleus, oral part and the red nucleus, respectively), while control IgG did not stain any structures (Fig. 4c). NT-4 immunoreactivity was found to be homogeneous in the cytoplasm and processes of large neurons in both nuclei. In addition, immunoreactivity was strong in large neurons of the locus coeruleus (Fig. 4f), moderate in cells of the mesencephalic and motor trigeminal nuclei, and weak in cells of the facial and vestibular nuclei (data not shown).

Figure 4.

Immunohistochemical staining of NT-4 in adult rat brain. (a) A drawing of a sagittal section through the adult rat brain. R, red nucleus; PnO, pontine reticular nucleus oral part; IMLFG, interstitial nucleus of the medial longitudinal fasciculus greater part. (b–f) Sagittal sections prepared from the brain of rats on post-natal day 40 were stained with pAb rNT-4 (b, d–f) or non-specific rabbit IgG (c). (b) The region enclosed by the square in (a). Arrow, R region; arrowhead, PnO region. (c and d) The PnO region; (e) the R region; (f) the locus coeruleus. Large neurons and fibers are stained. Bars, 0.5 mm in (b); 50 µm in (c–f).

Distribution in the rat embryo

Because NT-4 elicits growth of neurites from DRG neurons (Hallböök et al. 1991), sagittal sections of rat embryos at E13.5 were stained with antibodies. Figure 5(c) illustrates a DRG indicated by an arrow in Fig. 5(a) with pAb rNT-4 staining of some structures. In contrast, control IgG stained nothing (Fig. 5b). A high magnification of the part enclosed by the square in Fig. 5(c) indicates strong immunoreactivity in the cytoplasm and fibers of neurons (Fig. 5d; double long arrows). Similar staining was observed in other sensory ganglia, such as the trigeminal ganglia (data not shown). The intensity of staining in the DRG appeared stronger at E14.5 than at E13.5. NT-4 immunoreactivity in the brain was first found at E11.5 in the region of the pons/midbrain junction (Figs 6a and b). Strong immunoreactivity of NT-4 was observed in cells of the midbrain and pons (Fig. 6b). These two regions were also stained in embryos at E12.5–E15.5 (Figs 6c–h). However, the intensity was weak at E14.5 (Fig. 6h). At E13.5 (Fig. 6f), in addition to the midbrain and pons (an arrow), NT-4 was detected in the medulla oblongata and the cerebellar primordium (arrowhead and double arrows, respectively). By contrast, NT-4 immunoreactivity was not observed in control sections stained with non-specific IgG (see Fig. 6i). Granular reaction products were detected in the cytoplasm and apparently in the processes of cells in the pons and the medulla oblongata (Figs 6j and k, respectively). Cells stained in the pons seemed to be differentiated, just after they had divided in the ependymal layer and then migrated. On the other hand, only fibers were stained in the cerebellar primordium (Fig. 6l).

Figure 5.

Immunohistochemical staining of NT-4 in the DRG of a rat embryo at E13.5. (a) Sagittal section of the whole embryo. The arrow shows the DRG. (b–d) Sections stained with pAb rNT-4 (c and d) or non-specific rabbit IgG (b). (b) Control section; (c) a low magnification of the DRG indicated by the arrow in (a). Some structures are strongly stained with antibodies; (d) a high magnification picture of the square region in (c). NT-4 protein is present in the cytoplasm of neurons (double arrows) and their fibers. Bars, 1 mm in (a); 100 µm in (b and c); 25 µm in (d).

Figure 6.

Immunohistochemical staining of NT-4 in the brains of developing rat embryos. (a, c, e and g) Sagittal sections of whole embryos or their brain regions at E11.5 (a), E12.5 (c), E13.5 (e) and E14.5 (g). In (e), CP, cerebellar primordium; FB, forebrain; MB, midbrain; MO, medulla oblongata. (b, d, f, h and i–l) Sections stained with pAb rNT-4 (b, d, f, h and j–l) or non-specific rabbit IgG (i). (b, d, f and h) High magnifications of the squared regions in (a, c, e and g), respectively. At E11.5 (b) and E12.5 (d), cells in the midbrain/pons junction are stained. At E13.5 (f), immunoreactivity is observed in regions corresponding to the pons (arrow), MO (arrowhead), and CP (double arrows). The staining intensity was decreased at E14.5 (h, arrowhead) and E15.5 (data not shown). (I–l) High magnification of the pons (i and j), MO (k), and CP (l) at E13.5. The cytoplasm of cells is strongly stained in the pons and medulla oblongata while fibers are stained in the CP. No structures are stained in the control section (i). Bars, 1 mm in (a); 0.5 mm in (c, e and g); 200 µm in (b, d and f); 50 µm in (I and j); 20 µm in (h).

Next, we measured NT-4 levels in the hindbrain or the DRG (Fig. 7ai or aii, respectively) at E13.5 to confirm that the immunoreactivity is due to mature NT-4. In the hindbrain, all individual values determined were above the detection limit of our system. The NT-4 per hindbrain was 3.4 pg on the average (Fig. 7bi). The mean concentration in the DRG (0.72 ng/g) was about 10-fold higher than those in other brain regions from adult or young rats (Fig. 7bii). Western blot analysis demonstrated the protein immunoreacting with NT-4-specific antibodies to have a molecular mass of ∼14-kDa, the size of the mature form of NT-4, in both the hindbrain and DRG regions (Fig. 7c, lanes 4 and 3, respectively). No band corresponding to the precursor form (∼27 kDa) was observed. Detected amounts of mature NT-4 in the hindbrain and DRG were consistent with results obtained by the immunoassay method.

Figure 7.

Quantitation of NT-4 protein in the hindbrain and the DRG of rat embryos at E13.5. (a) Rat whole embryo. (ai and aii) Regions including the hindbrain (HB) and DRG. (b) Amounts of NT-4 in the HB (bi) and DRG (bii). Data are mean ± SD (bars) values from 13 and five embryos for the HB and DRG, respectively. (c) Western blotting analysis. Extracts were prepared from pooled HB and DRG regions of 92 embryos at E13.5. NT-4 protein in the two regions was analyzed as described in Fig. 2(c) except that immunoreactions were enhanced by the PAP method. Lane 1, Recombinant human NT-4 (1 ng); lane 2, recombinant human NT-4 (1 ng) recovered from beads; lanes 3 and 4, proteins from the DRG and HB, respectively; lane 5, proteins recovered from beads after they were incubated with the homogenizing buffer.

Effects of glutamate agonists on NT-4 levels

The synthesis of BDNF is known to be activity-dependent and glutamate agonists such as KA and AMPA stimulate BDNF production (McAllister et al. 1999; Katoh-Semba et al. 2001). Therefore, we investigated the effects of these glutamate agonists on NT-4 levels. KA increased levels of BDNF about sevenfold in the hippocampus and the entorhinal cortex (Table 1). Such increases were statistically significant (p < 0.01) although values varied considerably between rats (in the present study, from two- to 30-fold in the entorhinal cortex), because of the difference in a susceptibility to the action of this compound. By contrast, KA did not have any influence on NT-4 synthesis. Similar effects were observed in the hippocampus for AMPA, suggesting-non-activity-dependent NT-4 synthesis.


The present enzyme immunoassay system specific for NT-4 allowed demonstration of the mature form with a molecular mass of ∼14 kDa, with high concentrations of NT-4 protein in the testis, especially in rats at an early post-natal age, consistent with mRNA findings (Timmusk et al. 1993). Values determined using the present system were higher than those for testis at the same ages reported by Zhang et al. (1999; at 1 week of post-natal age, about 60-fold), indicating that our extraction procedure is more efficient, as earlier found to be the case for other neurotrophins (Katoh-Semba et al. 1998). Amounts of NT-4 protein in the brain, irrespective of the brain regions and post-natal ages, were far lower than those in the testis, consistent with the mRNA levels (Timmusk et al. 1993). Even though the levels are extremely low, the present results seem to suggest specific functions of NT-4 in the individual brain regions because significant changes in the concentration with post-natal development of the brain were observed in several regions.

Recently, cortical NT-4 infusion reportedly rescued lateral geniculate nucleus (LGN) neurons from monocular deprivation-induced atrophy (Riddle et al. 1995), suggesting retrograde action on thalamic afferent. Similarly, infusion of NT-4 into the visual cortex at the peak of the critical period of ocular dominance plasticity seems to accelerate LGN neuron growth (Wahle et al. 2003). At present, it is unclear whether thalamic NT-4 transported from the visual cortex can reach a peak level in the critical period in rats with normal vision. Here, the thalamus had the lowest NT-4 level among adult brain regions and immunoreactive neurons were not detected in the thalamus or neocortex of young adult rats examined. However, it should be noted that only the thalamus had the significantly higher NT-4 level at the neonatal period compared to the adult (see Fig. 3b). It has been reported that NT-4 dependence for a subset of cortical cells is observed only at the onset of the critical period of ocular dominance plasticity (Wahle et al. 2003), suggesting a developmentally regulated NT-4 function specific for the visual cortex of neonatal rats. Our findings may be noteworthy regarding in vivo transport of NT-4 in neonatal rats. In contrast to the decrease in the thalamus, the NT-4 level significantly increased with age in the neocortex. It is known that TrkB-expressing inhibitory neurons in the superior colliculus respond to cortical NT-4 infusion, where NT-4 is anterogradely transported (Spalding et al. 2002). Moreover, dendritic growth of neurons in layers 5 and 6 is reportedly regulated by NT-4 (McAllister et al. 1995). Thus, NT-4 may exhibit region-specific functions in subcortical regions besides the visual cortex, and the increase appears to be related to maturation of such neurons. Further investigations are necessary to elucidate transport of NT-4 in pathways between the thalamus and the neocortex in normal rats because there are many cortical and thalamic afferents and efferents.

Marked increases in levels of NT-4 with post-natal age are also observed in the olfactory bulb, the hypothalamus, and the brain stem although the functions of NT-4 in these regions still remain unclear. However, the presence of NT-4 protein in the red nucleus, interstitial nucleus of the medial longitudinal fasciculus, and pontine reticular nucleus, which are defined as part of the extrapyramidal tract system, suggests involvement in regulation of development of the motor regulatory system. Recently, it has been demonstrated that NT-4 infusion increases the time spent on non-rapid eye movement sleep in rabbits (Kushikata et al. 2002). Sleep-inducing neurons are considered to be controlled via monoaminergic neurons which are located in the locus coeruleus and raphe nucleus, and NT-4 is reported to support survival of noradrenergic neurons in the locus coeruleus (Friedman et al. 1993). Strong immunoreactivity of NT-4 in large neurons in the locus coeruleus points to a possible involvement of the protein in sleep control.

NT-4 synthesis is considered to be not activity-dependent, in contrast to the case with other neurotrophins (McAllister et al. 1999; Katoh-Semba et al. 2001). In fact, our previous observations indicated no change in the concentration of NT-4 protein in ferret visual cortex in response to binocular visual cortex deprivation (Ichisaka et al. 2003). This appears likely in regions including the hippocampus and entorhinal cortex because no developmental changes in NT-4 levels were observed in these regions. Injections of glutamate agonists gave more clear evidence in support of this suggestion (see Table 1).

The localization of NT-4 in the DRG and other peripheral ganglia was in line with expectations because the protein elicits growth of neurites from neurons in the DRG and nodose ganglia (Hallböök et al. 1991). In fact, the concentration of NT-4 in the DRG was higher than those in any brain regions. On the other hand, it was not expected that NT-4 should be distributed in limited brain regions of the rat embryo and that the regions are the midbrain and pons, where NT-4-immunoreactive large neurons are also present in young adult rats. This indicates that NT-4 can have some influence on cells in the midbrain and pons throughout life although the action seems to be developmentally regulated, in view of the weak staining of cells observed after E13.5 (see Fig. 6). NT-4 is reportedly a survival factor for cultured embryonic midbrain dopaminergic neurons (Hynes et al. 1994; Alexi and Hefti 1996) as is the case for cells of the adult midbrain (Lingor et al. 2000). Because strongly stained cells have appeared as differentiating cells migrating from the ependymal layer after division, NT-4 is likely to act in vivo as a survival factor for immature dopaminergic neurons in the embryonic period.


This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan.